Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2005-08-01 03:32:57 -0700 (Mon, 01 Aug 2005)
Revision: 1278
Log message:

      Somehow lm_hash.mli from the branch got added on the trunk _again_, preventing
      the trunk from compiling.
      
      Jason, when you do merging and then commit the merged stuff on the new brunch,
      please make sure that the added files have the branch tag as well.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2005-08-01 12:37:58 -0700 (Mon, 01 Aug 2005)
Revision: 1281
Log message:

      I believe this fixes bug #497.  The close_next function should
      indeed allow arguments after a word delimiter.
      
      Another way to solve this might be to use regular expression
      equivalences.
      
          \<\(r == \(\<r
      
      That is a little hard, so we go with expanding the cases close_next
      will consider.
      

Changes by: ( at unknown.email)
Date: 2005-08-01 16:38:02 -0700 (Mon, 01 Aug 2005)
Revision: 1284
Log message:

      This commit was manufactured by cvs2svn to create branch
      'omake_0_9_7_pre2'.

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2005-08-05 11:07:11 -0700 (Fri, 05 Aug 2005)
Revision: 1287
Log message:

      Exception was defined twice.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2005-08-05 13:51:25 -0700 (Fri, 05 Aug 2005)
Revision: 1288
Log message:

      Backported the new lib/ organization.
      

Changes by: ( at unknown.email)
Date: 2005-08-05 15:00:40 -0700 (Fri, 05 Aug 2005)
Revision: 1304
Log message:

      This commit was manufactured by cvs2svn to create branch
      'omake_0_9_7_pre3'.

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2005-08-06 08:52:38 -0700 (Sat, 06 Aug 2005)
Revision: 1305
Log message:

      Merge against the trunk.
      

Changes by: ( at unknown.email)
Date: 2005-08-07 19:10:18 -0700 (Sun, 07 Aug 2005)
Revision: 1322
Log message:

      This commit was manufactured by cvs2svn to create branch
      'omake_0_9_7_pre4'.

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2005-08-07 19:10:18 -0700 (Sun, 07 Aug 2005)
Revision: 1323
Log message:

      Merged against the trunk.  This also fixes a few minor linking issues.
      

Changes by: ( at unknown.email)
Date: 2005-08-07 19:10:18 -0700 (Sun, 07 Aug 2005)
Revision: 1324
Log message:

      This commit was manufactured by cvs2svn to create branch
      'omake_0_9_7_pre4'.

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2005-08-08 09:52:16 -0700 (Mon, 08 Aug 2005)
Revision: 1325
Log message:

      Once again files were added on the trunk instead of on a branch :-(( Fixing.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2005-08-09 14:11:50 -0700 (Tue, 09 Aug 2005)
Revision: 1338
Log message:

      Add lm_hash.mli on the trunk to avoid cvs problems.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2005-08-09 15:15:09 -0700 (Tue, 09 Aug 2005)
Revision: 1340
Log message:

      Add empty omake_virtual_id.ml{,i} to avoid cvs problems.
      

Changes by: ( at unknown.email)
Date: 2005-08-09 16:37:23 -0700 (Tue, 09 Aug 2005)
Revision: 1346
Log message:

      This commit was manufactured by cvs2svn to create branch
      'omake_0_9_7_pre5'.

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2005-08-09 16:56:51 -0700 (Tue, 09 Aug 2005)
Revision: 1347
Log message:

      Merged against the trunk.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2005-08-12 18:46:46 -0700 (Fri, 12 Aug 2005)
Revision: 1358
Log message:

      Bug 496 - attempting to fix "which" on Cygwin.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2005-08-12 18:55:45 -0700 (Fri, 12 Aug 2005)
Revision: 1359
Log message:

      Further fix for bug 496
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2005-08-15 20:59:47 -0700 (Mon, 15 Aug 2005)
Revision: 1361
Log message:

      Yet another attempt at fixing "which" on Cygwin (bug 496). Big thanks to Peter
      Jolly for his pointers to Cygwin information.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2005-08-15 21:05:10 -0700 (Mon, 15 Aug 2005)
Revision: 1362
Log message:

      The previous Cygwin fix is not quite right either (even when foo exists and is
      not executabe, Cygwin is happy to execute foo.exe in place of foo - if I
      understand the docs correctly).
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2005-08-19 16:20:34 -0700 (Fri, 19 Aug 2005)
Revision: 1408
Log message:

      This is some exploration into dynamic linking.  It contains:
         - a parser for C code     parse/C/{Parse,Lex}.om
         - a stub code generator   parse/C/Dll.om
      
      The contrib/gtk directory gives an demo for Gtk (I chose Gtk
      because it has nice header files, and it is sufficiently
      complicated).
      
      I'll write up some docs at some point, but here is the basic idea
      of how a DLL is generated.
      
         1. Parse the C header files.
         2. Extract the declarations.
         3. Generate C stubs for each of the declarations.
         4. Produce a (large) C file that can be compiled
            and then dynamically linked into your OCaml process.
      
      Doing this all from your OMakefile is pretty simple.  Mainly
      you need to specify which functions you want to include
      in your DLL.
      
      This is reasonably independent from omake, so you can use it as
      a generic way to generate wrappers around C functions.
      
      Note that the C file for GTK is perhaps 120k lines of code, so
      you get to stress your C compiler too.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2005-08-29 15:24:40 -0700 (Mon, 29 Aug 2005)
Revision: 1426
Log message:

      Here are some changes I made at TPHOLs--I'd like to finish this off soon.
      Most of these changes have to do with exports.
      
         -- Added exports with values,
      
                f() =
                    X = ...
                    export
                        1
                Y = $(f)
      
            This will export X, and define Y as 1.
      
         -- The value in the export block is defined in export
            mode.   So you no longer have to write that huge
            chain of exports.  Before we had code like:
      
                foreach(x => $(y))
                    if ...
                       z = ...
                       export
                    export
                export
      
            Now, you can just define it in export mode:
      
                export
                    foreach(x => $(y))
                        if ...
                            z = ...
      
          -- New variables defined in functions are now private.
             Private variables are not exported across function boundaries.
      
      Here is the current show-stopper (we always had this problem, we just
      never noticed it).  What happens with an export from a higher-order function?
      The context can be quite different, and the export doesn't always make sense.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2005-08-31 21:37:27 -0700 (Wed, 31 Aug 2005)
Revision: 1450
Log message:

      Backported a few "MAGICBEGIN/MAGICEND" unnotations (where it shouldn't
      create conflicts).