mono.git
13 years agoReport precise error about local variable hiding
Marek Safar [Thu, 9 Sep 2010 18:51:26 +0000 (19:51 +0100)]
Report precise error about local variable hiding

13 years agoDo all possible mistaken empty statement checks in the parser
Marek Safar [Thu, 9 Sep 2010 16:41:04 +0000 (17:41 +0100)]
Do all possible mistaken empty statement checks in the parser

13 years agoDon't recreate top-level destructor block, it carries precious stuff
Marek Safar [Thu, 9 Sep 2010 16:10:40 +0000 (17:10 +0100)]
Don't recreate top-level destructor block, it carries precious stuff

13 years agoVarious cleanup and addition
Andrew Jorgensen [Thu, 9 Sep 2010 15:57:59 +0000 (09:57 -0600)]
Various cleanup and addition

13 years agoUDP sockets allow connections to Any
Gonzalo Paniagua Javier [Thu, 9 Sep 2010 15:46:06 +0000 (11:46 -0400)]
UDP sockets allow connections to Any

UDP sockets that call Connect() to set a default address for datagrams
sent, can call Connect() again with IPv6Any or Any to 'disconnect'.

Fixes bug #639098.

13 years agoTry to recover after each wrong statement.
Marek Safar [Thu, 9 Sep 2010 14:58:13 +0000 (15:58 +0100)]
Try to recover after each wrong statement.

13 years ago__arglist can be used as an expression.
Marek Safar [Thu, 9 Sep 2010 14:00:50 +0000 (15:00 +0100)]
__arglist can be used as an expression.

13 years agoThe first push to make flow-analysis work for complex blocks.
Marek Safar [Thu, 9 Sep 2010 12:54:32 +0000 (13:54 +0100)]
The first push to make flow-analysis work for complex blocks.

All blocks are now strictly hierachical which greatly simplifies many
things and fixes many subtle bugs with local names. Removing the block
cross references helps scalling when compiling deeply nested blocks or
large method bodies including linq. Prepared possible dynamic statements
to work with variable declarators.

13 years agoIt should not fail to generate client proxy generation for having two "IsRequest...
Atsushi Eno [Thu, 9 Sep 2010 12:51:31 +0000 (21:51 +0900)]
It should not fail to generate client proxy generation for having two "IsRequest=false" messages (it is set only in ContractDescriptionGenerator).

13 years agoadd mono-abi-info to gitignore
Jb Evain [Thu, 9 Sep 2010 08:14:40 +0000 (10:14 +0200)]
add mono-abi-info to gitignore

13 years agoadd ignore file
Jb Evain [Thu, 9 Sep 2010 08:14:01 +0000 (10:14 +0200)]
add ignore file

13 years agoFix System.Core usage of bootstrap .sources from the normal .sources
Jb Evain [Thu, 9 Sep 2010 08:13:10 +0000 (10:13 +0200)]
Fix System.Core usage of bootstrap .sources from the normal .sources

13 years agoFix System usage of bootstrap .sources from the normal .sources
Jb Evain [Thu, 9 Sep 2010 08:12:47 +0000 (10:12 +0200)]
Fix System usage of bootstrap .sources from the normal .sources

13 years agoFix corlib usage of bootstrap .sources from the normal .sources
Jb Evain [Thu, 9 Sep 2010 08:12:29 +0000 (10:12 +0200)]
Fix corlib usage of bootstrap .sources from the normal .sources

13 years agoClean the aot log file too.
Zoltan Varga [Thu, 9 Sep 2010 03:42:11 +0000 (05:42 +0200)]
Clean the aot log file too.

