Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-01 20:08:09 -0700 (Thu, 01 Jun 2006)
Revision: 9242
Log message:

      Ignore the new fam_* symlinks.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-05 16:53:44 -0700 (Mon, 05 Jun 2006)
Revision: 9254
Log message:

      If a thread/process forked from the "main" OMake thread/process tries to save
      the DB, this means that something went wrong. In this case, print an error
      message and abort.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-06 07:15:20 -0700 (Tue, 06 Jun 2006)
Revision: 9257
Log message:

      Be more clean.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-06 13:47:38 -0700 (Tue, 06 Jun 2006)
Revision: 9258
Log message:

      "cmd_argv = []" no longer indicates an "empty" command, so commands with an
      empty cmd_argv should no longer be ignored.
      
      P.S. Should cmd_argv be renamed into cmd_args to emphasize that it does not
      include the executable itself?
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-06 14:54:29 -0700 (Tue, 06 Jun 2006)
Revision: 9260
Log message:

      Print built-in command lines differently from "proper" shell command lines.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-06 16:45:34 -0700 (Tue, 06 Jun 2006)
Revision: 9261
Log message:

      Make the .omakedb save interval configurable (both the compiled-in default via
      a .config option and the run-time value via the command-line flag). Also, a 0
      (or negative) value disables the intermediate .omakedb dumps altogether.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-06 17:58:23 -0700 (Tue, 06 Jun 2006)
Revision: 9263
Log message:

      It's FAM_KQUEUE instead of FAM_PSEUDO now.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-06 18:14:11 -0700 (Tue, 06 Jun 2006)
Revision: 9264
Log message:

      Bug 560 - print a better error message when an include file is not found.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-06 18:47:31 -0700 (Tue, 06 Jun 2006)
Revision: 9265
Log message:

      Actually, under the new scheme we need _both_ -DFAM_PSEUDO and -DFAM_KQUEUE
      when kqueue is used.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-07 14:45:38 -0700 (Wed, 07 Jun 2006)
Revision: 9267
Log message:

      Wrap all the code for the child processes in an unconditional
      "try ... with _ -> exit Omake_state.exn_error_code".
      
      TODO: figure out what was being raised and why.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-07 17:07:09 -0700 (Wed, 07 Jun 2006)
Revision: 9268
Log message:

      Call set_close_on_exec for the .omakedb.lock fd.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-07 18:08:53 -0700 (Wed, 07 Jun 2006)
Revision: 9270
Log message:

      Minor debugging changes

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-07 19:02:16 -0700 (Wed, 07 Jun 2006)
Revision: 9272
Log message:

      Backported the "while" loop.  It might be nice to have a "for"
      loop too, but I'm not exactly sure of a good syntax.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-07 19:49:36 -0700 (Wed, 07 Jun 2006)
Revision: 9273
Log message:

      Fall back to Root.om if OMakeroot does not exist.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-08 09:29:07 -0700 (Thu, 08 Jun 2006)
Revision: 9274
Log message:

      - Better handling of the "exit i" (the exit code used to be replaced with a
        "standard" one in several places).
      
      - Use the Omake_state.exn_error_code instead of hard-coded numbers in more
        places.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-08 11:40:33 -0700 (Thu, 08 Jun 2006)
Revision: 9275
Log message:

      - Small Shell.foo pipe formatting update.
      - Regenerated the documentation.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-08 16:08:48 -0700 (Thu, 08 Jun 2006)
Revision: 9276
Log message:

      Killing off an unused function.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-08 16:28:21 -0700 (Thu, 08 Jun 2006)
Revision: 9277
Log message:

      Killing off two more unused functions

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-08 16:30:35 -0700 (Thu, 08 Jun 2006)
Revision: 9278
Log message:

      Killing off two more unused functions

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-08 21:49:05 -0700 (Thu, 08 Jun 2006)
Revision: 9280
Log message:

      Added some (extensive) additional docs.  Rough draft.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-09 07:42:11 -0700 (Fri, 09 Jun 2006)
Revision: 9281
Log message:

      Regenerated the documentation.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-10 16:50:08 -0700 (Sat, 10 Jun 2006)
Revision: 9286
Log message:

      More documentation.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-11 12:49:43 -0700 (Sun, 11 Jun 2006)
Revision: 9287
Log message:

      Reorganized the documentation.
      
      This version uses HeVeA.  Don't get too grumpy, HeVeA is a lot
      better.  Some advantages:
         - macro definitions work
         - references work
         - commands work like you expect (including \verb/.../)
         - we get a TOC and index
      
      HeVeA also generates plain text and info files.  We should choose one
      of these for a man page.
      
