Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-01 09:28:09 -0700 (Fri, 01 Jun 2007)
Revision: 10854
Log message:

      Updating the download page for 0.9.8.3 release.

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-01 09:38:31 -0700 (Fri, 01 Jun 2007)
Revision: 10855
Log message:

      Rephrased a little.

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-01 11:55:44 -0700 (Fri, 01 Jun 2007)
Revision: 10857
Log message:

      Minor optimization.

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-01 12:49:25 -0700 (Fri, 01 Jun 2007)
Revision: 10858
Log message:

      Merge to rev 10853.

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-01 15:25:34 -0700 (Fri, 01 Jun 2007)
Revision: 10859
Log message:

      Do the correct translation to absolute name (for things like the history file
     name) when a relative directory is given via the --dotomake flag.

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-01 15:50:38 -0700 (Fri, 01 Jun 2007)
Revision: 10861
Log message:

      Our current algorithm for the "generic" case-sensitivity test was:
         1. Try the name being created itself.
         2. Try looking at the directory entries.
         3. Create a dummy file, and test.
         4. Test the parent.
     
     Step 2 would start with the Unix.opendir call and if that failed we would
     abvance to step 3. 
     
     This commit changes it so that if the Unix.opendir error code tells us the
     directory does not exist (either of: ENOENT, ENOTDIR, ELOOP, or ENAMETOOLONG),
     then we advance to step 4 right away.
     
     Note that in the case of a permission error in Unix.opendir, we still go to
     step 3 (as a directory with "wx" permission, but no "r" would allow step 3,
     but not opendir).

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-01 15:56:01 -0700 (Fri, 01 Jun 2007)
Revision: 10862
Log message:

      - Merged the 10854:10861 from the 0.9.8.x branch (this includes the changes
       aimed at reducing the cases where OMake tried to create a
       case-sensitivity-testing temp file in a "weird" place).
     
     - Bumped the version number to 0.9.8.3-2

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-01 16:11:53 -0700 (Fri, 01 Jun 2007)
Revision: 10863
Log message:

      $(OCAMLDEPFLAGS) should be passed to "ocamldep -modules"

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-01 16:14:41 -0700 (Fri, 01 Jun 2007)
Revision: 10864
Log message:

      $(OCAMLDEPFLAGS) should be passed to "ocamldep -modules" (merges r10863 off
     0.9.8.3)

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 09:56:21 -0700 (Mon, 04 Jun 2007)
Revision: 10870
Log message:

      There was a Unix.closedir missing along one of the control paths.

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 09:56:48 -0700 (Mon, 04 Jun 2007)
Revision: 10871
Log message:

      There was a Unix.closedir missing along one of the control paths.
     (This is the same as rev. 10870 on 0.9.8.x)

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 10:02:52 -0700 (Mon, 04 Jun 2007)
Revision: 10872
Log message:

      Updated the version to 
     "0.9.8.3+ (working towards 0.9.8.3-3, or 0.9.8.4, or moving to a newer branch)"

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 12:46:05 -0700 (Mon, 04 Jun 2007)
Revision: 10877
Log message:

      - In Omake_cache.exists, only do the Unix stat, but do not attempt to compute the digest.
     
     - In Omake_cache.{stat_unix,is_dir,exists,exists_dir}, allow passing an
       optional "force" flag that would cause the cache to re-stat a file that it
       thinks does not exist.
     
     - When looking up a doing a "node" -> "executable" lookup on Windows (which
       cycles through potential suffixes), do a "forced" existence check for the
       node itself (but not the suffix-appended versions). Without this, some of my
       code was failing on Windows...

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-04 13:04:17 -0700 (Mon, 04 Jun 2007)
Revision: 10878
Log message:

      Rearrange the LaTeX .STATIC code a bit.
     
     Static sections are actually created as objects first, then folded into the global
     environment.  This means that lazy expressions in a .STATIC block escape the object
     in which they were defined, so code like this will fail:
     
         TETEX2_ENABLED = ...
         LATEXFLAGS =
             BASE = ...
             value $`(if $(TETEX2_ENABLED), ...)
     
     It is unclear to me whether this is an issue.

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 13:24:03 -0700 (Mon, 04 Jun 2007)
Revision: 10879
Log message:

      - In Omake_cache.exists, only do the Unix stat, but do not attempt to compute
       the digest.
     
     - In Omake_cache.{stat_unix,is_dir,exists,exists_dir}, allow passing an
       optional "force" flag that would cause the cache to re-stat a file that it
       thinks does not exist.
     
     - When looking up a doing a "node" -> "executable" lookup on Windows (which
       cycles through potential suffixes), do a "forced" existence check for the
       node itself (but not the suffix-appended versions). Without this, some of my
       code was failing on Windows...
     
     This merges rev 10877 off the "old" 0.9.8.3 branch.

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-04 14:13:58 -0700 (Mon, 04 Jun 2007)
Revision: 10880
Log message:

      Merge to rev 10878.

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 15:46:47 -0700 (Mon, 04 Jun 2007)
Revision: 10883
Log message:

      Preparing for 0.9.8.4

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 15:46:47 -0700 (Mon, 04 Jun 2007)
Revision: 10883
Log message:

      Preparing for 0.9.8.4

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 15:52:23 -0700 (Mon, 04 Jun 2007)
Revision: 10884
Log message:

      0.9.8.4

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 15:57:51 -0700 (Mon, 04 Jun 2007)
Revision: 10885
Log message:

      Preparing the 0.9.8.4-1 release

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 15:59:03 -0700 (Mon, 04 Jun 2007)
Revision: 10886
Log message:

      Use the correct branch.

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-04 16:01:45 -0700 (Mon, 04 Jun 2007)
Revision: 10887
Log message:

      Posting the 0.9.8.4-1 release on the web.

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-04 18:47:59 -0700 (Mon, 04 Jun 2007)
Revision: 10890
Log message:

      Updated some of the documentation 
         - private/this/public qualifiers
         - export

Changes by: Nathaniel Gray (n8gray at cs.caltech.edu)
Date: 2007-06-05 11:53:53 -0700 (Tue, 05 Jun 2007)
Revision: 10892
Log message:

      Automatically append the architecture to the pkg/dmg filenames.

Changes by: Nathaniel Gray (n8gray at cs.caltech.edu)
Date: 2007-06-05 11:57:18 -0700 (Tue, 05 Jun 2007)
Revision: 10893
Log message:

      Added an upload target.  Now the whole thing's truly brainless:
         make_dmg all && make_dmg upload

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-06 14:55:50 -0700 (Wed, 06 Jun 2007)
Revision: 10902
Log message:

      Restore the current directory after executing a .SUBDIRS,
     addressing bug #670.
     
     That is, the following does not export the current directory.
     
         .SUBDIRS: a
             export
     
     But the following does.
     
         section
             cd a
             export

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-07 15:43:35 -0700 (Thu, 07 Jun 2007)
Revision: 10909
Log message:

      Added dcoumentation for .STATIC

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-07 15:51:25 -0700 (Thu, 07 Jun 2007)
Revision: 10910
Log message:

      Add some macros for version numbers.

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-08 14:01:32 -0700 (Fri, 08 Jun 2007)
Revision: 10913
Log message:

      Mark .STATIC as \newinreorg, which now says "will be introduced".

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-10 16:49:20 -0700 (Sun, 10 Jun 2007)
Revision: 10925
Log message:

      Merged Nathan's "make_dmg upload" implementation (rev 10893 off the 0.9.8.x
     branch).

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-10 16:52:53 -0700 (Sun, 10 Jun 2007)
Revision: 10926
Log message:

      Bumping the version number (to 0.9.8.4.0+)

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-10 18:33:29 -0700 (Sun, 10 Jun 2007)
Revision: 10927
Log message:

      FS case sensitivity improvements:
     - If the name does not ocontain any upper case characters, abortt the case
       sensitivity early (as we do not really care).
     - Test with Unix.access before trying to create test entries.
     - Test for all the 3 pairs of ranges that Char.uppercase/Char.lowercase
       recognizes (not just a single 'A'..'Z'/'a'..'z').

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-10 18:37:30 -0700 (Sun, 10 Jun 2007)
Revision: 10928
Log message:

      FS case sensitivity improvements:
     - If the name does not ocontain any upper case characters, abortt the case
       sensitivity early (as we do not really care).
     - Test with Unix.access before trying to create test entries.
     - Test for all the 3 pairs of ranges that Char.uppercase/Char.lowercase
       recognizes (not just a single 'A'..'Z'/'a'..'z').
     
     (This merges rev. 10927 off the 0.9.8.4 branch).

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-10 18:44:24 -0700 (Sun, 10 Jun 2007)
Revision: 10929
Log message:

      Will package the current state as 0.9.8.4-2 for the Gentoo people to try.

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-11 14:10:16 -0700 (Mon, 11 Jun 2007)
Revision: 10930
Log message:

      Even if OCAMLDEP_MODULES_ENABLED is set, ocamldep needs to be passed
     the OCAMLDEPFLAGS, because they might indicate how to preprocess the
     file (thanks David Noblet for catching this).
     
     Do we also need to use ocamlfind?  I'm not sure...

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-12 14:38:27 -0700 (Tue, 12 Jun 2007)
Revision: 10937
Log message:

      Even if OCAMLDEP_MODULES_ENABLED is set, ocamldep needs to be passed
     the OCAMLDEPFLAGS, because they might indicate how to preprocess the
     file (thanks David Noblet for catching this).
     
     (Merges rev 10930 from he 0.9.8.4 branch)

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-13 11:05:00 -0700 (Wed, 13 Jun 2007)
Revision: 10947
Log message:

      Minor: Addressed one of Jason's "ZZZ" comments.

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-13 11:06:14 -0700 (Wed, 13 Jun 2007)
Revision: 10948
Log message:

      Trivial spelling fix.

Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-06-13 11:17:49 -0700 (Wed, 13 Jun 2007)
Revision: 10949
Log message:

      Minor error reporting improvement.

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-13 12:07:43 -0700 (Wed, 13 Jun 2007)
Revision: 10950
Log message:

      Documenting the meanings of the qualifiers:
         private: statically scoped
         this: field in the current object
         public: global variable
         global: same as public
         protected:
            - in 0.9.8, same as "this."
            - in 0.9.9, local to the current file/object
                 this.protected.x = ...
                 global.protected.x = ...
              or, the correct qualifier is usually inferred
                 protected.x = ...

Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-06-13 12:53:23 -0700 (Wed, 13 Jun 2007)
Revision: 10952
Log message:

      Merge to rev 10950.