13 years agoAvoid freeing the allocated monitors on shutdown, it seems to cause crashes
Zoltan Varga [Wed, 8 Sep 2010 23:12:09 +0000 (01:12 +0200)]
Avoid freeing the allocated monitors on shutdown, it seems to cause crashes
when running with SGEN, since SGEN writes to those arrays in:

==29095==    at 0x6844E5: add_or_remove_disappearing_link (sgen-gc.c:4143)
==29095==    by 0x6845F8: mono_gc_register_disappearing_link (sgen-gc.c:4167)
==29095==    by 0x67DE35: clear_domain_process_object (sgen-gc.c:1384)
==29095==    by 0x67DEA3: clear_domain_process_major_object_callback (sgen-gc.c:1400)
==29095==    by 0x6DE2DC: major_iterate_objects (sgen-marksweep.c:692)
==29095==    by 0x67E00B: mono_gc_clear_domain (sgen-gc.c:1459)

13 years agoFix the monodroid and monotouch profile builds
Jb Evain [Wed, 8 Sep 2010 16:37:27 +0000 (18:37 +0200)]
Fix the monodroid and monotouch profile builds

13 years agoFix the monodroid profile's defines
Jb Evain [Wed, 8 Sep 2010 16:35:01 +0000 (18:35 +0200)]
Fix the monodroid profile's defines

13 years agoFactorize the System.Core monotouch and monodroid sources into a mobile.sources
Jb Evain [Wed, 8 Sep 2010 15:37:56 +0000 (17:37 +0200)]
Factorize the System.Core monotouch and monodroid sources into a mobile.sources

13 years agoFactorize the System.Xml monotouch and monodroid sources into a mobile.sources
Jb Evain [Wed, 8 Sep 2010 15:36:34 +0000 (17:36 +0200)]
Factorize the System.Xml monotouch and monodroid sources into a mobile.sources

13 years agoFactorize the System.Runtime.Serialization monotouch and monodroid sources into a...
Jb Evain [Wed, 8 Sep 2010 15:35:23 +0000 (17:35 +0200)]
Factorize the System.Runtime.Serialization monotouch and monodroid sources into a mobile.sources

13 years agoFix the mobile_System.ServiceModel.Web include
Jb Evain [Wed, 8 Sep 2010 15:34:02 +0000 (17:34 +0200)]
Fix the mobile_System.ServiceModel.Web include

13 years agoFix the mobile_System.Data include
Jb Evain [Wed, 8 Sep 2010 15:33:44 +0000 (17:33 +0200)]
Fix the mobile_System.Data include

13 years agoFactorize the System.Web.Services monotouch and monodroid sources into a mobile.sources
Jb Evain [Wed, 8 Sep 2010 15:31:10 +0000 (17:31 +0200)]
Factorize the System.Web.Services monotouch and monodroid sources into a mobile.sources

13 years agoFactorize the System.Data monotouch and monodroid sources into a mobile.sources
Jb Evain [Wed, 8 Sep 2010 15:29:27 +0000 (17:29 +0200)]
Factorize the System.Data monotouch and monodroid sources into a mobile.sources

13 years agoFactorize the System.ServiceModel.Web monotouch and monodroid sources into a mobile...
Jb Evain [Wed, 8 Sep 2010 15:28:25 +0000 (17:28 +0200)]
Factorize the System.ServiceModel.Web monotouch and monodroid sources into a mobile.sources

13 years agoFactorize the System.ServiceModel monotouch and monodroid sources into a mobile.sources
Jb Evain [Wed, 8 Sep 2010 15:27:02 +0000 (17:27 +0200)]
Factorize the System.ServiceModel monotouch and monodroid sources into a mobile.sources

13 years agoFactorize the monotouch and monodroid sources into a mobile.sources
Jb Evain [Wed, 8 Sep 2010 15:25:11 +0000 (17:25 +0200)]
Factorize the monotouch and monodroid sources into a mobile.sources

13 years agoEnable the mcs monodroid build from mono's configure
Jb Evain [Wed, 8 Sep 2010 15:21:15 +0000 (17:21 +0200)]
Enable the mcs monodroid build from mono's configure

13 years agoOrchestrate the monodroid build bootstrap
Jb Evain [Wed, 8 Sep 2010 15:20:36 +0000 (17:20 +0200)]
Orchestrate the monodroid build bootstrap

13 years agoFactorize the monotouch build into a mobile build for monodroid and monotouch
Jb Evain [Wed, 8 Sep 2010 15:20:14 +0000 (17:20 +0200)]
Factorize the monotouch build into a mobile build for monodroid and monotouch

13 years agoEnable the Mono.Data.Sqlite build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:19:08 +0000 (17:19 +0200)]
Enable the Mono.Data.Sqlite build for monodroid

