Rodrigo Kumpera [Mon, 23 May 2011 04:11:44 +0000 (01:11 -0300)]
Reorder fields to improve object layout since the runtime can't do it for corlib types.
Zoltan Varga [Sat, 21 May 2011 13:32:31 +0000 (15:32 +0200)]
Fix some warnings.
Zoltan Varga [Sat, 21 May 2011 13:25:23 +0000 (15:25 +0200)]
Disable the -Wunused-by-set-variable warning supported by newer gcc versions.
Zoltan Varga [Sat, 21 May 2011 12:52:40 +0000 (14:52 +0200)]
Add an assert.
Mark Probst [Fri, 20 May 2011 21:05:10 +0000 (23:05 +0200)]
[tests] Add test case for large vtypes in SGen.
Mark Probst [Fri, 20 May 2011 20:27:17 +0000 (22:27 +0200)]
[sgen] Handle all cases in major_scan_vtable().
Mark Probst [Fri, 20 May 2011 19:35:38 +0000 (21:35 +0200)]
[sgen] Don't pass vtables to FOREACH macros when the descriptor is enough.
Mark Probst [Fri, 20 May 2011 19:24:36 +0000 (21:24 +0200)]
[sgen] Remove superfluous arguments from minor_scan_vtype().
Mark Probst [Fri, 20 May 2011 19:16:09 +0000 (21:16 +0200)]
[sgen] Remove small bitmap descriptor type.
The large bitmap can handle everything the small bitmap does and
there's no performance penalty, so there is no reason to keep it
around.
Rodrigo Kumpera [Fri, 20 May 2011 18:41:08 +0000 (15:41 -0300)]
Fix aligninig copy and improve overlap detection
Mark Probst [Fri, 20 May 2011 17:21:37 +0000 (19:21 +0200)]
[sgen] Fix a compiler warning.
Marek Safar [Fri, 20 May 2011 16:59:26 +0000 (17:59 +0100)]
Add more checks for async modifier
Marek Safar [Fri, 20 May 2011 16:55:36 +0000 (17:55 +0100)]
Rename Task<T> implementation file
Rodrigo Kumpera [Fri, 20 May 2011 16:32:47 +0000 (13:32 -0300)]
Apparently I'm too stupid to implement memmove correctly, unbreak the build while I learn how to do it.
Marek Safar [Fri, 20 May 2011 08:35:15 +0000 (09:35 +0100)]
Pass first async test
Miguel de Icaza [Fri, 20 May 2011 01:18:52 +0000 (18:18 -0700)]
Merge pull request #102 from konrad-kruczynski/fix_bug_635971
[Fix] #635971 Fixed throwing exception in case of using zero timeout. Now
Rodrigo Kumpera [Thu, 19 May 2011 22:16:33 +0000 (19:16 -0300)]
Fix mono_gc_memmove backward copy.
Rodrigo Kumpera [Thu, 19 May 2011 20:28:25 +0000 (17:28 -0300)]
Fix tp_poll_wait to work with sgen
Rodrigo Kumpera [Thu, 19 May 2011 20:28:11 +0000 (17:28 -0300)]
Add mono_ptr_array_reset and mono_ptr_array_clear
Rodrigo Kumpera [Thu, 19 May 2011 19:42:45 +0000 (16:42 -0300)]
Use new mono_gc_(memmove|bzero) in places messing with managed memory.
Rodrigo Kumpera [Wed, 18 May 2011 19:04:24 +0000 (16:04 -0300)]
Add new API to copy managed memory.
* gc.c: Add mono_gc_bzero and mono_gc_memmove. Those
functions do word sized load/stores so they are GC safe.
Those oddly seeing functions are required as libc equivalents
don't ensure that copies and done at least in word granularity
so we can end up pausing a thread while is has, for example,
zeroed just the first byte of a managed reference, causing the
GC to see an invalid pointer.
This happens on OSX but can potentially happen on other systems.
Marek Safar [Thu, 19 May 2011 16:30:14 +0000 (17:30 +0100)]
New tests.