      Some choices I made:
         - the document is documentclass{book}, so we start with
           \chapter, not \section.
      
         - there are multiple indexes, 1) for everything, 2) for vars,
           3) for functions, and 4) for targets.
      
           Use the macro \var{id} when you define a var.  This will
           add it to the index, and you can refer to it with
           \ref{var:id}.  \fun{id} and \target{id} work too.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-11 13:04:30 -0700 (Sun, 11 Jun 2006)
Revision: 9288
Log message:

      Oops, accidentally wiped the index.html.  Restored it.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-11 13:09:48 -0700 (Sun, 11 Jun 2006)
Revision: 9289
Log message:

      I like the ToC on each page better.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-11 14:10:52 -0700 (Sun, 11 Jun 2006)
Revision: 9290
Log message:

      Minor documentation changes.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-12 12:04:25 -0700 (Mon, 12 Jun 2006)
Revision: 9291
Log message:

      Added the --all-dependencies option.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-12 12:30:09 -0700 (Mon, 12 Jun 2006)
Revision: 9292
Log message:

      Added --verbose-dependencies.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-12 13:38:56 -0700 (Mon, 12 Jun 2006)
Revision: 9293
Log message:

      Man pages are no longer maintained.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-12 14:48:34 -0700 (Mon, 12 Jun 2006)
Revision: 9294
Log message:

      Added the hevea directory to TEXINPUTS.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-12 17:45:55 -0700 (Mon, 12 Jun 2006)
Revision: 9295
Log message:

      The info files are split into parts, and there is no way to know in advance
      how many parts we'd have. Hardcode the number of parts for now :-(
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-12 18:02:05 -0700 (Mon, 12 Jun 2006)
Revision: 9296
Log message:

      - Use Hevea's -fix option instead of resorting to manually repeating the hevea
        call.
      
      - Made the "clean" target more clean.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-12 18:09:38 -0700 (Mon, 12 Jun 2006)
Revision: 9297
Log message:

      Added the explicit list of files that will be generated by the hacha splitter.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-12 18:24:51 -0700 (Mon, 12 Jun 2006)
Revision: 9298
Log message:

      - It seems that the index package needs to go after the hyperref one in order
        for things to work correctly.
      
      - Made the "omake clean" mode clean.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-13 01:14:54 -0700 (Tue, 13 Jun 2006)
Revision: 9299
Log message:

      Added support for "v" and "h" options to built-in grep function.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-13 01:48:18 -0700 (Tue, 13 Jun 2006)
Revision: 9300
Log message:

      - Minor documentation fixes.
      
      - Made the documentation commands much less verbose:
        - Added \batchmode on top of omake-doc.tex
        - Pass "-s" to hevea (unless VERBOSE is set)
        - Pass "-q" to dvips (unless VERBOSE is set)
        - Pipe dvips output to a command that would drop all the stupid warnings
          that are generated because the .dvi file was built using the "dvipdfm"
          mode of hyperref (again, unless VERBOSE is set).
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-13 01:50:44 -0700 (Tue, 13 Jun 2006)
Revision: 9301
Log message:

      Minor message update.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-13 02:44:25 -0700 (Tue, 13 Jun 2006)
Revision: 9302
Log message:

      Added to "target" index:
      - Scoping sections for .SCANNER, .PHONY
      
      - The "file-sort" section that introduces the .ORDER and .BUILDORDER
        pseudo-targets.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-13 10:36:53 -0700 (Tue, 13 Jun 2006)
Revision: 9303
Log message:

      Added TARGETS documentation.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-13 12:05:24 -0700 (Tue, 13 Jun 2006)
Revision: 9304
Log message:

      Branching temporarily, so that I can configure a fixed ocamldep
      without disruption.  I want to get it working before I post
      the ocamldep patch.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-13 14:17:40 -0700 (Tue, 13 Jun 2006)
Revision: 9305
Log message:

      Initial version with a fixed ocamldep.
      The ocamldep bytecode is included in the ocamldep directory.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-13 16:30:16 -0700 (Tue, 13 Jun 2006)
Revision: 9306
Log message:

      Added support for Menhir.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-13 17:19:34 -0700 (Tue, 13 Jun 2006)
Revision: 9308
Log message:

      Pass the --ocamlc and --ocamldep options to menhir.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-13 17:53:19 -0700 (Tue, 13 Jun 2006)
Revision: 9309
Log message:

      Slight update to comments.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-13 18:45:06 -0700 (Tue, 13 Jun 2006)
Revision: 9310
Log message:

      Oops, ocaml is 3.09.2
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-14 13:07:35 -0700 (Wed, 14 Jun 2006)
Revision: 9311
Log message:

      Made the "+ <value>" and "+ <exp>" lines more verbose.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-14 15:32:40 -0700 (Wed, 14 Jun 2006)
Revision: 9312
Log message:

      Add a section:logic
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-14 15:39:27 -0700 (Wed, 14 Jun 2006)
Revision: 9313
Log message:

      Minor reference change.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-14 16:13:51 -0700 (Wed, 14 Jun 2006)
Revision: 9314
Log message:

      Added MenhirMulti, and some documentation.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-14 16:28:56 -0700 (Wed, 14 Jun 2006)
Revision: 9315
Log message:

      Fix the PREFIXED_OCAMLPACKS definition (forgot about it before).
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-14 17:56:42 -0700 (Wed, 14 Jun 2006)
Revision: 9317
Log message:

      Copied over the OCAMLPACKS fix (rev 9315 from the 0.9.8.ocamldep branch).
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-15 11:32:54 -0700 (Thu, 15 Jun 2006)
Revision: 9319
Log message:

      Various fixes for accurate dependencies:
         - .SCANNER targets should use
              :value: $(find-targets-in-path-optional ... $&)
           not
              :value: $(find-in-path-optional ... $&)
      
         - Call menhir with the --raw-depend option
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-15 14:01:08 -0700 (Thu, 15 Jun 2006)
Revision: 9321
Log message:

      Be more careful about escaping filenames.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-15 21:11:30 -0700 (Thu, 15 Jun 2006)
Revision: 9324
Log message:

      Added an icon.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-15 21:16:49 -0700 (Thu, 15 Jun 2006)
Revision: 9325
Log message:

      Small change to page headers.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-15 22:55:56 -0700 (Thu, 15 Jun 2006)
Revision: 9326
Log message:

      A slightly different take on the HTML header/footer.
      
      Jason, what do you think?
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-15 23:45:03 -0700 (Thu, 15 Jun 2006)
Revision: 9327
Log message:

      Small change to the header.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-16 10:18:41 -0700 (Fri, 16 Jun 2006)
Revision: 9329
Log message:

      It is ok to pass MENHIR_FLAGS even during scanning.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-16 10:26:38 -0700 (Fri, 16 Jun 2006)
Revision: 9330
Log message:

      Split the Menhir --ocamlc and --ocamldep options so we don't always pass both.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-16 10:46:34 -0700 (Fri, 16 Jun 2006)
Revision: 9331
Log message:

      Oops, it should be $+, not $^.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-16 11:47:07 -0700 (Fri, 16 Jun 2006)
Revision: 9333
Log message:

      Be more careful not to expand targets to strings when processing rules.
      
      Also, added a documentation section on the value->filename->string
      conversion in rules.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-16 13:29:35 -0700 (Fri, 16 Jun 2006)
Revision: 9335
Log message:

      "make all" should not call "omake doc"
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-16 13:34:26 -0700 (Fri, 16 Jun 2006)
Revision: 9336
Log message:

      The "jyh: this option needs to be documented in Chapter~\ref{chapter:rules}"
      is no longer relevant.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-16 14:05:51 -0700 (Fri, 16 Jun 2006)
Revision: 9338
Log message:

      Fixed the issue with implicit rules and filenames with whitespace.
      The wildcard % should be ValData, not ValString.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-16 14:26:20 -0700 (Fri, 16 Jun 2006)
Revision: 9339
Log message:

      [0.9.8.ocamldep merge] Fixing a documentation typo.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-16 14:27:35 -0700 (Fri, 16 Jun 2006)
Revision: 9340
Log message:

      [0.9.8.ocamldep merge] Adding new built-in functions: find-targets-in-path,
      find-targets-in-path-optional, string-escaped.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-17 09:38:18 -0700 (Sat, 17 Jun 2006)
Revision: 9341
Log message:

      Print all the errors for "multiple ways to build target".
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-17 10:26:45 -0700 (Sat, 17 Jun 2006)
Revision: 9342
Log message:

      Move the quick description to the Quickstart guide.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-18 11:15:52 -0700 (Sun, 18 Jun 2006)
Revision: 9343
Log message:

      Pass rule options as a Map, not an array of length-2 arrays
      (which is not a sensible value anyway).
      
