Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 10:02:29 -0700 (Mon, 02 Jul 2007)
Revision: 11185
Log message:
Implement the --configure for the .STATIC sections correctly. Previous
implementation would simply fail to consult the Omake_cache when --configure
is given. However, the Omake_cache is used both for the caching of the .STATIC
sections to disk and for the "memoization" of the .STATIC sections, while only
the former function is supposed to be affected by --configure.
The new implementation only checks for the --configure option when the cache
is loaded from disk - and drops the loaded value table if --configure is
present.
This commit also pushes the processing of all the cache-related options
(including -U and --depend) into Omake_cache.from_channel, instead of having
each caller of Omake_cache.from_channel (currently we have two) have to do it
independently.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 11:20:23 -0700 (Mon, 02 Jul 2007)
Revision: 11186
Log message:
Added the .MEMO rules that are very similar to the .STATIC ones, except they
are not cached in .omakedb
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 11:40:43 -0700 (Mon, 02 Jul 2007)
Revision: 11187
Log message:
- Run all tests with OMAKEFLAGS environment variable set to "--no--progress --no--print-status"
in order to avoid any spurious errors due to variations in status output.
- Run the .STATIC tests using omake instead of osh to test the .omakedb
caching. Run omake twice (omake --configure; omake) for complete testing.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 11:45:27 -0700 (Mon, 02 Jul 2007)
Revision: 11188
Log message:
Run all the .MEMO tests using omake instead of osh to test any possible
misinteractions with the .omakedb caching. Run omake twice (omake
--configure; omake) for complete testing.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 12:12:29 -0700 (Mon, 02 Jul 2007)
Revision: 11189
Log message:
Added Omake_util.pp_time that would print the time as
"%0.2f sec", or
"%d min %05.2f sec", or
"%d hrs %02d min %05.2f sec"
depending on how big it is. This follows Sam's (sds at gnu.org) suggestion
from bug #680.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 13:42:03 -0700 (Mon, 02 Jul 2007)
Revision: 11190
Log message:
When processing the :key: dependency, we need to call the value_of_values in
order to support the situation, where the key is a sequence containing arrays.
Changes | Path |
+7 -8 | omake-branches/0.9.8.x/src/builtin/omake_builtin_target.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 13:56:31 -0700 (Mon, 02 Jul 2007)
Revision: 11191
Log message:
When processing interactive commands, define the __FILE__ variable (to a fake
phony) in order for the .STATIC/.MEMO sections to work.
Changes | Path |
+1 -0 | omake-branches/0.9.8.x/src/main/omake_shell.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 14:14:28 -0700 (Mon, 02 Jul 2007)
Revision: 11192
Log message:
Test case for bug 682.
Changes | Path |
Copied | omake-branches/0.9.8.x/test/memo/Test4/ (from rev 11186, omake-branches/0.9.8.x/test/memo/Test1) |
+1 -1 | omake-branches/0.9.8.x/test/memo/Test4/run.osh |
Deleted | omake-branches/0.9.8.x/test/memo/Test4/static.om |
+3 -7 | omake-branches/0.9.8.x/test/memo/Test4/stdout.expected |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 14:18:04 -0700 (Mon, 02 Jul 2007)
Revision: 11193
Log message:
Forgot to commit the actual test.
Changes | Path |
Added | omake-branches/0.9.8.x/test/memo/Test4/memo.om |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 14:41:17 -0700 (Mon, 02 Jul 2007)
Revision: 11194
Log message:
Regenerated the documentation.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 14:50:57 -0700 (Mon, 02 Jul 2007)
Revision: 11195
Log message:
Fixing a typo.
Changes | Path |
+1 -1 | omake-branches/0.9.8.x/doc/src/omake-rules.tex |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-02 17:20:37 -0700 (Mon, 02 Jul 2007)
Revision: 11197
Log message:
Minor correction in an error message.
Changes | Path |
+1 -1 | omake-branches/0.9.8.x/src/eval/omake_value.ml |
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-02 18:28:37 -0700 (Mon, 02 Jul 2007)
Revision: 11198
Log message:
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-03 09:57:17 -0700 (Tue, 03 Jul 2007)
Revision: 11203
Log message:
Adding a simple test (it passes fine, I just want to make sure it does so in
the future).