Atsushi Eno [Thu, 19 May 2011 04:15:46 +0000 (13:15 +0900)]
Allow DocumentType in ImportNode().
http://lists.ximian.com/pipermail/mono-devel-list/2011-May/037575.html
Marek Habersack [Thu, 19 May 2011 01:35:42 +0000 (03:35 +0200)]
[asp.net] Ignore JavaScript blocks enclosed in HTML comments
Marek Habersack [Wed, 18 May 2011 19:54:56 +0000 (21:54 +0200)]
[runtime] When shadow-copying assemblies, use the io portability function to find the file.
This is necessary in situations when managed code pases case-insenstive file name and the
file on disk uses different case. stat(2) would return an error in such situation and, in effect,
the stat structure would get invalid/random values in its fields.
Andreia Gaita [Wed, 18 May 2011 01:12:58 +0000 (02:12 +0100)]
[runtime] fix the partial version check, it should only check the first 4 chars
Marek Habersack [Tue, 17 May 2011 19:54:35 +0000 (21:54 +0200)]
Support colors on 256-color terminals
Rodrigo Kumpera [Tue, 17 May 2011 18:34:39 +0000 (11:34 -0700)]
Merge pull request #103 from XTZGZoReX/master
Mono.Tasklets: Suppress Continuation finalization when explicitly disposed.
XTZGZoReX [Tue, 17 May 2011 09:56:50 +0000 (11:56 +0200)]
Mono.Tasklets: Suppress Continuation finalization when explicitly disposed.
Atsushi Eno [Tue, 17 May 2011 16:09:21 +0000 (01:09 +0900)]
Check invalid recursion of xs:group use. Fixed bug #650117.
Miguel de Icaza [Tue, 17 May 2011 13:29:47 +0000 (09:29 -0400)]
Add C launcher for Mono-shipped binaries
Marek Safar [Tue, 17 May 2011 08:09:12 +0000 (09:09 +0100)]
[693947] Ignore delegate declaration when looking for type name
Konrad M. Kruczynski [Mon, 16 May 2011 20:21:35 +0000 (22:21 +0200)]
[Fix] #635971 Fixed throwing exception in case of using zero timeout. Now it is legal.
For details, please see bug description and comments on bugzilla.
Marek Safar [Mon, 16 May 2011 15:41:39 +0000 (16:41 +0100)]
[692149] Fixed ordering of named arguments list when starting with positional argument
Marek Habersack [Mon, 16 May 2011 14:13:20 +0000 (16:13 +0200)]
[asp.net] Fix for bug #650847. Output both tagged and untagged UpdatePanel names in AJAX responses.
Also, update panels which rendered empty string are included in the output.
Marek Safar [Mon, 16 May 2011 10:57:44 +0000 (11:57 +0100)]
Initial refactoring to accommodate async
Marek Safar [Mon, 16 May 2011 08:45:07 +0000 (09:45 +0100)]
New test
Atsushi Eno [Mon, 16 May 2011 09:45:08 +0000 (18:45 +0900)]
Do not import xs:any member as XmlElement[]. Create a wrapper class instead.
Part of bugfix #670945.
Atsushi Eno [Mon, 16 May 2011 07:35:10 +0000 (00:35 -0700)]
Merge pull request #101 from occ/master
Decimal support to XmlBinaryDictionaryReader
Atsushi Eno [Mon, 16 May 2011 07:31:14 +0000 (16:31 +0900)]
Attachable property setter was not working with XamlObjectWriter.
Fixed bug #693779.
Flose Maier [Sun, 15 May 2011 11:25:38 +0000 (13:25 +0200)]
Add test for bug 686486.
Zoltan Varga [Sun, 15 May 2011 11:04:02 +0000 (13:04 +0200)]
Avoid a crash in Field:GetRawConstantValue () if the field type couldn't be loaded. Fixes #693894.
Zoltan Varga [Fri, 13 May 2011 20:16:15 +0000 (22:16 +0200)]
Fix warnings + the windows build.