      Also, we weren't expanding :value: dependencies in
      Omake_rule.eval_rule.  Suppose we had a rule like the
      following.
      
         a: b :value: $X :exists: foo
            ...
      
      The dependency is parsed as <sequence $`X " ">.  This is
      a primitive value, so the $`X was never evaluated (and
      so changing the value of X had no effect).
      
      To do it right, flatten the values with values_of_value,
      then take the primitive values.  You will see that the
      value counts are now much larger.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-18 11:25:34 -0700 (Sun, 18 Jun 2006)
Revision: 9344
Log message:

      Regenerate the documentation.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-18 19:37:52 -0700 (Sun, 18 Jun 2006)
Revision: 9345
Log message:

      
      Added --divert options.
      
         --divert : divert output from commands so it can be printed later.
         --divert-repeat : repeat diverted output when a command is finished.
         --divert-only : print only diverted output, omake will be quiet otherwise.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-19 12:37:00 -0700 (Mon, 19 Jun 2006)
Revision: 9347
Log message:

      A number of documentation changes:
       - Made subsubsections numbered and added them to table of contents.
      
       - Since the "Build Rules" chapter ("omake-root") is nested one level deeper,
         change the \fun,\var,etc functions to generate \subsubsection instead of
         \subsection while inside this chapter.
      
       - Added macros \twofuns{...}{...} and \threefuns{...}{...}{...} for the case
         of a \fun section that introduces several functions at once.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-20 13:10:23 -0700 (Tue, 20 Jun 2006)
Revision: 9349
Log message:

      Oops, add back the output options to OMakeFlags.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-20 16:07:36 -0700 (Tue, 20 Jun 2006)
Revision: 9353
Log message:

      Following Aleksey's suggestion, added the --divert-discard-success option.
      
      Also added the -o option so you can specify output control in short form.
      See the manual for more info.
      
         # Be very, very, quiet
         % omake -o 0
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-20 16:09:33 -0700 (Tue, 20 Jun 2006)
Revision: 9354
Log message:

      Oops, put a line in the wrong chapter.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-20 17:16:35 -0700 (Tue, 20 Jun 2006)
Revision: 9356
Log message:

      Be more clean on "omake clean"
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-20 20:42:59 -0700 (Tue, 20 Jun 2006)
Revision: 9361
Log message:

      I have a lot of trouble making the "ocamldep -modules" stuff (0.9.8.ocamldep
      branch marge) to work, so I'll create a new branch with my current state of
      things to explore it further.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-20 20:47:24 -0700 (Tue, 20 Jun 2006)
Revision: 9362
Log message:

      Merged the 0.9.8.ocamlde branch (with LOTS of manual changes).
      
      Some very weird things are happening when I try using it - so I want to
      checkpoint what I currently have to make it easier to debug it.
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-20 20:48:18 -0700 (Tue, 20 Jun 2006)
Revision: 9363
Log message:

      The old 0.9.8.ocamldep branch is now closed.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-22 12:31:30 -0700 (Thu, 22 Jun 2006)
Revision: 9366
Log message:

      Added a print statement and rearranged some code.
      
      I would say "elseif" is stylistically preferable to
      
         else
            if
      
      if only to keep the indentation under control.
      
      Note the following:
         $(target-exists lm_printf.mli) == true
         $(target-exists lm_printf.cmi) == false
      

Changes by: Nathaniel Gray (n8gray at caltech.edu)
Date: 2006-06-22 15:45:32 -0700 (Thu, 22 Jun 2006)
Revision: 9371
Log message:

      Added a .PHONY section to the default OMakefile.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-22 21:19:33 -0700 (Thu, 22 Jun 2006)
Revision: 9373
Log message:

      Resolved the "strage" bug (see the discussion).
      
      http://lists.metaprl.org/pipermail/omake-devel/2006-June/000261.html
      
      Here were the issues:
         1. target-exists and friends should not be called until the
            build phase (after all OMakefiles are evaluated).  This
            was not the source of the error.
      
         2. When testing whether a target exists, we must switch to
            the environment (the directory) for the target.
      
      For #2, suppose we want to build a/a.boo, and
      a/OMakefile contains a rule
      
         # a/OMakefile
         %.boo: ../b/%.boo
      
      and b/OMakefile has
      
         # b/OMakefile
         %.boo: %.moo
            ...
      
      Basically, we must use the appropriate OMakefile when determining
      whether a target can be built so a/a.boo <- b/b.boo <- b/b.moo
      (assuming b.moo can be built).
      
      As usual for implicit rules, if the dependencies are cyclic,
      omake will loop.  Say, b/OMakefile instead contains
      
         %.boo: ../a/%.boo
             ...
      