Changes | Path |
Properties | omake-branches/0.9.8.x/test/simple/ |
Added | omake-branches/0.9.8.x/test/simple/Test5 |
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-03 19:53:00 -0700 (Tue, 03 Jul 2007)
Revision: 11206
Log message:
Moved in hooks for where I'll put my server. It looks like it will take nontrivial effort to weed out the old way of doing things.
Changes | Path |
+3 -3 | omake-research/omake-distrib/remote_schema |
+74 -0 | omake-research/omake-distrib/src/exec/omake_exec_remote.ml |
+0 -0 | omake-research/omake-distrib/src/exec/omake_exec_type.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-05 12:54:15 -0700 (Thu, 05 Jul 2007)
Revision: 11210
Log message:
Bug 684 - fixing the sorting function.
Changes | Path |
+45 -58 | omake-branches/0.9.8.x/src/build/omake_build_util.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-06 10:35:09 -0700 (Fri, 06 Jul 2007)
Revision: 11214
Log message:
Rev. 11158 follow-up. Jason have missed "public.Sequence = Sequence".
Quoting from Jason's rev 11158 log message:
This is an annoying naming problem. It does not exist
in var3 onward.
Changes | Path |
+1 -0 | omake-branches/0.9.8.x/lib/Pervasives.om |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-06 10:59:16 -0700 (Fri, 06 Jul 2007)
Revision: 11215
Log message:
Bug: the Array.foreach function does not preserve exports!
Changes | Path |
Added | omake-branches/0.9.8.x/test/object/Test3 |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-06 11:02:48 -0700 (Fri, 06 Jul 2007)
Revision: 11216
Log message:
Bug: the Map.foreach function does not preserve exports!
Changes | Path |
+1 -0 | omake-branches/0.9.8.x/test/object/Test3 |
Copied | omake-branches/0.9.8.x/test/object/Test4 (from rev 11215, omake-branches/0.9.8.x/test/object/Test3) |
+12 -8 | omake-branches/0.9.8.x/test/object/Test4 (from rev 11215, omake-branches/0.9.8.x/test/object/Test3) |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-06 11:29:15 -0700 (Fri, 06 Jul 2007)
Revision: 11217
Log message:
Make sure the Object.object-foreach, Array.foreach and Map.foreach preserve
the exported environment.
Changes | Path |
+5 -2 | omake-branches/0.9.8.x/lib/Pervasives.om |
+1 -1 | omake-branches/0.9.8.x/src/builtin/omake_builtin_object.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-06 11:45:15 -0700 (Fri, 06 Jul 2007)
Revision: 11218
Log message:
A follow-up fix for Object.object-foreach and Map.foreach - the previous one
was incomplete.
Changes | Path |
+5 -5 | omake-branches/0.9.8.x/src/builtin/omake_builtin_object.ml |
Added | omake-branches/0.9.8.x/test/object/Test5 |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-12 15:15:01 -0700 (Thu, 12 Jul 2007)
Revision: 11245
Log message:
Adding a string-length function.
Changes | Path |
+29 -0 | omake-branches/0.9.8.x/src/builtin/omake_builtin_base.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-12 15:30:17 -0700 (Thu, 12 Jul 2007)
Revision: 11247
Log message:
Bug 685 - sprintf is buggy.
Changes | Path |
Added | omake-branches/0.9.8.x/test/simple/Test6 |
Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-07-15 09:52:43 -0700 (Sun, 15 Jul 2007)
Revision: 11272
Log message:
Do not force evaluation on variable references.
That is, 0-arity applications should not be forced
in Omake_eval.eval_var_export and eval_var.
This addresses bug #682.
Changes | Path |
+3 -2 | omake-branches/0.9.8.x/src/eval/omake_eval.ml |
Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-07-15 10:56:10 -0700 (Sun, 15 Jul 2007)
Revision: 11275
Log message:
Add test/simple/Test7 to test sprintf truncation with
a somewhat easier to understand output.
Changes | Path |
+1 -1 | omake-branches/0.9.8.x/src/builtin/omake_builtin_io.ml |
Added | omake-branches/0.9.8.x/test/simple/Test7 |
Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-07-15 12:21:56 -0700 (Sun, 15 Jul 2007)
Revision: 11276
Log message:
Document the .MEMO and :key:
Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-07-15 14:30:00 -0700 (Sun, 15 Jul 2007)
Revision: 11277
Log message:
Small optimization in Omake_cache. For the index, use (NodeSet.t IntTable.t)
instead of (Node.t IntMTable.t), since the former is less likely to leak
space.
Changes | Path |
+25 -10 | omake-branches/0.9.8.x/src/ir/omake_cache.ml |
Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-07-16 11:50:03 -0700 (Mon, 16 Jul 2007)
Revision: 11278
Log message:
inotify-test.c doesn't exist anymore?
Changes | Path |
+1 -3 | omake-branches/0.9.8.4/src/Makefile |
+1 -3 | omake-branches/0.9.8.4/src/Makefile.nt |
+1 -1 | omake-branches/0.9.8.4/src/clib/OMakefile |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-16 13:44:08 -0700 (Mon, 16 Jul 2007)
Revision: 11281
Log message:
Made the locking of the .omc files even more finegrained than before.
(I was having a locking problem, which turned out to be caused by
misinteractions of the NLM with the firewall, but this change is probably a
good idea anyway).
Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-07-16 16:14:35 -0700 (Mon, 16 Jul 2007)
Revision: 11283
Log message:
Merge 11106 -> 11281. Untested.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-17 15:01:15 -0700 (Tue, 17 Jul 2007)
Revision: 11305
Log message:
- [Bug 683] Added the FILENAME and FNR variables to awk.
- Added a channel-name and Channel.name functions.
Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-07-17 15:21:26 -0700 (Tue, 17 Jul 2007)
Revision: 11307
Log message:
Added the tests, without .svnignore for now.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-18 09:59:55 -0700 (Wed, 18 Jul 2007)
Revision: 11316
Log message:
Use "delayed:memo" instead of "delayed:static" in value printouts of
static/memo variables.
Changes | Path |
+1 -1 | omake-branches/0.9.8.x/src/ir/omake_value_print.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-18 11:32:43 -0700 (Wed, 18 Jul 2007)
Revision: 11320
Log message:
- Make sure that Sys.Break (and some other exceptions) are never wrapped in an
UncaughtException
- When the build is interrupted with Sys.Break, always report the
exn_error_code, _except_ when this have happened during the -P polling.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-18 11:35:53 -0700 (Wed, 18 Jul 2007)
Revision: 11321
Log message:
Oops, my previous commit was broken, sorry.
Changes | Path |
+1 -1 | omake-branches/0.9.8.x/src/build/omake_build.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-18 15:12:48 -0700 (Wed, 18 Jul 2007)
Revision: 11333
Log message:
[Bug 535] Some improvements of Ctrl-C handling on Windows. Often it will still break in weird ways, but now it works correctly more often than before.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-18 15:24:57 -0700 (Wed, 18 Jul 2007)
Revision: 11334
Log message:
Adding a -debug-thread operation.
Changes | Path |
+2 -0 | omake-branches/0.9.8.x/src/main/omake_main.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-18 17:30:29 -0700 (Wed, 18 Jul 2007)
Revision: 11337
Log message:
Need to define public.Node before var3 lands.
Changes | Path |
+1 -0 | omake-branches/0.9.8.x/lib/Pervasives.om |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-18 17:31:14 -0700 (Wed, 18 Jul 2007)
Revision: 11338
Log message:
There is still some nastiness with map-map :-(
Changes | Path |
Added | omake-branches/0.9.8.x/test/object/Test6 |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-18 17:39:40 -0700 (Wed, 18 Jul 2007)
Revision: 11342
Log message:
This hits a really strange bug under 0.9.8.4, but works correctly under
0.9.8.x
Changes | Path |
Added | omake-branches/0.9.8.x/test/object/Test7 |
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-18 23:51:49 -0700 (Wed, 18 Jul 2007)
Revision: 11349
Log message:
I think this may work, a few bugs still...
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-18 23:53:42 -0700 (Wed, 18 Jul 2007)
Revision: 11350
Log message:
Whoops, forgot this one
Changes | Path |
Added | omake-research/omake-distrib/src/shell/omake_shell_servers.ml |
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-18 23:55:21 -0700 (Wed, 18 Jul 2007)
Revision: 11351
Log message:
Err, sorry, and another
Changes | Path |
Added | omake-research/omake-distrib/src/main/omake_main_args.ml |
Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-07-19 09:06:49 -0700 (Thu, 19 Jul 2007)
Revision: 11352
Log message:
Exports were not being hoisted correctly. This fixes objects/Test6
Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-07-19 09:15:27 -0700 (Thu, 19 Jul 2007)
Revision: 11353
Log message:
Fix export hoisting for specific vars too.
Changes | Path |
+3 -2 | omake-branches/0.9.8.x/src/env/omake_env.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-19 09:51:20 -0700 (Thu, 19 Jul 2007)
Revision: 11354
Log message:
Another example, where map-map still does not work right :-(
Changes | Path |
Added | omake-branches/0.9.8.x/test/object/Test8 |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-19 12:19:39 -0700 (Thu, 19 Jul 2007)
Revision: 11356
Log message:
Trying to debug the deadlocks I am seeing on Windows when -p is enabled. What I am seeing is that the worker threads are being created, but they never start (even the child part of the Lm_thread.create_thread is never executed). I have no idea how the -p causes this...
In this commit I:
- Added a lot of thread-related debugging code
- Made the -debug-thread command-line option also activate Lm_thread.debug_lock and Lm_thread_code.debug_mutex (the latter variable is new).
- Changed the Lm_thread_core_null so that Lm_thread.Mutex.try_lock would return a value consisntent with the real thread implementation.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-19 12:34:35 -0700 (Thu, 19 Jul 2007)
Revision: 11358
Log message:
Choose between omake_shell_sys_unix and omake_shell_sys_win32 based on which
OS is being used, not based on whether the threads are enabled (since we might
want to enable them on Unix - e.g. for debugging).
Changes | Path |
+1 -1 | omake-branches/0.9.8.x/src/shell/OMakefile |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-19 13:13:39 -0700 (Thu, 19 Jul 2007)
Revision: 11359
Log message:
Minor no-op change - have Lm_notify.file_descr return an option and leave it
up to Omake_exec_notify to fake it with a Unix.stdin, if necessary.
Changes | Path |
+3 -4 | libmojave/unix/lm_notify.ml |
+3 -6 | libmojave/unix/lm_notify.mli |
+7 -1 | omake-branches/0.9.8.x/src/exec/omake_exec_notify.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-26 09:30:29 -0700 (Thu, 26 Jul 2007)
Revision: 11448
Log message:
Fixes the object/Test8 by having the hoisting check that the head variable was
not already shadowed (as I proposed on the mailing list).
This might not be the right way - please feel free to revert if you disagree
with this approach.
Changes | Path |
+20 -4 | omake-branches/0.9.8.x/src/env/omake_env.ml |
+1 -1 | omake-branches/0.9.8.x/src/env/omake_env.mli |
+1 -1 | omake-branches/0.9.8.x/src/eval/omake_eval.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-26 09:36:54 -0700 (Thu, 26 Jul 2007)
Revision: 11449
Log message:
In OCamlPackage function, check whether the .mli corresponding to the package
is buildable. If so, try doing the right thing (as requested by Mike Furr,
furr at cs.umd.edu). Untested.
Changes | Path |
+36 -14 | omake-branches/0.9.8.x/lib/build/OCaml.om |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-26 10:12:09 -0700 (Thu, 26 Jul 2007)
Revision: 11450
Log message:
Regenrated the documentation.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-26 11:00:04 -0700 (Thu, 26 Jul 2007)
Revision: 11451
Log message:
Added a -regex option to find.
Changes | Path |
+11 -3 | omake-branches/0.9.8.x/src/builtin/omake_builtin_test.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-26 11:01:01 -0700 (Thu, 26 Jul 2007)
Revision: 11452
Log message:
- Allow test names to have more to have any number of digits, not just one.
- Added object/Test09 and object/Test10; Test10 currently fails :-(
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-26 11:03:07 -0700 (Thu, 26 Jul 2007)
Revision: 11454
Log message:
Another test that currently fails :-(
Changes | Path |
Added | omake-branches/0.9.8.x/test/object/Test11 |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-26 11:04:43 -0700 (Thu, 26 Jul 2007)
Revision: 11455
Log message:
Ignore result.log
Changes | Path |
Properties | omake-branches/0.9.8.x/test/parse/C/ |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-26 16:13:18 -0700 (Thu, 26 Jul 2007)
Revision: 11460
Log message:
Minor documentation fix.
Changes | Path |
+1 -3 | omake-branches/0.9.8.x/lib/Pervasives.om |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-26 16:28:42 -0700 (Thu, 26 Jul 2007)
Revision: 11461
Log message:
Fixed the TeTeX 2 autodetection.
Changes | Path |
+3 -3 | omake-branches/0.9.8.x/lib/build/LaTeX.om |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-26 16:56:22 -0700 (Thu, 26 Jul 2007)
Revision: 11462
Log message:
Adding support for building shared libraries (.so/.dll). Only minimally
tested.
Changes | Path |
+69 -21 | omake-branches/0.9.8.x/lib/build/C.om |
+4 -1 | omake-branches/0.9.8.x/lib/build/Common.om |
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-27 12:08:26 -0700 (Fri, 27 Jul 2007)
Revision: 11468
Log message:
Everything seems to work except pick_server, though I'm sure once that's working more bugs will appear.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-27 12:34:17 -0700 (Fri, 27 Jul 2007)
Revision: 11469
Log message:
Wrap more calls to Lm_notify in Lm_thread_pool.blocking_section
Changes | Path |
+8 -13 | omake-branches/0.9.8.x/src/exec/omake_exec_notify.ml |
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-27 14:16:45 -0700 (Fri, 27 Jul 2007)
Revision: 11474
Log message:
Switched server list to a queue, butthis didn't seem to fix anything.
Changes | Path |
+9 -21 | omake-research/omake-distrib/src/shell/omake_shell_servers.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-27 14:35:52 -0700 (Fri, 27 Jul 2007)
Revision: 11475
Log message:
- When exiting from -P via Ctrl-C, print a message.
- A test showing that caching may need to be a bit more precise.
Changes by: Jason J. Hickey (jyh at cs.caltech.edu)
Date: 2007-07-27 15:21:18 -0700 (Fri, 27 Jul 2007)
Revision: 11477
Log message:
Merge to rev #11462. The test/object/Test0X got lost, but it is very hard
to propagate additions to all jumbo branches. Probably we should just do
svn rm XXX/test/object
svn cp 0.9.8.x/test/object XXX/test/object
for each of the jumbos.
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-30 11:32:57 -0700 (Mon, 30 Jul 2007)
Revision: 11480
Log message:
It appears to work. Next, time to port this to ocaml 3.10
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-30 12:44:11 -0700 (Mon, 30 Jul 2007)
Revision: 11482
Log message:
Fixed a few things to deal with a new libmojave
Changes | Path |
+6 -1 | omake-research/omake-distrib/src/exec/omake_exec_notify.ml |
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-30 14:25:53 -0700 (Mon, 30 Jul 2007)
Revision: 11483
Log message:
Added the hack to make the execserver try to thread ALL the jobs rather than just those on localhost, so our scheduler actually gets the proper number of requests
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-30 15:05:47 -0700 (Mon, 30 Jul 2007)
Revision: 11484
Log message:
Err, these were missing
Changes | Path |
Added | omake-research/omake-distrib/src/aux/nfs_flushr.ml |
Added | omake-research/omake-distrib/src/aux/nfs_flushw.ml |
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-30 15:48:17 -0700 (Mon, 30 Jul 2007)
Revision: 11485
Log message:
Possibly an error in my test.
Changes | Path |
+1 -1 | omake-research/file-consistency/filetestlib.ml |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-31 10:42:56 -0700 (Tue, 31 Jul 2007)
Revision: 11493
Log message:
Merged the "SharedCLibrary" mechanism that I've added on the 0.9.8.x branch
with Jason's DynamicCLibaray mechanism. The merged functionality still uses
the SharedCLibrary name schema and is compatible with the 0.9.8.x branch,
except that the *Copy and *Install functions will now take a keyword "version"
argument (defaults to empty, which would cause the functions to only build the
unversioned DLLs). The version argument will be ignored on Windows.