Rodrigo Kumpera [Fri, 13 May 2011 20:09:57 +0000 (17:09 -0300)]
More win32 fixing
Rodrigo Kumpera [Fri, 13 May 2011 18:18:45 +0000 (15:18 -0300)]
Fix win32
Onur Can Cakmak [Fri, 13 May 2011 18:11:55 +0000 (14:11 -0400)]
Add decimal support to XmlBinaryDictionaryReader
Atsushi Eno [Fri, 13 May 2011 16:28:03 +0000 (01:28 +0900)]
Implement XmlSerializerMessageContractImporter, round 2.
This time it is XmlSerializer-based, got it somewhat working.
Rodrigo Kumpera [Thu, 12 May 2011 22:40:19 +0000 (19:40 -0300)]
Fix variant iface verification.
class.c (mono_gparam_is_reference_conversible): Short-circuit
if arguments are the same type.
* class.c (mono_class_is_variant_compatible): Ditto.
* verify.c (verifier_class_is_assignable_from): Ditto.
* verify.c (verifier_class_is_assignable_from): GTD and open
types interface expansion is not very regular, so we need
to check them more throughtly.
Fixes #693266
Zoltan Varga [Fri, 13 May 2011 15:44:56 +0000 (17:44 +0200)]
Fix the build.
Zoltan Varga [Fri, 13 May 2011 14:07:50 +0000 (07:07 -0700)]
Merge pull request #100 from Unity-Technologies/master-leak-descriptors
[Fix] Don't leak function descriptors on ia64 and ppc64.
Tak [Fri, 13 May 2011 13:45:54 +0000 (15:45 +0200)]
Don't ifdef ftnptrs_hash.
Mark Probst [Fri, 13 May 2011 13:02:02 +0000 (15:02 +0200)]
[sgen] Use critical region instead of lock in mono_gc_wbarrier_value_copy().
Since the card table is lock-free we don't need the full GC lock in
the value copy wbarrier - the critical region, which prevents the GC
from interrupting, is sufficient.
Atsushi Eno [Fri, 13 May 2011 12:52:31 +0000 (21:52 +0900)]
Do not take host part into consideration on comparing URLs.
It blocks access by different host name (e.g. localhost vs. IP).
This was part of bugfix #687580 (forgot to commit).
Atsushi Eno [Fri, 13 May 2011 12:46:24 +0000 (21:46 +0900)]
Implement WsdlContractConversionContext, and ignore message importer for now.
Tak [Fri, 13 May 2011 12:10:14 +0000 (14:10 +0200)]
[Fix] Don't leak function descriptors on ia64 and ppc64.
* metadata/domain.c:
* metadata/domain-internals.h: Cache function descriptors for domain.
* mini.c: Use cache when creating function pointer descriptors.
License: MIT/X11
Mark Probst [Thu, 12 May 2011 21:54:36 +0000 (23:54 +0200)]
Merge branch 'sgen-stage-finalize'
Mark Probst [Wed, 11 May 2011 20:54:26 +0000 (22:54 +0200)]
[sgen] Staged weak link registration.
Do the staging that we've done for finalizers for weak links as
well. Both these optimizations speed up the HttpListener test
significantly.
Mark Probst [Tue, 10 May 2011 22:40:43 +0000 (00:40 +0200)]
[sgen] Move weak link hash code to separate file.
Mark Probst [Tue, 10 May 2011 20:57:15 +0000 (22:57 +0200)]
[sgen] Staging of finalizer registrations.
Registering a finalizer required the GC lock to be taken. To remove
this need we introduce an array where finalizer registrations are
staged in a lock-free way. Only when the array (currently holding
1024 elements) is full is the GC lock taken and the finalizers are
inserted into the hash. Of course this also needs to happen when a
collection starts and when a domain is unloaded.
Mark Probst [Tue, 10 May 2011 16:27:44 +0000 (18:27 +0200)]
[sgen] Move finalization hash code to separate file.
Rodrigo Kumpera [Thu, 12 May 2011 19:01:48 +0000 (12:01 -0700)]
Merge pull request #82 from Unity-Technologies/master-gc-race
[Fix] Fix intermittent initialization race crash with boehm gc.