13 years agoEnable the Mono.Security build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:18:48 +0000 (17:18 +0200)]
Enable the Mono.Security build for monodroid

13 years agoEnable the System.Transactions build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:18:24 +0000 (17:18 +0200)]
Enable the System.Transactions build for monodroid

13 years agoMake HttpUtility compilable on MonoDroid as well as MonoTouch
Jb Evain [Wed, 8 Sep 2010 15:18:00 +0000 (17:18 +0200)]
Make HttpUtility compilable on MonoDroid as well as MonoTouch

13 years agoEnable the System.Runtime.Serialization build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:16:59 +0000 (17:16 +0200)]
Enable the System.Runtime.Serialization build for monodroid

13 years agoEnable the System.ServiceModel.Web build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:16:32 +0000 (17:16 +0200)]
Enable the System.ServiceModel.Web build for monodroid

13 years agoEnable the System.ServiceModel build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:16:07 +0000 (17:16 +0200)]
Enable the System.ServiceModel build for monodroid

13 years agoEnable the System.Web.Services build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:14:57 +0000 (17:14 +0200)]
Enable the System.Web.Services build for monodroid

13 years agoEnable the System.Data build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:12:18 +0000 (17:12 +0200)]
Enable the System.Data build for monodroid

13 years agoEnable the System.Xml build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:10:55 +0000 (17:10 +0200)]
Enable the System.Xml build for monodroid

13 years agoEnable the System.Core build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:08:48 +0000 (17:08 +0200)]
Enable the System.Core build for monodroid

13 years agoEnable the System build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:07:54 +0000 (17:07 +0200)]
Enable the System build for monodroid

13 years agoEnable the corlib build for monodroid
Jb Evain [Wed, 8 Sep 2010 15:06:18 +0000 (17:06 +0200)]
Enable the corlib build for monodroid

13 years agoDefine MOBILE for the monotouch build as a shared symbol between MonoTouch and MonoDroid
Jb Evain [Wed, 8 Sep 2010 15:02:32 +0000 (17:02 +0200)]
Define MOBILE for the monotouch build as a shared symbol between MonoTouch and MonoDroid

13 years agoAdd the monodroid profile
Jb Evain [Wed, 8 Sep 2010 15:01:30 +0000 (17:01 +0200)]
Add the monodroid profile

13 years agoRemove no-op code
Jb Evain [Wed, 8 Sep 2010 14:24:01 +0000 (16:24 +0200)]
Remove no-op code

13 years agoSimplify ifdefs specifics to MOONLIGHT
Jb Evain [Wed, 8 Sep 2010 13:58:31 +0000 (15:58 +0200)]
Simplify ifdefs specifics to MOONLIGHT

13 years ago[mdoc-update] Delegate types should get <returns/>, not <value/>.
Jonathan Pryor [Wed, 8 Sep 2010 12:20:36 +0000 (08:20 -0400)]
[mdoc-update] Delegate types should get <returns/>, not <value/>.

Oops, that's a long-standing bug...  Found by Joel Marcey from ECMA.

13 years agoAdd tests for g_strdupv
Gonzalo Paniagua Javier [Tue, 7 Sep 2010 22:29:33 +0000 (18:29 -0400)]
Add tests for g_strdupv

Add test for g_strdupv and check for NULL in g_strdupv.

13 years agoAdd a few more remaps.
Zoltan Varga [Tue, 7 Sep 2010 21:42:57 +0000 (23:42 +0200)]
Add a few more remaps.

13 years agoRemap g_strdupv too.
Zoltan Varga [Tue, 7 Sep 2010 21:26:39 +0000 (23:26 +0200)]
Remap g_strdupv too.

13 years ago[asp.net] Fix for bug #636709. ClientID must be reset when ID is assigned.
Marek Habersack [Tue, 7 Sep 2010 21:20:28 +0000 (23:20 +0200)]
[asp.net] Fix for bug #636709. ClientID must be reset when ID is assigned.

13 years ago[asp.net] Fix for bug #636475. A typo caused ads not to be matched in ChooseAd
Marek Habersack [Tue, 7 Sep 2010 20:09:48 +0000 (22:09 +0200)]
[asp.net] Fix for bug #636475. A typo caused ads not to be matched in ChooseAd

The second loop over the ads hashtable must not reject an app if the keyword filter is empty.

13 years agoAdd wsdl help generator to the 4.0 files
Gonzalo Paniagua Javier [Tue, 7 Sep 2010 18:20:31 +0000 (14:20 -0400)]
Add wsdl help generator to the 4.0 files

Ditto.

Fixes bug #636841.

13 years agoAdd profile docs
Miguel de Icaza [Tue, 7 Sep 2010 15:06:56 +0000 (11:06 -0400)]
Add profile docs

13 years agoAdd DateTimeOffsetConverter.cs and DateTimeOffsetConverterTests.cs to our build.
Carlos Alberto Cortez [Tue, 7 Sep 2010 02:46:37 +0000 (04:46 +0200)]
Add DateTimeOffsetConverter.cs and DateTimeOffsetConverterTests.cs to our build.

* System.dll.sources: Add DateTimeOffsetConverter.cs.
* System_test.dll.sources: Add DateTimeOffsetConverterTests.cs.

13 years agoAdd new DateTimeOffsetConverterTests.cs test file.
Carlos Alberto Cortez [Tue, 7 Sep 2010 02:45:32 +0000 (04:45 +0200)]
Add new DateTimeOffsetConverterTests.cs test file.

* DateTimeOffsetConverterTests.cs: New file.

13 years agoImplement System.ComponentModel.DateTimeOffsetConverter.
Carlos Alberto Cortez [Tue, 7 Sep 2010 02:39:18 +0000 (04:39 +0200)]
Implement System.ComponentModel.DateTimeOffsetConverter.

* DateTimeOffsetConverter.cs: Implement new 4.0 class.

13 years agoAvoid crash for use of deprecated InstallTrackingHandler in Snow Leopard.
Atsushi Eno [Tue, 7 Sep 2010 03:25:14 +0000 (12:25 +0900)]
Avoid crash for use of deprecated InstallTrackingHandler in Snow Leopard.

13 years agoFix build.
Rodrigo Kumpera [Mon, 6 Sep 2010 21:15:48 +0000 (18:15 -0300)]
Fix build.

* bug-348522.2.cs: gmcs now complains about a thinko here.

13 years agoFix 64bits builds.
Rodrigo Kumpera [Mon, 6 Sep 2010 21:15:14 +0000 (18:15 -0300)]
Fix 64bits builds.

* sgen-cardtable.h (sgen_card_table_get_card_address): This
needs marking under SGEN_HAVE_OVERLAPPING_CARDS too.

13 years agoCheck for a parent extending a generic instantiation of the child.
Rodrigo Kumpera [Mon, 6 Sep 2010 21:07:24 +0000 (18:07 -0300)]
Check for a parent extending a generic instantiation of the child.

* class.c (mono_class_create_from_typedef): If if any parent
doesn't extend the generic instantiation.

Fixes #604270

13 years ago[mdoc-update] Clean up System.Object craziness in ILAsm generation.
Jonathan Pryor [Mon, 6 Sep 2010 20:12:37 +0000 (16:12 -0400)]
[mdoc-update] Clean up System.Object craziness in ILAsm generation.

The problem: ILAsm has 'object' as a keyword alias for System.Object, and
99.99% of the time should use 'object'...except in 'extends' declarations,
where it remains System.Object and not 'object'.

The solution: Have an ILFullMemberFormatter.WithinExtends "flag" which
controls whether we generate System.Object or object.

The problem: this is flaky as hell, and results 'class A:List<object>'
generating:

.class public auto ansi beforefieldinit MyList
extends System.Collections.Generic.List`1<System.Object>

This is wrong, as it should be List`1<object>, not List`1<System.Object>.
(At least, monodis(1) produces List`1<object>...)

So, instead of the craziness that is WithinExtends, just check check for
System.Object as the base type immediately after the 'extends' generation.  If
it's System.Object, keep it, otherwise call full.GetName() (and remove the
leading "class " that's generated from GetName).