Jason, could you please propagate this to child branches? Thanks!
P.S. There were a number of things that appear broken - possibly due to
incomplete "jumbo-branch" propagation (for example, many configure test leave
their temporary outputs around). I did not try to investigate those - it will
probably be easier to fix them as "jumbo patches" land.
Changes by: Matthew Maurer (maurer at caltech.edu)
Date: 2007-07-31 12:01:40 -0700 (Tue, 31 Jul 2007)
Revision: 11494
Log message:
Most things working. Still have the problems with libmojave and hevea, but it works on the test setup.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-31 13:10:41 -0700 (Tue, 31 Jul 2007)
Revision: 11496
Log message:
[Bug 535] Added a hack to address the "Sys.Break does not get raised inside C
blocking sections" problem (see
http://caml.inria.fr/pub/ml-archives/caml-list/2007/07/3662ad69f77253674f580b174c85dfbb.en.html
for detail).
My "solution" is that when threads are enabled, the Lm_notify.next_event would
run in a child thread, while the master thread periodically (currently - at
0.1s intervals) checks whether it's done. If Ctrl-C is pressed, OCaml will
interrupt the master thread with Sys.Break, which is what we want.
Note that simply having the master thread wait for the child one does not work
- we need the master thread to execute caml_leave_blocking_section()
regularly!
P.S. I have only tested this on Unix with threads enabled, will test on
Windows next.
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-31 14:42:53 -0700 (Tue, 31 Jul 2007)
Revision: 11500
Log message:
Fixed a problem with static environment being wiped out by hoisting.
Changes | Path |
+1 -1 | omake-branches/0.9.8.x/src/env/omake_env.ml |
+2 -0 | omake-branches/0.9.8.x/test/object/Test09 |
Copied | omake-branches/0.9.8.x/test/object/Test13 (from rev 11479, omake-branches/0.9.8.x/test/object/Test09) |
+3 -1 | omake-branches/0.9.8.x/test/object/Test13 (from rev 11479, omake-branches/0.9.8.x/test/object/Test09) |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-31 14:53:57 -0700 (Tue, 31 Jul 2007)
Revision: 11501
Log message:
This is a variation of the Test8, where the variable being tested is a
function argument (in "protected" scope), not a global one. In this case, we
still get "Error: expected new, got old" with 0.9.8.x :-(
Changes | Path |
Added | omake-branches/0.9.8.x/test/object/Test14 |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-31 15:12:41 -0700 (Tue, 31 Jul 2007)
Revision: 11503
Log message:
Made it a bit more obvious what this is about.
Changes | Path |
+2 -1 | omake-branches/0.9.8.x/test/object/Test14 |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-31 16:36:53 -0700 (Tue, 31 Jul 2007)
Revision: 11504
Log message:
Solves the "this diambiguation" problem (destinguishing between the "this"
variables that came from the original environment and should stay there and the
"this" variables that are object fields and should be hoisted). The
implementation is probably inefficient, will try to optimize it next.
P.S. test/object/Test14 still fails as there is a clash between the map
variable and the "map" field of the Array object. A similar
test/object/Test15, which uses "M" instead of "map" now succeeds.
Changes | Path |
+43 -20 | omake-branches/0.9.8.x/src/env/omake_env.ml |
+1 -1 | omake-branches/0.9.8.x/src/eval/omake_eval.ml |
Copied | omake-branches/0.9.8.x/test/object/Test15 (from rev 11503, omake-branches/0.9.8.x/test/object/Test14) |
+5 -4 | omake-branches/0.9.8.x/test/object/Test15 (from rev 11503, omake-branches/0.9.8.x/test/object/Test14) |
Changes by: Aleksey Nogin (nogin at metaprl.org)
Date: 2007-07-31 16:42:31 -0700 (Tue, 31 Jul 2007)
Revision: 11505
Log message:
The environment part of the PathVar (which used to contain the original value
of the object) is never actually needed, so deleting it from the type.