Rodrigo Kumpera [Thu, 12 May 2011 19:00:54 +0000 (12:00 -0700)]
Merge pull request #92 from konrad-kruczynski/master
[Fix] Patch addressing bug #688940
Rodrigo Kumpera [Thu, 12 May 2011 18:51:40 +0000 (11:51 -0700)]
Merge pull request #96 from ualtinok/master
[FIX] Addition to resolution of bug #677539
Veerapuram Varadhan [Thu, 12 May 2011 18:06:26 +0000 (11:06 -0700)]
Merge pull request #99 from IgorTkachev/master
SqlDbType missing values.
Igor Tkachev [Thu, 12 May 2011 03:09:38 +0000 (23:09 -0400)]
SqlDbType missing value.
Rodrigo Kumpera [Wed, 11 May 2011 21:20:13 +0000 (18:20 -0300)]
Properly handle compressed interface bitmaps on the array store wrapper.
* marshal.c (get_virtual_stelemref_kind): Don't emit optimized
stelemref wrapper if compressed interface bitmaps are enabled.
Rodrigo Kumpera [Wed, 11 May 2011 21:14:16 +0000 (18:14 -0300)]
Fix windows build
Rodrigo Kumpera [Wed, 11 May 2011 20:31:34 +0000 (17:31 -0300)]
Enable the new interrupt machinery on x86 + linux|mach.
* mono-threads.c (mono_thread_info_new_interrupt_enabled):
Enable it on the targets that have been ported and tested.
Igor Tkachev [Wed, 11 May 2011 20:26:15 +0000 (16:26 -0400)]
SqlDbType missing values.
Rodrigo Kumpera [Wed, 11 May 2011 20:09:55 +0000 (17:09 -0300)]
Implement thread dump using new machinery.
* gc.c: Thread dump now uses the finalizer thread
to be more reliable.
* threads.c (mono_threads_perform_thread_dump): New
function that dumps all threads using the new machinery.
* mini-posix.c (sigquit_signal_handler): Use new machinery
if available.
Rodrigo Kumpera [Wed, 11 May 2011 19:24:50 +0000 (16:24 -0300)]
Make sure the GC takes the appropriate locks during STW.
* sgen-gc.c:
* boehm-gc.c: The collector must takes the interruption
lock before starting STW to avoid deadlocks. This is specially
problematic in the case of hosts that are not signal based.
For example, say we have threads 1,2,3. Thread 1 is interrupting
thread 2 while thread 3 begins to Stop-The-World. Now given the
following ordered events:
-thread 1 suspends thread 2, which is in a gc unsafe area.
-thread 3 suspends thread 1 and 2.
-thread 3 tries to restart thread 2 since it is in a gc unsafe
area, but since its suspend count is 2, it fails to do so and
deadlock.
An alternative would be to make the suspended state not have
a counter, but that would make the STW code trickier as it would
need to restore the thread to its previous state and not simply
resume it.
This could work with sgen, but would require some surgery on
boehm, which is completely undesirable given how state it is.
Atsushi Eno [Wed, 11 May 2011 09:40:15 +0000 (18:40 +0900)]
Revert previous XmlSerializerMessageContractImporter implementation.
XsdDataContractImporter is not XmlSerializer-based, while this class is.
Atsushi Eno [Wed, 11 May 2011 07:54:52 +0000 (16:54 +0900)]
Add PipeException. fixed bug #692987 by Eberhard Beilharz.
Atsushi Eno [Wed, 11 May 2011 07:34:45 +0000 (16:34 +0900)]
Fixed orders in XmlSchemaObjectTable and got some S.R.Serialization tests working.
Rodrigo Kumpera [Tue, 10 May 2011 18:57:00 +0000 (15:57 -0300)]
Implement sync suspend based interruption
Rodrigo Kumpera [Tue, 10 May 2011 17:28:39 +0000 (14:28 -0300)]
work around full aot not working
Rodrigo Kumpera [Tue, 10 May 2011 17:28:21 +0000 (14:28 -0300)]
Remove dead variable
Neale Ferguson [Tue, 10 May 2011 18:14:22 +0000 (14:14 -0400)]
Merge branch 'master' of github.com:mono/mono
Neale Ferguson [Tue, 10 May 2011 18:13:45 +0000 (14:13 -0400)]
Add MONO_ARCH_HAVE_CONTEXT_SET_INT_REG for s390x
Rodrigo Kumpera [Tue, 10 May 2011 14:48:14 +0000 (11:48 -0300)]
Pass mono_install_handler_block_guard in the eh callback table
Atsushi Eno [Tue, 10 May 2011 13:02:20 +0000 (22:02 +0900)]
CanImport() just returns false, without throwing error.
Atsushi Eno [Tue, 10 May 2011 12:45:45 +0000 (21:45 +0900)]
support import of xs:anyType, as well as make couple of minor fixes.
Atsushi Eno [Tue, 10 May 2011 10:58:17 +0000 (19:58 +0900)]
Fix common issue in MetadataSet - call MoveToContent() while iterating nodes.
Atsushi Eno [Tue, 10 May 2011 10:33:52 +0000 (19:33 +0900)]
Add new MetadataSetTest and fix couple of minor issues.
Atsushi Eno [Tue, 10 May 2011 10:33:15 +0000 (19:33 +0900)]
Revive this test by implementing mock web server.
Atsushi Eno [Tue, 10 May 2011 10:32:25 +0000 (19:32 +0900)]
Add (not-working) callback instance context mode test and FIXME comment.
Atsushi Eno [Tue, 10 May 2011 08:40:35 +0000 (17:40 +0900)]
Update XsdDataContractImporter.CanImport and implement ImportXmlType.
Rodrigo Kumpera [Tue, 10 May 2011 01:46:52 +0000 (22:46 -0300)]
Enable async invoke on x86.
* exceptions-x86.c (mono_arch_handle_exception): Extract
mono_arch_setup_async_callback which is used to setup
async calls.
* mini-x86.h: Enable async calls on x86.
Sebastien Pouliot [Mon, 9 May 2011 17:47:35 +0000 (13:47 -0400)]
Add (disabled) unit tests for sync use of CommunicationObject
* System.ServiceModel_test.dll.sources: Add new files to the tests
* Test/NUnitMoonHelper.cs: Helper to re-use Moonlight unit tests on NUnit
* Test/System.ServiceModel.Channels/CommunicationObjectSyncTest.cs: New.
Atsushi Eno [Mon, 9 May 2011 13:58:18 +0000 (22:58 +0900)]
ongoing XmlSerializerMessageContractImporter work.
It is part of bugfix #670945.
Atsushi Eno [Mon, 9 May 2011 10:09:19 +0000 (19:09 +0900)]
Multiple schemas in wsdl could share same SourceUri, so check more identity.
Part of bugfix #670945.
Atsushi Eno [Mon, 9 May 2011 04:50:38 +0000 (13:50 +0900)]
Allow xsi:nil for void. Fixed client side of bug #687580.
Atsushi Eno [Mon, 9 May 2011 04:40:16 +0000 (13:40 +0900)]
Now our HttpListener can listen to wildcard prefix safely.
Fixed server part of bug #687580.
Flose Maier [Sun, 8 May 2011 10:59:01 +0000 (12:59 +0200)]
Avoid crash when setting the DialogResult in the FormClosing event method. Fixes bug 686486.
Mark Probst [Fri, 6 May 2011 23:39:00 +0000 (01:39 +0200)]
[utils] Fix lock-free data structures.
Hazard pointers didn't check for the third pointer.
Lots of memory barriers missing.
Small bug in LLS.
Comments and asserts.
Rodrigo Kumpera [Fri, 6 May 2011 19:43:17 +0000 (16:43 -0300)]
More win32 fixes
Neale Ferguson [Fri, 6 May 2011 16:53:12 +0000 (12:53 -0400)]
Merge branch 'master' of github.com:mono/mono
Neale Ferguson [Fri, 6 May 2011 16:52:19 +0000 (12:52 -0400)]
Implement soft debugger for s390x and fix context macro for s390x
Lluis Sanchez [Fri, 6 May 2011 16:37:19 +0000 (18:37 +0200)]
[Web Services] Fix regression caused by previous commit
Fixes a regression caused by the implementation of support
for optional parameters. GlobalIndex was not being set
when using SOAP RPC protocol.
Konrad M. Kruczynski [Fri, 6 May 2011 13:52:06 +0000 (15:52 +0200)]
Small thing reverted.