mono.git
9 years agoMerge pull request #1109 from adbre/iss358
Alex Rønne Petersen [Thu, 19 Jun 2014 00:34:53 +0000 (02:34 +0200)]
Merge pull request #1109 from adbre/iss358

Override registry storage with MONO_CURRENTUSER_REGISTRY_PATH

9 years ago[ppc] Bump instruction sizes after 51c454f6b03b468273bb93db2aa9b76e330a7e57.
Alex Rønne Petersen [Wed, 18 Jun 2014 19:30:07 +0000 (21:30 +0200)]
[ppc] Bump instruction sizes after 51c454f6b03b468273bb93db2aa9b76e330a7e57.

9 years agoMerge pull request #1014 from txdv/master
Jonathan Pryor [Wed, 18 Jun 2014 19:21:05 +0000 (15:21 -0400)]
Merge pull request #1014 from txdv/master

Expose full epoll_ctl functionality.

9 years agoMerge pull request #1110 from tpokorra/FixImageListStreamerGeneratingValidResourcesFo...
Alex Rønne Petersen [Wed, 18 Jun 2014 19:02:38 +0000 (21:02 +0200)]
Merge pull request #1110 from tpokorra/FixImageListStreamerGeneratingValidResourcesForMSNet

Fixing ImageListStreamer generating winform resource files that work on MS.Net

9 years agoOverride registry storage with MONO_CURRENTUSER_REGISTRY_PATH
Iain Lane [Mon, 2 Jul 2012 12:59:51 +0000 (14:59 +0200)]
Override registry storage with MONO_CURRENTUSER_REGISTRY_PATH

https://bugzilla.xamarin.com/show_bug.cgi?id=358

9 years agoMerge pull request #1111 from akoeplinger/add-gtk-dllmap
Alex Rønne Petersen [Wed, 18 Jun 2014 15:21:03 +0000 (17:21 +0200)]
Merge pull request #1111 from akoeplinger/add-gtk-dllmap

Added dllmap for libgtk-x11 and libgdk-x11

9 years agoAdded dllmap for libgtk-x11 and libgdk-x11
Alexander Köplinger [Wed, 18 Jun 2014 15:02:32 +0000 (17:02 +0200)]
Added dllmap for libgtk-x11 and libgdk-x11

Commit licensed under MIT/X11.

9 years agoExpose full epoll_ctl functionality.
Andrius Bentkus [Fri, 2 May 2014 14:18:45 +0000 (16:18 +0200)]
Expose full epoll_ctl functionality.

Now the user has access to all struct elements in EpollEvent when
using epoll_ctl. This is in particular very intersting when the
user wants to access the ptr field.

The user can use GCHandle.ToIntPtr to give the struct a GCHandle
and then handle the pointer later on when using epoll_wait.

9 years agoImageListStreamer generates winform resource files that work on Mono but fail on...
Timotheus Pokorra [Wed, 18 Jun 2014 14:08:04 +0000 (16:08 +0200)]
ImageListStreamer generates winform resource files that work on Mono but fail on Microsoft.net
with exception System.InvalidOperationException: Loading of the ImageList did not succeed.
see bug https://bugzilla.xamarin.com/show_bug.cgi?id=3841 for detailed analysis

9 years ago[ppc] Add missing `sync` instructions to atomic opcodes.
Alex Rønne Petersen [Wed, 18 Jun 2014 14:06:34 +0000 (16:06 +0200)]
[ppc] Add missing `sync` instructions to atomic opcodes.

Note that lwarx and stwcx don't produce the necessary
load/store barrier that's needed to conform to the
`Interlocked` API.

9 years agoFixed another 8080 test port
Alexander Köplinger [Wed, 18 Jun 2014 08:01:36 +0000 (10:01 +0200)]
Fixed another 8080 test port

I missed that one in f396d37c5493455f0b7fe726b408d74ba1c8fb5f.

9 years agoSilence a warning in file-mmap-posix.c.
Alex Rønne Petersen [Wed, 18 Jun 2014 00:38:52 +0000 (02:38 +0200)]
Silence a warning in file-mmap-posix.c.

9 years agoMerge pull request #1108 from akoeplinger/change-test-ports
Alex Rønne Petersen [Tue, 17 Jun 2014 21:05:04 +0000 (23:05 +0200)]
Merge pull request #1108 from akoeplinger/change-test-ports

[System.ServiceModel] Changed test port from 8080 to 30158

9 years ago[System.ServiceModel] Changed test port from 8080 to 30158
Alexander Köplinger [Tue, 17 Jun 2014 21:00:42 +0000 (23:00 +0200)]
[System.ServiceModel] Changed test port from 8080 to 30158

Port 8080 is widely used and resulted in address conflicts on monojenkins, so we use a different port instead.

Commit licensed under MIT/X11.

9 years ago[aot] Add support for overriding the alignment calculations done by the runtime durin...
Zoltan Varga [Tue, 17 Jun 2014 18:43:44 +0000 (20:43 +0200)]
[aot] Add support for overriding the alignment calculations done by the runtime during AOT. This is needed because the alignment of some types, notable i8/double can be different between the host and the target platform.

9 years agoDisable generic-xdomain.2.exe on CI as it occasionally hangs.
Alex Rønne Petersen [Tue, 17 Jun 2014 15:55:29 +0000 (17:55 +0200)]
Disable generic-xdomain.2.exe on CI as it occasionally hangs.

9 years agoMerge pull request #1106 from akoeplinger/fix-winforms-tests
Alex Rønne Petersen [Tue, 17 Jun 2014 15:47:45 +0000 (17:47 +0200)]
Merge pull request #1106 from akoeplinger/fix-winforms-tests

[Managed.Windows.Forms] Fixed System.Resources tests that relied on DummyAssembly

9 years agoMerge pull request #1107 from akoeplinger/fix-numericupdown
Alex Rønne Petersen [Tue, 17 Jun 2014 15:46:56 +0000 (17:46 +0200)]
Merge pull request #1107 from akoeplinger/fix-numericupdown

[Managed.Windows.Forms] Fix 'minimum' default value of NumericUpDown

9 years ago[Managed.Windows.Forms] Fix 'minimum' default value of NumericUpDown
Alexander Köplinger [Tue, 17 Jun 2014 15:35:46 +0000 (17:35 +0200)]
[Managed.Windows.Forms] Fix 'minimum' default value of NumericUpDown

The 'minimum' default value of NumericUpDown was set to 0.0M, which caused the ToString() to format it as "0.0" instead of "0" as it does on .NET.
A similar change was done for the 'maximum' default value already in f6c8d48a3aed377b8a11e38c96c0d4c774664279.

Commit licensed under MIT/X11.

9 years ago[mcs] Implements finally execution for await with try block exit.
Marek Safar [Tue, 17 Jun 2014 14:43:29 +0000 (16:43 +0200)]
[mcs] Implements finally execution for await with try block exit.

9 years ago[Managed.Windows.Forms] Fixed System.Resources tests that relied on DummyAssembly
Alexander Köplinger [Tue, 17 Jun 2014 11:45:45 +0000 (13:45 +0200)]
[Managed.Windows.Forms] Fixed System.Resources tests that relied on DummyAssembly

Several System.Resources tests rely on a DummyAssembly.dll, which was removed in 016b0b5170fca72b977f952ee817fe06e32890a9 (possibly by accident?).
This commit adds the assembly back.

Commit licensed under MIT/X11.

9 years agoRevert "[msbuild] fix some weird local result variable initialization issue."
Atsushi Eno [Tue, 17 Jun 2014 04:59:01 +0000 (13:59 +0900)]
Revert "[msbuild] fix some weird local result variable initialization issue."

This reverts commit 76b08f12ef69c180733a0cad99ea89b2a225ec51.

9 years agoMerge pull request #1103 from nathanaeljones/patch-1
Zoltan Varga [Mon, 16 Jun 2014 21:26:28 +0000 (23:26 +0200)]
Merge pull request #1103 from nathanaeljones/patch-1

Fix AddCacheDependency signature to use params

9 years ago[runtime] Add a workaround to file-mmap-posix.c to fix the android build.
Zoltan Varga [Mon, 16 Jun 2014 21:10:13 +0000 (23:10 +0200)]
[runtime] Add a workaround to file-mmap-posix.c to fix the android build.

9 years ago[aot] Use g_print instead of printf in one place since the former works on android.
Zoltan Varga [Mon, 16 Jun 2014 20:25:54 +0000 (22:25 +0200)]
[aot] Use g_print instead of printf in one place since the former works on android.

9 years ago[arm] Fix alignment problems when AOTing from osx to arm linux.
Zoltan Varga [Mon, 16 Jun 2014 20:25:02 +0000 (22:25 +0200)]
[arm] Fix alignment problems when AOTing from osx to arm linux.

9 years agoMerge pull request #1104 from tpokorra/TarballIncludeMoreDllsInMonoLite
Alex Rønne Petersen [Mon, 16 Jun 2014 19:49:08 +0000 (21:49 +0200)]
Merge pull request #1104 from tpokorra/TarballIncludeMoreDllsInMonoLite

Tarball: include System.Security.dll and System.Configuration.dll in monolite directory

9 years agoTarball: include System.Security.dll and System.Configuration.dll in monolite directory
Timotheus Pokorra [Mon, 16 Jun 2014 19:11:26 +0000 (21:11 +0200)]
Tarball: include System.Security.dll and System.Configuration.dll in monolite directory
this is related to commit https://github.com/mono/mono/commit/993f4773638bcb36abf8b9660ee737d6445c5200

9 years ago[bcl] Fix the mobile build.
Zoltan Varga [Mon, 16 Jun 2014 17:06:41 +0000 (19:06 +0200)]
[bcl] Fix the mobile build.

9 years agoFix AddCacheDependency signature to use params
Nathanael Jones [Mon, 16 Jun 2014 16:43:28 +0000 (18:43 +0200)]
Fix AddCacheDependency signature to use params

The .NET AddCacheDependency method signature is `public void AddCacheDependency(params CacheDependency[] dependencies);` Mono should mirror the signature (even if it isn't implemented) to prevent build errors. Conditional code, of course, will need to be used for an alternative mono path.

9 years agoMerge pull request #1102 from tpokorra/TarballFixAvoidCDataTextReaderTests
Alex Rønne Petersen [Mon, 16 Jun 2014 14:11:43 +0000 (16:11 +0200)]
Merge pull request #1102 from tpokorra/TarballFixAvoidCDataTextReaderTests

Remove a reference to the file AvoidCDataTextReaderTests.cs so that tarballs can be built again

9 years agoRemove a reference to the file AvoidCDataTextReaderTests.cs so that tarballs can...
Timotheus Pokorra [Mon, 16 Jun 2014 14:03:46 +0000 (16:03 +0200)]
Remove a reference to the file AvoidCDataTextReaderTests.cs so that tarballs can be built again.
The file was removed in https://github.com/mono/mono/commit/d8d1b0be6baf6f8881871592b040d85be790926c

9 years agoMerge pull request #1101 from jariq/master
Alex Rønne Petersen [Mon, 16 Jun 2014 12:54:30 +0000 (14:54 +0200)]
Merge pull request #1101 from jariq/master

Support both unencrypted PKCS#12 and PKCS#12 encrypted with an empty pas...

9 years ago[mcs] Run new initializer await checks only in async context. Fixes #20614
Marek Safar [Mon, 16 Jun 2014 12:12:19 +0000 (14:12 +0200)]
[mcs] Run new initializer await checks only in async context. Fixes #20614

9 years agoMerge pull request #1099 from jsportaro/master
Marek Safar [Mon, 16 Jun 2014 10:52:45 +0000 (12:52 +0200)]
Merge pull request #1099 from jsportaro/master

Added fix for Bug 11630

9 years ago[system.data] Parse numeric expressions using invariant culture. Fixes #20582
Marek Safar [Mon, 16 Jun 2014 10:07:25 +0000 (12:07 +0200)]
[system.data] Parse numeric expressions using invariant culture. Fixes #20582

9 years ago[mcs] Reachability analysis should ignore local variables declarations. Fixes #20603
Marek Safar [Mon, 16 Jun 2014 09:56:57 +0000 (11:56 +0200)]
[mcs] Reachability analysis should ignore local variables declarations. Fixes #20603

9 years agoMerge pull request #1100 from BrzVlad/master
Zoltan Varga [Sun, 15 Jun 2014 10:34:45 +0000 (12:34 +0200)]
Merge pull request #1100 from BrzVlad/master

[x86] Fix initialization of stackalloc-ed memory in no pushes mode.

9 years ago[runtime] Enable support for arm64-darwin in dolt.
Zoltan Varga [Sat, 14 Jun 2014 22:41:28 +0000 (00:41 +0200)]
[runtime] Enable support for arm64-darwin in dolt.

9 years agoSupport both unencrypted PKCS#12 and PKCS#12 encrypted with an empty password in...
Jaroslav Imrich [Sat, 14 Jun 2014 22:32:15 +0000 (00:32 +0200)]
Support both unencrypted PKCS#12 and PKCS#12 encrypted with an empty password in X509Certificate2 constructor

9 years ago[x86] Align stackalloc-ed memory in no pushes mode.
Brezae Vlad [Sat, 14 Jun 2014 21:05:42 +0000 (00:05 +0300)]
[x86] Align stackalloc-ed memory in no pushes mode.

9 years ago[arm] Add an enum for gsharedvt argument marshalling conventions.
Zoltan Varga [Sat, 14 Jun 2014 20:40:16 +0000 (22:40 +0200)]
[arm] Add an enum for gsharedvt argument marshalling conventions.

9 years ago[x86] Fix initialization of stackalloc-ed memory in no pushes mode.
Brezae Vlad [Fri, 13 Jun 2014 21:43:12 +0000 (00:43 +0300)]
[x86] Fix initialization of stackalloc-ed memory in no pushes mode.

9 years agoMonodoc: Fix for CDATA bug (https://bugzilla.xamarin.com/show_bug.cgi?id=17850)....
Larry O'Brien [Thu, 12 Jun 2014 23:25:44 +0000 (13:25 -1000)]
Monodoc: Fix for CDATA bug (https://bugzilla.xamarin.com/show_bug.cgi?id=17850). Reverts 3f361e1a263fc23680beab62eadde0cd3e4ac64d

Removed code that stripped CDATA statements (but not blocks) from input when showing a namespace page:

The function Ecma2Html.WrapStream(TextReader,Dictionary<String,String>) very explicitly returns a new `AvoidCDataTextReader` when the system is showing a namespace. The `AvoidCDataTextReader` strips the CDATA statement, but not it's internal contents. The intent, according to AvoidCDataTextReaderTests, is to allow class summaries of this form:

"<elements><summary>Addressbook APIs.</summary><remarks /><class name=""ABAddressBook"" fullname=""MonoTouch.AddressBook.ABAddressBook"" assembly=""monotouch""><summary><![CDATA[
      Provides access to the system Address Book.]]>
    </summary></class></elements>"

But that form is not desired.

The bug arose because the input to `WrapStream` is very large (it actually includes the entire *description* for each class and member in the namespace!). If any CDATA-block contained entities such as, e.g., < or >, the resulting XML would (obviously) be ill-formed:

"<xml>This should be valid <![CDATA[<T>]]> </xml>"

would throw because the <T> would be attempted to be parsed as XML. Obviously, samples that contain generics or lambda or XAML would cause this problem.

Although an XMLException was thrown, Macdoc swallows the exception in DocTools.GetHTML(). That is why this caused a silent failure on macdoc and was visible on Webdoc.

9 years ago[System.Core] Add new tests for MemoryMappedFile.
Rodrigo Kumpera [Fri, 13 Jun 2014 18:47:54 +0000 (14:47 -0400)]
[System.Core] Add new tests for MemoryMappedFile.

9 years ago[runtime] Fix share/private flags when doing file mmap.
Rodrigo Kumpera [Fri, 13 Jun 2014 18:35:55 +0000 (14:35 -0400)]
[runtime] Fix share/private flags when doing file mmap.

9 years ago[runtime] Rewrite MemoryMappedFile to support named mappings.
Rodrigo Kumpera [Fri, 13 Jun 2014 16:16:31 +0000 (12:16 -0400)]
[runtime] Rewrite MemoryMappedFile to support named mappings.

9 years agoAdded fix for Bug 11630
Joseph Portaro [Fri, 13 Jun 2014 15:55:37 +0000 (11:55 -0400)]
Added fix for Bug 11630

9 years ago[mcs] Always store hidden catch variable when clause contains await
Marek Safar [Fri, 13 Jun 2014 15:34:43 +0000 (17:34 +0200)]
[mcs] Always store hidden catch variable when clause contains await

9 years ago[corlib] Set AggregateException message text in all overloads
Marek Safar [Fri, 13 Jun 2014 15:33:43 +0000 (17:33 +0200)]
[corlib] Set AggregateException message text in all overloads

9 years ago[mcs] Emit correct value type collection initialization when second initializer uses...
Marek Safar [Fri, 13 Jun 2014 14:18:06 +0000 (16:18 +0200)]
[mcs] Emit correct value type collection initialization when second initializer uses await

9 years ago[mcs] Keep stack populated when array initializer contains await
Marek Safar [Fri, 13 Jun 2014 12:31:02 +0000 (14:31 +0200)]
[mcs] Keep stack populated when array initializer contains await

9 years ago[x86] Pop the vtype return addr from the stack in no pushes mode too.
Zoltan Varga [Fri, 13 Jun 2014 14:23:24 +0000 (16:23 +0200)]
[x86] Pop the vtype return addr from the stack in no pushes mode too.

9 years agoMerge pull request #1098 from tpokorra/TarMissing_vasprintfh
Alex Rønne Petersen [Fri, 13 Jun 2014 14:21:20 +0000 (16:21 +0200)]
Merge pull request #1098 from tpokorra/TarMissing_vasprintfh

tarball is missing file vasprintf.h

9 years agosince gerror.c requires vasprintf.h, we need to add that file so that it becomes...
Timotheus Pokorra [Fri, 13 Jun 2014 14:15:19 +0000 (16:15 +0200)]
since gerror.c requires vasprintf.h, we need to add that file so that it becomes part of the tarball

9 years ago[x86] Reduce some code duplication in the x86 call opcodes.
Zoltan Varga [Fri, 13 Jun 2014 13:55:38 +0000 (15:55 +0200)]
[x86] Reduce some code duplication in the x86 call opcodes.

9 years agoDisable push-less x86 code for now.
Alex Rønne Petersen [Fri, 13 Jun 2014 13:32:42 +0000 (15:32 +0200)]
Disable push-less x86 code for now.

Breaks stuff: https://wrench.mono-project.com/Wrench/ViewLane.aspx?lane_id=10&host_id=8&revision_id=68534

9 years agoMerge pull request #1097 from BrzVlad/master
Zoltan Varga [Fri, 13 Jun 2014 11:15:03 +0000 (13:15 +0200)]
Merge pull request #1097 from BrzVlad/master

[jit] Fix bug when using pushless code and reenable the feature.

9 years ago[mcs] Handle catch filter with await-ed rethrow and exception filters
Marek Safar [Fri, 13 Jun 2014 09:22:03 +0000 (11:22 +0200)]
[mcs] Handle catch filter with await-ed rethrow and exception filters

9 years ago[jit] Fix bug when using pushless code and reenable the feature.
Brezae Vlad [Thu, 12 Jun 2014 18:48:51 +0000 (21:48 +0300)]
[jit] Fix bug when using pushless code and reenable the feature.

9 years ago[mcs] Add support for await in finally clause
Marek Safar [Thu, 12 Jun 2014 15:46:18 +0000 (17:46 +0200)]
[mcs] Add support for await in finally clause

9 years ago[Mono.Debugger.Soft] Handle cases where the instruction list is empty
Jeffrey Stedfast [Wed, 11 Jun 2014 22:27:15 +0000 (18:27 -0400)]
[Mono.Debugger.Soft] Handle cases where the instruction list is empty

9 years ago[jit] Don't emit a write barrier when writing to the delegate->method field, its...
Zoltan Varga [Wed, 11 Jun 2014 17:09:01 +0000 (19:09 +0200)]
[jit] Don't emit a write barrier when writing to the delegate->method field, its not needed.

9 years ago[Http]: Fix a potential race condition in the WebConnectionGroup's connection list.
Martin Baulig [Wed, 11 Jun 2014 16:50:47 +0000 (18:50 +0200)]
[Http]: Fix a potential race condition in the WebConnectionGroup's connection list.

Remove the "ConnectionState" from the connection list before closing the connection,
this should avoid ServicePoint.SendRequest() ever crashing with a NullReferenceException
due to getting a null return value from WebConnectionGroup.GetConnection().

9 years ago[mcs] Check for missing dependencies of container types. Fixes #20382
Marek Safar [Wed, 11 Jun 2014 13:13:00 +0000 (15:13 +0200)]
[mcs] Check for missing dependencies of container types. Fixes #20382

9 years ago[mcs] Resolve event expression when testing left side expression. Fixes #20493
Marek Safar [Wed, 11 Jun 2014 12:24:34 +0000 (14:24 +0200)]
[mcs] Resolve event expression when testing left side expression. Fixes #20493

9 years ago[mcs] Inflate nested pointer containers too.
Marek Safar [Wed, 11 Jun 2014 10:02:33 +0000 (12:02 +0200)]
[mcs] Inflate nested pointer containers too.

9 years ago[mcs] Flow analysis of binary expressions not using logical operators. Fixes #20515
Marek Safar [Wed, 11 Jun 2014 08:07:20 +0000 (10:07 +0200)]
[mcs] Flow analysis of binary expressions not using logical operators. Fixes #20515

9 years agoMerge pull request #890 from xoofx/master
Alex Rønne Petersen [Wed, 11 Jun 2014 07:22:29 +0000 (09:22 +0200)]
Merge pull request #890 from xoofx/master

[Mono.Linker] Add support for <Module> with static method and pre-mark t...

9 years ago[Mono.Linker] Add support for using the linker with assemblies referencing WPF
Alexandre Mutel [Fri, 7 Feb 2014 15:33:23 +0000 (00:33 +0900)]
[Mono.Linker] Add support for using the linker with assemblies referencing WPF

9 years ago[Mono.Linker] Add support for <Module> with static method and pre-mark them if necessary
Alexandre Mutel [Thu, 6 Feb 2014 22:19:07 +0000 (07:19 +0900)]
[Mono.Linker] Add support for <Module> with static method and pre-mark them if necessary

9 years ago[sgen] Fix an infinite loop in find_previous_pointer_fragment due to remove CAS'ing...
Rodrigo Kumpera [Tue, 10 Jun 2014 21:37:40 +0000 (17:37 -0400)]
[sgen] Fix an infinite loop in find_previous_pointer_fragment due to remove CAS'ing the wrong value.

The fragment list has one invariant, the next pointer is tagged if the current fragment is been deleted,
which means the list head can never point to a tagged pointer.

If this invariant is broken, the code in find_previous_pointer_fragment will end up in an infinite loop.

The infinite loop happens in this simplified fragment:

try_again:
prev = &allocator->alloc_head;
1) cur = unmask (*prev);
while (1) {
2) if (*prev != cur)
goto try_again;

So assume allocator->alloc_head points to 0x121, which is tagged.

So execution goes like this:

1) cur is set to 0x120
2) *prev (value 0x121) is different than cur (value 0x120) so retry.

The above will loop forever as the fixup code never touches allocator->alloc_head.

Now that we understand the infinite loop, let's see how the remove code can trigger it.

Use use the following notation in the event sequence:

-> a non tagged pointer
-*> a tagged pointer
first CAS: InterlockedCompareExchangePointer ((volatile gpointer*)&frag->next, mask (next, 1), next)
second CAS: InterlockedCompareExchangePointer ((volatile gpointer*)prev_ptr, next, frag)

Assuming the following initial fragment list and two threads removing two separate fragments:

fragment list; -> A -> B -> C

thread 1 removes A
thread 2 removes B

thread 2 succeed the first CAS
-> A -> B -*> C

thread 1 succeed the first CAS
-> A -*> B -*> C

thread 2 fail the second CAS because of thread 1
-> A -*> B -*> C

thread 1 succeed the second CAS
-> B -*> C

thread 2 succeed the second CAS and set prev_ptr to next,
which at this point is a tagged pointer. Leaving the fragment list like this:
-*> C

After the last CAS, we end up with alloc_head with a tagged pointer, which will cause
the infinite loop we previously explained.

The fix is to unmask the value we store in the second CAS, as the objective is to link
two live fragments and not flag the previous fragment for deletion.

The exact same bug exists in mono-linked-list-set and was fixed there.

9 years ago[Http]: Cosmetic, remove some dead code.
Martin Baulig [Tue, 10 Jun 2014 21:05:35 +0000 (23:05 +0200)]
[Http]: Cosmetic, remove some dead code.

The is no reason to ever invoke the callback on the ThreadPool,
it would in fact only cause trouble.

Marek's commit db92449f was correct, I removed the commented out
code and the no longer used method.

9 years ago[mcs] Implement rethrow support in awaited catch clause
Marek Safar [Tue, 10 Jun 2014 19:30:50 +0000 (21:30 +0200)]
[mcs] Implement rethrow support in awaited catch clause

9 years ago[Http]: Fix digest authentication (bug #18799).
Martin Baulig [Tue, 10 Jun 2014 19:22:11 +0000 (21:22 +0200)]
[Http]: Fix digest authentication (bug #18799).

We did not correctly parse key-value pairs without quotes in the
challenge.

9 years ago[System.Net.Http]: Always add "Content-Length" in HttpClientHandler.SendAsync().
Martin Baulig [Tue, 10 Jun 2014 18:29:35 +0000 (20:29 +0200)]
[System.Net.Http]: Always add "Content-Length" in HttpClientHandler.SendAsync().

This has been reported on the forums:
http://forums.xamarin.com/discussion/17770/length-required-error-in-http-post-since-latest-release

9 years ago[Mono.Debugger.Soft] Improved ILInterpreter to eval properties that return a primitiv...
Jeffrey Stedfast [Tue, 10 Jun 2014 19:15:56 +0000 (15:15 -0400)]
[Mono.Debugger.Soft] Improved ILInterpreter to eval properties that return a primitive constant

9 years ago[windows] Return early in the vectored exception handler if the thread is not managed...
Joao Matos [Tue, 10 Jun 2014 19:14:09 +0000 (20:14 +0100)]
[windows] Return early in the vectored exception handler if the thread is not managed by the runtime.

9 years ago[Http]: Only recycle ServicePoints from the idle timer (fixes #19823).
Martin Baulig [Tue, 10 Jun 2014 17:13:10 +0000 (19:13 +0200)]
[Http]: Only recycle ServicePoints from the idle timer (fixes #19823).

Thanks a lot for Martin Potter to find and report the race condition
which was causing the problem.

ServicePointManager.RecycleServicePoints() was a left-over from the
old times when we didn't have an idle-timer based approach.

9 years ago[Mono.Debugger.Soft] minor code cleanup
Jeffrey Stedfast [Tue, 10 Jun 2014 16:54:48 +0000 (12:54 -0400)]
[Mono.Debugger.Soft] minor code cleanup

9 years agoFix mono exception handler crashing on windows.
Joao Matos [Tue, 10 Jun 2014 16:03:30 +0000 (17:03 +0100)]
Fix mono exception handler crashing on windows.

Here's the scenario that causes the crash:

Exception is fired, which invokes mono's vectored exception handler. It
is not an exception mono recognizes, so falls through to the default:
case in the exception handler. The problem is the return value has been
initialized to EXCEPTION_CONTINUE_EXECUTION. This causes SEH not to
process any other handlers, and just attempt to continue execution at
the point the exception was thrown.

The problem is much of windows internals will catch various exceptions
internally and handle them. So instead of the exception being handled,
we would just crash in the bowels of windows.

The solution is to set mono_win_chained_exception_needs_run to TRUE in
the default case - this is what the various exception handlers do if the
exception is not in JIT'd code. This causes the handler to return
EXCEPTION_CONTINUE_SEARCH, which causes SEH to search for the next
handler, and no crash is produced.

Both x86 and x64 had this bug, so I fixed it in both places.

Fix contributed by Steve Anichini.