13 years agoAdd regression test for #633248
Rodrigo Kumpera [Mon, 6 Sep 2010 18:41:15 +0000 (15:41 -0300)]
Add regression test for #633248

13 years agoProperly handle stack merges of leave.
Rodrigo Kumpera [Mon, 6 Sep 2010 17:58:08 +0000 (14:58 -0300)]
Properly handle stack merges of leave.

* verify.c (do_leave): Set the target of leave.

* verify.c (mono_method_verify): Merge the resulting empty
stack of a leave instruction with it's target.

Fixes #633248

13 years agoFix NetworkInterface marshaling size on OSX and avoid crash.
Atsushi Eno [Mon, 6 Sep 2010 07:29:26 +0000 (16:29 +0900)]
Fix NetworkInterface marshaling size on OSX and avoid crash.

13 years agoFix the build.
Zoltan Varga [Sun, 5 Sep 2010 18:29:01 +0000 (20:29 +0200)]
Fix the build.

13 years agoNew test for CultureInfoConverter.GetCultureName.
Carlos Alberto Cortez [Sun, 5 Sep 2010 03:50:01 +0000 (05:50 +0200)]
New test for CultureInfoConverter.GetCultureName.

* CultureInfoConverterTest.cs: New test for GetCultureName.

13 years agoImplement CultureInfoConverter.GetCultureName.
Carlos Alberto Cortez [Sun, 5 Sep 2010 03:36:59 +0000 (05:36 +0200)]
Implement CultureInfoConverter.GetCultureName.

* CultureInfoConverter.cs: Implement GetCultureName.

13 years agoRemove the NET_2_0 directives from CultureInfoConverter.cs
Carlos Alberto Cortez [Fri, 3 Sep 2010 04:26:12 +0000 (06:26 +0200)]
Remove the NET_2_0 directives from CultureInfoConverter.cs

* CultureInfoConverter.cs: Remove the NET_2_0 directive.

13 years agoAdd AttributeCollectionTest.cs to System_test.dll.sources.
Carlos Alberto Cortez [Fri, 3 Sep 2010 04:16:50 +0000 (06:16 +0200)]
Add AttributeCollectionTest.cs to System_test.dll.sources.

* System_test.dll.sources: Add AttributeCollectionTest.cs.

13 years agoAdd new AttributeCollectionTest file.
Carlos Alberto Cortez [Fri, 3 Sep 2010 04:16:08 +0000 (06:16 +0200)]
Add new AttributeCollectionTest file.

* AttributeCollectionTest.cs: New test file.

13 years agoAdd 4.0 Attributes property and protected ctor to AttributeCollection.
Carlos Alberto Cortez [Fri, 3 Sep 2010 04:12:46 +0000 (06:12 +0200)]
Add 4.0 Attributes property and protected ctor to AttributeCollection.

* AttributeCollection.cs: Add 4.0 Attributes property as well
as a new parameterless ctor.

13 years agoRemove the NET_2_0 directives from AttributeCollection.cs.
Carlos Alberto Cortez [Fri, 3 Sep 2010 03:42:12 +0000 (05:42 +0200)]
Remove the NET_2_0 directives from AttributeCollection.cs.

* AttributeCollection.cs: Remove the NET_2_0 directives, as well
as the else case for it (we are always NET_2_0 now).

13 years agoMerge branch 'sgen_cardtable_optimization_work'
Rodrigo Kumpera [Sat, 4 Sep 2010 20:33:39 +0000 (17:33 -0300)]
Merge branch 'sgen_cardtable_optimization_work'

13 years agoFix a race condition in barriers using cardtable.
Rodrigo Kumpera [Sat, 4 Sep 2010 20:26:28 +0000 (17:26 -0300)]
Fix a race condition in barriers using cardtable.

* sgen-gc.c: Introduce dummy_use function that forces
the compiler to keep a given value alive until the call
to it.

* sgen-gc.c (mono_gc_wbarrier_set_field): Use dummy_use to
avoid race condition between the store and the barrier.

* sgen-gc.c (mono_gc_wbarrier_set_arrayref): Ditto.

* sgen-gc.c (mono_gc_wbarrier_arrayref_copy): Ditto.