      The implicit dep computation is Turing complete, so a complete
      check for termination is undecidable.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-23 12:58:58 -0700 (Fri, 23 Jun 2006)
Revision: 9376
Log message:

      Added the bounds-check.  Implicit calculations are not guaranteed
      to terminate in general, but simple cycles will be rejected.
      A simple cycle is one where a file depends on itself (through
      some chain of dependencies).
      

Changes by: Aleksey Nogin (nogin at cs.caltech.edu)
Date: 2006-06-23 14:00:18 -0700 (Fri, 23 Jun 2006)
Revision: 9380
Log message:

      Make sure the OCaml generated files are created before any of the Makefiles.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-23 14:30:21 -0700 (Fri, 23 Jun 2006)
Revision: 9382
Log message:

      We still have some bogus filter-targets lying around.  Temporarily disable
      the check.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-24 17:35:56 -0700 (Sat, 24 Jun 2006)
Revision: 9388
Log message:

      Support single-letter lazy-variables $`x, $`@ (issue #581).
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-24 18:37:20 -0700 (Sat, 24 Jun 2006)
Revision: 9389
Log message:

      which/where should use Omake_cache, not Lm_filename_util (issue #467).
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-25 12:01:41 -0700 (Sun, 25 Jun 2006)
Revision: 9390
Log message:

      More minor cleanup, addressing issue #516.
      
      This adds the phony targets
         .BUILD_BEGIN
         .BUILD_SUCCESS
         .BUILD_FAILURE
      
      .BUILD_BEGIN is built before anything else in your project.
      One of .BUILD_SUCCESS or .BUILD_FAILURE is built when the
      build for your project terminates.
      
      These are normal targets/rules.  If you want to use them, you
      should probably add them to your ~/.omakerc, rather than adding
      them directly to each of the projects you work on.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-25 12:50:56 -0700 (Sun, 25 Jun 2006)
Revision: 9391
Log message:

      Added some documentation for glob syntax (issue #517).
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-25 13:19:18 -0700 (Sun, 25 Jun 2006)
Revision: 9392
Log message:

      Adjusted documentation of -name (issue #552).
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-25 18:32:25 -0700 (Sun, 25 Jun 2006)
Revision: 9393
Log message:

      Don't open the diversion on 0-byte writes.  This is what was causing
      the progress bar to flash.
      
      Also, save the .omakedb after the .BUILD_* targets (it seems like a
      better idea, in case they do some other building).
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-25 18:59:07 -0700 (Sun, 25 Jun 2006)
Revision: 9394
Log message:

      Also divert stderr from .SCANNER commands.
      This is why "-o 0" was still printing output from successful
      LaTeX commands.
      
      BTW, the choice of "-o" is a little weird.  Maybe "-v"
      or "-V" or "-d"?
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-25 19:40:29 -0700 (Sun, 25 Jun 2006)
Revision: 9395
Log message:

      Merged the ocamldep2 branch.  It seems stable with the implicit-deps
      commit:
      
      http://svn.metaprl.org/commitlogs/omake/2006-06.html#06/06/22.21:19:33
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-25 19:48:41 -0700 (Sun, 25 Jun 2006)
Revision: 9397
Log message:

      Add OCAMLDEP_MODULES_ENABLED to the .config.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-26 19:21:21 -0700 (Mon, 26 Jun 2006)
Revision: 9401
Log message:

      Added variable qualifiers in an attempt to bring the .om files
      closer to the ones in 0.9.9.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-26 19:30:21 -0700 (Mon, 26 Jun 2006)
Revision: 9402
Log message:

      Merged
      
          svn merge -r 9083:9392 svn://svn.metaprl.org/svnroot/mojave/omake-branches/0.9.8.x
      
      with many hand edits.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-26 20:15:50 -0700 (Mon, 26 Jun 2006)
Revision: 9403
Log message:

      More compatibility changes with 0.9.9
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-26 20:58:51 -0700 (Mon, 26 Jun 2006)
Revision: 9404
Log message:

      Regenerate the documentation.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-28 11:09:50 -0700 (Wed, 28 Jun 2006)
Revision: 9407
Log message:

      Forgot the omake-detail.tex file (Issue #583), but it is useful.
      Added it back, with some editing to remove duplicate text.
      

Changes by: Jason Hickey (jyh at cs.caltech.edu)
Date: 2006-06-28 13:47:47 -0700 (Wed, 28 Jun 2006)
Revision: 9408
Log message: