------------------------------------------------------------------------
r1730 | kkunal | 2006-05-25 13:41:20 -0400 (Thu, 25 May 2006) | 1 line

bug fix in diff3.ml 
------------------------------------------------------------------------
r1729 | kkunal | 2006-05-25 13:40:39 -0400 (Thu, 25 May 2006) | 1 line

added more examples
------------------------------------------------------------------------
r1727 | kkunal | 2006-05-24 19:21:21 -0400 (Wed, 24 May 2006) | 1 line


------------------------------------------------------------------------
r1726 | kkunal | 2006-05-24 19:20:56 -0400 (Wed, 24 May 2006) | 1 line


------------------------------------------------------------------------
r1725 | kkunal | 2006-05-24 19:04:03 -0400 (Wed, 24 May 2006) | 1 line


------------------------------------------------------------------------
r1724 | kkunal | 2006-05-24 18:56:29 -0400 (Wed, 24 May 2006) | 1 line


------------------------------------------------------------------------
r1723 | bcpierce | 2006-05-24 18:47:36 -0400 (Wed, 24 May 2006) | 1 line

Small improvement to change printing
------------------------------------------------------------------------
r1722 | kkunal | 2006-05-24 18:31:11 -0400 (Wed, 24 May 2006) | 1 line

added more examples for testing cycle merger
------------------------------------------------------------------------
r1721 | kkunal | 2006-05-24 17:53:29 -0400 (Wed, 24 May 2006) | 1 line

fixed typo
------------------------------------------------------------------------
r1720 | bcpierce | 2006-05-24 17:47:30 -0400 (Wed, 24 May 2006) | 1 line

oops
------------------------------------------------------------------------
r1719 | bcpierce | 2006-05-24 17:44:20 -0400 (Wed, 24 May 2006) | 1 line

Beginnings of unit tests for cycle merge
------------------------------------------------------------------------
r1718 | kkunal | 2006-05-24 17:26:32 -0400 (Wed, 24 May 2006) | 1 line

added cycle merge - test cases need to be modified such that the replicas are a permutation of archive and no element is repeated 
------------------------------------------------------------------------
r1712 | jnfoster | 2006-05-23 00:42:06 -0400 (Tue, 23 May 2006) | 1 line

use bcp.bib from repository instead of wget'ing it
------------------------------------------------------------------------
r1711 | bcpierce | 2006-05-22 22:30:28 -0400 (Mon, 22 May 2006) | 1 line

Final tidying in diff3.ml
------------------------------------------------------------------------
r1708 | jnfoster | 2006-05-22 16:58:45 -0400 (Mon, 22 May 2006) | 1 line

turning off all hot keys except for C-s until I can figure out how to eliminate interaction with input forms
------------------------------------------------------------------------
r1707 | jnfoster | 2006-05-22 16:41:50 -0400 (Mon, 22 May 2006) | 1 line

quick hack. set the title of any element with $k_title in demos.php. e.g., to override the name of 'replica #1' set $demo[r1_title]='Concrete'. i made appropriate changes to examples/lenses/
------------------------------------------------------------------------
r1706 | bcpierce | 2006-05-22 16:17:32 -0400 (Mon, 22 May 2006) | 1 line

Better version of diff3.ml -- sorry for not committing it sooner (the prev commit was KK -- not sure what happened with the commit message)
------------------------------------------------------------------------
r1705 | jnfoster | 2006-05-22 16:14:23 -0400 (Mon, 22 May 2006) | 10 lines

Fixed control key stuff to use AccessKey, and added JavaScript hack so
that focus does not remain on textareas. (If I understand the spec,
browsers should do this, but neither Safari nor Firefox does :-/)
 
diff3.ml wasn't compiling after the last commit (which, incidentally,
did not send a commit email. Whoever sent it, did it appear to go
through ok? weird!)  I did what looked reasonable to get it to
compile, but please double check, especially around line 180.


------------------------------------------------------------------------
r1704 | kkunal | 2006-05-22 14:40:11 -0400 (Mon, 22 May 2006) | 1 line

modified diff3.ml to take care of printing problems
------------------------------------------------------------------------
r1702 | bcpierce | 2006-05-22 10:30:43 -0400 (Mon, 22 May 2006) | 1 line

More refinements of list sync action printing... still working on fixing the 'chunks too large' problem, which is also what is causing the 'Change instead of Add/Delete' problem
------------------------------------------------------------------------
r1701 | bcpierce | 2006-05-22 09:32:30 -0400 (Mon, 22 May 2006) | 1 line

Added a bunch of unit tests for list sync -- even more coming
------------------------------------------------------------------------
r1697 | bcpierce | 2006-05-21 16:00:43 -0400 (Sun, 21 May 2006) | 1 line

List sync demo is ready for others to play with
------------------------------------------------------------------------
r1696 | bcpierce | 2006-05-21 15:45:16 -0400 (Sun, 21 May 2006) | 1 line

Added section headers to the demos (so that it's easier to see when you're moving from one section to another)
------------------------------------------------------------------------
r1695 | bcpierce | 2006-05-21 15:30:17 -0400 (Sun, 21 May 2006) | 1 line

Checkpoint, with a unit test showing one more printing bug
------------------------------------------------------------------------
r1694 | jnfoster | 2006-05-21 15:17:46 -0400 (Sun, 21 May 2006) | 1 line

set svn:keywords (to make $Id$ expansion work) on all of trunk
------------------------------------------------------------------------
r1693 | bcpierce | 2006-05-21 15:02:51 -0400 (Sun, 21 May 2006) | 1 line

Checkpoint
------------------------------------------------------------------------
r1692 | jnfoster | 2006-05-21 13:33:23 -0400 (Sun, 21 May 2006) | 1 line

fix sort checking error in listexamples
------------------------------------------------------------------------
r1691 | bcpierce | 2006-05-21 13:19:48 -0400 (Sun, 21 May 2006) | 1 line

Add listexamples demo
------------------------------------------------------------------------
r1690 | bcpierce | 2006-05-21 13:18:37 -0400 (Sun, 21 May 2006) | 1 line

Checkpoint
------------------------------------------------------------------------
r1689 | jnfoster | 2006-05-21 13:13:00 -0400 (Sun, 21 May 2006) | 4 lines

nuking fancy changelog stuff, .version hackery. just use `svn commit`
to checkin, and `svn log` to see ChangeLogs


------------------------------------------------------------------------
r1688 | bcpierce | 2006-05-21 12:21:36 -0400 (Sun, 21 May 2006) | 15 lines

* Finished printing for list sync (modulo one small off-by-one bug
  that I need to talk to Keshav about)

* Since the diff3-based list sync seems to be working now, I've turned
  it on by default; to disable it again, pass '-diff3=false' on the
  command line.

* Deleted the Spinelist module (which was a hack to get around the
  lack of proper list sync)

* Moved the Xmi examples to /experimental (they are a bit dubious
  anyway, and now some test cases are breaking)



------------------------------------------------------------------------
r1686 | bcpierce | 2006-05-20 10:25:31 -0400 (Sat, 20 May 2006) | 2 lines

* List sync printing is now working pretty well

------------------------------------------------------------------------
r1685 | jnfoster | 2006-05-19 17:51:49 -0400 (Fri, 19 May 2006) | 4 lines

Fix another small bug in web demo. It was not updating the replicas or
archive textareas on conflicts.

Also removed Prelude.Native.iter.
------------------------------------------------------------------------
r1672 | jnfoster | 2006-05-15 05:18:42 -0400 (Mon, 15 May 2006) | 1 line

Housekeeping / tidying
------------------------------------------------------------------------
r1671 | jnfoster | 2006-05-12 16:02:38 -0400 (Fri, 12 May 2006) | 12 lines

Fixed version # tracking in ChangeLog, turning on mailing of commit
messages to harmony-hackers again.

To harmony-hackers: you may notice that revision #s do not go up
consecutively. The reason is that /trunk is only one directory in a
larger repository that contains our research papers etc. Commits to
those other directories bump up the revision # but do not send mail to
this list. Sorry :( 

I may also occasionally turn off the the mail hooks when I'm debugging
the commit scripts (to avoid spamming you all with uninteresting
messages).
------------------------------------------------------------------------
r1670 | jnfoster | 2006-05-12 15:58:55 -0400 (Fri, 12 May 2006) | 4 lines

Fixed ChangeLog so that versioning info gets added to commit messages. 

This is a little bit useless because (a) we don't have a numbering scheme beyond SVN revision #s and (b) those #s are global for the whole repository, not just for trunk, so they won't always be consecutive. However, they give folks a reference for changes to trunk.

------------------------------------------------------------------------
r1669 | jnfoster | 2006-05-12 15:21:53 -0400 (Fri, 12 May 2006) | 1 line

checkpoint
------------------------------------------------------------------------
r1668 | jnfoster | 2006-05-12 15:19:30 -0400 (Fri, 12 May 2006) | 1 line

checkpoint
------------------------------------------------------------------------
r1667 | jnfoster | 2006-05-12 10:32:04 -0400 (Fri, 12 May 2006) | 21 lines

Demo, website, and Makefile cleanups:

  * added Javascript key board shortcuts to the web demo. 
       - C-S runs the synchronizer
       - C-h-r resets the demo
       - C-h-e toggles expert mode
       - C-h-i toggles showing instructions
       - C-h-a toggles showing abstract trees
       - C-h-c toggles showing concrete trees
       - C-h-l toggles showing lenses
       - C-h-t toggles showing schema
       - C-h-o toggles showing the archive
       - C-h-h toggles showing harmony output

  * added instructions for anonymous SVN access to website

  * tidied Makefile for exporting (now done nightly)

  * stole unison's Makefile code for updating a ChangeLog. please
    commit to trunk using "make checkin", after editing the file
    logmsg.
------------------------------------------------------------------------
r1665 | jnfoster | 2006-05-12 04:56:13 -0400 (Fri, 12 May 2006) | 1 line

test commit access
------------------------------------------------------------------------
r1659 | jnfoster | 2006-05-11 03:32:13 -0400 (Thu, 11 May 2006) | 1 line

trunk commits -> harmony-hackers, everything else -> harmony
------------------------------------------------------------------------
r1658 | jnfoster | 2006-05-11 03:04:52 -0400 (Thu, 11 May 2006) | 8 lines

Removed done items from TODO:
  * control keys in web demo

  * make'ing examples on a clean source works in branches/schemas, at
    least (the problem was that the bakery was not included in the
    dependencies for making harmony as a library)


------------------------------------------------------------------------
r1656 | bcpierce | 2006-05-10 18:14:00 -0400 (Wed, 10 May 2006) | 1 line

Misc
------------------------------------------------------------------------
r1651 | bcpierce | 2006-05-09 11:07:07 -0400 (Tue, 09 May 2006) | 1 line

TODO tidying
------------------------------------------------------------------------
r1631 | jnfoster | 2006-05-05 09:45:43 -0400 (Fri, 05 May 2006) | 1 line

fix demo
------------------------------------------------------------------------
r1630 | bcpierce | 2006-05-05 08:08:11 -0400 (Fri, 05 May 2006) | 1 line

Checkpoint
------------------------------------------------------------------------
r1629 | bcpierce | 2006-05-05 06:28:26 -0400 (Fri, 05 May 2006) | 1 line

Small (and incomplete) fixes to the logging functionality, following Conrad's change the other day
------------------------------------------------------------------------
r1618 | jnfoster | 2006-05-04 09:02:09 -0400 (Thu, 04 May 2006) | 5 lines

filled out inner inductions in iterator wb proof. 

coded up primitive lens, fast_reverse


------------------------------------------------------------------------
r1616 | bcpierce | 2006-05-04 04:49:29 -0400 (Thu, 04 May 2006) | 1 line

Conrad's errfile patch
------------------------------------------------------------------------
r1590 | bcpierce | 2006-04-25 11:48:57 -0400 (Tue, 25 Apr 2006) | 1 line

Apply Conrad's patch
------------------------------------------------------------------------
r1581 | bcpierce | 2006-04-22 09:07:35 -0400 (Sat, 22 Apr 2006) | 1 line

Action printing for list sync is now minimally finished -- it's not pretty, but it shows more or less what happened.  Still seeing some strange behavior from diff3 itself, though.
------------------------------------------------------------------------
r1576 | harmony | 2006-04-20 11:30:53 -0400 (Thu, 20 Apr 2006) | 1 line

checkpoint few bugfixes to export
------------------------------------------------------------------------
r1575 | jnfoster | 2006-04-20 11:17:30 -0400 (Thu, 20 Apr 2006) | 20 lines

Fixed as per Conrad's bug reports:

- renamed TMP and TMPDIR in trunk/Makefile to avoid clashes with
  standard POSIX environment variables.

- replaced unsafe shell call to "cp" with a function using the
  Pervasives file I/O primitives.

- replaced unsafe shell call to "rm" with Sys.remove

- replaced open_out and open_in with open_out_bin and open_in_bin in
  many places. this should not affect OS X folks; it definitely helps
  Conrad.

- Safelist.rev'd the include path for Focal files. 

- added Misc.is_file and added Conrad's code to Registry so that we
  don't stop at a directory when looking for Focal code.


------------------------------------------------------------------------
r1570 | bcpierce | 2006-04-19 17:39:12 -0400 (Wed, 19 Apr 2006) | 1 line

Printing of sync actions is fixed again.  (I had decided to rip it all out and just have the sync function print out what it's doing, but it turns out this is a bad idea, mainly because you don't know, as you're recursing down the tree, whether you're going to encounter changes, so you're not sure whether to print the path you are traversing (so that, when changes are encountered, it will be clear where they are!).  Much better to create a separate data structure and print that later.  The cost of this was quite a bit of fooling around with OCaml's new recursive module features... :-)
------------------------------------------------------------------------
r1569 | bcpierce | 2006-04-19 15:51:05 -0400 (Wed, 19 Apr 2006) | 1 line

More progress on list synchronization.  Things are more or less working now, but I've somewhat broken the printing of what happens during synchronization.  Will fix that next, but wanted to make a checkpoint of what's working now.  Check out experimental/diff3/diff3.fcl to see what we can do...
------------------------------------------------------------------------
r1568 | jnfoster | 2006-04-19 12:28:43 -0400 (Wed, 19 Apr 2006) | 1 line

checkpoint. some more profiling-suggested optimizations
------------------------------------------------------------------------
r1566 | jnfoster | 2006-04-19 06:26:37 -0400 (Wed, 19 Apr 2006) | 1 line

checkpoint with myself for profiling; tidy oc from trunk
------------------------------------------------------------------------
r1559 | jnfoster | 2006-04-15 09:38:27 -0400 (Sat, 15 Apr 2006) | 16 lines

Removed omega from trunk for now. 

Finished "obvious" quantifier elimination for Presburger
formulas. Unfortunately for the full benefits (i.e., eliminate as many
existentials as possible) we have to eliminate until we reach a fixed
point :-( It might be easier just to write our own quantifier
elimination procedure and avoid using Omega.

The cause of the segfaults I was getting from oc when trying to solve
the huge formulas in addr.fcl seems to be a lexer bug (!). Fiddled the
formatter with an ugly hack so that it prints newlines
periodically. Addr.fcl is processing on my machine now w/o
problems. However, the big formulas really take a long time (2-3min
each in 'oc' alone!)...


------------------------------------------------------------------------
r1558 | jnfoster | 2006-04-15 06:20:23 -0400 (Sat, 15 Apr 2006) | 1 line

fix typo in trunk/examples/bookmarks/mozilla.fcl, checkpoint some easy optimizations, not quite done but don't break anything, in schemas branch
------------------------------------------------------------------------
r1557 | kkunal | 2006-04-15 02:46:07 -0400 (Sat, 15 Apr 2006) | 1 line

modified diff3.ml
------------------------------------------------------------------------
r1554 | bcpierce | 2006-04-14 16:01:09 -0400 (Fri, 14 Apr 2006) | 1 line

Preliminary integration of Keshav's diff3 code (not completely working yet, but controlled by a flag that is false by default, so should not affect normal working)
------------------------------------------------------------------------
r1497 | jnfoster | 2006-04-04 04:55:26 -0400 (Tue, 04 Apr 2006) | 1 line

checkpointing a few small bugfixes for presburger stuff
------------------------------------------------------------------------
r1494 | jnfoster | 2006-04-03 11:08:16 -0400 (Mon, 03 Apr 2006) | 1 line

checkpointing presburger stuff. clunky implementation of satisfiable : formula -> bool using the Omega Calculator
------------------------------------------------------------------------
r1492 | jnfoster | 2006-04-03 08:46:49 -0400 (Mon, 03 Apr 2006) | 1 line

make omega_calc build quietly!
------------------------------------------------------------------------
r1491 | jnfoster | 2006-04-03 07:34:14 -0400 (Mon, 03 Apr 2006) | 1 line

fixed presburger module, standardized omega makefile to build and clean correctly, added oc to src/Makefile
------------------------------------------------------------------------
r1490 | jnfoster | 2006-04-03 07:06:28 -0400 (Mon, 03 Apr 2006) | 1 line

small fixes to compile on stock OS X
------------------------------------------------------------------------
r1489 | jnfoster | 2006-04-03 06:09:27 -0400 (Mon, 03 Apr 2006) | 1 line

added Bill Pugh's Omega Calculator as an external tool
------------------------------------------------------------------------
r1481 | jnfoster | 2006-03-25 11:05:54 -0500 (Sat, 25 Mar 2006) | 1 line

checkpoint
------------------------------------------------------------------------
r1435 | jnfoster | 2006-03-15 08:51:58 -0500 (Wed, 15 Mar 2006) | 1 line

detect localhost for demos
------------------------------------------------------------------------
r1419 | jnfoster | 2006-03-06 23:36:34 -0500 (Mon, 06 Mar 2006) | 1 line

some old hacking i started a while ago, checkpointing for backup, etc
------------------------------------------------------------------------
r1402 | jnfoster | 2006-02-27 18:43:32 -0500 (Mon, 27 Feb 2006) | 1 line

streamlined concat, added definition to lens paper, wb and totality proofs. still needs text
------------------------------------------------------------------------
r1401 | jnfoster | 2006-02-27 16:24:07 -0500 (Mon, 27 Feb 2006) | 1 line

checkpoint; streamlined groupby2, finished wb and totality proofs
------------------------------------------------------------------------
r1282 | bcpierce | 2006-02-05 16:39:34 -0500 (Sun, 05 Feb 2006) | 1 line

Fix address book lens using new flatten
------------------------------------------------------------------------
r1264 | schmitta | 2006-02-02 09:40:21 -0500 (Thu, 02 Feb 2006) | 7 lines

introducing flatten_op, something that's almost List.map l; flatten, except 
that the mapping between concrete and abstract elements is by key instead of by 
position (for the List.map part).

See the manual for a quick explanation.


------------------------------------------------------------------------
r1263 | schmitta | 2006-02-02 08:57:08 -0500 (Thu, 02 Feb 2006) | 3 lines

switch to code at the beginning of the file (which was ignored otherwise)


------------------------------------------------------------------------
r1238 | jnfoster | 2006-01-30 11:53:33 -0500 (Mon, 30 Jan 2006) | 1 line

added Stephane's homepage
------------------------------------------------------------------------
r1225 | harmony | 2006-01-24 14:17:41 -0500 (Tue, 24 Jan 2006) | 1 line

checkpoint small Makefile tweak for sync -nate
------------------------------------------------------------------------
r1224 | harmony | 2006-01-24 14:14:35 -0500 (Tue, 24 Jan 2006) | 1 line

separated papers into genuine papers/talks/other, fixed links in demos
------------------------------------------------------------------------
r1223 | jnfoster | 2006-01-24 13:47:35 -0500 (Tue, 24 Jan 2006) | 1 line

checkpoint for sync
------------------------------------------------------------------------
r1222 | jnfoster | 2006-01-24 11:51:06 -0500 (Tue, 24 Jan 2006) | 1 line

checkpoint to sync with myself
------------------------------------------------------------------------
r1219 | jnfoster | 2006-01-24 01:23:26 -0500 (Tue, 24 Jan 2006) | 1 line

small Makefile fix
------------------------------------------------------------------------
r1218 | jnfoster | 2006-01-24 01:21:57 -0500 (Tue, 24 Jan 2006) | 3 lines

incorporated xml2moz and moz2xml into harmonize-bookmarks. it still depends on plutil, so i should also write some code to fail gracefully if it can't be found and the safari input is not plain xml.

also a few tweaks to web
------------------------------------------------------------------------
r1217 | bcpierce | 2006-01-23 23:44:57 -0500 (Mon, 23 Jan 2006) | 1 line

More better version
------------------------------------------------------------------------
r1216 | bcpierce | 2006-01-23 23:27:32 -0500 (Mon, 23 Jan 2006) | 1 line

New harmonize-plist application.  Also some internal reorganization (the abstract plist type used to represent dictionaries as lists of lists, but I decided that was silly and just flattened everything); I hope I got the safari lens back in a working condition after this change (the unit tests do pass)
------------------------------------------------------------------------
r1214 | jnfoster | 2006-01-23 18:53:28 -0500 (Mon, 23 Jan 2006) | 1 line

updated logo for demo, made div clickable
------------------------------------------------------------------------
r1213 | jnfoster | 2006-01-23 18:25:57 -0500 (Mon, 23 Jan 2006) | 1 line

build harmonize-calendars as stand-alone application (but src/ does not depend on extern/iCalendar). this is the way forward! i think we should release osx ppc and linux x86binaries on the webpage soon.
------------------------------------------------------------------------
r1211 | jnfoster | 2006-01-23 16:07:49 -0500 (Mon, 23 Jan 2006) | 1 line

small tweaks
------------------------------------------------------------------------
r1210 | jnfoster | 2006-01-23 15:15:21 -0500 (Mon, 23 Jan 2006) | 1 line

checkpoint; checking out sources as 'harmony' user on halfdome
------------------------------------------------------------------------
r1209 | jnfoster | 2006-01-23 11:34:32 -0500 (Mon, 23 Jan 2006) | 1 line

small HTML fixes for validation
------------------------------------------------------------------------
r1208 | jnfoster | 2006-01-23 11:01:51 -0500 (Mon, 23 Jan 2006) | 13 lines

More cleaning up. 

The distinction between .src and .fcl is gone. It made the registry
and code that walks the filesystem looking for Focal sources more
confusing than they had to be. All .fcl sources are "literate" and
start in SRC mode (escape to TEX with #{@} as usual). I left some
files with the .src extension if they are used to generate TeX and
non-Focal code (e.g., the parser and lexer in src/). Makefiles still
need some tidying.

Fixed web pages with suggestions from BCP, AS. 


------------------------------------------------------------------------
r1205 | jnfoster | 2006-01-22 20:33:45 -0500 (Sun, 22 Jan 2006) | 1 line

few missing files
------------------------------------------------------------------------
r1204 | jnfoster | 2006-01-22 20:32:42 -0500 (Sun, 22 Jan 2006) | 1 line

revamped webpages. check out http://www.seas.upenn.edu/~harmony/
------------------------------------------------------------------------
r1201 | jnfoster | 2006-01-22 14:11:50 -0500 (Sun, 22 Jan 2006) | 1 line

halfdome install checkpoint
------------------------------------------------------------------------
r1199 | jnfoster | 2006-01-21 23:13:54 -0500 (Sat, 21 Jan 2006) | 1 line

more shuffling. checkpoint
------------------------------------------------------------------------
r1198 | jnfoster | 2006-01-21 22:39:04 -0500 (Sat, 21 Jan 2006) | 1 line

added php
------------------------------------------------------------------------
r1197 | jnfoster | 2006-01-21 22:37:56 -0500 (Sat, 21 Jan 2006) | 1 line

checkpoint
------------------------------------------------------------------------
r1196 | bcpierce | 2006-01-21 11:38:21 -0500 (Sat, 21 Jan 2006) | 1 line

Small improvements and fixups -- 'make test' now runs successfully on OCaml 3.09
------------------------------------------------------------------------
r1191 | bcpierce | 2006-01-19 16:59:50 -0500 (Thu, 19 Jan 2006) | 1 line

Address demo working
------------------------------------------------------------------------
r1190 | jnfoster | 2006-01-19 11:29:30 -0500 (Thu, 19 Jan 2006) | 1 line

dumb typo
------------------------------------------------------------------------
r1189 | schmitta | 2006-01-19 11:03:06 -0500 (Thu, 19 Jan 2006) | 5 lines

to_list now return a list in alphabetical order
this makes flatten add elements in alphabetical order if using ocaml more 
recent than 3.8.04


------------------------------------------------------------------------
r1188 | jnfoster | 2006-01-18 18:50:17 -0500 (Wed, 18 Jan 2006) | 1 line

no need for demo.php wrapper
------------------------------------------------------------------------
r1187 | jnfoster | 2006-01-18 18:50:04 -0500 (Wed, 18 Jan 2006) | 1 line

need to commit this before web changeover.
------------------------------------------------------------------------
r1185 | jnfoster | 2006-01-18 15:58:39 -0500 (Wed, 18 Jan 2006) | 17 lines

Bookmark hacking. 

We now handle the FEEDURL attribute in Mozilla Firefox bookmark
files. It's a little messy because the formats are messy. Safari
stores RSS feeds as regular old links with URIs of the form
"feed://..." whereas Firefox stores them under an attribute
FEEDURL="http://..." The way I decided to handle this is to rewrite
the Firefox-style entries to "feed://..." (using explode, implode, and
some conditionals. FUN!!)

Safari seems to keep some meta-data on RSS feeds. I haven't bothered
to worry about this. When I populated my Mozilla bookmarks with an RSS
feed and synchronized, it appeared in my Safari bookmark file even
though this section was not filled in... But if there are bugs, it's
something to keep in mind.


------------------------------------------------------------------------
r1184 | bcpierce | 2006-01-18 11:12:53 -0500 (Wed, 18 Jan 2006) | 1 line

Lots of work on addresses
------------------------------------------------------------------------
r1182 | bcpierce | 2006-01-16 23:35:48 -0500 (Mon, 16 Jan 2006) | 1 line

Minor changes
------------------------------------------------------------------------
r1181 | bcpierce | 2006-01-07 23:28:08 -0500 (Sat, 07 Jan 2006) | 1 line

Rewrote instructions for using Harmony with Unison; added 'make install'
------------------------------------------------------------------------
r1180 | bcpierce | 2006-01-07 23:27:40 -0500 (Sat, 07 Jan 2006) | 1 line

Rewrote instructions for using Harmony with Unison; added 'make install'
------------------------------------------------------------------------
r1179 | bcpierce | 2006-01-07 23:01:24 -0500 (Sat, 07 Jan 2006) | 1 line

Checkpoint
------------------------------------------------------------------------
r1178 | bcpierce | 2006-01-07 13:37:22 -0500 (Sat, 07 Jan 2006) | 1 line

Checkpoint and release locks for now; I'll continue working on the manual later
------------------------------------------------------------------------
r1177 | bcpierce | 2006-01-07 12:39:08 -0500 (Sat, 07 Jan 2006) | 1 line

Working on installation instructions, manual, etc. -- checkpoint
------------------------------------------------------------------------
r1176 | jnfoster | 2006-01-06 19:47:31 -0500 (Fri, 06 Jan 2006) | 1 line

set keywords property
------------------------------------------------------------------------
r1175 | jnfoster | 2006-01-06 19:44:08 -0500 (Fri, 06 Jan 2006) | 21 lines

implemented "baking in" of focal libraries to the Harmony binary,
along the lines that benjamin suggested.

baker.ml scans the list of files and/or directories given on the
command line, recursively searches for all the .fcl and .src files it
can find (choosing .src over .fcl where both files are defined and
clobbering sources with the same prefix arbitrarily), and emits code
to initialize a hashtable with a bool (to indicate if it is fcl or
src) and the contents of the file. 

when searching for a module, the registry checks the filesystem first
(using the search paths as before). if none is found, it checks the
Bakery.items hash. 

minimally tested, but compare these (make sure FOCALPATH is unset)

  harmony -check Prelude -verbose

  harmony -check Prelude -verbose -I path/to/lenses


------------------------------------------------------------------------
r1172 | jnfoster | 2006-01-05 04:26:30 -0500 (Thu, 05 Jan 2006) | 1 line

bib changes
------------------------------------------------------------------------
r1119 | bcpierce | 2005-12-20 23:28:00 -0500 (Tue, 20 Dec 2005) | 1 line

Fixed the install bugs that Geoff found
------------------------------------------------------------------------
r899 | jnfoster | 2005-11-18 14:34:29 -0500 (Fri, 18 Nov 2005) | 1 line

testing svn commit
------------------------------------------------------------------------
r889 | geoffw | 2005-11-08 11:03:37 -0500 (Tue, 08 Nov 2005) | 2 lines

Testing commit with LANG=en_US.

------------------------------------------------------------------------
r888 | jnfoster | 2005-11-08 10:56:53 -0500 (Tue, 08 Nov 2005) | 1 line

test CETS svn solution
------------------------------------------------------------------------
r884 | jnfoster | 2005-10-27 22:01:48 -0400 (Thu, 27 Oct 2005) | 1 line

much improved bookmark demo
------------------------------------------------------------------------
r882 | jnfoster | 2005-10-25 13:49:29 -0400 (Tue, 25 Oct 2005) | 1 line

testing new 1.2.3 client
------------------------------------------------------------------------
r881 | jnfoster | 2005-10-25 12:53:37 -0400 (Tue, 25 Oct 2005) | 1 line

incorporating BCP's r812 from Friday into new repository
------------------------------------------------------------------------
r880 | jnfoster | 2005-10-25 12:00:55 -0400 (Tue, 25 Oct 2005) | 1 line

testing with en_US.UTF-8
------------------------------------------------------------------------
r879 | jnfoster | 2005-10-25 11:58:11 -0400 (Tue, 25 Oct 2005) | 1 line

testing Metissian SVN 1.1.2 binary
------------------------------------------------------------------------
r878 | jnfoster | 2005-10-24 16:26:44 -0400 (Mon, 24 Oct 2005) | 1 line

test commit
------------------------------------------------------------------------
r877 | jnfoster | 2005-10-24 16:25:28 -0400 (Mon, 24 Oct 2005) | 1 line

remove ie bookmark folder
------------------------------------------------------------------------
r876 | jnfoster | 2005-10-24 16:18:53 -0400 (Mon, 24 Oct 2005) | 1 line

testing from baydemir computer
------------------------------------------------------------------------
r875 | jnfoster | 2005-10-24 16:07:06 -0400 (Mon, 24 Oct 2005) | 1 line

testing commit
------------------------------------------------------------------------
r874 | jnfoster | 2005-10-21 13:26:09 -0400 (Fri, 21 Oct 2005) | 1 line

test commit
------------------------------------------------------------------------
r873 | jnfoster | 2005-10-21 13:20:36 -0400 (Fri, 21 Oct 2005) | 1 line

clean up repository
------------------------------------------------------------------------
r594 | jnfoster | 2005-10-21 10:50:17 -0400 (Fri, 21 Oct 2005) | 1 line

enable logging
------------------------------------------------------------------------
r593 | jnfoster | 2005-10-20 19:56:07 -0400 (Thu, 20 Oct 2005) | 1 line

test commit from home
------------------------------------------------------------------------
r592 | jnfoster | 2005-10-20 19:49:22 -0400 (Thu, 20 Oct 2005) | 1 line

testing svn commit from halfdome
------------------------------------------------------------------------
r591 | jnfoster | 2005-10-20 12:38:41 -0400 (Thu, 20 Oct 2005) | 1 line

test commit with 1.2.3 client
------------------------------------------------------------------------
r590 | bohannon | 2005-10-20 11:08:01 -0400 (Thu, 20 Oct 2005) | 3 lines

Dummy commit.


------------------------------------------------------------------------
r589 | jnfoster | 2005-10-19 18:27:31 -0400 (Wed, 19 Oct 2005) | 1 line

testing svn update for JNF II
------------------------------------------------------------------------
r588 | jnfoster | 2005-10-19 18:26:40 -0400 (Wed, 19 Oct 2005) | 1 line

testing svn update for JNF
------------------------------------------------------------------------
r587 | jnfoster | 2005-10-18 17:34:24 -0400 (Tue, 18 Oct 2005) | 12 lines

demo checkpoint!!

all but one demo working (except xmi which is broken anyways; i
commented it out). aaron: the last demo in relational is broken and i
couldn't figure out why. you can probably fix it faster than i can?

did a pass over all content. not the final word, but incremental
improvements i hope. bookmarks still needs more fleshing out.

a few small warts i know about, but i think we're basically there!


------------------------------------------------------------------------
r586 | jnfoster | 2005-10-13 22:47:13 -0400 (Thu, 13 Oct 2005) | 1 line

testing svn
------------------------------------------------------------------------
r585 | bcpierce | 2005-10-12 19:40:35 -0400 (Wed, 12 Oct 2005) | 7 lines

Implemented Alan's new semantics for synchronization (no delete/delete
conflicts any more).

A little progress toward getting the XMI demo working.  (But commented
it out for now from the demo driver.)


------------------------------------------------------------------------
r584 | jnfoster | 2005-10-11 13:54:48 -0400 (Tue, 11 Oct 2005) | 1 line

checkpoint
------------------------------------------------------------------------
r583 | jnfoster | 2005-10-11 13:52:18 -0400 (Tue, 11 Oct 2005) | 1 line

another checkpoint with aaron
------------------------------------------------------------------------
r582 | jnfoster | 2005-10-11 13:48:07 -0400 (Tue, 11 Oct 2005) | 1 line

checkpoint with aaron
------------------------------------------------------------------------
r581 | jnfoster | 2005-10-11 13:44:54 -0400 (Tue, 11 Oct 2005) | 1 line

oops. added some dummy images
------------------------------------------------------------------------
r580 | jnfoster | 2005-10-11 13:34:24 -0400 (Tue, 11 Oct 2005) | 1 line

checkpoint with aaron. don't svn update to this revision if you require a 100% working demo
------------------------------------------------------------------------
r579 | bcpierce | 2005-10-11 12:09:22 -0400 (Tue, 11 Oct 2005) | 3 lines

Some more cosmetic improvements to printing.


------------------------------------------------------------------------
r578 | bcpierce | 2005-10-11 11:31:51 -0400 (Tue, 11 Oct 2005) | 17 lines

Preliminary support for printing of actions containing spined lists.  

There are still some significant shortcomings.  In particular, because
of the way synchronization of spined lists works, changing an element
in the middle of a spined list will be synchronized (which we
intended) and printed (which is a problem) as a replacement of the
entire tail of the list.

A smaller shortcoming is that I piggybacked the spined list printing
on the list printing in a simple way, so the outer brackets print as
[...] rather than [|...|].

Thoughts?

(If you want to play with it, use the Playground area of the demo.)


------------------------------------------------------------------------
r577 | bcpierce | 2005-10-11 10:36:41 -0400 (Tue, 11 Oct 2005) | 7 lines

Printing of action lists now uses "Replace..." instead of Add/Delete whenever
   - the two replicas have disjoint domains
   - and one is equal to the archive

Comments, please...


------------------------------------------------------------------------
r576 | bcpierce | 2005-10-08 10:12:20 -0400 (Sat, 08 Oct 2005) | 16 lines

First cut at better pretty printing of actions for lists
  (Unfortunately, I ran into a wall on spinelists -- there is no way, in
  general, given the information currently passed to Sync.format, to
  tell whether we are looking at a list node or a spinelist node!  One
  possible solution for this is to pass the intended schema to
  Sync.format...)

Added a new demo called Playground, which provides raw access to the
  Harmony top level (it is similar to the lens tutorial demo at the
  moment, but the latter is going to get some text added pretty soon)

Added keynote slides for WG2.8 meeting in Estonia

Some random tidying


------------------------------------------------------------------------
r575 | bcpierce | 2005-10-08 07:38:05 -0400 (Sat, 08 Oct 2005) | 6 lines

Fixed a small bug in structured text demo.
Changed license to LGPL
Started on pretty printing of "list actions" in user interface



------------------------------------------------------------------------
r574 | schmitta | 2005-09-30 11:03:03 -0400 (Fri, 30 Sep 2005) | 3 lines

example of syntax for spined list


------------------------------------------------------------------------
r573 | schmitta | 2005-09-30 11:02:42 -0400 (Fri, 30 Sep 2005) | 3 lines

We now deal with spined list


------------------------------------------------------------------------
r572 | schmitta | 2005-09-29 10:41:12 -0400 (Thu, 29 Sep 2005) | 3 lines

Ugly inefficiency when converting structures to lists


------------------------------------------------------------------------
r571 | schmitta | 2005-09-29 02:05:00 -0400 (Thu, 29 Sep 2005) | 3 lines

Correct spinelist schema, and a lens to transform from a real list


------------------------------------------------------------------------
r570 | schmitta | 2005-09-27 05:21:54 -0400 (Tue, 27 Sep 2005) | 3 lines

oops, demo logging got commented out ;)


------------------------------------------------------------------------
r569 | schmitta | 2005-09-27 05:20:34 -0400 (Tue, 27 Sep 2005) | 4 lines

fixes to harmony demo basics n.11
ideas on how to really do spined lists


------------------------------------------------------------------------
r568 | bcpierce | 2005-09-27 05:18:04 -0400 (Tue, 27 Sep 2005) | 1 line

Tweaks for demo
------------------------------------------------------------------------
r567 | jnfoster | 2005-08-31 12:36:49 -0400 (Wed, 31 Aug 2005) | 1 line

small fix to skip plutil on linux: make harmonize-bookmarks recognize .xml as an xml encoding of plist (could also add a command-line flag if we later add other xml formats)
------------------------------------------------------------------------
r566 | jnfoster | 2005-08-29 09:21:44 -0400 (Mon, 29 Aug 2005) | 1 line

bookmark demos from DBPL. Thanks to Alan for fixing the forcer1 bug interplay between demos.php + toplevel :-)
------------------------------------------------------------------------
r565 | schmitta | 2005-08-28 11:15:24 -0400 (Sun, 28 Aug 2005) | 6 lines

we now honor the forcer1 preference when doing preprocessing (we don't), and 
when doing the put after sync (we put in None instead of the original tree).

This fixes the last addresses demo and the first calendars php demos


------------------------------------------------------------------------
r564 | jnfoster | 2005-08-25 12:07:32 -0400 (Thu, 25 Aug 2005) | 1 line

oops. enable logging by default
------------------------------------------------------------------------
r563 | jnfoster | 2005-08-25 12:06:41 -0400 (Thu, 25 Aug 2005) | 1 line

make web demos friendlier for offline usage
------------------------------------------------------------------------
r562 | lescuyer | 2005-08-22 11:41:39 -0400 (Mon, 22 Aug 2005) | 1 line

removed intersection function from schema (to the intersection note), removed old sfork, added doc for bfork.
------------------------------------------------------------------------
r561 | jnfoster | 2005-08-19 09:37:50 -0400 (Fri, 19 Aug 2005) | 1 line

some fixes to bookmark lenses; mozilla now keeps hr elements properly
------------------------------------------------------------------------
r560 | bohannon | 2005-08-16 11:29:51 -0400 (Tue, 16 Aug 2005) | 3 lines

Included a more complex example of relational lenses to start off the demo.


------------------------------------------------------------------------
r559 | bcpierce | 2005-08-14 20:59:15 -0400 (Sun, 14 Aug 2005) | 1 line

Some polishing of the demos
------------------------------------------------------------------------
r558 | lescuyer | 2005-08-13 20:55:28 -0400 (Sat, 13 Aug 2005) | 1 line

Added the section in the Harmony manual to explain how to use Unison with it. Progress in the sfork note (the end is in a very draft state right now).
------------------------------------------------------------------------
r557 | lescuyer | 2005-08-12 16:37:54 -0400 (Fri, 12 Aug 2005) | 1 line

Fixed a bug in sfork, added unit tests and documentation for sfork
------------------------------------------------------------------------
r556 | bcpierce | 2005-08-12 15:03:28 -0400 (Fri, 12 Aug 2005) | 1 line

Today's progress on demos
------------------------------------------------------------------------
r555 | lescuyer | 2005-08-12 14:46:01 -0400 (Fri, 12 Aug 2005) | 1 line

Typo..
------------------------------------------------------------------------
r554 | lescuyer | 2005-08-12 14:44:04 -0400 (Fri, 12 Aug 2005) | 1 line

And a third demo on XMI, pointing to Argo and to the example file (because there is not much to say about the behaviour of the lens, but to actually try it on real files)
------------------------------------------------------------------------
r553 | lescuyer | 2005-08-12 14:22:29 -0400 (Fri, 12 Aug 2005) | 1 line

Cleaning
------------------------------------------------------------------------
r552 | lescuyer | 2005-08-12 14:19:46 -0400 (Fri, 12 Aug 2005) | 1 line

The second XMI demo, with colourized explanation
------------------------------------------------------------------------
r551 | lescuyer | 2005-08-12 11:21:00 -0400 (Fri, 12 Aug 2005) | 1 line

Filling in the XMI demo, added an images directory to html/
------------------------------------------------------------------------
r550 | lescuyer | 2005-08-11 19:35:44 -0400 (Thu, 11 Aug 2005) | 1 line

An optimistic sfork (that is, without checks of splittability or orthogonality)
------------------------------------------------------------------------
r549 | lescuyer | 2005-08-11 12:25:37 -0400 (Thu, 11 Aug 2005) | 1 line

Made the building of the iCalendar viewer automatic when making the calendars example
------------------------------------------------------------------------
r548 | bohannon | 2005-08-11 11:44:46 -0400 (Thu, 11 Aug 2005) | 3 lines

Made the instructions less technical.


------------------------------------------------------------------------
r547 | lescuyer | 2005-08-11 10:59:10 -0400 (Thu, 11 Aug 2005) | 1 line

Yet another demo, on XMI files. It's actually added to demo.php, but I'm not very comfortable with that (not because of the behaviour of the lens, which is astonishly ok, but because even the simplest xmi files are really clunky to edit manually). There is a makefile based demo, though.
------------------------------------------------------------------------
r546 | bcpierce | 2005-08-10 22:43:32 -0400 (Wed, 10 Aug 2005) | 1 line

Final progress on demos for today
------------------------------------------------------------------------
r545 | lescuyer | 2005-08-10 19:59:39 -0400 (Wed, 10 Aug 2005) | 1 line

Added bfork for the XMI purpose. bfork.get splits a tree according to the schema under the edges and bfork.put splits the tree according to a set of names. This would be part of a sfork lens of course, but this one's not ready yet, and I need at least this to go on with xmi
------------------------------------------------------------------------
r544 | lescuyer | 2005-08-10 19:57:41 -0400 (Wed, 10 Aug 2005) | 2 lines

A working XMI lens, and some examples to play with. Type make clean to reinitialize the files, make some modifications to example1.ics and example2.ics, and type make sync to see what happens. \ 
I will demo-ize it to examples tomorrow
------------------------------------------------------------------------
r543 | bcpierce | 2005-08-10 15:18:25 -0400 (Wed, 10 Aug 2005) | 1 line

Demo checkpoint
------------------------------------------------------------------------
r542 | bohannon | 2005-08-10 15:18:15 -0400 (Wed, 10 Aug 2005) | 3 lines

Filled out relational demo.


------------------------------------------------------------------------
r541 | bcpierce | 2005-08-10 14:04:51 -0400 (Wed, 10 Aug 2005) | 1 line

Missing files
------------------------------------------------------------------------
r540 | bcpierce | 2005-08-10 13:59:23 -0400 (Wed, 10 Aug 2005) | 1 line

Progress on demos
------------------------------------------------------------------------
r539 | bcpierce | 2005-08-10 13:23:53 -0400 (Wed, 10 Aug 2005) | 1 line

Some polishing of text in Stephane's demos
------------------------------------------------------------------------
r538 | lescuyer | 2005-08-10 13:16:09 -0400 (Wed, 10 Aug 2005) | 1 line

yet another demo with calendars
------------------------------------------------------------------------
r537 | bohannon | 2005-08-10 11:03:37 -0400 (Wed, 10 Aug 2005) | 3 lines

Added a PHP demo for relations.


------------------------------------------------------------------------
r536 | bohannon | 2005-08-09 23:17:45 -0400 (Tue, 09 Aug 2005) | 3 lines

The version has some basic examples suitable for a demonstration.


------------------------------------------------------------------------
r535 | bcpierce | 2005-08-09 18:42:57 -0400 (Tue, 09 Aug 2005) | 1 line

Oops
------------------------------------------------------------------------
r534 | lescuyer | 2005-08-09 18:12:59 -0400 (Tue, 09 Aug 2005) | 1 line

Added newlines in php code at the end of calendars, for the ical viewer
------------------------------------------------------------------------
r533 | bcpierce | 2005-08-09 16:53:20 -0400 (Tue, 09 Aug 2005) | 1 line

More refinement of demos
------------------------------------------------------------------------
r532 | lescuyer | 2005-08-09 16:46:00 -0400 (Tue, 09 Aug 2005) | 1 line

Ahem.. unescaped too much actually, in my last checkin
------------------------------------------------------------------------
r531 | lescuyer | 2005-08-09 15:50:56 -0400 (Tue, 09 Aug 2005) | 1 line

Added the calendars demo to the demogroup
------------------------------------------------------------------------
r530 | lescuyer | 2005-08-09 15:41:44 -0400 (Tue, 09 Aug 2005) | 1 line

Removed useless escaping in the icalendar objects
------------------------------------------------------------------------
r529 | lescuyer | 2005-08-09 15:40:41 -0400 (Tue, 09 Aug 2005) | 1 line

PHPized the calendar demo - Benjamin, if you don't mind giving it a try...
------------------------------------------------------------------------
r528 | lescuyer | 2005-08-09 11:59:49 -0400 (Tue, 09 Aug 2005) | 1 line

Demo 3
------------------------------------------------------------------------
r527 | lescuyer | 2005-08-09 11:38:18 -0400 (Tue, 09 Aug 2005) | 1 line

Demo n?\194?\1762
------------------------------------------------------------------------
r526 | bcpierce | 2005-08-09 11:36:56 -0400 (Tue, 09 Aug 2005) | 1 line

Some missing files
------------------------------------------------------------------------
r525 | bcpierce | 2005-08-09 11:26:12 -0400 (Tue, 09 Aug 2005) | 1 line

More progress on demos
------------------------------------------------------------------------
r524 | bcpierce | 2005-08-09 11:22:06 -0400 (Tue, 09 Aug 2005) | 1 line

More progress on demos; another commit is coming soon
------------------------------------------------------------------------
r523 | lescuyer | 2005-08-09 11:01:21 -0400 (Tue, 09 Aug 2005) | 1 line

Progress towards iCalendar demo
------------------------------------------------------------------------
r522 | bcpierce | 2005-08-08 23:30:15 -0400 (Mon, 08 Aug 2005) | 17 lines

LOTS more progress on the web demos.  

I got frustrated with the fragility and hackiness of the
makefile-based demos, so I finally just junked them in favor of a
pure-PHP solution.  (I'm a halfway decent PHP programmer now. :-)

I still don't have all this installed on a web server where you can
play with it, but if you look at examples/structuredtext/demos.php you
should get a clear idea what's going on.  

I still need to add a little more functionality so that people can
play with lenses in addition to synchronization (this is also needed
for some of Aaron's demos).  And I hope some CSS hacker can help make
it a little prettier.  :-) But basically it is now working very well.
I'm going to use it for the demo on Wednesday to the DB group.


------------------------------------------------------------------------
r521 | lescuyer | 2005-08-08 15:35:10 -0400 (Mon, 08 Aug 2005) | 1 line

Just a tiny correction to the fmodify implementation.
------------------------------------------------------------------------
r520 | lescuyer | 2005-08-08 15:21:31 -0400 (Mon, 08 Aug 2005) | 1 line

Committed stuff for intersection of schemas back, after Nate removed it. There is a new lens, just for debugging, such that intersect A B behaves like id but prints out if the A inter B is empty or not. Also, added two other primitive lenses fconst and fmodify, used in iCalendar to handle dtstamps and lastmods : the first one is a const taking a command as a default tree, the other one is like a prune, but uses its default not only in case of creation, but also when the abstract view has changed.
------------------------------------------------------------------------
r519 | bcpierce | 2005-08-08 12:03:06 -0400 (Mon, 08 Aug 2005) | 1 line

Missing files
------------------------------------------------------------------------
r518 | bcpierce | 2005-08-08 12:01:53 -0400 (Mon, 08 Aug 2005) | 12 lines

Progress on a web-based demo.  It's not quite ready for prime time yet
(and I can't actually put it on the SEAS web server until CETS helps
me out with some strange Apache/PHP problems), but if you want to take
a look you can run it yourself (if you have a webserver that's running
PHP -- note that you need to edit the configuration file on OSX to
make it do this).  If anybody wants to do this, give me a yell and
I'll help.  Or stay tuned for the web version sometime soon, I hope.

(Oops -- I see some files missing.  I'll add them next.)



------------------------------------------------------------------------
r517 | schmitta | 2005-08-04 06:25:29 -0400 (Thu, 04 Aug 2005) | 9 lines

viewers: the XML printer escapes ' and " only in attributes
structuredtext.fcl: we add a dummy header every time the header first "header" 
is not one
structuredtext/Makefile: do the unit test before the demos (makes it simpler to 
debug). I hope I did not break the fragile Makefile infrastructure moving the 
"include" lines a couple lines below
addresses/examples: small fixes


------------------------------------------------------------------------
r516 | bcpierce | 2005-08-04 00:33:36 -0400 (Thu, 04 Aug 2005) | 19 lines

* I reinstated the testing code in Top.Makefile that Alan removed and
  (after LONG struggles with make!) managed to actually make it work.
  Now when you type 'make test' all the defined demos automatically
  get run.  (The results are not checked, but at least if the demo
  falls over the make will fail.)

* Added a "progress" lens that just prints out a given name when it is
  called.  Haven't used it for anything serious yet.  (Also, it
  doesn't have quite the right type: it should really take both a name
  and a sub-lens, so that we can print out the name *before* we execute
  the lens in both get and put directions.)

* Added some text to the address demos.  Check it out!

* Removed dependency of "all" on "test" in main top-level makefile.
  (Too annoying!)



------------------------------------------------------------------------
r515 | jnfoster | 2005-08-02 14:55:20 -0400 (Tue, 02 Aug 2005) | 1 line

checkpoint on lang documentation. also did the small hack to print Info.ts when asserts fail...
------------------------------------------------------------------------
r514 | bcpierce | 2005-08-01 11:50:28 -0400 (Mon, 01 Aug 2005) | 1 line

Sent release message to a few more people
------------------------------------------------------------------------
r513 | jnfoster | 2005-08-01 00:40:12 -0400 (Mon, 01 Aug 2005) | 1 line

checkpoint
------------------------------------------------------------------------
r512 | jnfoster | 2005-08-01 00:25:35 -0400 (Mon, 01 Aug 2005) | 1 line

checkpoint small tweaks
------------------------------------------------------------------------
r511 | jnfoster | 2005-07-31 14:54:36 -0400 (Sun, 31 Jul 2005) | 3 lines

prettier pretty printing. 

Stephane: I removed your stuff from schemas.ml, because I couldn't make sense of it. Can you add it back in? Note that I changed Wildcards. They upper bound, which used to be an int option is now just a bool --- whether or not it's n bangs and a star, or just n bangs. make sense?
------------------------------------------------------------------------
r510 | lescuyer | 2005-07-29 20:00:41 -0400 (Fri, 29 Jul 2005) | 1 line

Progress towards checking if two schemas intersect or not (need it for sfork).
------------------------------------------------------------------------
r509 | lescuyer | 2005-07-29 19:58:05 -0400 (Fri, 29 Jul 2005) | 1 line

Added the lens that simply ignores the date-time metadata, in order to have somthing that works while I won't be here next week. It is now used by harmonize-calendars.
------------------------------------------------------------------------
r508 | lescuyer | 2005-07-29 13:27:22 -0400 (Fri, 29 Jul 2005) | 1 line

Added support for TODO entries to the iCalendar parser. By the way, Alan, your parser really has a great design, it was easy to figure out where and how to add things. Next step, journal entries.
------------------------------------------------------------------------
r507 | lescuyer | 2005-07-29 13:25:14 -0400 (Fri, 29 Jul 2005) | 26 lines

Progress with iCalendar :
* the lens handles the TODO entries
* added a calendar sample with todos instead of only events
* a new ML program, harmonize-calendars, to be called from Unison as a
merge command : the program converts the ics files in meta, calls
harmony with the appropriate arguments, and converts the results back
into ics files
* a Unison profile template for using Unison to synchronize your
calendars, called 'default2'
* the makefile has a 'sync' and a 'modifs' target, that won't work as
they are on your machines, 'sync' applies unison in text mode with the
default2.prf (but it has to be in the .unison dir, and the root names
should be changed for your machines also), while the 'modifs' target
calls an organizer (in my case, korganizer, but you should put ical
instead if you want to try) on both files for viewing and/or making
updates.

The only problem with synchronization right now are the dtstamp and
lastmod properties, who always yield schema-conflict. A lazy solution
is to filter them, but then the resulting ics files have wrong dtstamp
and lastmod contents. I haven't figured out yet how to handle this
properly. Still, if you make make some modifs and then make sync,
after setting up everything for your own system, you can see in the
output how information between entries gets properly synchronized, and
for each entry where something has been added (resp. modified), the
dtstamp (resp. lastmod) throwing a schema-conflict.
------------------------------------------------------------------------
r506 | schmitta | 2005-07-29 05:35:08 -0400 (Fri, 29 Jul 2005) | 3 lines

Changed the unit tests to make them work.


------------------------------------------------------------------------
r505 | schmitta | 2005-07-28 13:00:23 -0400 (Thu, 28 Jul 2005) | 8 lines

Makefiles: changing R2ORIG to R2FORMAT, which is meta if unspecified
example1.xml: first try at a documented simple address demo, there are other to 
do here
viewers.ml: beautyfication of xml output: '\n' are not converted to code point 
(which make pcdata with newlines readable), closing tags are put on a new line 
if needed.


------------------------------------------------------------------------
r504 | lescuyer | 2005-07-26 17:46:56 -0400 (Tue, 26 Jul 2005) | 17 lines

Progress with iCalendar :
* I tried to create a file with KOrganizer, and after skipping the
to-dos and making the parser happy with empty lines, synchronized it
with the old ics sample that was in the repository. Looks fine ! (It
added a dummy anniversary party that occurs every week on Thursday and
lasts for 365 hours, do you remember if that's right Alan ? because
that's  also what happens if I open the old calendar but it sounds
weird though, like hours should be changed to days... perhaps a
compatibility issue ?)
* To synchronize, I aligned the entries according to their uids (tzids
for Timezone entries). A problem though : according to the RFC, uids
are *not* required (but can appear only once... pfew :)). I guess
since iCal and KOrganizer write uids, every decent Organizer must do
so, but that's a necessary assumption if we want to align correctly.

TODO : add support for TODO :-) because it's definitely supported by
KOrganizer and Evolution.
------------------------------------------------------------------------
r503 | bcpierce | 2005-07-26 15:43:06 -0400 (Tue, 26 Jul 2005) | 1 line

A few random improvements, mainly to docs
------------------------------------------------------------------------
r502 | schmitta | 2005-07-26 11:01:20 -0400 (Tue, 26 Jul 2005) | 6 lines

Some progress, we now may have a "category" information in the xcard format, we
deal with "fax" and "other" 'phone' numbers from a palm, and we allow the gaps
in phone numbers on the palm side (but do not deal with too many phone numbers
to write back).


------------------------------------------------------------------------
r501 | schmitta | 2005-07-26 08:29:04 -0400 (Tue, 26 Jul 2005) | 11 lines

Worked on the demo infrastructure:
- if R2ORIG is not set, then the demo does not try to copy it over (and fail)
  and basically does a DEMOCMD on R1 only (which by default is a get whose
  output goes to stdout)
- if NOEDIT is set to something, then the EDITOR value will be ignored (useful
  if one of the replica is binary)
- ported test0 and test10 of addresses. You should try demo2 there (previously
  known as test10): it's fun to edit xcard and csv and see the results
  propagated.


------------------------------------------------------------------------
r500 | schmitta | 2005-07-26 05:29:22 -0400 (Tue, 26 Jul 2005) | 7 lines

It's now possible to run "make test10" twice in a row.
The problem was in the csv lens: it did not ensure that there were entries for
Home, Work, etc... Phone numbers.
What I did is generalize "handle_nonunique_field" to take as argument what to
do if the field is there, and use it to ensure the presence of required fields


------------------------------------------------------------------------
r499 | lescuyer | 2005-07-25 18:51:12 -0400 (Mon, 25 Jul 2005) | 12 lines

* First successful getput of the iCalendar lens ! (unfortunately, the
  ical I have on my distribution is awful and though I can open the file
  properly, I cannot really check its contents ; anyway it's well
  formed ics and that's enough for now)
  The good news is that the compiler is ok with the huge schema, and
  it is around 2^40 times faster than before :)

* Actually, printing just the place where the assertion was failing
  was a bit confusing, since when there are a lot of asserts in one's 
  code, one doesn't know which one has failed. So assert prints both
  now.

------------------------------------------------------------------------
r498 | lescuyer | 2005-07-25 15:52:33 -0400 (Mon, 25 Jul 2005) | 1 line

Fixed the make clean error in extern
------------------------------------------------------------------------
r497 | lescuyer | 2005-07-25 15:50:45 -0400 (Mon, 25 Jul 2005) | 1 line

Oops, forgot to remove my fancy sfork stuff before committing.
------------------------------------------------------------------------
r496 | lescuyer | 2005-07-25 15:40:25 -0400 (Mon, 25 Jul 2005) | 1 line

Changed the membership test so that we can return the first part of the tree (and the associated schema) where the assertion failed. Schema.member hasn't changed, but there is a new function that returns some tree and view instead of a boolean. The assert lens now uses this function to tell the user where the assertion failed. Also, some progress on experimental demos.
------------------------------------------------------------------------
r495 | bcpierce | 2005-07-25 11:49:11 -0400 (Mon, 25 Jul 2005) | 1 line

Fixed Alan's bugs as proposed
------------------------------------------------------------------------
r494 | schmitta | 2005-07-25 10:47:53 -0400 (Mon, 25 Jul 2005) | 4 lines

when the result of synchronization is missing, the default post process
behavior is to delete the replica (or archive)


------------------------------------------------------------------------
r493 | schmitta | 2005-07-25 10:46:40 -0400 (Mon, 25 Jul 2005) | 3 lines

I realized I never committed these small bugfixes


------------------------------------------------------------------------
r492 | bcpierce | 2005-07-25 08:59:49 -0400 (Mon, 25 Jul 2005) | 1 line

Added missing file
------------------------------------------------------------------------
r491 | bcpierce | 2005-07-24 19:52:53 -0400 (Sun, 24 Jul 2005) | 1 line

Added a few words to the quick start
------------------------------------------------------------------------
r490 | bcpierce | 2005-07-24 19:43:11 -0400 (Sun, 24 Jul 2005) | 13 lines

Finished off a first draft of the demo infrastructure and put together
a tutorial-style demo for the structuredtext example (and did some
cute hacks to the lenses involved).  

Please try the demo!!  Just go to the structuredtext directory and
type 'make demo1'.  (This will be more fun if you set your EDITOR
variable to something like 'emacs' or 'vim' first.)

Nate: I made some tiny changes to toplevel.ml -- they should not
conflict with anything you've been doing there, I hope.



------------------------------------------------------------------------
r489 | jnfoster | 2005-07-24 17:33:37 -0400 (Sun, 24 Jul 2005) | 1 line

two small fixes for 'make export'-ing
------------------------------------------------------------------------
r488 | jnfoster | 2005-07-23 10:09:45 -0400 (Sat, 23 Jul 2005) | 7 lines

Fixed the *source* of the bug Benjamin found yesterday afternoon. 

The compiler now throws an exception when it reaches a dummy
value. (According to the semantics, it should diverge, but why not
trap and print a helpful error message instead?) 


------------------------------------------------------------------------
r487 | bcpierce | 2005-07-23 08:57:19 -0400 (Sat, 23 Jul 2005) | 1 line

Little bit more work on the demo infrastructure
------------------------------------------------------------------------
r486 | jnfoster | 2005-07-22 16:50:56 -0400 (Fri, 22 Jul 2005) | 1 line

fixed symptom of test4 in bookmarks. still need to fix compiler so that un-protected values loop...
------------------------------------------------------------------------
r485 | bcpierce | 2005-07-22 16:24:43 -0400 (Fri, 22 Jul 2005) | 1 line

Demo infrastructure working better now.
------------------------------------------------------------------------
r484 | jnfoster | 2005-07-22 16:17:47 -0400 (Fri, 22 Jul 2005) | 13 lines

Fixed a bug in schema implementation. 

(The explanation is long and boring. Summary: because (1) we compile
blocks of schemas/lets by dummy-allocating and backpatching, (2) we
check schema well-formedness of variables by expanding the variable's
definitionand (3) we use CEnvs to represent recursive schemas, we need
to delay checking of schemas until the backpatching has happened; when
schema decls are nested, this gets tricky)

Compiler-specific unit tests should live in lenses/units.src for now...



------------------------------------------------------------------------
r483 | bcpierce | 2005-07-22 16:06:49 -0400 (Fri, 22 Jul 2005) | 13 lines

First cut at demo/testing infrastructure.  To play, cd to
examples/bookmarks and do 'make demo1'.

However, not all is happy.  Doing 'make demo4' for example yields this:

./harmonize-bookmarks  -I /Users/bcpierce/current/harmony/examples/bookmarks/../../lenses -I /Users/bcpierce/current/harmony/examples/bookmarks -compress=false r1.html r2.meta -ar ar.meta 
Fatal error: dummy  was not overwritten.
Fatal error: exception Assert_failure("/Users/bcpierce/current/harmony/src/../src/value.ml", 178, 1)

Any idea why??



------------------------------------------------------------------------
r482 | schmitta | 2005-07-22 11:16:21 -0400 (Fri, 22 Jul 2005) | 3 lines

we convert Safari bookmarks back to binary plist after sync


------------------------------------------------------------------------
r481 | schmitta | 2005-07-22 10:59:18 -0400 (Fri, 22 Jul 2005) | 3 lines

do not exit before files are postprocessed if sync is successful


------------------------------------------------------------------------
r480 | bcpierce | 2005-07-22 09:51:50 -0400 (Fri, 22 Jul 2005) | 1 line

Fix for latest makefile silliness
------------------------------------------------------------------------
r479 | jnfoster | 2005-07-22 00:39:00 -0400 (Fri, 22 Jul 2005) | 1 line

checkpoint
------------------------------------------------------------------------
r478 | bcpierce | 2005-07-21 21:44:44 -0400 (Thu, 21 Jul 2005) | 1 line

FINALLY fixed the makefiles
------------------------------------------------------------------------
r477 | bcpierce | 2005-07-20 14:33:52 -0400 (Wed, 20 Jul 2005) | 1 line

More tidying of toplevel and friends
------------------------------------------------------------------------
r476 | jnfoster | 2005-07-20 14:12:44 -0400 (Wed, 20 Jul 2005) | 1 line

fixes to toplevel
------------------------------------------------------------------------
r475 | lescuyer | 2005-07-20 12:17:01 -0400 (Wed, 20 Jul 2005) | 1 line

Another tiny one.
------------------------------------------------------------------------
r474 | lescuyer | 2005-07-20 12:12:09 -0400 (Wed, 20 Jul 2005) | 1 line

Tweak to toplevel.ml, error output was just a -
------------------------------------------------------------------------
r473 | bcpierce | 2005-07-20 11:26:44 -0400 (Wed, 20 Jul 2005) | 1 line

More improvements
------------------------------------------------------------------------
r472 | bcpierce | 2005-07-20 11:08:03 -0400 (Wed, 20 Jul 2005) | 1 line

all tests now running again
------------------------------------------------------------------------
r471 | bcpierce | 2005-07-20 11:00:03 -0400 (Wed, 20 Jul 2005) | 1 line

Add toplevel.mli
------------------------------------------------------------------------
r470 | bcpierce | 2005-07-20 10:56:05 -0400 (Wed, 20 Jul 2005) | 1 line

Move /copy to /experimental, since it involves a non-projectible schema
------------------------------------------------------------------------
r469 | bcpierce | 2005-07-20 10:53:45 -0400 (Wed, 20 Jul 2005) | 1 line

Front end nearly finished, but this checkpoint doesn't quite work -- don't upgrade yet
------------------------------------------------------------------------
r468 | jnfoster | 2005-07-19 12:08:43 -0400 (Tue, 19 Jul 2005) | 36 lines


Fixed the bug with the missing schemas in harmonize-addresses (and
bookmarks, according to Alan).

The bug appeared because we switched from linking directly against the
object code to linking against harmony.cmxa. Here's how ocamlopt uses
libraries (from the manpage):

> Arguments  ending  in  .cmxa  are taken to be libraries of object code.
> Such a library packs in two files lib.cmxa and lib.a a  set  of  object
> files  (.cmx/.o  files).  Libraries are build with ocamlopt -a (see the
> description of the -a option below). The object files contained in  the
> library  are  linked  as  regular  .cmx files (see above), in the order
> specified when the library was built. The only difference is that if an
> object  file  contained  in a library is not referenced anywhere in the
> program, then it is not linked in.

Note that modules that are not referenced do not get linked
in. Harmony is a little funny because if you chase the modules from
the top level, (at least) viewers.cmx, prelude.cmx, and compiler.cmx
are not directly referenced. E.g., the compiler is only used to
compile dynamically-loaded .fcl files via a backpatch reference in the
registry; Focal modules (like prelude.src) expect Registry.get_library
to contain values for QIds like Native.Prelude.xfork, but do not
actually link against the OCaml module... and so on.

So if we're going to use an OCaml library to package up the Harmony
code for linking against by front-ends (and I'm not sure this is a
good idea, because of these complications), we need to be sure to
explicitly load these otherwise only implicitly-referenced modules. 

I've added [init ()] functions for Compiler and Prelude, and applied
those functions in Toplevel. This is enough to fix the bug with "make
test10" in examples/addresses. 


------------------------------------------------------------------------
r467 | schmitta | 2005-07-19 11:01:08 -0400 (Tue, 19 Jul 2005) | 3 lines

undup now takes care of 3 element lists


------------------------------------------------------------------------
r466 | bcpierce | 2005-07-19 08:35:56 -0400 (Tue, 19 Jul 2005) | 1 line

Adding harmonize-addresses.ml
------------------------------------------------------------------------
r465 | jnfoster | 2005-07-19 00:58:18 -0400 (Tue, 19 Jul 2005) | 16 lines


Fixed a bug in the schema back-end: 

(For the curious: the previous version of the code hid the schema
 datatype and included some constructors. These functions ensured that
 each schema satisfied certain well-formedness constraints:
 disjointness of concatenations, projectability,
 contractiveness. Unfortunately, for recursive schemas, we sometimes
 had to expand the very variable being compiled. And while we only
 needed local information, e.g., the domain set of that schema,
 looking up that variable yielded a dummy value because the compiler
 had not yet backpatched the environment with the actual schema. 

 The workaround is to validate schemas only after they're compiled.)


------------------------------------------------------------------------
r464 | bcpierce | 2005-07-18 23:05:37 -0400 (Mon, 18 Jul 2005) | 1 line

Small checkpoint
------------------------------------------------------------------------
r463 | schmitta | 2005-07-18 14:05:27 -0400 (Mon, 18 Jul 2005) | 4 lines

tiny new test for join
testing my ability to commit


------------------------------------------------------------------------
r462 | bcpierce | 2005-07-18 13:38:52 -0400 (Mon, 18 Jul 2005) | 1 line

Checkpoint
------------------------------------------------------------------------
r461 | bcpierce | 2005-07-18 13:31:20 -0400 (Mon, 18 Jul 2005) | 1 line

Checkpoint
------------------------------------------------------------------------
r460 | jnfoster | 2005-07-18 13:21:31 -0400 (Mon, 18 Jul 2005) | 1 line

combine bug
------------------------------------------------------------------------
r459 | jnfoster | 2005-07-18 12:50:01 -0400 (Mon, 18 Jul 2005) | 1 line

new schema module
------------------------------------------------------------------------
r458 | bcpierce | 2005-07-18 10:33:29 -0400 (Mon, 18 Jul 2005) | 1 line

Checkpoint for reintegration -- NOT COMPLETELY WORKING AT THE MOMENT (problems with makfiles, and with order of initialization of modules... I think)
------------------------------------------------------------------------
r457 | bcpierce | 2005-07-17 17:39:39 -0400 (Sun, 17 Jul 2005) | 1 line

Checkpoint after a major hack attack.  Address sync mostly done.
------------------------------------------------------------------------
r456 | lescuyer | 2005-07-15 16:48:05 -0400 (Fri, 15 Jul 2005) | 1 line

Changed the iCalViewer in order to use the harmony library, as Benjamin did with bookmarks
------------------------------------------------------------------------
r455 | bcpierce | 2005-07-15 14:33:12 -0400 (Fri, 15 Jul 2005) | 15 lines

Makefile reorganization.  

The make process for a tool like harmonizebookmarks is now separated
into two steps: first build harmony.cmxa, then build one more
top-level module and link it against this library.

I haven't fixed the Makefiles in other places (e.g., iCalendar), but
we should do this right away.  In particular, I would like to get rid
of Common.Makefile (fold its contents into src/Makefile and not
include it from anyplace else).

Stephane: can we work on the iCalendar makefile later today?



------------------------------------------------------------------------
r454 | lescuyer | 2005-07-14 16:55:34 -0400 (Thu, 14 Jul 2005) | 1 line

Added a schema for XMI files (not only UML). The assertion fails for some reason in the compiler, I am committing it on debugging purpose.
------------------------------------------------------------------------
r453 | bcpierce | 2005-07-12 12:38:16 -0400 (Tue, 12 Jul 2005) | 3 lines

First successful bookmark sync driven by Unison!!


------------------------------------------------------------------------
r452 | bcpierce | 2005-07-11 22:41:08 -0400 (Mon, 11 Jul 2005) | 5 lines

Mozilla example working... hooray!  

I think we're very close to something that we can use with Unison.  


------------------------------------------------------------------------
r451 | lescuyer | 2005-07-11 16:23:44 -0400 (Mon, 11 Jul 2005) | 1 line

Updated the experimental/iCalendar test to use the extern/iCalViewer stuff. NB : the test will fail right now, because the put way of the lens cannot work properly without using the actual schemas, ie those that take too long right now.
------------------------------------------------------------------------
r450 | lescuyer | 2005-07-11 15:49:18 -0400 (Mon, 11 Jul 2005) | 1 line

Externalized the iCalendar viewer. ics files should be converted with extern/iCalendar/iCalViewer and wont be recognized by harmony anymore
------------------------------------------------------------------------
r449 | schmitta | 2005-07-11 10:00:04 -0400 (Mon, 11 Jul 2005) | 4 lines

put the csv target before the native_code or byte_code target (for all:) as it
needs to be built first


------------------------------------------------------------------------
r448 | schmitta | 2005-07-11 09:28:08 -0400 (Mon, 11 Jul 2005) | 3 lines

HARMONYBIN defined twice


------------------------------------------------------------------------
r447 | jnfoster | 2005-07-11 00:08:28 -0400 (Mon, 11 Jul 2005) | 3 lines

tiny changes as workaround to DD problem:
  * The moz2xml lexer now projects away the DD fields. This is NOT the right thing to do for homogeneous sync but keeping it and extending the mozilla lens to do the right thing would take a little doing. This makes BCP's example work. 
   * Also changed the bit of lens code to rename the bookmarks toolbar folder. BCP's Firefox and my Firefox were using different names. This lens just clobbers the name of the first folder and makes it "bookmarks toolbar". I don't know if every Mozilla keeps the toolbar in the first folder, but I haven't found one that doesn't...
------------------------------------------------------------------------
r446 | bcpierce | 2005-07-10 18:44:11 -0400 (Sun, 10 Jul 2005) | 3 lines

Progress on bookmarks.


------------------------------------------------------------------------
r445 | bcpierce | 2005-07-08 18:16:42 -0400 (Fri, 08 Jul 2005) | 16 lines

First working version of harmonizebookmarks script.

Makefile-ology is not working very well yet:
  - building harmonizebookmarks after a 'make clean' works, but 
    a second one does not
  - 'make clean' removes all the stuff in /src

Only safari-safari sync is working.  Mozilla still needs to be added.
(And IE6, eventually.  And maybe we should revive IE5?)

To play, do 'make test-script' in examples/bookmarks.

   - B



------------------------------------------------------------------------
r444 | lescuyer | 2005-07-08 14:40:39 -0400 (Fri, 08 Jul 2005) | 1 line

Completed synchronization schema from scratch, for Calendar
------------------------------------------------------------------------
r443 | kimthang | 2005-07-08 14:34:40 -0400 (Fri, 08 Jul 2005) | 1 line

make it more perfect
------------------------------------------------------------------------
r442 | bcpierce | 2005-07-08 12:06:29 -0400 (Fri, 08 Jul 2005) | 1 line

Checkin to make sure I really fixed the infinite makefile loop...
------------------------------------------------------------------------
r441 | lescuyer | 2005-07-08 11:34:29 -0400 (Fri, 08 Jul 2005) | 1 line

Added schema and lens for recur elements ; 'recordize' things as more as possible in the sync schema
------------------------------------------------------------------------
r440 | lescuyer | 2005-07-07 18:22:48 -0400 (Thu, 07 Jul 2005) | 1 line

A first try for a icalendar lens, from scratch, and the beginning of another one, straightly translated from Alan's
------------------------------------------------------------------------
r439 | bcpierce | 2005-07-06 14:22:33 -0400 (Wed, 06 Jul 2005) | 1 line

Working shell script for harmonizing safari bookmarks
------------------------------------------------------------------------
r438 | bcpierce | 2005-07-05 23:52:40 -0400 (Tue, 05 Jul 2005) | 1 line

Oops -- here's the rest
------------------------------------------------------------------------
r437 | jnfoster | 2005-07-05 22:52:18 -0400 (Tue, 05 Jul 2005) | 1 line

code to map bookmarksbar to something canonical. we may want to change the abstract schema down the road, but this works for safari/mozilla sync at least...
------------------------------------------------------------------------
r436 | jnfoster | 2005-07-05 15:46:58 -0400 (Tue, 05 Jul 2005) | 1 line

checkpoint bookmark demo
------------------------------------------------------------------------
r435 | lescuyer | 2005-06-30 19:07:12 -0400 (Thu, 30 Jun 2005) | 12 lines

Finished writing the schema for concrete views, as delivered by Alan's viewer.
The CompProp schema for calendar components is a schema of 40 or so optional 
children, so an assert cannot be run without risk with this schema :) 
To make it ok for now and keep the code, I just redefine CompProp as Any at the
end, waiting for better days to remove this line :)

Note that when trying with the full schema, this almost freezes the computer
because most of the time is not spent in Harmony but in the memory management ;
my system ran out of memory after 20 sec and then the swap began to fill up
very quickly. It looks like we're deep in some recursive loop that transports 
a growing amount of data on its way... and anyway 2^40 is way too big :D

------------------------------------------------------------------------
r434 | lescuyer | 2005-06-30 16:36:11 -0400 (Thu, 30 Jun 2005) | 1 line

Some progress into writing the schema for Alan's concrete iCal views
------------------------------------------------------------------------
r433 | lescuyer | 2005-06-27 18:34:33 -0400 (Mon, 27 Jun 2005) | 1 line

Ok, re-added lexer and parser generated files to the repository, and removed them from the make clean
------------------------------------------------------------------------
r432 | lescuyer | 2005-06-27 18:29:12 -0400 (Mon, 27 Jun 2005) | 1 line

Oops, removing the source files generated by yacc and lex from the repository
------------------------------------------------------------------------
r431 | lescuyer | 2005-06-27 18:22:20 -0400 (Mon, 27 Jun 2005) | 16 lines

After the discussion with Alan & Benjamin, ported the iCalendar files
to Harmony4 and added them in a iCalendar subdirectory in src.

I also added a iCalendar directory in experimental and a test on a
ics files. There is no lens at work right now, just the reader and 
this test is just to check that everything is going well when the 
file is	converted into meta.

Now, be careful with making clean... because 95% of the time spent to
compile the whole Harmony is spent in iCalendarlex.ml (looking at the
file, it's no wonder !). I hesitated adding iCalendar/ or not to the
SUBDIRSCLEANONLY variable in src/Makefile, and I decided to put it for
the sake of coherence. If the compilation time bothers you, consider
removing it and typing make clean in src/iCalendar/ manually when
really needed.

------------------------------------------------------------------------
r430 | bohannon | 2005-06-27 13:52:30 -0400 (Mon, 27 Jun 2005) | 3 lines

Corrected a typo.


------------------------------------------------------------------------
r429 | bcpierce | 2005-06-27 13:52:33 -0400 (Mon, 27 Jun 2005) | 1 line

Forgot to add these files
------------------------------------------------------------------------
r428 | bohannon | 2005-06-27 13:03:24 -0400 (Mon, 27 Jun 2005) | 3 lines

New versions of relational lenses (and updated documentation and examples).


------------------------------------------------------------------------
r427 | bcpierce | 2005-06-24 15:00:56 -0400 (Fri, 24 Jun 2005) | 1 line

Some progress on addresses...
------------------------------------------------------------------------
r426 | schmitta | 2005-06-24 11:44:48 -0400 (Fri, 24 Jun 2005) | 3 lines

this now compiles


------------------------------------------------------------------------
r425 | schmitta | 2005-06-23 17:38:28 -0400 (Thu, 23 Jun 2005) | 3 lines

tiny shell script to synchronize addresses


------------------------------------------------------------------------
r424 | bcpierce | 2005-06-23 17:24:49 -0400 (Thu, 23 Jun 2005) | 1 line

CSV sample
------------------------------------------------------------------------
r423 | bohannon | 2005-06-23 17:22:03 -0400 (Thu, 23 Jun 2005) | 3 lines

Minor fixes.


------------------------------------------------------------------------
r422 | bcpierce | 2005-06-23 17:18:11 -0400 (Thu, 23 Jun 2005) | 1 line

Samples
------------------------------------------------------------------------
r421 | bohannon | 2005-06-23 17:18:02 -0400 (Thu, 23 Jun 2005) | 3 lines

A viewer for CSV databases (directories of csv files).


------------------------------------------------------------------------
r420 | bcpierce | 2005-06-23 17:12:31 -0400 (Thu, 23 Jun 2005) | 1 line

Forgot to add Makefile
------------------------------------------------------------------------
r419 | bcpierce | 2005-06-23 17:10:53 -0400 (Thu, 23 Jun 2005) | 1 line

Sample addresses for alan
------------------------------------------------------------------------
r418 | bohannon | 2005-06-23 17:10:14 -0400 (Thu, 23 Jun 2005) | 3 lines

CSV database converter module.


------------------------------------------------------------------------
r417 | bohannon | 2005-06-23 17:09:34 -0400 (Thu, 23 Jun 2005) | 3 lines

CSV view for loading tables.


------------------------------------------------------------------------
r416 | bcpierce | 2005-06-23 16:46:10 -0400 (Thu, 23 Jun 2005) | 1 line

Generalized readers so that they can take either a filename or a string as argument; writers put things directly in files, not strings
------------------------------------------------------------------------
r415 | bcpierce | 2005-06-23 15:43:36 -0400 (Thu, 23 Jun 2005) | 1 line

Finished makefile support for cvs library
------------------------------------------------------------------------
r414 | bohannon | 2005-06-23 15:33:05 -0400 (Thu, 23 Jun 2005) | 3 lines

Update makefiles to build CSV library.


------------------------------------------------------------------------
r413 | bohannon | 2005-06-23 15:31:59 -0400 (Thu, 23 Jun 2005) | 3 lines

Added the csv libarary.


------------------------------------------------------------------------
r412 | bcpierce | 2005-06-23 14:36:24 -0400 (Thu, 23 Jun 2005) | 6 lines

Added capability to read and write stdin/stdout by giving - as filename.
Makefile fixes.  (Esp for export.)
Some minor tweaks.



------------------------------------------------------------------------
r411 | bcpierce | 2005-06-21 20:09:47 -0400 (Tue, 21 Jun 2005) | 1 line

Small fixes and additions
------------------------------------------------------------------------
r410 | bcpierce | 2005-06-20 14:09:06 -0400 (Mon, 20 Jun 2005) | 3 lines

Small fixes.  Ready for release...!


------------------------------------------------------------------------
r409 | bcpierce | 2005-06-20 13:28:20 -0400 (Mon, 20 Jun 2005) | 1 line

Documentation update.  I added a new form of unit test that prints out the result instead of comparing it to anything (put a question mark on the rhs of the =)
------------------------------------------------------------------------
r408 | bcpierce | 2005-06-20 13:24:38 -0400 (Mon, 20 Jun 2005) | 1 line

Quick start finished.  Whew.
------------------------------------------------------------------------
r407 | kimthang | 2005-06-20 12:13:51 -0400 (Mon, 20 Jun 2005) | 8 lines

With a "new technique" :-) , the process of synchronization in Vcard is very
fast now. I have transformed a view:
      { ?n1 = V, ?n2 = V, ..., ?nk = V } to {info = V} 
 (using some lens: to_list, concat, explode and implode; see detail in lens.fcl).
 So, the synchronization is ... 2^k times faster. (In the case of Vcard, it's 
2^14 times faster). 
 I think, we can apply this "technique" to synchronize a view with a
big number of nodes in a same level. 
------------------------------------------------------------------------
r406 | lescuyer | 2005-06-17 18:40:53 -0400 (Fri, 17 Jun 2005) | 1 line

unit tests and explanations for fold_right. fold_right now doesn't put its init argument into a list anymore, but the default when missing is still [], because we want the concrete view to be a list. I changed what was then broken in structuredtext, but I don't have time to check if there is something wrong some place else, I must go and catch a friend from NY in Chinatown. I'm sorry for that, but the fix is easy, put the init argument into brackets.
------------------------------------------------------------------------
r405 | bohannon | 2005-06-17 18:25:25 -0400 (Fri, 17 Jun 2005) | 3 lines

Still working on examples.


------------------------------------------------------------------------
r404 | bcpierce | 2005-06-17 18:19:07 -0400 (Fri, 17 Jun 2005) | 1 line

checkpoint (and releasing all locks for the evening)
------------------------------------------------------------------------
r403 | bcpierce | 2005-06-17 16:30:22 -0400 (Fri, 17 Jun 2005) | 1 line

checkpoint
------------------------------------------------------------------------
r402 | jnfoster | 2005-06-17 14:24:31 -0400 (Fri, 17 Jun 2005) | 1 line

final commit
------------------------------------------------------------------------
r401 | jnfoster | 2005-06-17 13:54:32 -0400 (Fri, 17 Jun 2005) | 1 line

oops. accidental commit in Thang's stuff earlier, second time
------------------------------------------------------------------------
r400 | jnfoster | 2005-06-17 13:54:00 -0400 (Fri, 17 Jun 2005) | 1 line

oops. accidental commit in Thang's stuff earlier
------------------------------------------------------------------------
r399 | jnfoster | 2005-06-17 12:10:41 -0400 (Fri, 17 Jun 2005) | 1 line

checkpoint for PLClub
------------------------------------------------------------------------
r398 | bohannon | 2005-06-17 11:58:18 -0400 (Fri, 17 Jun 2005) | 3 lines

Fixed outer join--or, more specifically, made it more flexible.


------------------------------------------------------------------------
r397 | kimthang | 2005-06-17 11:40:08 -0400 (Fri, 17 Jun 2005) | 1 line

put back Vcard schema
------------------------------------------------------------------------
r396 | bohannon | 2005-06-17 11:25:02 -0400 (Fri, 17 Jun 2005) | 3 lines

Combined database example into one file.


------------------------------------------------------------------------
r395 | jnfoster | 2005-06-17 11:12:17 -0400 (Fri, 17 Jun 2005) | 1 line

moved type -> schema :-)
------------------------------------------------------------------------
r394 | bohannon | 2005-06-17 11:05:26 -0400 (Fri, 17 Jun 2005) | 3 lines

Here it is.


------------------------------------------------------------------------
r393 | jnfoster | 2005-06-17 11:03:31 -0400 (Fri, 17 Jun 2005) | 1 line

fix to type sort routine
------------------------------------------------------------------------
r392 | bohannon | 2005-06-17 11:03:01 -0400 (Fri, 17 Jun 2005) | 3 lines

Added a larger relational example.  Why aren't the files listed here?


------------------------------------------------------------------------
r391 | bohannon | 2005-06-17 10:08:16 -0400 (Fri, 17 Jun 2005) | 4 lines

Added a newline so that errors don't run directly into whatever follows, e.g.
my prompt.


------------------------------------------------------------------------
r390 | jnfoster | 2005-06-17 10:07:12 -0400 (Fri, 17 Jun 2005) | 1 line

checkpoint. coming into Levine
------------------------------------------------------------------------
r389 | kimthang | 2005-06-16 18:00:30 -0400 (Thu, 16 Jun 2005) | 1 line

Small changes to make it works in new definitions of PCDATA, CHILDREN and ?n
------------------------------------------------------------------------
r388 | bcpierce | 2005-06-16 17:37:53 -0400 (Thu, 16 Jun 2005) | 1 line

Last commit for the day.  Looks like we need to postpone the release date till tomorrow morning... we're almost there, though!
------------------------------------------------------------------------
r387 | bohannon | 2005-06-16 17:07:55 -0400 (Thu, 16 Jun 2005) | 3 lines

Adjusted documentation.


------------------------------------------------------------------------
r386 | kimthang | 2005-06-16 17:01:33 -0400 (Thu, 16 Jun 2005) | 31 lines

Use:

  "make"         :  to synchronize vcard1.xml and vcard1.vcf
  "make xml-xml" :  to synchronize vcard1.xml and vcard2.xml
  "make vcf-vcf" :  to synchronize vcard1.vcf and vcard2.vcf

It takes about 11 minutes in "make" and "make xml-xml", about 5 minutes in "make vcf-vcf"  
(It's so slow because of the schema of vCard; I don't like it :-) )

Resume the process of demoVcard:
 _ According to the definition of vCard in xml, its format is not a view. 
   The lexer xml2view.mll translates this format to a view (in the format xml). 
    For exmaple: in the original vCard, it could have two keyword of the same name 
                 "TEL" (in the same depth of tree) so it is not a view. The lexer 
	         combines all subtrees of two "TEL" to a subtree of only "TEL":
                    {..., TEL = tree1, ..., TEL = tree2, ... } 
	   	  => { ..., TEL = {tree1, tree2}, ...}
   The lexer view2xml.mll does the contrary.
 
 _ The lexers vcf2xml.mll and xml2vcf translate the format vcf to xml and xml to vcf 
   respectively.  

 _ In the module Lens, I use two new lens: to_list and "general" concat (Remark: in 
   the lens concat I change character space (" ") bettwen 2 lists by character semicolon 
   because it's more appropriate when we synchronize xml and vcf). The lenses "vcf" and 
   "xml" change vCard in each format to the view of schema Vcard that I define in schemas.fcl
 
 _ I write the schema Vcard less general than the definition of vCard in condition of speed.
   I keep all the field I think it's important or more general in both format. If you want,
   you can add any field (I've defined in the same file) to the schema Vcard. Certainly, it 
   makes the process of synchronization slower. 
------------------------------------------------------------------------
r385 | kimthang | 2005-06-16 16:50:30 -0400 (Thu, 16 Jun 2005) | 1 line

Prepare to commit the whole Vcard
------------------------------------------------------------------------
r384 | jnfoster | 2005-06-16 16:41:01 -0400 (Thu, 16 Jun 2005) | 1 line

parser changes : ?! --> ?
------------------------------------------------------------------------
r383 | jnfoster | 2005-06-16 16:10:55 -0400 (Thu, 16 Jun 2005) | 1 line

oops. parser.src was broken
------------------------------------------------------------------------
r382 | jnfoster | 2005-06-16 16:08:54 -0400 (Thu, 16 Jun 2005) | 16 lines

Changed syntax of `?'s in types. 

	?n=T
	?!=T
instead of 
	n?=T
	!?=T
(sorry Thang)

Parser documentation is legal EBNF, I think.

Random cleanups of code from hacking with Stephane today. 

Included hyperrefs in the document. Do we like?


------------------------------------------------------------------------
r381 | bcpierce | 2005-06-16 15:48:44 -0400 (Thu, 16 Jun 2005) | 1 line

Some significant changes to Xml module -- introduced predefined names for CHILDREN and PCDATA
------------------------------------------------------------------------
r380 | bcpierce | 2005-06-16 15:00:20 -0400 (Thu, 16 Jun 2005) | 1 line

Nearly done with list.src -- left some questions in relational.src
------------------------------------------------------------------------
r379 | bohannon | 2005-06-16 13:17:25 -0400 (Thu, 16 Jun 2005) | 3 lines

First complete draft of docs and test cases for relations is finished.


------------------------------------------------------------------------
r378 | jnfoster | 2005-06-15 15:09:13 -0400 (Wed, 15 Jun 2005) | 1 line

contractiveness check gone
------------------------------------------------------------------------
r377 | lescuyer | 2005-06-15 11:37:57 -0400 (Wed, 15 Jun 2005) | 1 line

the schema value was defined as a single child pointing to Any tree, which I think was a typo, so I changed it.
------------------------------------------------------------------------
r376 | jnfoster | 2005-06-15 09:44:00 -0400 (Wed, 15 Jun 2005) | 3 lines

Dropped trace, changed subsection headers near the bottom.


------------------------------------------------------------------------
r375 | bcpierce | 2005-06-15 09:36:15 -0400 (Wed, 15 Jun 2005) | 1 line

Prelude documentation finished!  (releasing all locks for the next couple hours)
------------------------------------------------------------------------
r374 | bohannon | 2005-06-14 18:04:50 -0400 (Tue, 14 Jun 2005) | 3 lines

Improvements on documentation for lenses--almost done.


------------------------------------------------------------------------
r373 | jnfoster | 2005-06-14 17:52:00 -0400 (Tue, 14 Jun 2005) | 1 line

oops. below is gone. replaced with a wmap
------------------------------------------------------------------------
r372 | jnfoster | 2005-06-14 17:42:42 -0400 (Tue, 14 Jun 2005) | 9 lines

* Took away Lens memo hit stats (Lens.rate ()) will give it to you if
  you want to check it)

* Brought bookmarks stuff into a compiling state. I will continue
  developing it back into a full-demo tonight.

  Off to play volleyball for a little while :-)


------------------------------------------------------------------------
r371 | jnfoster | 2005-06-14 17:23:11 -0400 (Tue, 14 Jun 2005) | 9 lines


Fixed formatting of error messages for Emacs users :-) 

My line numbers look OK now. Aaron, can you check that yours are?

Added the default pre-processor to the XML viewer (strips away
DOCTYPE). The old code also did this.


------------------------------------------------------------------------
r370 | bcpierce | 2005-06-14 16:50:39 -0400 (Tue, 14 Jun 2005) | 1 line

More progress on prelude.src
------------------------------------------------------------------------
r369 | jnfoster | 2005-06-14 15:31:07 -0400 (Tue, 14 Jun 2005) | 1 line

preliminary almost-EBNF for parser
------------------------------------------------------------------------
r368 | bcpierce | 2005-06-14 15:03:42 -0400 (Tue, 14 Jun 2005) | 1 line

yet more on prelude.src
------------------------------------------------------------------------
r367 | jnfoster | 2005-06-14 14:02:04 -0400 (Tue, 14 Jun 2005) | 5 lines

IMPORTANT CHANGE:

We now REQUIRE sort annotations on let declarations. 


------------------------------------------------------------------------
r366 | bcpierce | 2005-06-14 13:58:11 -0400 (Tue, 14 Jun 2005) | 1 line

more on prelude.src; makefile fix
------------------------------------------------------------------------
r365 | jnfoster | 2005-06-14 13:51:15 -0400 (Tue, 14 Jun 2005) | 1 line

checkpoint
------------------------------------------------------------------------
r364 | jnfoster | 2005-06-14 13:46:33 -0400 (Tue, 14 Jun 2005) | 1 line

small bug fix: need to call t2nf before calculating tdoms
------------------------------------------------------------------------
r363 | jnfoster | 2005-06-14 12:17:49 -0400 (Tue, 14 Jun 2005) | 1 line

fixed printing of line directives in src2f. New format is '# nnn "filename"'
------------------------------------------------------------------------
r362 | jnfoster | 2005-06-14 11:30:41 -0400 (Tue, 14 Jun 2005) | 1 line

preliminary pass simplifying error handling. Format.printfs are not all properly boxed yet. Thought I should do it in two phases
------------------------------------------------------------------------
r361 | bcpierce | 2005-06-13 18:25:13 -0400 (Mon, 13 Jun 2005) | 1 line

More progress on documenting prelude.src
------------------------------------------------------------------------
r360 | jnfoster | 2005-06-13 18:18:56 -0400 (Mon, 13 Jun 2005) | 1 line

removed probe from spinelist sync test
------------------------------------------------------------------------
r359 | bcpierce | 2005-06-13 17:10:19 -0400 (Mon, 13 Jun 2005) | 12 lines

BIG CHANGE!  Proceed with caution...

* "type" becomes "schema" everywhere (we still need to change type.ml to 
  schema.ml, I guess 

* "view" becomes "tree"

This was harder than I expected!  There are probably a few remaining
ones to catch, in comments and such, and certainly some in
experimental/*


------------------------------------------------------------------------
r358 | bohannon | 2005-06-13 16:27:36 -0400 (Mon, 13 Jun 2005) | 1 line

quick and dirty hack to get rid of the first # in lexer and parser sources
------------------------------------------------------------------------
r357 | jnfoster | 2005-06-13 15:51:33 -0400 (Mon, 13 Jun 2005) | 1 line

moved parser and lexer to .src files, preliminary lexer documentation (mostly stolen from Pict :-) ) and keyword changes: view -> tree and type -> schema.
------------------------------------------------------------------------
r356 | bohannon | 2005-06-13 15:50:40 -0400 (Mon, 13 Jun 2005) | 3 lines

Added some test cases.  Commented code implementing outer join.


------------------------------------------------------------------------
r355 | bcpierce | 2005-06-13 15:28:21 -0400 (Mon, 13 Jun 2005) | 1 line

tweaked etags support
------------------------------------------------------------------------
r354 | bcpierce | 2005-06-13 15:26:14 -0400 (Mon, 13 Jun 2005) | 1 line

etags support
------------------------------------------------------------------------
r353 | jnfoster | 2005-06-13 14:42:46 -0400 (Mon, 13 Jun 2005) | 1 line

todos
------------------------------------------------------------------------
r352 | bcpierce | 2005-06-13 14:12:21 -0400 (Mon, 13 Jun 2005) | 1 line

Broke main.tex up into several parts (with .src suffixes)
------------------------------------------------------------------------
r351 | bohannon | 2005-06-13 13:05:39 -0400 (Mon, 13 Jun 2005) | 3 lines

Inner join was broken, as I discovered when adding docs and unit tests for it.  So I fixed it.


------------------------------------------------------------------------
r350 | bcpierce | 2005-06-13 12:30:19 -0400 (Mon, 13 Jun 2005) | 1 line

Prelude.explode/implode and split/unsplit moved to List module.  Also, split/unsplit renamed to lines/unlines.
------------------------------------------------------------------------
r349 | bcpierce | 2005-06-13 12:23:02 -0400 (Mon, 13 Jun 2005) | 1 line

Prelude.flatten and Prelude.join moved to List module
------------------------------------------------------------------------
r348 | bcpierce | 2005-06-11 09:36:15 -0400 (Sat, 11 Jun 2005) | 1 line

Some notes
------------------------------------------------------------------------
r347 | bohannon | 2005-06-10 21:47:02 -0400 (Fri, 10 Jun 2005) | 5 lines

Found bug in the spec for inner join.  Discovered fix and implemented it in
the code.  (PutGet show now hold.)  Wrote some docs and test cases for inner
join.


------------------------------------------------------------------------
r346 | jnfoster | 2005-06-10 21:14:54 -0400 (Fri, 10 Jun 2005) | 1 line

last checkpoint of the week (just trying to help us make it to rev 500 really, Benjamin promised we'd make it by today)
------------------------------------------------------------------------
r345 | jnfoster | 2005-06-10 20:33:18 -0400 (Fri, 10 Jun 2005) | 19 lines

Fixed bug in type normalization (was trying to normalize applications
at every level, rather than stoppping one level down. All tests now
work. 

Got rid of extra messages I accidentally committed earlier today
(e.g. "Welcome to sync") :-)

Fixed Spinelist.T, which wasn't path consistent (try projecting the old definition:
            T X = { !\(`List.TL) = X, `List.TL = (T X) } | []
on nil: you get {} and X). 

Fixed the sync test in spinelist. The specification didn't match the
actual results. Benjamin, or Alan, is this what you intended? (I can
also check it later, but figured you guys might have a quick answer;
the old spec is in a comment below).

Whew! I think all the code we have works now :-)


------------------------------------------------------------------------
r344 | bohannon | 2005-06-10 18:02:03 -0400 (Fri, 10 Jun 2005) | 3 lines

First stab at outer join.


------------------------------------------------------------------------
r343 | schmitta | 2005-06-10 17:44:08 -0400 (Fri, 10 Jun 2005) | 3 lines

first useful use of copy ;-)


------------------------------------------------------------------------
r342 | jnfoster | 2005-06-10 17:32:41 -0400 (Fri, 10 Jun 2005) | 1 line

parity error on -no-assert flag
------------------------------------------------------------------------
r341 | schmitta | 2005-06-10 17:30:32 -0400 (Fri, 10 Jun 2005) | 3 lines

a few printing issues


------------------------------------------------------------------------
r340 | schmitta | 2005-06-10 16:21:17 -0400 (Fri, 10 Jun 2005) | 3 lines

svn ignore for ubase


------------------------------------------------------------------------
r339 | schmitta | 2005-06-10 16:19:42 -0400 (Fri, 10 Jun 2005) | 8 lines

ignoring some files

To do the same:
svn propedit svn:ignore .
and edit the files you want to ignore in the current directory (such as
"*.cmo") in your default editor


------------------------------------------------------------------------
r338 | jnfoster | 2005-06-10 14:45:52 -0400 (Fri, 10 Jun 2005) | 1 line

normalization for types
------------------------------------------------------------------------
r337 | kimthang | 2005-06-10 14:40:32 -0400 (Fri, 10 Jun 2005) | 1 line

I think general concat is pretty working now
------------------------------------------------------------------------
r336 | schmitta | 2005-06-10 14:00:00 -0400 (Fri, 10 Jun 2005) | 3 lines

checkpoint for this lens


------------------------------------------------------------------------
r335 | jnfoster | 2005-06-10 12:07:01 -0400 (Fri, 10 Jun 2005) | 8 lines


write types like this (but wait for the new type normalizer, maybe):

	n? = X
	!? \ (x,y,z) = X
	+ \ (x,y,z) = X


------------------------------------------------------------------------
r334 | bohannon | 2005-06-10 11:53:05 -0400 (Fri, 10 Jun 2005) | 3 lines

Shuffled around some text.


------------------------------------------------------------------------
r333 | jnfoster | 2005-06-10 11:50:07 -0400 (Fri, 10 Jun 2005) | 1 line

+
------------------------------------------------------------------------
r332 | schmitta | 2005-06-10 11:39:30 -0400 (Fri, 10 Jun 2005) | 3 lines

oops, the test was failing


------------------------------------------------------------------------
r331 | schmitta | 2005-06-10 11:38:26 -0400 (Fri, 10 Jun 2005) | 4 lines

compiler: better error reporting when unit test fails
addresses: moving towards more abstract schema


------------------------------------------------------------------------
r330 | bohannon | 2005-06-10 11:26:06 -0400 (Fri, 10 Jun 2005) | 3 lines

Updated syntax for relational.src.  Some debugging code added.


------------------------------------------------------------------------
r329 | jnfoster | 2005-06-10 09:59:13 -0400 (Fri, 10 Jun 2005) | 1 line

cleaned up debugging a little bit. commented out debugging of every registry lookup (so -debug all is a little more managable). Hid [loading foo.fcl] stuff under a -verbose (or -v) flag
------------------------------------------------------------------------
r328 | jnfoster | 2005-06-10 09:40:51 -0400 (Fri, 10 Jun 2005) | 1 line

fixed the bugs holding up spinelist (type domain stuff was still commented out in type.ml) and ported relational.src to new syntax (whew)
------------------------------------------------------------------------
r327 | bcpierce | 2005-06-10 08:38:10 -0400 (Fri, 10 Jun 2005) | 1 line

Tiny tweaks
------------------------------------------------------------------------
r326 | jnfoster | 2005-06-10 08:23:48 -0400 (Fri, 10 Jun 2005) | 1 line

small tweaks to compiler, error printing
------------------------------------------------------------------------
r325 | bcpierce | 2005-06-10 07:13:03 -0400 (Fri, 10 Jun 2005) | 1 line

Small fixes to make ocamlc happy
------------------------------------------------------------------------
r324 | jnfoster | 2005-06-10 01:15:11 -0400 (Fri, 10 Jun 2005) | 17 lines

Here's the big commit. New parser, new compiler, ported lens library. 

Everything in examples works EXCEPT for the very last sync test in
structuredtext. An explode is failing because it is presented with a
view with multiple children.

Rather than record all the changes, I think it's just easiest for us
to sync up in the AM. I plan on begin at Penn no later than 10am.

Still to do:
      - improve checking of recursive variables
          * automatic delaying of lenses

      - uncomment/port type normalizer, test for non-termination in
        type operations


------------------------------------------------------------------------
r323 | schmitta | 2005-06-09 16:58:19 -0400 (Thu, 09 Jun 2005) | 3 lines

some small effort at getting an address demo


------------------------------------------------------------------------
r322 | bohannon | 2005-06-09 16:26:30 -0400 (Thu, 09 Jun 2005) | 3 lines

Updated project docs.


------------------------------------------------------------------------
r321 | bohannon | 2005-06-09 16:24:20 -0400 (Thu, 09 Jun 2005) | 4 lines

Some new documentation for Relational.project.  Implemented lexocographic
sorting for the output of relational operations.


------------------------------------------------------------------------
r320 | bohannon | 2005-06-09 15:39:12 -0400 (Thu, 09 Jun 2005) | 4 lines

Included a hack (list reversal) to get unit tests to work without normalizing
relational views.


------------------------------------------------------------------------
r319 | bohannon | 2005-06-09 15:38:17 -0400 (Thu, 09 Jun 2005) | 3 lines

Slightly modified tests.


------------------------------------------------------------------------
r318 | bcpierce | 2005-06-09 14:59:01 -0400 (Thu, 09 Jun 2005) | 1 line

Docs for relational lenses
------------------------------------------------------------------------
r317 | bohannon | 2005-06-09 13:00:17 -0400 (Thu, 09 Jun 2005) | 4 lines

Revised and complete unit tests for rename, union, unionl, unionr, inter,
interl, interr, diff, diffl, diffr.  Bugfixes.


------------------------------------------------------------------------
r316 | schmitta | 2005-06-09 12:53:18 -0400 (Thu, 09 Jun 2005) | 3 lines

fixed Xml.squash_elt


------------------------------------------------------------------------
r315 | schmitta | 2005-06-09 12:35:26 -0400 (Thu, 09 Jun 2005) | 3 lines

added "squash_elt" in xml


------------------------------------------------------------------------
r314 | bcpierce | 2005-06-09 12:23:27 -0400 (Thu, 09 Jun 2005) | 1 line

Improvements to xml.src
------------------------------------------------------------------------
r313 | bcpierce | 2005-06-09 09:17:21 -0400 (Thu, 09 Jun 2005) | 1 line

More progress on XML -- holding no locks for next few hours
------------------------------------------------------------------------
r312 | bcpierce | 2005-06-08 18:15:58 -0400 (Wed, 08 Jun 2005) | 1 line

Some cool XML processing lenses
------------------------------------------------------------------------
r311 | bohannon | 2005-06-08 16:23:03 -0400 (Wed, 08 Jun 2005) | 3 lines

Additional unit tests and bug fixes for relations.


------------------------------------------------------------------------
r310 | schmitta | 2005-06-08 15:22:32 -0400 (Wed, 08 Jun 2005) | 4 lines

the output of Meta may now be parsed directly
use Misc.whack_ident s to create a string that is quoted if need be


------------------------------------------------------------------------
r309 | bcpierce | 2005-06-08 14:21:12 -0400 (Wed, 08 Jun 2005) | 9 lines

Added switch for controlling whether we compile using ocamlc (fast
compiling) or ocamlopt (fast object code).

To compile fast, add a file .harmony.mk to your home directory
containing this line:

   COMPILEHARMONYASBYTECODE=yes


------------------------------------------------------------------------
r308 | schmitta | 2005-06-08 14:17:55 -0400 (Wed, 08 Jun 2005) | 3 lines

xml unit test and encoding working


------------------------------------------------------------------------
r307 | bcpierce | 2005-06-08 14:16:14 -0400 (Wed, 08 Jun 2005) | 1 line

Checkpoint
------------------------------------------------------------------------
r306 | schmitta | 2005-06-08 13:54:02 -0400 (Wed, 08 Jun 2005) | 3 lines

changing representation of XML slightly: the toplevel view is a list, not a "" -> view


------------------------------------------------------------------------
r305 | bcpierce | 2005-06-08 13:52:59 -0400 (Wed, 08 Jun 2005) | 1 line

Added xml.src
------------------------------------------------------------------------
r304 | bcpierce | 2005-06-08 13:46:57 -0400 (Wed, 08 Jun 2005) | 1 line

Checkpoint for Alan
------------------------------------------------------------------------
r303 | bcpierce | 2005-06-08 13:27:03 -0400 (Wed, 08 Jun 2005) | 1 line

Checkpoint for Nate
------------------------------------------------------------------------
r302 | bohannon | 2005-06-08 09:46:42 -0400 (Wed, 08 Jun 2005) | 3 lines

Implemented a valid comparison function on views.


------------------------------------------------------------------------
r301 | bcpierce | 2005-06-07 23:03:05 -0400 (Tue, 07 Jun 2005) | 1 line

More progress on prelude.src docs
------------------------------------------------------------------------
r300 | bcpierce | 2005-06-07 19:36:11 -0400 (Tue, 07 Jun 2005) | 1 line

Whew^2
------------------------------------------------------------------------
r299 | bcpierce | 2005-06-07 19:33:34 -0400 (Tue, 07 Jun 2005) | 1 line

Whew
------------------------------------------------------------------------
r298 | schmitta | 2005-06-07 19:20:05 -0400 (Tue, 07 Jun 2005) | 3 lines

some cleanup


------------------------------------------------------------------------
r297 | bohannon | 2005-06-07 18:58:26 -0400 (Tue, 07 Jun 2005) | 3 lines

Relational lenses now return more useful error messages.


------------------------------------------------------------------------
r296 | bcpierce | 2005-06-07 18:53:55 -0400 (Tue, 07 Jun 2005) | 1 line

Checkpoint
------------------------------------------------------------------------
r295 | bcpierce | 2005-06-07 18:49:51 -0400 (Tue, 07 Jun 2005) | 1 line

/lenses working
------------------------------------------------------------------------
r294 | bcpierce | 2005-06-07 18:46:55 -0400 (Tue, 07 Jun 2005) | 1 line

DO NOT UPDATE -- broken version
------------------------------------------------------------------------
r293 | bcpierce | 2005-06-07 18:38:07 -0400 (Tue, 07 Jun 2005) | 1 line

added section header to spinelist
------------------------------------------------------------------------
r292 | bcpierce | 2005-06-07 18:35:04 -0400 (Tue, 07 Jun 2005) | 1 line

fixed spinelist
------------------------------------------------------------------------
r291 | bcpierce | 2005-06-07 18:34:13 -0400 (Tue, 07 Jun 2005) | 1 line

changed lists.fcl and relational.fcl to .src
------------------------------------------------------------------------
r290 | bohannon | 2005-06-07 18:30:30 -0400 (Tue, 07 Jun 2005) | 3 lines

Update error handling for relations.


------------------------------------------------------------------------
r289 | bcpierce | 2005-06-07 18:28:58 -0400 (Tue, 07 Jun 2005) | 1 line

Checkpoint
------------------------------------------------------------------------
r288 | schmitta | 2005-06-07 18:02:21 -0400 (Tue, 07 Jun 2005) | 3 lines

added a `Space formatting thingie


------------------------------------------------------------------------
r287 | jnfoster | 2005-06-07 16:55:34 -0400 (Tue, 07 Jun 2005) | 1 line

making the parser a little more rigid :)
------------------------------------------------------------------------
r286 | jnfoster | 2005-06-07 16:42:17 -0400 (Tue, 07 Jun 2005) | 1 line

make the parser a tiny bit more flexible
------------------------------------------------------------------------
r285 | bohannon | 2005-06-07 16:30:27 -0400 (Tue, 07 Jun 2005) | 3 lines

Corrected error messeages for unit tests.


------------------------------------------------------------------------
r284 | jnfoster | 2005-06-07 16:17:42 -0400 (Tue, 07 Jun 2005) | 1 line

reorganized testing stuff
------------------------------------------------------------------------
r283 | bcpierce | 2005-06-07 16:16:21 -0400 (Tue, 07 Jun 2005) | 1 line

Checkpoint
------------------------------------------------------------------------
r282 | schmitta | 2005-06-07 16:03:53 -0400 (Tue, 07 Jun 2005) | 6 lines

syntax for multiline string:
foo = "hello
      |world"
will be lexed as "hello\nworld"


------------------------------------------------------------------------
r281 | bcpierce | 2005-06-07 15:18:34 -0400 (Tue, 07 Jun 2005) | 1 line

checkpoint
------------------------------------------------------------------------
r280 | bcpierce | 2005-06-07 12:23:06 -0400 (Tue, 07 Jun 2005) | 1 line

Checkpoint
------------------------------------------------------------------------
r279 | jnfoster | 2005-06-07 12:03:57 -0400 (Tue, 07 Jun 2005) | 1 line

syncing with BCP
------------------------------------------------------------------------
r278 | bcpierce | 2005-06-07 11:48:05 -0400 (Tue, 07 Jun 2005) | 1 line

Another checkpoint for Nate
------------------------------------------------------------------------
r277 | schmitta | 2005-06-07 11:28:24 -0400 (Tue, 07 Jun 2005) | 3 lines

oops


------------------------------------------------------------------------
r276 | bcpierce | 2005-06-07 11:26:03 -0400 (Tue, 07 Jun 2005) | 1 line

Checkpoint for Nate
------------------------------------------------------------------------
r275 | jnfoster | 2005-06-07 11:02:11 -0400 (Tue, 07 Jun 2005) | 5 lines

Two bits of housecleaning:
    * Merged Info.ts for stuffing into syntax trees in the parser.
    * Eliminated Info.t from sort data type (it was just clutter)


------------------------------------------------------------------------
r274 | jnfoster | 2005-06-06 18:40:47 -0400 (Mon, 06 Jun 2005) | 1 line

oops. my change to the lexer vanished. here it is
------------------------------------------------------------------------
r273 | jnfoster | 2005-06-06 18:38:01 -0400 (Mon, 06 Jun 2005) | 1 line

fixed off-by-1 bug in Lexer Info.t extraction. Merge info in unit test parsing
------------------------------------------------------------------------
r272 | bcpierce | 2005-06-06 17:38:21 -0400 (Mon, 06 Jun 2005) | 1 line

Makefile tweaks
------------------------------------------------------------------------
r271 | bcpierce | 2005-06-06 17:37:35 -0400 (Mon, 06 Jun 2005) | 1 line

TODO notes
------------------------------------------------------------------------
r270 | schmitta | 2005-06-06 17:06:07 -0400 (Mon, 06 Jun 2005) | 3 lines

moved fold_right to list module


------------------------------------------------------------------------
r269 | schmitta | 2005-06-06 16:59:49 -0400 (Mon, 06 Jun 2005) | 5 lines

silent coercion of variables standing for views to singleton type when a type is needed.

structuredtext now defines and use fold_right


------------------------------------------------------------------------
r268 | bohannon | 2005-06-06 14:45:25 -0400 (Mon, 06 Jun 2005) | 5 lines

Finished implementing all basic relational lenses: rename, union, inter, diff,
select, select_eq, project, join, along with the asymmetric versions of union,
inter, diff, and join.  Unit tests and bugfixes will follow.


------------------------------------------------------------------------
r267 | bohannon | 2005-06-06 12:05:37 -0400 (Mon, 06 Jun 2005) | 3 lines

Added the interface file for Treedb.


------------------------------------------------------------------------
r266 | bohannon | 2005-06-06 12:04:28 -0400 (Mon, 06 Jun 2005) | 3 lines

Added documentation for Dblens.


------------------------------------------------------------------------
r265 | bohannon | 2005-06-06 11:15:53 -0400 (Mon, 06 Jun 2005) | 4 lines

Moved the Treedb module to the src directory, which is a more appropriate
place.  Added an interface file for it.


------------------------------------------------------------------------
r264 | bohannon | 2005-06-06 10:54:20 -0400 (Mon, 06 Jun 2005) | 8 lines

Fixed relations to properly deal with missing concrete views in the put
direction.

Fixed the behavior of join to match the spec.

Added documentation to the Relation module.


------------------------------------------------------------------------
r263 | jnfoster | 2005-06-06 02:45:05 -0400 (Mon, 06 Jun 2005) | 12 lines

Got rid of V.Error (V.Illformed still needs to be zapped; we need to
think about Errors carefully. I want to talk about it).

Added native lenses
      read : name -> name
      load : name -> name -> view
      load_file : name -> view

load_view can almost be defined using load and read, but we need to
get the encoding!


------------------------------------------------------------------------
r262 | jnfoster | 2005-06-05 23:57:43 -0400 (Sun, 05 Jun 2005) | 1 line

correcting keywords
------------------------------------------------------------------------
r261 | jnfoster | 2005-06-05 23:56:14 -0400 (Sun, 05 Jun 2005) | 1 line

more tidying
------------------------------------------------------------------------
r260 | jnfoster | 2005-06-05 23:49:27 -0400 (Sun, 05 Jun 2005) | 1 line

some tidying
------------------------------------------------------------------------
r259 | bcpierce | 2005-06-05 18:24:54 -0400 (Sun, 05 Jun 2005) | 1 line

Sync unit tests seem to work (see experimental/structuredtext/structuredtext.fcl)
------------------------------------------------------------------------
r258 | schmitta | 2005-06-05 18:15:36 -0400 (Sun, 05 Jun 2005) | 3 lines

back to compiling state


------------------------------------------------------------------------
r257 | bcpierce | 2005-06-05 18:02:45 -0400 (Sun, 05 Jun 2005) | 1 line

Checkpoint
------------------------------------------------------------------------
r256 | schmitta | 2005-06-05 17:50:21 -0400 (Sun, 05 Jun 2005) | 3 lines

first try at sync unit test


------------------------------------------------------------------------
r255 | bcpierce | 2005-06-05 17:46:50 -0400 (Sun, 05 Jun 2005) | 1 line

Tweak
------------------------------------------------------------------------
r254 | bcpierce | 2005-06-05 17:30:30 -0400 (Sun, 05 Jun 2005) | 1 line

More structured text; syntax (no content yet) for unit testing of sync
------------------------------------------------------------------------
r253 | bcpierce | 2005-06-05 16:05:50 -0400 (Sun, 05 Jun 2005) | 1 line

Oops -- here come the files...
------------------------------------------------------------------------
r252 | bcpierce | 2005-06-05 15:43:25 -0400 (Sun, 05 Jun 2005) | 1 line

First working version of structured text lens (hooray)
------------------------------------------------------------------------
r251 | schmitta | 2005-06-05 14:23:52 -0400 (Sun, 05 Jun 2005) | 3 lines

fixed bijection testing


------------------------------------------------------------------------
r250 | schmitta | 2005-06-05 14:17:31 -0400 (Sun, 05 Jun 2005) | 3 lines

now using the "put a buffer around format" trick


------------------------------------------------------------------------
r249 | schmitta | 2005-06-05 13:46:03 -0400 (Sun, 05 Jun 2005) | 3 lines

added V.format_msg_as_string


------------------------------------------------------------------------
r248 | schmitta | 2005-06-05 12:49:03 -0400 (Sun, 05 Jun 2005) | 8 lines

we can now do bijective tests:
test foo / \ concrete = abstract
will result in
test foo / concrete = abstract
and
test foo \ abstract missing = concrete


------------------------------------------------------------------------
r247 | bcpierce | 2005-06-04 14:43:27 -0400 (Sat, 04 Jun 2005) | 1 line

Checkpoint and release locks for the rest of today.  I've pretty much finished a first pass over the .mli files and ocamldocs.  Also decided to consolidate the two viewers into a single file, since it sounded like we were going to write some more viewers and I didn't want to get a proliferation of itty bitty modules...
------------------------------------------------------------------------
r246 | bcpierce | 2005-06-04 14:08:36 -0400 (Sat, 04 Jun 2005) | 1 line

Checkpoint -- more progress on .mli docs
------------------------------------------------------------------------
r245 | bcpierce | 2005-06-04 13:17:28 -0400 (Sat, 04 Jun 2005) | 1 line

Forgot to add these...
------------------------------------------------------------------------
r244 | schmitta | 2005-06-04 12:43:22 -0400 (Sat, 04 Jun 2005) | 3 lines

breaking for lunch


------------------------------------------------------------------------
r243 | jnfoster | 2005-06-04 12:15:53 -0400 (Sat, 04 Jun 2005) | 9 lines

D'oh. Of course, we need a stack of lexing info. 

This is the quick hack to get it working (Stack of string ref * int
ref * int ref ). But we don't need those ref cells. Should also
check/catch Empty exceptions.

-nate


------------------------------------------------------------------------
r242 | schmitta | 2005-06-04 12:13:16 -0400 (Sat, 04 Jun 2005) | 3 lines

slowly getting the hang of the way the documentation feature works


------------------------------------------------------------------------
r241 | bcpierce | 2005-06-04 11:53:38 -0400 (Sat, 04 Jun 2005) | 1 line

Checkpoint -- going out for an hour
------------------------------------------------------------------------
r240 | bcpierce | 2005-06-04 11:14:48 -0400 (Sat, 04 Jun 2005) | 1 line

Checkpoint, just in case
------------------------------------------------------------------------
r239 | jnfoster | 2005-06-04 11:13:04 -0400 (Sat, 04 Jun 2005) | 24 lines

Fixed missing file info bug. 

Lexing info including line and column number and file name are stored in ref cells.
They are set when you call

    Lexer.setup fn

and reset to their previous values when you call

    Lexer.finish ()

We need to do this caching / restoring of the old lexing info because
compiling one Focal module may trigger compilation of another module
at any point (when the Registry comes across a QID it hasn't loaded
yet and goes off to load that module).

There were two bugs:

      (1) I was calling finish () too early, which restored the "" filename
      (2) I wasn't caching line and column numbers. 

Fixed now.


------------------------------------------------------------------------
r238 | jnfoster | 2005-06-04 10:59:44 -0400 (Sat, 04 Jun 2005) | 1 line

fixed Native.id bug
------------------------------------------------------------------------
r237 | bcpierce | 2005-06-04 10:58:56 -0400 (Sat, 04 Jun 2005) | 1 line

Oops -- fix syntax error
------------------------------------------------------------------------
r236 | bcpierce | 2005-06-04 10:58:16 -0400 (Sat, 04 Jun 2005) | 1 line

Some tidying of documentation in /src (Nate: Am I in danger of stepping on things that you are doing if I continue with this?)
------------------------------------------------------------------------
r235 | bcpierce | 2005-06-04 09:50:06 -0400 (Sat, 04 Jun 2005) | 1 line

This should be better.  N.b.: the gnu make PWD variable does not do quite what you might think.
------------------------------------------------------------------------
r234 | bcpierce | 2005-06-04 09:40:17 -0400 (Sat, 04 Jun 2005) | 1 line

Fixed some other problems that I noticed while discussing these... :-)  Also, made generated .fcl files read-only, to avoid problems in future...
------------------------------------------------------------------------
r233 | jnfoster | 2005-06-04 09:33:10 -0400 (Sat, 04 Jun 2005) | 1 line

two more bug fixes
------------------------------------------------------------------------
r232 | jnfoster | 2005-06-04 09:29:37 -0400 (Sat, 04 Jun 2005) | 1 line

fixed Native.xxxx bug
------------------------------------------------------------------------
r231 | bcpierce | 2005-06-04 09:20:28 -0400 (Sat, 04 Jun 2005) | 1 line

Tiny bit more makefile tweaking
------------------------------------------------------------------------
r230 | jnfoster | 2005-06-04 03:03:05 -0400 (Sat, 04 Jun 2005) | 1 line

last few interfaces. To avoid cutting and pasting, I had to give Mapplus.Map's signature a name. Is there a more elegant way? All these still need ocamldoc'ing
------------------------------------------------------------------------
r229 | jnfoster | 2005-06-03 17:19:45 -0400 (Fri, 03 Jun 2005) | 1 line

oops. good to include the file
------------------------------------------------------------------------
r228 | jnfoster | 2005-06-03 17:18:46 -0400 (Fri, 03 Jun 2005) | 1 line

added interface for syntax.ml
------------------------------------------------------------------------
r227 | schmitta | 2005-06-03 17:17:30 -0400 (Fri, 03 Jun 2005) | 4 lines

we now generate doc
converted octal chars to decimal in tools


------------------------------------------------------------------------
r226 | bohannon | 2005-06-03 16:51:52 -0400 (Fri, 03 Jun 2005) | 3 lines

Moved relation module to src.


------------------------------------------------------------------------
r225 | jnfoster | 2005-06-03 16:48:56 -0400 (Fri, 03 Jun 2005) | 1 line

cleaned up naming of native lenses in Prelude
------------------------------------------------------------------------
r224 | bcpierce | 2005-06-03 16:48:22 -0400 (Fri, 03 Jun 2005) | 1 line

More tidying of makefiles
------------------------------------------------------------------------
r223 | schmitta | 2005-06-03 16:43:36 -0400 (Fri, 03 Jun 2005) | 3 lines

fixing "illegal escape" issues in tools


------------------------------------------------------------------------
r222 | schmitta | 2005-06-03 16:38:32 -0400 (Fri, 03 Jun 2005) | 3 lines

some makefile tweaks


------------------------------------------------------------------------
r221 | jnfoster | 2005-06-03 16:34:59 -0400 (Fri, 03 Jun 2005) | 1 line

Forgot that Meta moved
------------------------------------------------------------------------
r220 | jnfoster | 2005-06-03 16:30:47 -0400 (Fri, 03 Jun 2005) | 1 line

Nate's reorganization
------------------------------------------------------------------------
r219 | bcpierce | 2005-06-03 16:12:44 -0400 (Fri, 03 Jun 2005) | 1 line

Finishing reorg
------------------------------------------------------------------------
r218 | bcpierce | 2005-06-03 15:43:51 -0400 (Fri, 03 Jun 2005) | 1 line

Checkpoint
------------------------------------------------------------------------
r217 | bcpierce | 2005-06-03 15:38:23 -0400 (Fri, 03 Jun 2005) | 1 line

Moved examples/demo to experimental
------------------------------------------------------------------------
r216 | bcpierce | 2005-06-03 15:38:16 -0400 (Fri, 03 Jun 2005) | 1 line

Moved examples/demo to experimental
------------------------------------------------------------------------
r215 | schmitta | 2005-06-03 15:35:18 -0400 (Fri, 03 Jun 2005) | 3 lines

make ocamldoc happy


------------------------------------------------------------------------
r214 | bcpierce | 2005-06-03 15:29:40 -0400 (Fri, 03 Jun 2005) | 1 line

Reorganized directories
------------------------------------------------------------------------
r213 | bohannon | 2005-06-03 15:00:43 -0400 (Fri, 03 Jun 2005) | 3 lines

Made the type Lens.t polymorphic in the concrete and abstract domains.


------------------------------------------------------------------------
r212 | bcpierce | 2005-06-03 14:59:42 -0400 (Fri, 03 Jun 2005) | 1 line

Misc updates before moving things around...
------------------------------------------------------------------------
r211 | lescuyer | 2005-06-03 10:34:07 -0400 (Fri, 03 Jun 2005) | 1 line

Some more comments in src. Finished with existing *.mli, I am going to add mli interfaces for single ml files.
------------------------------------------------------------------------
r210 | jnfoster | 2005-06-02 23:39:43 -0400 (Thu, 02 Jun 2005) | 1 line

cleaned up misc.ml some more; removed Bad and replaced with Error.Fatal_error everywhere
------------------------------------------------------------------------
r209 | bcpierce | 2005-06-02 22:11:45 -0400 (Thu, 02 Jun 2005) | 1 line

TODO note
------------------------------------------------------------------------
r208 | bohannon | 2005-06-02 18:43:25 -0400 (Thu, 02 Jun 2005) | 3 lines

Added some asymmetric relational lenses.


------------------------------------------------------------------------
r207 | lescuyer | 2005-06-02 18:04:14 -0400 (Thu, 02 Jun 2005) | 1 line

Finished (pfew) doc for misc.mli. ocamldoc reports irrelevant syntax errors in misc.mli (line 0, there is no line 0...) and yet makes the job well. If someone sees why...
------------------------------------------------------------------------
r206 | jnfoster | 2005-06-02 17:33:29 -0400 (Thu, 02 Jun 2005) | 4 lines

A small change: Makefile hackery to inform ocamlfind to use the .opt versions of the standard tools. 
My running time for (cd src; make clean; make) went down from ~121 to 26 seconds :-)


------------------------------------------------------------------------
r205 | jnfoster | 2005-06-02 16:35:11 -0400 (Thu, 02 Jun 2005) | 1 line

bug fix in visual
------------------------------------------------------------------------
r204 | lescuyer | 2005-06-02 16:30:54 -0400 (Thu, 02 Jun 2005) | 1 line

test 3
------------------------------------------------------------------------
r203 | jnfoster | 2005-06-02 16:25:29 -0400 (Thu, 02 Jun 2005) | 1 line

testing commit emails
------------------------------------------------------------------------
r202 | lescuyer | 2005-06-02 16:22:15 -0400 (Thu, 02 Jun 2005) | 1 line

test 2
------------------------------------------------------------------------
r201 | jnfoster | 2005-06-02 16:17:24 -0400 (Thu, 02 Jun 2005) | 1 line

testing commit messages through halfdome with improved commit-email.pl script (looks for svnlook in several places)
------------------------------------------------------------------------
r200 | lescuyer | 2005-06-02 16:06:41 -0400 (Thu, 02 Jun 2005) | 1 line

test
------------------------------------------------------------------------
r199 | bcpierce | 2005-06-02 16:01:56 -0400 (Thu, 02 Jun 2005) | 1 line

Test subversion commit messages (take 3)
------------------------------------------------------------------------
r198 | bcpierce | 2005-06-02 15:59:13 -0400 (Thu, 02 Jun 2005) | 1 line

Test subversion commit messages
------------------------------------------------------------------------
r197 | jnfoster | 2005-06-02 15:50:23 -0400 (Thu, 02 Jun 2005) | 1 line

testing again
------------------------------------------------------------------------
r196 | jnfoster | 2005-06-02 15:44:52 -0400 (Thu, 02 Jun 2005) | 1 line

testing commit messages
------------------------------------------------------------------------
r195 | schmitta | 2005-06-02 15:40:43 -0400 (Thu, 02 Jun 2005) | 3 lines

Removed the "well formed" information from the view, as it's not used


------------------------------------------------------------------------
r194 | bohannon | 2005-06-02 15:35:09 -0400 (Thu, 02 Jun 2005) | 3 lines

Migrated relational code.


------------------------------------------------------------------------
r193 | jnfoster | 2005-06-02 15:27:08 -0400 (Thu, 02 Jun 2005) | 1 line

two bug fixes
------------------------------------------------------------------------
r192 | jnfoster | 2005-06-02 15:12:20 -0400 (Thu, 02 Jun 2005) | 10 lines

Moved utility functions from harmony.ml to surveyor.ml

Cleaned up front end

Added a real "load" function that takes a name (like "abc.tmp:meta"),
locates it in the search path and uses the surveyor to select a
viewer. It fails (1) if it can't find a unique viewer or (2) no file
is found in the path, or (3) it can't be parsed using the viewer.


------------------------------------------------------------------------
r191 | bcpierce | 2005-06-02 14:06:22 -0400 (Thu, 02 Jun 2005) | 1 line

TODO notes
------------------------------------------------------------------------
r190 | lescuyer | 2005-06-02 11:46:46 -0400 (Thu, 02 Jun 2005) | 1 line

tutudutudu
------------------------------------------------------------------------
r189 | lescuyer | 2005-06-02 11:02:47 -0400 (Thu, 02 Jun 2005) | 1 line

youpi
------------------------------------------------------------------------
r188 | lescuyer | 2005-06-02 10:57:16 -0400 (Thu, 02 Jun 2005) | 1 line

__
------------------------------------------------------------------------
r187 | lescuyer | 2005-06-02 10:50:14 -0400 (Thu, 02 Jun 2005) | 1 line

Commit test, cause my messages don't go through anymore :(
------------------------------------------------------------------------
r186 | lescuyer | 2005-06-02 10:48:32 -0400 (Thu, 02 Jun 2005) | 1 line

Fixed lenses in bookmarks for the new syntax
------------------------------------------------------------------------
r185 | jnfoster | 2005-06-02 09:40:55 -0400 (Thu, 02 Jun 2005) | 8 lines

rolled concat back to concat2 (Thang, your version is commented out in
list.fcl. I didn't know where you'd like to keep it). 

Added unit tests for concat2. Its type would require (if we calculated
it) that the concrete view be a list of lists of length exactly 2 and
that the abstract view be a list with exactly one element {" "}.


------------------------------------------------------------------------
r184 | jnfoster | 2005-06-02 03:38:13 -0400 (Thu, 02 Jun 2005) | 31 lines

A whole bunch of bug fixes and improvements. 

  * Added all the old unit tests in to prelude.fcl and list.fcl

  * Can check (force loading of) a module like this:
    	harmony check -m Prelude
  * Can run unit tests (without forcing load) for modules using these flags
        -test Prelude   run tests in Prelude, if it gets loaded
	-test-all       run tests in any module that gets loaded

	e.g. running
	     harmony check -m List -test-all
	ends up running all the unit tests in Prelude and List; running
	     harmony check -m List -test Prelude
	would only run the unit tests from Prelude
	
	if M has nested modules M.O and M.P
		-test M runs tests in M, M.O, M.P
		-test M.O runs tests in M.O only

    Thoughts on this interface?

  * Added load_meta : name -> view
     - but I want to rip it out first thing in the morning with a more
       general "load" that uses the surveyor to find an appropriate
       viewer. Too tired to do it just now. So don't start using this
       yet :)

  * Thang, I left a few queries for you in list.fcl's concat lens. 


------------------------------------------------------------------------
r183 | lescuyer | 2005-06-01 18:55:31 -0400 (Wed, 01 Jun 2005) | 1 line

Test
------------------------------------------------------------------------
r182 | jnfoster | 2005-06-01 18:46:55 -0400 (Wed, 01 Jun 2005) | 1 line

zapped concat
------------------------------------------------------------------------
r181 | lescuyer | 2005-06-01 18:36:52 -0400 (Wed, 01 Jun 2005) | 2 lines

Dummy commit, for a commit message test

------------------------------------------------------------------------
r180 | jnfoster | 2005-06-01 18:33:44 -0400 (Wed, 01 Jun 2005) | 8 lines


big commit

will break everyone's code

made the switch from names as strings to names as idents, added antiquotation, and unit tests

added Thang's concat to list.fcl
------------------------------------------------------------------------
r179 | lescuyer | 2005-06-01 18:18:37 -0400 (Wed, 01 Jun 2005) | 1 line

a dummy commit, just to see if my commit messages will go thru now
------------------------------------------------------------------------
r178 | lescuyer | 2005-06-01 17:51:28 -0400 (Wed, 01 Jun 2005) | 1 line

Added the mli files to the SOURCES in the Makefile, so that typing 'make doc' calls ocamldoc with all these mli files (btw, the result is amazing, assuming how easy it is to write the comments in ocamldoc 'syntax', ocamldoc is really great \!)
------------------------------------------------------------------------
r177 | lescuyer | 2005-06-01 17:37:49 -0400 (Wed, 01 Jun 2005) | 1 line

Commiting comments in several .mli files (not finished yet)
------------------------------------------------------------------------
r176 | lescuyer | 2005-06-01 11:21:26 -0400 (Wed, 01 Jun 2005) | 1 line

Gone through v.mli and grep'ed all values. Commented code and definitions for unused values, and marked definitions of values that do not appear outside of v.ml
------------------------------------------------------------------------
r175 | jnfoster | 2005-05-31 19:07:27 -0400 (Tue, 31 May 2005) | 1 line

testing mail hooks
------------------------------------------------------------------------
r174 | lescuyer | 2005-05-31 18:50:02 -0400 (Tue, 31 May 2005) | 1 line

Commented everything in v.mli for ocamldoc purpose. Proof-reading required :)
------------------------------------------------------------------------
r173 | kimthang | 2005-05-27 16:24:17 -0400 (Fri, 27 May 2005) | 1 line

I think concat works well right now. If you want the write the empty list, write it like []. I can't explain myself why we have to write it in that way
------------------------------------------------------------------------
r172 | jnfoster | 2005-05-27 09:48:32 -0400 (Fri, 27 May 2005) | 1 line

fixed vcard.fcl to work with new location of List types
------------------------------------------------------------------------
r171 | jnfoster | 2005-05-27 08:11:28 -0400 (Fri, 27 May 2005) | 1 line

checkpoint of demo; split into 3 distinct phases
------------------------------------------------------------------------
r170 | jnfoster | 2005-05-26 18:14:48 -0400 (Thu, 26 May 2005) | 8 lines

Checkpoint for tomorrow's bookmark demo. It may not work on your
machines, but it works on ours :-)

*!*!* ONE MAJOR CHANGE *!*!* 
      List.List, List.Cons, and List.Nil now live in Prelude. 
      (Stephane convinced me)


------------------------------------------------------------------------
r169 | kimthang | 2005-05-26 14:48:19 -0400 (Thu, 26 May 2005) | 1 line

make the lexer a little more general
------------------------------------------------------------------------
r168 | kimthang | 2005-05-26 14:03:29 -0400 (Thu, 26 May 2005) | 1 line

Some changes of lens
------------------------------------------------------------------------
r167 | kimthang | 2005-05-26 11:53:04 -0400 (Thu, 26 May 2005) | 1 line

Done for format vcf of vcard
------------------------------------------------------------------------
r166 | jnfoster | 2005-05-26 11:09:57 -0400 (Thu, 26 May 2005) | 1 line

checkpoint; some fixes to visual and compiler stuff; added a pad lens to extend lists to length power of two
------------------------------------------------------------------------
r165 | lescuyer | 2005-05-26 11:03:13 -0400 (Thu, 26 May 2005) | 1 line

Added some comments, and a 'all' target to the Makefile, for sync AND test (31sec \!)
------------------------------------------------------------------------
r164 | lescuyer | 2005-05-26 10:49:29 -0400 (Thu, 26 May 2005) | 1 line

Some code tidying in IE and Safari lenses, plus comments here n there
------------------------------------------------------------------------
r163 | jnfoster | 2005-05-26 01:32:34 -0400 (Thu, 26 May 2005) | 12 lines

OK. Visualizer is usable now. 

Error display is not very fancy. I had to turn off the tagging
(yellow syntax highlighting) that Malo added because the arithmetic is
wrong and it's causing GTK assertions to blow up at the C level. 

But basic usage is OK. 

Don't forget to pass a -I flag on the commandline, or you won't have
your Prelude :-)


------------------------------------------------------------------------
r162 | jnfoster | 2005-05-26 00:41:33 -0400 (Thu, 26 May 2005) | 16 lines

Standardized error messages throughout compiler. The backend raises these errors

    Syntax_error of i * string * string 
    Sort_error of i * string * string
    Run_error of i * string * string 
    Fatal_error of string 

and UIs handle appropriately. E.g., 

    - the text UI prints out the filename, info, and message in a format
      that Emacs compile-mode understands

    - the GTK UI highlights various text GUI components and prints the
      error message in the scroller on the bottom pane. 


------------------------------------------------------------------------
r161 | jnfoster | 2005-05-25 17:12:59 -0400 (Wed, 25 May 2005) | 1 line

checkpointing a few small fixes
------------------------------------------------------------------------
r160 | lescuyer | 2005-05-25 16:17:57 -0400 (Wed, 25 May 2005) | 1 line

Forgot the ml program... type make testie if you only want to test IE6, and nothing or test for all 3 browsers
------------------------------------------------------------------------
r159 | lescuyer | 2005-05-25 16:14:27 -0400 (Wed, 25 May 2005) | 1 line

Added all the stuff for IE6 : Focal lenses and schema, utility for converting IE folders to and fro views, a sample directory for Favorites (it is ugly because specific svn files in this folder are also parsed and included in the IE6 concrete view, but it's necessary since most of you do not have a Windows installed, and on the other hand the process still works well with the svn files around, which is good :) ), and the corresponding test in the makefile. BTW this means we have a end to end sync between safari and IE (unfortunately, not very convenient to test because none of us have both of these browsers on their laptop :)) , as for mozilla, I still have to figure out what is wrong for firefox with the well formed html delivered by tidy
------------------------------------------------------------------------
r158 | kimthang | 2005-05-25 15:20:17 -0400 (Wed, 25 May 2005) | 1 line

Done lexers .vcf => .xml and .xml => .vcf 
------------------------------------------------------------------------
r157 | jnfoster | 2005-05-25 15:00:17 -0400 (Wed, 25 May 2005) | 6 lines

Checkpoitn of the "it compiles" resurrected GTK visualizer. 

Not quite usable yet---needs different error handlings from Compiler and viewers. 



------------------------------------------------------------------------
r156 | lescuyer | 2005-05-25 10:52:41 -0400 (Wed, 25 May 2005) | 1 line

Demo growing bigger : now includes IExplorer and 5 synchronizations to synchronize all the samples (takes about 12s)
------------------------------------------------------------------------
r155 | kimthang | 2005-05-25 10:02:15 -0400 (Wed, 25 May 2005) | 1 line

Make the lexer more compact and more elegant the format .xml (from format .vcf) 
------------------------------------------------------------------------
r154 | lescuyer | 2005-05-24 19:02:00 -0400 (Tue, 24 May 2005) | 1 line

A working tiny demo \! Completed all schemas, including the flattening to a bush, and thanks to Nate's improvements of the speed, we can synchronize two big heterogeneous files from scratch (type make sync, results are mixed*.xml). For those interested, the synchronization schema is Schemas.BushAbstract and is easy to read ; in particular, a bookmark view always contains links and folders children, eventually empty, and link's names are stored under the corresponding urls, whereas folder contents are stored in a list under the folder name.
------------------------------------------------------------------------
r153 | jnfoster | 2005-05-24 18:28:29 -0400 (Tue, 24 May 2005) | 7 lines

weakened the sort of projectability we require for concatenations in member; we only need consistency across the atoms in concatenations, not the full-blown notion stated in teh sync paper. 
Later on, we could write a more efficient member test for consistent types (no backtracking for unions) by projecting out each name in the tree and quotienting the rest of the type... but that's for later.

Btw, I needed this type to write the following type for fast_reverse
type PaddedList = [{pad_name=Any}] | { HD = Any, TL = PaddedList }


------------------------------------------------------------------------
r152 | jnfoster | 2005-05-24 17:28:03 -0400 (Tue, 24 May 2005) | 1 line

fixed bug in flatten
------------------------------------------------------------------------
r151 | lescuyer | 2005-05-24 16:52:00 -0400 (Tue, 24 May 2005) | 1 line

make sync now takes two real bookmark files
------------------------------------------------------------------------
r150 | jnfoster | 2005-05-24 16:49:26 -0400 (Tue, 24 May 2005) | 1 line

bug fixes
------------------------------------------------------------------------
r149 | jnfoster | 2005-05-24 16:13:34 -0400 (Tue, 24 May 2005) | 1 line

faster compiler, fast types (?)
------------------------------------------------------------------------
r148 | bcpierce | 2005-05-24 16:05:00 -0400 (Tue, 24 May 2005) | 1 line

Started on ocamldocification of sources; some tidying of v.ml
------------------------------------------------------------------------
r147 | lescuyer | 2005-05-24 15:43:11 -0400 (Tue, 24 May 2005) | 1 line

Safari lens fixed for correct proxy's handling, added real safari bookmark file (should only be used with Nate's enhancements to the compiler \!)
------------------------------------------------------------------------
r146 | jnfoster | 2005-05-24 13:46:45 -0400 (Tue, 24 May 2005) | 1 line

removed quick debugging hack
------------------------------------------------------------------------
r145 | jnfoster | 2005-05-24 13:44:45 -0400 (Tue, 24 May 2005) | 1 line

fixed PXP bug; we now parse documents using utf8; some cleanups to Common.Makefile
------------------------------------------------------------------------
r144 | jnfoster | 2005-05-23 17:33:24 -0400 (Mon, 23 May 2005) | 1 line

removed extra printout
------------------------------------------------------------------------
r143 | kimthang | 2005-05-23 17:21:29 -0400 (Mon, 23 May 2005) | 1 line

A small lexer for vcard in format vcf
------------------------------------------------------------------------
r142 | lescuyer | 2005-05-23 17:00:16 -0400 (Mon, 23 May 2005) | 1 line

Fixed sync in harmony.ml, fixed plist.fcl and safari.fcl when concrete is missing, 'make sync' in bookmarks/ now does two synchronizations (actually, two 'conversions')
------------------------------------------------------------------------
r141 | lescuyer | 2005-05-20 16:35:41 -0400 (Fri, 20 May 2005) | 1 line

tiny cleaning of comments
------------------------------------------------------------------------
r140 | lescuyer | 2005-05-20 16:30:26 -0400 (Fri, 20 May 2005) | 1 line

Added sync in harmony.ml, see harmony.ml or experimental/bookmarks/Makefile for the syntax of the command
------------------------------------------------------------------------
r139 | lescuyer | 2005-05-19 17:51:09 -0400 (Thu, 19 May 2005) | 1 line

Finished Safari lenses, works well for now, but still has to be stressed on big files. 'make' now tests get & put of safari lens for heterogeneous sync, btw put is slightly slow (the sample is one link in one folder...)
------------------------------------------------------------------------
r138 | lescuyer | 2005-05-19 17:26:28 -0400 (Thu, 19 May 2005) | 1 line

Restarted safari.fcl from scratch, lens for homogeneous sync is working, the concrete.xml sample is a little biger than before (added root & folder info, as in a raw safari bookmarks file)
------------------------------------------------------------------------
r137 | lescuyer | 2005-05-18 18:10:48 -0400 (Wed, 18 May 2005) | 1 line

added some stuff for schemas, and asserts in the lenses
------------------------------------------------------------------------
r136 | lescuyer | 2005-05-18 16:53:49 -0400 (Wed, 18 May 2005) | 1 line

Mozilla lens working both ways, added schemas.fcl for the def of the common abstract schema, still waiting for more speed before including the tests in the makefile. BTW, the make test now does a getput and untidies the xml file to an html file taht one can directly import in Firefox or open with any browser
------------------------------------------------------------------------
r135 | kimthang | 2005-05-17 11:33:52 -0400 (Tue, 17 May 2005) | 1 line

I throw away some probe in concat.fcl
------------------------------------------------------------------------
r134 | kimthang | 2005-05-17 11:28:31 -0400 (Tue, 17 May 2005) | 1 line

 Concat
------------------------------------------------------------------------
r133 | kimthang | 2005-05-17 11:17:57 -0400 (Tue, 17 May 2005) | 1 line

I have made something wrong in Makefile of Vcard. I change it now
------------------------------------------------------------------------
r132 | kimthang | 2005-05-17 11:08:16 -0400 (Tue, 17 May 2005) | 1 line

General concat has been done
------------------------------------------------------------------------
r131 | kimthang | 2005-05-13 17:59:59 -0400 (Fri, 13 May 2005) | 1 line

Vcard with concat (Thanks Nate)
------------------------------------------------------------------------
r130 | lescuyer | 2005-05-13 17:51:57 -0400 (Fri, 13 May 2005) | 1 line

tiny mod in the Misc.whack, to escape only ' and "
------------------------------------------------------------------------
r129 | lescuyer | 2005-05-13 17:51:03 -0400 (Fri, 13 May 2005) | 1 line

Uncommented the put part of the mozilla test
------------------------------------------------------------------------
r128 | jnfoster | 2005-05-13 17:13:22 -0400 (Fri, 13 May 2005) | 1 line

tiny cleanup
------------------------------------------------------------------------
r127 | jnfoster | 2005-05-13 16:48:01 -0400 (Fri, 13 May 2005) | 1 line

fixed the problem stephane noticed. The external view parser (meta) and pretty printer (V.format) only require/print quotes around the symbols used in the lexer: {, }, [, ], =, :, ',', and whitespace (note that colon must now be enclosed in a string)
------------------------------------------------------------------------
r126 | lescuyer | 2005-05-13 14:38:54 -0400 (Fri, 13 May 2005) | 1 line

Mozilla lens.put works, test added to Makefile but commented for now, because meta views as written by a harmony get cannot be used as inputs in harmony put (that's to be fixed, but I have to understand how Format works... :))
------------------------------------------------------------------------
r125 | jnfoster | 2005-05-13 12:11:42 -0400 (Fri, 13 May 2005) | 1 line

tidied error messages a bit
------------------------------------------------------------------------
r124 | jnfoster | 2005-05-13 12:10:48 -0400 (Fri, 13 May 2005) | 10 lines

* Added a constant for Any, with a short-circuit member test. Seems to
 help efficiency a bit. Still need to get rid of all these thunks and
 save results...

* split out the meta lexer and parser (for Thang). Names in external
 views can be any string that is not a symbol. See metal.mll for the
 full definition.



------------------------------------------------------------------------
r123 | lescuyer | 2005-05-13 11:15:24 -0400 (Fri, 13 May 2005) | 1 line

Tiny mod in plunge, prints the name of the expected child (helpful for debugging the 'put') in case of error
------------------------------------------------------------------------
r122 | lescuyer | 2005-05-12 17:01:32 -0400 (Thu, 12 May 2005) | 1 line

Removed all the trace probes from the previous code, added a lens l2 that also discards descriptions, yielding a view that therefore is in the bookmarks schema described in the paper
------------------------------------------------------------------------
r121 | lescuyer | 2005-05-12 16:27:07 -0400 (Thu, 12 May 2005) | 1 line

Done with mozilla lenses for now. Descriptions in folders & links are handled correctly (and kept for eventual sync with another mozilla bookmark). The only thing to complete is to keep the separators, but it's a pain in the neck cause they can be *EVERY*where, even out of dl tags. Also, I introduced lightweight versions of the bookmarks to speed up the make.
------------------------------------------------------------------------
r120 | kimthang | 2005-05-12 15:57:36 -0400 (Thu, 12 May 2005) | 1 line

change in put of harmony by Nate
------------------------------------------------------------------------
r119 | kimthang | 2005-05-12 14:49:17 -0400 (Thu, 12 May 2005) | 1 line

Vcard with concat in get direction. Add explode and implode in native.ml
------------------------------------------------------------------------
r118 | jnfoster | 2005-05-12 13:22:26 -0400 (Thu, 12 May 2005) | 7 lines

Added the following lenses:
      - trace n: a quieter probe
      - assert T: the identity, with a membership check
      - invert l: flip the get and put functions; only for bijective lenses!
        (e.g., ungroupby2 can be written as invert (groupby2) :-) )


------------------------------------------------------------------------
r117 | kimthang | 2005-05-12 11:50:05 -0400 (Thu, 12 May 2005) | 1 line

Vcard with the quasi-concat without using recursive function
------------------------------------------------------------------------
r116 | lescuyer | 2005-05-12 10:46:20 -0400 (Thu, 12 May 2005) | 1 line

Ported mozilla lens for the new List module - Now VERY VERY SLOW, 4m30s of CPU time for bookmarks.xml ! :(
------------------------------------------------------------------------
r115 | jnfoster | 2005-05-12 02:48:59 -0400 (Thu, 12 May 2005) | 18 lines


Finally, a faster list reverse. Here's the algorithm (suggested by Sampath):
  
  - (1) map plunge x over the list to ensure that our list's type
    is not itself a list.
  - (2) repeatedly groupby2 until the tree has the form 
    	Any::Any::[]
    or
	Any::[]
  - (3) swap the order of the elements at the top, and so on, as long
        as their is list structure
  - (4) repeatedly ungroupby2 until we read a list of non-lists.

I've implemented this algorithm, but it only works so far for even powers of
two. (The reason is that ungroupby2 expects the odd element to be at
the end, but after swapping, the odd elt can be at the beginning.)


------------------------------------------------------------------------
r114 | jnfoster | 2005-05-12 01:04:21 -0400 (Thu, 12 May 2005) | 1 line

some list lenses (reverse, append, filter)
------------------------------------------------------------------------
r113 | jnfoster | 2005-05-12 00:06:34 -0400 (Thu, 12 May 2005) | 1 line

experimental/bookmarks (safari) back into a working state. Some *hs and *ts were hardcoded in and I missed them on my first pass.
------------------------------------------------------------------------
r112 | jnfoster | 2005-05-11 21:25:14 -0400 (Wed, 11 May 2005) | 1 line

a few bug fixes
------------------------------------------------------------------------
r111 | jnfoster | 2005-05-11 19:11:02 -0400 (Wed, 11 May 2005) | 29 lines

OOPS. I tried to kill that last commit, but it went through
anyways. Here's the full message I wanted to write.

A few major changes:
  * moved list.fcl up to the top-level lib/lenses directory
     - moved list stuff out of prelude.fcl and into list.fcl

  * changed the native prelude to "Native" from "Prelude" this way
    native stuff is not present in the default open module unless
    explicitly re-exported by Prelude. 

  * registered hd_tag, tl_tag, nil_tag, Cons, and Nil in Native. These
    are used by the parser to rewrite type expressions so I don't want
    to get into a situation where they are defined in prelude.fcl, so
    it can't use them, but as soon as prelude.fcl is compiled, they're
    available. Much cleaner to just have the parser use baked in versions.

  * Got rid of the DOUBLECOLON token and used productions to 
    enforce right associativity of :: (no %right directive)

  * Moved probe stuff up out of lens to native, where it probably
    belongs.  
      - Added a call to Format.flush in probe, which seems to have fixed
        the issue of disappearing probes in aconds (hurray!)

  * NOTICED that harmony suddenly got VERY VERY slow, even for the
    template. Will do some lightweight profiling tonight to diagnose.


------------------------------------------------------------------------
r110 | jnfoster | 2005-05-11 19:04:05 -0400 (Wed, 11 May 2005) | 13 lines


A few major changes:
  * moved list.fcl up to the top-level lib/lenses directory
     - moved list stuff out of prelude.fcl and into list.fcl

  * changed the native prelude to "Native" from "Prelude" this way
    native stuff is not present in the default open module unless
    explicitly re-exported by Prelude. 

  * registered hd_tag, tl_tag, nil_tag, Cons, and Nil in Native. These
    are used by the parser to rewrite type expressions 


------------------------------------------------------------------------
r109 | lescuyer | 2005-05-11 17:52:05 -0400 (Wed, 11 May 2005) | 1 line

tidying of the folder lens - Question : do we want to harmonize the hr separators or we simply ignore them ?
------------------------------------------------------------------------
r108 | lescuyer | 2005-05-11 17:42:16 -0400 (Wed, 11 May 2005) | 1 line

A BIG improvement in the mozilla lens, still folders with description TODO (for the moment, discarded). But though the bookmarks' sample is not big at all, the get is already very slow... :(
------------------------------------------------------------------------
r107 | lescuyer | 2005-05-11 17:01:30 -0400 (Wed, 11 May 2005) | 1 line

the :: right associativity, with a %right :?\194?\176) Nate, beware cause I also changed COLON COLON in the external view parser
------------------------------------------------------------------------
r106 | lescuyer | 2005-05-11 15:34:44 -0400 (Wed, 11 May 2005) | 1 line

not perfect, but at least there's no error anymore (Thx Nate !!)
------------------------------------------------------------------------
r105 | lescuyer | 2005-05-11 15:30:10 -0400 (Wed, 11 May 2005) | 1 line

Added the files related to the bookmark lens, did some mods to the makefile then, and used a Mozilla.mk to keep Makefile as clean as possible. For the moment, I commented the use of tidy, because it seems it doesn't work with the macOS/X version of tidy, so we use an already tidied file instead.       NB : the lens doesn't work for now, but it shoudn't be too long
------------------------------------------------------------------------
r104 | jnfoster | 2005-05-11 15:09:32 -0400 (Wed, 11 May 2005) | 1 line

fixed a bug in delayed evaluation of type operators
------------------------------------------------------------------------
r103 | jnfoster | 2005-05-11 13:12:26 -0400 (Wed, 11 May 2005) | 1 line

fixed bug in compiler
------------------------------------------------------------------------
r102 | jnfoster | 2005-05-11 11:46:16 -0400 (Wed, 11 May 2005) | 1 line

replaced listutil with list
------------------------------------------------------------------------
r101 | kimthang | 2005-05-11 10:48:08 -0400 (Wed, 11 May 2005) | 1 line

Vcard is good now
------------------------------------------------------------------------
r100 | bcpierce | 2005-05-11 10:46:58 -0400 (Wed, 11 May 2005) | 1 line

Checkpoint -- got to work on other things for a while now
------------------------------------------------------------------------
r99 | jnfoster | 2005-05-11 10:30:40 -0400 (Wed, 11 May 2005) | 1 line

I thought '::' notation would be nice...
------------------------------------------------------------------------
r98 | bcpierce | 2005-05-11 10:17:29 -0400 (Wed, 11 May 2005) | 1 line

Forgot a couple files
------------------------------------------------------------------------
r97 | bcpierce | 2005-05-11 10:14:36 -0400 (Wed, 11 May 2005) | 1 line

Checkpoint, for nate :-)
------------------------------------------------------------------------
r96 | bcpierce | 2005-05-11 10:07:55 -0400 (Wed, 11 May 2005) | 1 line

Working plist lens!  And some miscellaneous changes.
------------------------------------------------------------------------
r95 | jnfoster | 2005-05-11 00:12:27 -0400 (Wed, 11 May 2005) | 1 line

fixed bug in external view parser
------------------------------------------------------------------------
r94 | jnfoster | 2005-05-11 00:05:35 -0400 (Wed, 11 May 2005) | 6 lines

Fixed another small compiler bug (recursive variables were not being
cleared out for let-expressions).

Replaced groupby2 with a simpler, working version. 


------------------------------------------------------------------------
r93 | jnfoster | 2005-05-10 23:10:50 -0400 (Tue, 10 May 2005) | 11 lines

Fixed the bug in the compiler that Benjamin found this afternoon.

The check for recursive variables was using Safelist.mem on
Syntax.qids... structural equality is clearly WRONG because qids
contain Info.t's!

A big d'oh from my end. (An earlier version of the compiler defined a
new data type "qn" that didn't contain any Info.t. I thought it was
cleaner to keep it around.)


------------------------------------------------------------------------
r92 | jnfoster | 2005-05-10 16:17:44 -0400 (Tue, 10 May 2005) | 1 line

fixed bug in type.ml and V.format
------------------------------------------------------------------------
r91 | kimthang | 2005-05-10 16:15:06 -0400 (Tue, 10 May 2005) | 1 line

delete vcard in template
------------------------------------------------------------------------
r90 | kimthang | 2005-05-10 15:59:24 -0400 (Tue, 10 May 2005) | 1 line

I have made a mistake in commit 
------------------------------------------------------------------------
r89 | bcpierce | 2005-05-10 15:51:53 -0400 (Tue, 10 May 2005) | 1 line

Next checkpoint for nate
------------------------------------------------------------------------
r88 | kimthang | 2005-05-10 15:46:11 -0400 (Tue, 10 May 2005) | 1 line


------------------------------------------------------------------------
r87 | jnfoster | 2005-05-10 15:41:01 -0400 (Tue, 10 May 2005) | 1 line

grabbed newest OCamlMakefile. seems to eliminate some of the "inconsistent assumptions over foo.cmi" errors. also made a small improvement to type.ml.
------------------------------------------------------------------------
r86 | kimthang | 2005-05-10 15:40:02 -0400 (Tue, 10 May 2005) | 1 line

A simple vcard
------------------------------------------------------------------------
r85 | bcpierce | 2005-05-10 15:21:49 -0400 (Tue, 10 May 2005) | 1 line

Checkpoint for Nate
------------------------------------------------------------------------
r84 | jnfoster | 2005-05-10 14:43:08 -0400 (Tue, 10 May 2005) | 1 line

added a sanity check to test for projectability on the domain of v in (member v t)
------------------------------------------------------------------------
r83 | jnfoster | 2005-05-10 14:09:10 -0400 (Tue, 10 May 2005) | 21 lines

Finished porting normalization code and member checking. SIMPLE aconds now work for me. 

I'm committing this now, so that others can get on with their work,
but it is still very much alpha. In particular, make sure that you
write down "obviously projectable" types: i.e., at each level in a
type, for each name n, the type under n should be the exact same type. So

      {!="X",*\(n,m)="X"} 

is OK but

      {!="X",*\(n,m)="Y"} 

is not. 

With this assumption, model checking for types is simple. Later on
(maybe now) we should do the more general thing that constructs the
bipartite graph and tries to find a perfect matching. It wouldn't be
that inefficient...


------------------------------------------------------------------------
r82 | lescuyer | 2005-05-09 15:53:12 -0400 (Mon, 09 May 2005) | 1 line

Adding the set of names for lists
------------------------------------------------------------------------
r81 | jnfoster | 2005-05-09 11:15:16 -0400 (Mon, 09 May 2005) | 1 line

ported the rest of the code for sync; I don't think this is functional yet, but it compiles
------------------------------------------------------------------------
r80 | bcpierce | 2005-05-09 08:19:39 -0400 (Mon, 09 May 2005) | 1 line

Checkpoint, for Nate to look at...
------------------------------------------------------------------------
r79 | jnfoster | 2005-05-08 23:11:11 -0400 (Sun, 08 May 2005) | 8 lines

Adopted Benjamin's suggestion for exception lists:
	*\n = ...
	*\n,m = ...
	*\(n) = ...
	*\(n,m) = ...
all parse fine.


------------------------------------------------------------------------
r78 | jnfoster | 2005-05-08 21:22:18 -0400 (Sun, 08 May 2005) | 10 lines

A much improved parser. The quirky exp_core is gone. 

Also gone are the two modes for writing types. The one where you'd write:
     n{X},*(n){Y} | m{Z} 
is gone. Instead, always write in view-style:
     { n=X, *(n){Y} } | m{Z}
(I'm still not happy with the way exception lists look---maybe \(n) or
\{n}?---but they are easy to change)


------------------------------------------------------------------------
r77 | jnfoster | 2005-05-07 22:21:29 -0400 (Sat, 07 May 2005) | 13 lines

Fixes for the dreaded wmap loop. (For those interested, merging the
checking and the compiling is a BAD idea unless you really only want
run-time checking. I can explain more on Monday.)

Also fixed several other bugs yesterday, they're included in this
commit but I can't remember the details.

The parsing of types is still a little fragile. I spent a long time
figting with ocamlyacc, but didn't resolve it yet. For now, be careful
where you stick identifiers, and if you get a parse error, check out
the types section of parser.mly. It's pretty straightforward.


------------------------------------------------------------------------
r76 | bcpierce | 2005-05-06 22:24:48 -0400 (Fri, 06 May 2005) | 1 line

checkpoint
------------------------------------------------------------------------
r75 | jnfoster | 2005-05-06 14:46:18 -0400 (Fri, 06 May 2005) | 1 line

many cleanups to parser and error reporting. map fix coming soon
------------------------------------------------------------------------
r74 | jnfoster | 2005-05-05 19:36:25 -0400 (Thu, 05 May 2005) | 9 lines

Moved from Pervasives.Native and Pervasives to just "Prelude"

Also re-exporeted all the definitions (with their types) in
lib/lenses/prelude.fcl to better record exactly what's in Prelude.

The template test works; bookmarks doesn't (was it in a working state?
i.e., did I break something?)


------------------------------------------------------------------------
r73 | jnfoster | 2005-05-05 19:29:54 -0400 (Thu, 05 May 2005) | 1 line

making subversion happy
------------------------------------------------------------------------
r72 | jnfoster | 2005-05-05 19:04:15 -0400 (Thu, 05 May 2005) | 1 line

parser fixes
------------------------------------------------------------------------
r71 | bcpierce | 2005-05-05 15:43:00 -0400 (Thu, 05 May 2005) | 1 line

Checkpoint for nate -- DO NOT UPDATE (not compiling right now)
------------------------------------------------------------------------
r70 | bcpierce | 2005-05-05 14:06:57 -0400 (Thu, 05 May 2005) | 1 line

checkpoint a few miscellaneous changes
------------------------------------------------------------------------
r69 | bcpierce | 2005-05-05 13:48:34 -0400 (Thu, 05 May 2005) | 1 line

Print syntax errors so that emacs can understand them. :-)
------------------------------------------------------------------------
r68 | bcpierce | 2005-05-04 20:33:19 -0400 (Wed, 04 May 2005) | 1 line

Beginnings of bookmark example
------------------------------------------------------------------------
r67 | bcpierce | 2005-05-04 20:31:36 -0400 (Wed, 04 May 2005) | 1 line

Reorg finished; typing 'make' in the top level directory now builds src/harmony and runs some tests from /examples
------------------------------------------------------------------------
r66 | bcpierce | 2005-05-04 20:19:32 -0400 (Wed, 04 May 2005) | 1 line

Next phase of reorg: newsrc moved to src
------------------------------------------------------------------------
r65 | bcpierce | 2005-05-04 20:14:06 -0400 (Wed, 04 May 2005) | 1 line

(Cross your fingers)  Moving old /src out of the way...
------------------------------------------------------------------------
r64 | bcpierce | 2005-05-04 20:13:02 -0400 (Wed, 04 May 2005) | 1 line

Small tweak to parser: top-level module doesnt require an END at the end or an IN between OPENs and other declarations
------------------------------------------------------------------------
r63 | bcpierce | 2005-05-04 14:53:02 -0400 (Wed, 04 May 2005) | 1 line

Added examples directory with a template showing how to run the new harmony on a very simple test case
------------------------------------------------------------------------
r62 | jnfoster | 2005-05-04 13:48:02 -0400 (Wed, 04 May 2005) | 1 line

small fix to registry
------------------------------------------------------------------------
r61 | jnfoster | 2005-05-04 13:46:50 -0400 (Wed, 04 May 2005) | 1 line

Makefile fix
------------------------------------------------------------------------
r60 | jnfoster | 2005-05-04 12:49:13 -0400 (Wed, 04 May 2005) | 1 line

some progress porting old Schema functions to new datatype for Schemas (now just called Types)
------------------------------------------------------------------------
r59 | bcpierce | 2005-05-04 12:44:01 -0400 (Wed, 04 May 2005) | 1 line

Removed config.ml and distributed its contents to where they were needed
------------------------------------------------------------------------
r58 | bcpierce | 2005-05-04 12:43:24 -0400 (Wed, 04 May 2005) | 1 line

Checkpoint (working on harmony.ml)
------------------------------------------------------------------------
r57 | bcpierce | 2005-05-01 15:27:50 -0400 (Sun, 01 May 2005) | 1 line

First two sections are now finished, I think.
------------------------------------------------------------------------
r56 | jnfoster | 2005-05-01 14:26:21 -0400 (Sun, 01 May 2005) | 1 line

tiny fix: a meaningful error message for meta parser
------------------------------------------------------------------------
r55 | jnfoster | 2005-05-01 14:09:14 -0400 (Sun, 01 May 2005) | 1 line

Bug fixes for compiler
------------------------------------------------------------------------
r54 | jnfoster | 2005-05-01 13:52:07 -0400 (Sun, 01 May 2005) | 5 lines

Sometimes, you just gotta start over. 

Here's a mostly new compiler.


------------------------------------------------------------------------
r53 | jnfoster | 2005-04-28 17:10:58 -0400 (Thu, 28 Apr 2005) | 1 line

oops. forgot to add the files (thanks to stephane for pointing it out :) )
------------------------------------------------------------------------
r52 | jnfoster | 2005-04-28 16:54:49 -0400 (Thu, 28 Apr 2005) | 1 line

some adjustments to teh front end, and a meta parser
------------------------------------------------------------------------
r51 | jnfoster | 2005-04-28 11:53:11 -0400 (Thu, 28 Apr 2005) | 1 line

add parser.output to list of files to remove on make clean. someone seems to have done this with ._d directories already.
------------------------------------------------------------------------
r50 | jnfoster | 2005-04-28 11:49:13 -0400 (Thu, 28 Apr 2005) | 1 line

fixed problem with compiling recursive types
------------------------------------------------------------------------
r49 | jnfoster | 2005-04-28 01:04:32 -0400 (Thu, 28 Apr 2005) | 8 lines

 - Fixed parsing/checking/compilation of types... almost.  Still need
   to think through representing the recursive types. Going to bed,
   will pick up again in the AM.

 - added conditionals. Had to make acond native again, because we
   don't have a NOT operation on types :(


------------------------------------------------------------------------
r48 | jnfoster | 2005-04-27 18:28:35 -0400 (Wed, 27 Apr 2005) | 37 lines


Many large changes:

     * reorganized source code. Stuff that can be plugged in: viewers
       (meta, xml), interpreted lenses, native lens code now lives in
       lib. We can move this up a level to suit Benjamin's
       sensabilities later :)

     * reworked the parser again to get rid of anti-quotation. names
       are strings, full stop. bare identifiers are always parsed as
       variables. 

       this makes names in views a little clunky because you have to write
          { "x" = ["1", "2", "3"], "y"=["4", "5", "6"] }
       but it's much easier to make sense of, IMHO.

     * fixed parsing of list-views and a few other small bugs.

Currently broken:
      
      * type definitions with parameters (List T). The old checker
        imposed a quite draconian restriction which, when I removed
        it, broke several other things. 

	Will start fixing now.

To do:

      * fix cond like we talked about on Monday.

      * add native unit tests back in

To do later:

      * parse FocalDoc from source, include in registry.


------------------------------------------------------------------------
r47 | jnfoster | 2005-04-26 16:09:58 -0400 (Tue, 26 Apr 2005) | 1 line

fixed bug in parser, derived lenses in pervasives.fcl
------------------------------------------------------------------------
r46 | jnfoster | 2005-04-26 15:42:59 -0400 (Tue, 26 Apr 2005) | 1 line

Safelist-ized checker and compiler. Removed some appends.
------------------------------------------------------------------------
r45 | lescuyer | 2005-04-26 13:16:28 -0400 (Tue, 26 Apr 2005) | 1 line

added registration of the probe lens
------------------------------------------------------------------------
r44 | jnfoster | 2005-04-25 17:17:41 -0400 (Mon, 25 Apr 2005) | 1 line

a few small tweaks to harmony
------------------------------------------------------------------------
r43 | lescuyer | 2005-04-25 17:00:32 -0400 (Mon, 25 Apr 2005) | 1 line

flattened structure, all src files in one directory
------------------------------------------------------------------------
r42 | jnfoster | 2005-04-25 16:39:56 -0400 (Mon, 25 Apr 2005) | 1 line

syncing up, changes to harmony.ml front-end
------------------------------------------------------------------------
r41 | lescuyer | 2005-04-25 16:21:30 -0400 (Mon, 25 Apr 2005) | 1 line

add xml to common
------------------------------------------------------------------------
r40 | jnfoster | 2005-04-25 14:36:06 -0400 (Mon, 25 Apr 2005) | 1 line

more makefile hacking
------------------------------------------------------------------------
r39 | lescuyer | 2005-04-25 14:25:53 -0400 (Mon, 25 Apr 2005) | 1 line

some stuff
------------------------------------------------------------------------
r38 | bcpierce | 2005-04-25 14:20:42 -0400 (Mon, 25 Apr 2005) | 1 line

Adding some papers (that I had actually added a long time ago!)
------------------------------------------------------------------------
r37 | lescuyer | 2005-04-25 13:22:42 -0400 (Mon, 25 Apr 2005) | 1 line

Prefs module modifications + Move surveyor from sync to common + Add Harmony.ml
------------------------------------------------------------------------
r36 | jnfoster | 2005-04-25 01:49:05 -0400 (Mon, 25 Apr 2005) | 1 line

oops. one more bug fix
------------------------------------------------------------------------
r35 | jnfoster | 2005-04-25 01:46:50 -0400 (Mon, 25 Apr 2005) | 1 line

Final bug fixes for the night. Everything from Pervasives parses/checks/compiles except for cond (and values derived from it)
------------------------------------------------------------------------
r34 | jnfoster | 2005-04-25 01:04:02 -0400 (Mon, 25 Apr 2005) | 1 line

Pervasives.Derived module, in Focal :)
------------------------------------------------------------------------
r33 | jnfoster | 2005-04-25 00:39:43 -0400 (Mon, 25 Apr 2005) | 6 lines

A few more bug fixes. 

Added all the native lenses from pervasives_plugin except for cond...
  - xfork now takes views as arguments, not (name -> bool) functions.


------------------------------------------------------------------------
r32 | jnfoster | 2005-04-24 22:56:21 -0400 (Sun, 24 Apr 2005) | 1 line

small bug fixes
------------------------------------------------------------------------
r31 | jnfoster | 2005-04-24 20:26:30 -0400 (Sun, 24 Apr 2005) | 10 lines


Another cleanup pass over everything. Tidied code a bit. More importantly:

  - fixed several bugs with nested modules in checker/compiler
  - added real search paths 
  - added check that module M is defined in a file M.fcl or m.fcl

Now onto porting all the Pervasives...


------------------------------------------------------------------------
r30 | jnfoster | 2005-04-24 14:54:16 -0400 (Sun, 24 Apr 2005) | 1 line

oops. a few files were missing ;)
------------------------------------------------------------------------
r29 | jnfoster | 2005-04-24 14:52:35 -0400 (Sun, 24 Apr 2005) | 1 line

another cleanup pass; checkpointing for backup
------------------------------------------------------------------------
r28 | jnfoster | 2005-04-24 01:20:06 -0400 (Sun, 24 Apr 2005) | 1 line

cleanup and reorg. checkpointing for backup.
------------------------------------------------------------------------
r27 | jnfoster | 2005-04-23 16:56:34 -0400 (Sat, 23 Apr 2005) | 11 lines

A few more tweaks and bug fixes. 

The thing works now and:
    - dynamically loads .fcl files from a (for now, single, static) search path
    - registers bindings with their module names
    - fixed a bug with nested modules. 

I'm now going to make some significant tidying changes now. Checking
in to mark a checkpoint so I can go back to a good state if needed.


------------------------------------------------------------------------
r26 | jnfoster | 2005-04-23 00:51:15 -0400 (Sat, 23 Apr 2005) | 1 line

Changed environments again. The library is now hidden; checker and compiler use Value.envs for local variables/parameters only. I still have a couple of things to clean up in the AM, but back to a working state so comitting...
------------------------------------------------------------------------
r25 | jnfoster | 2005-04-22 18:09:20 -0400 (Fri, 22 Apr 2005) | 1 line

added Common.Makefile
------------------------------------------------------------------------
r24 | bcpierce | 2005-04-22 18:01:29 -0400 (Fri, 22 Apr 2005) | 3 lines

Adding surveyor.mli (updated/simplified) and .ml (not yet updated)


------------------------------------------------------------------------
r23 | lescuyer | 2005-04-22 16:34:51 -0400 (Fri, 22 Apr 2005) | 1 line

new options using prefs, sync dir added to newsrc
------------------------------------------------------------------------
r22 | jnfoster | 2005-04-22 12:56:24 -0400 (Fri, 22 Apr 2005) | 1 line

added concat back in
------------------------------------------------------------------------
r21 | jnfoster | 2005-04-22 11:59:07 -0400 (Fri, 22 Apr 2005) | 1 line

load-on-demand (almost)
------------------------------------------------------------------------
r20 | jnfoster | 2005-04-22 02:31:29 -0400 (Fri, 22 Apr 2005) | 1 line

fixed a few small things, added syntax for describing finite maps
------------------------------------------------------------------------
r19 | jnfoster | 2005-04-22 00:47:13 -0400 (Fri, 22 Apr 2005) | 1 line

checkpoint
------------------------------------------------------------------------
r18 | jnfoster | 2005-04-21 19:17:07 -0400 (Thu, 21 Apr 2005) | 1 line

checkpoint before biking home
------------------------------------------------------------------------
r17 | jnfoster | 2005-04-21 12:19:45 -0400 (Thu, 21 Apr 2005) | 1 line

testing mailer from eniac-l
------------------------------------------------------------------------
r16 | jnfoster | 2005-04-21 12:12:06 -0400 (Thu, 21 Apr 2005) | 1 line

testing mailer
------------------------------------------------------------------------
r15 | jnfoster | 2005-04-21 12:10:25 -0400 (Thu, 21 Apr 2005) | 1 line

testing mailer
------------------------------------------------------------------------
r14 | lescuyer | 2005-04-21 11:34:24 -0400 (Thu, 21 Apr 2005) | 1 line

a misspelling in the Common.makefile, sorry for that...
------------------------------------------------------------------------
r13 | lescuyer | 2005-04-21 11:30:37 -0400 (Thu, 21 Apr 2005) | 1 line

and the modification in the makefile, else addressbook isn't compiled..
------------------------------------------------------------------------
r12 | lescuyer | 2005-04-21 11:26:34 -0400 (Thu, 21 Apr 2005) | 1 line

addressbook plugin directory (and an addressbook example)
------------------------------------------------------------------------
r11 | lescuyer | 2005-04-21 11:04:47 -0400 (Thu, 21 Apr 2005) | 1 line

adding implode & explode lenses
------------------------------------------------------------------------
r10 | jnfoster | 2005-04-21 10:35:13 -0400 (Thu, 21 Apr 2005) | 1 line

testing commit mail from eniac-l checked out sources
------------------------------------------------------------------------
r9 | bcpierce | 2005-04-21 08:24:15 -0400 (Thu, 21 Apr 2005) | 3 lines

Committing some miscellaneous changes.


------------------------------------------------------------------------
r8 | schmitta | 2005-04-21 03:22:44 -0400 (Thu, 21 Apr 2005) | 3 lines

Added a TODO line, to test for commit and commit emails


------------------------------------------------------------------------
r7 | jnfoster | 2005-04-21 01:30:23 -0400 (Thu, 21 Apr 2005) | 1 line

testing mailer scripts
------------------------------------------------------------------------
r6 | jnfoster | 2005-04-21 01:28:47 -0400 (Thu, 21 Apr 2005) | 1 line

testing mailer scripts
------------------------------------------------------------------------
r5 | jnfoster | 2005-04-21 01:25:56 -0400 (Thu, 21 Apr 2005) | 1 line

testing mailer
------------------------------------------------------------------------
r4 | jnfoster | 2005-04-21 01:05:11 -0400 (Thu, 21 Apr 2005) | 1 line

testing commit emails
------------------------------------------------------------------------
r3 | jnfoster | 2005-04-21 00:59:15 -0400 (Thu, 21 Apr 2005) | 1 line

testing commit mailer, updated TODO with stuff that has been done
------------------------------------------------------------------------
r2 | jnfoster | 2005-04-21 00:32:38 -0400 (Thu, 21 Apr 2005) | 2 lines

migrating to subversion

------------------------------------------------------------------------
r1 | jnfoster | 2005-10-21 11:49:19 -0400 (Fri, 21 Oct 2005) | 1 line

top-level directory structure
------------------------------------------------------------------------