* sgen-gc.c (mono_gc_wbarrier_arrayref_copy): Ditto.

* sgen-gc.c (mono_gc_wbarrier_generic_store): Ditto.

This pattern of usage is the same we do with managed code
write barrier. The race condition is extremely race and
requires the thread to be interrupted in between the store
and the barrier, which is a very tiny narrow window.

13 years agoOptimize cardtable unmanaged array copy.
Rodrigo Kumpera [Sat, 4 Sep 2010 19:31:23 +0000 (16:31 -0300)]
Optimize cardtable unmanaged array copy.

* sgen-gc.c (mono_gc_wbarrier_arrayref_copy): Avoid locking
if using cardtables. Even thou the new code looks slower
than a plain memmove, it has the advantage of not taking
locks, which means it's much more scalable.

Array copy under sgen/cardtables was 1.6x slower than with
boehm. With this patch it reduces to 1.54x. The serial
performance speedup is nice, but it is much more helpfull
in multi-threaded workloads.

13 years agoFix a warning
Rodrigo Kumpera [Fri, 3 Sep 2010 19:17:06 +0000 (16:17 -0300)]
Fix a warning

13 years agoOptimize the cardtable path a bit more.
Rodrigo Kumpera [Fri, 3 Sep 2010 19:14:57 +0000 (16:14 -0300)]
Optimize the cardtable path a bit more.

* sgen-gc.c: Avoid taking locks for single pointer barriers.

13 years agoOptimize major cardtable scanning by about 15%.
Rodrigo Kumpera [Fri, 3 Sep 2010 18:39:28 +0000 (15:39 -0300)]
Optimize major cardtable scanning by about 15%.

* sgen-cardtable.c: Move a bunch of stuff to new header.

* sgen-cardtable.h: New header with functions meant to
be inlined.

* sgen-gc.c: Use new header.

* sgen-marksweep.c (major_scan_card_table): Introduce fast
version of some macros that avoid a few indirections.

13 years agoBetter cardtable stats gathering codebase.
Rodrigo Kumpera [Fri, 3 Sep 2010 18:38:36 +0000 (15:38 -0300)]
Better cardtable stats gathering codebase.

13 years agoAdd notes when to use cardtable and remset.
Rodrigo Kumpera [Fri, 3 Sep 2010 17:27:02 +0000 (14:27 -0300)]
Add notes when to use cardtable and remset.

13 years agoUse los scanning function for big objects on major heap.
Rodrigo Kumpera [Fri, 3 Sep 2010 17:03:49 +0000 (14:03 -0300)]
Use los scanning function for big objects on major heap.

* sgen-cardtable.c (sgen_cardtable_scan_object): The
source of the cardtable can now be supplied, this is needed
since major blocks can't use regular card checking code here.

* sgen-los.c (los_scan_card_table): Adjust to signature
change in sgen_cardtable_scan_object.

* sgen-marksweep.c (major_scan_card_table): If the block
size is larger than a card it makes more sense to loop
around objects and not cards since objects might be scanned
twice since a card can be shared between two of them.

In such case the cards of a block must be copied to a side
buffer since we are no longer looping one by one. This ia
a problem when two objects share a marked card as the first
can cause it to be left cleared and not have the second one
scanned.

Another advantage of this patch is that it enables partial
scanning of larger array in the major heap.

13 years agoMove core of los scanning into cardtable codebase.
Rodrigo Kumpera [Fri, 3 Sep 2010 14:17:42 +0000 (11:17 -0300)]
Move core of los scanning into cardtable codebase.

* sgen-los.c (los_scan_card_table): Move meat of
this function into a separate one.

* sgen-cardtable.c: New function sgen_cardtable_scan_object.

* sgen-gc.h: Add sgen_cardtable_scan_object.

13 years agoMicro optimize cardtable scanning
Rodrigo Kumpera [Fri, 3 Sep 2010 14:10:04 +0000 (11:10 -0300)]
Micro optimize cardtable scanning

* sgen-marksweep.c (major_scan_card_table): Don't load
some fields if the block has no references.

13 years agoAdd stats for cardtable processing time
Rodrigo Kumpera [Thu, 2 Sep 2010 21:59:11 +0000 (18:59 -0300)]
Add stats for cardtable processing time

13 years agoFix return value of items_read
Gonzalo Paniagua Javier [Sat, 4 Sep 2010 18:24:54 +0000 (14:24 -0400)]
Fix return value of items_read

items_read was not being initialized and it was not being set in the
fast common case of g_utf8_to_utf16().

13 years agoFixed GetLogicalDriveString
Gonzalo Paniagua Javier [Sat, 4 Sep 2010 18:02:17 +0000 (14:02 -0400)]
Fixed GetLogicalDriveString

When calling GetLogicalDriveString, icall.c was passing the
length in the wrong units.

GetLogicalDriveString now translates any octal sequence found
into its corresponding byte. Use the UTF16 length
('items_written') instead of the UTF8 input string length
('items_read') when computing the total size.

13 years ago[mdoc-update] Properly map System.IntPtr to 'native int' for ILAsm.
Jonathan Pryor [Sat, 4 Sep 2010 16:57:42 +0000 (12:57 -0400)]
[mdoc-update] Properly map System.IntPtr to 'native int' for ILAsm.

Similarly, System.UIntPtr maps to 'native unsigned int'.

13 years ago[mdoc-update] Add support ILAsm language output.
Jonathan Pryor [Sat, 4 Sep 2010 02:02:58 +0000 (22:02 -0400)]
[mdoc-update] Add support ILAsm language output.

13 years ago[mdoc-update] Add support for multiple languages in type/member signatures.
Jonathan Pryor [Fri, 3 Sep 2010 15:04:29 +0000 (11:04 -0400)]
[mdoc-update] Add support for multiple languages in type/member signatures.

We're planning on adding IL signature support for <TypeSignature/> and
<MemberSignature/>, so having arrays of MemberFormatters instead of having
single instances of them will make supporting IL output easier.

13 years agoUse UTF8 instead of ASCII when reading mount info
Gonzalo Paniagua Javier [Fri, 3 Sep 2010 21:26:16 +0000 (17:26 -0400)]
Use UTF8 instead of ASCII when reading mount info

This allows for mount points with UTF8 characters in them.

13 years agoLinux /proc/mounts might have escaped characters
Gonzalo Paniagua Javier [Fri, 3 Sep 2010 20:01:42 +0000 (16:01 -0400)]
Linux /proc/mounts might have escaped characters

Translate escaped characters in mount paths.

Fixes bug #637078.

13 years agoAllow ordinal entry points in windows.
Gonzalo Paniagua Javier [Fri, 3 Sep 2010 18:58:50 +0000 (14:58 -0400)]
Allow ordinal entry points in windows.

Windows allows entry point names like "#10" which refer to the Nth
function (n > 1 and less than the highest ordinal value in the .def
file).

Fixes bug #636966.

13 years agoFix a crash when the signature of the Main method could not be loaded. Fixes #636794.
Zoltan Varga [Fri, 3 Sep 2010 17:17:41 +0000 (19:17 +0200)]
Fix a crash when the signature of the Main method could not be loaded. Fixes #636794.

13 years agoOngoing Discovery duplex (UDP) implementation (not working yet).
Atsushi Eno [Fri, 3 Sep 2010 09:45:15 +0000 (18:45 +0900)]
Ongoing Discovery duplex (UDP) implementation (not working yet).

13 years agoRemove workaround for moonlight now that System.dll's WebExceptionStatus is properly...
Jb Evain [Fri, 3 Sep 2010 09:30:58 +0000 (11:30 +0200)]
Remove workaround for moonlight now that System.dll's WebExceptionStatus is properly compiled

13 years agoFix the visiblity of WebExceptionStatus for moonlight in the same way as WebException
Jb Evain [Fri, 3 Sep 2010 09:29:32 +0000 (11:29 +0200)]
Fix the visiblity of WebExceptionStatus for moonlight in the same way as WebException

13 years agoFix another non-null RemoteAddress dependency.
Atsushi Eno [Fri, 3 Sep 2010 07:18:59 +0000 (16:18 +0900)]
Fix another non-null RemoteAddress dependency.