9 years agoChanged mono_win_chained_exception_needs_run to be a JIT TLS variable.
Joao Matos [Tue, 10 Jun 2014 15:55:24 +0000 (16:55 +0100)]
Changed mono_win_chained_exception_needs_run to be a JIT TLS variable.

This fixes a potential race condition that might happen if two threads try to handle an exception at the same time.

9 years ago[sys.xaml] fix TypeId and get broken WindowsBase tests pass.
Atsushi Eno [Tue, 10 Jun 2014 12:49:05 +0000 (21:49 +0900)]
[sys.xaml] fix TypeId and get broken WindowsBase tests pass.

9 years agoAttempt to sanitize the `vasprintf` situation in eglib.
Alex Rønne Petersen [Tue, 10 Jun 2014 07:40:22 +0000 (09:40 +0200)]
Attempt to sanitize the `vasprintf` situation in eglib.

The problem is that on some platforms we call the function
without having provided a prototype for it. This breaks when
building with `-Werror`.

Now we declare it in the private `vasprintf.h` header when
the platform doesn't have a declaration in `stdio.h`. This
should also let us remove the hack in `glib.h` for MSVC.

This may or may not break the build on some platforms; I've
tried on those that I have access to with success. Let me
know if this breaks anything.

9 years agoAdd build system plumbing for building with -Werror.
Alex Rønne Petersen [Tue, 10 Jun 2014 07:34:01 +0000 (09:34 +0200)]
Add build system plumbing for building with -Werror.

9 years agoMerge pull request #1094 from LRFalk01/master
Alex Rønne Petersen [Tue, 10 Jun 2014 01:29:10 +0000 (03:29 +0200)]
Merge pull request #1094 from LRFalk01/master

Identities not being set in ClaimsPrincipal Ctor

9 years agoIdentities not being set in ClaimsPrincipal Ctor
Lucas [Tue, 10 Jun 2014 01:20:35 +0000 (21:20 -0400)]
Identities not being set in ClaimsPrincipal Ctor

Class' identities member is not set in constructor ClaimsPrincipal
(IEnumerable<ClaimsIdentity> identities)

9 years agoRemove unnecessary MPH_INTERNAL applied to a function definition.
Alex Rønne Petersen [Tue, 10 Jun 2014 01:15:16 +0000 (03:15 +0200)]
Remove unnecessary MPH_INTERNAL applied to a function definition.

This should be applied to prototypes, and it already is for
this function (in mph.h).

9 years agoRemove the architecture conditions for using `perf` in the log profiler.
Alex Rønne Petersen [Mon, 9 Jun 2014 23:45:25 +0000 (01:45 +0200)]
Remove the architecture conditions for using `perf` in the log profiler.

We now use generic atomics which work on all platforms, so there's
no reason to keep this constraint here. This allows the `perf`-backed
profiler to be used on mips, ppc, s390x, etc.

9 years agoRevert "Disable `perf` support in the profiler when running on Android (for now)."
Alex Rønne Petersen [Mon, 9 Jun 2014 23:44:18 +0000 (01:44 +0200)]
Revert "Disable `perf` support in the profiler when running on Android (for now)."

This reverts commit 8e35b7d01e14371851a806dad6196b36d9163c91.

The fallback implementation (as used on other platforms) appears to hang
the runtime. Having a maybe-working `perf`-backed sample profiler is better
than a not-working sample profiler.

9 years agoRevert "[x86] Reenable the no pushes code."
Zoltan Varga [Mon, 9 Jun 2014 15:37:20 +0000 (17:37 +0200)]
Revert "[x86] Reenable the no pushes code."

This reverts commit 5093efd55883e8a8a093fdb43d938d7b9dde7a4e.

Revert this as it still breaks the build.

9 years ago[x86] Fix the build.
Zoltan Varga [Mon, 9 Jun 2014 15:11:51 +0000 (17:11 +0200)]
[x86] Fix the build.

9 years ago[x86] Reenable the no pushes code.
Zoltan Varga [Mon, 9 Jun 2014 14:35:39 +0000 (16:35 +0200)]
[x86] Reenable the no pushes code.

9 years ago[sys.xaml] remove extraneous NET_2_1 and use __MOBILE__ to make it build for XA/XiOS.
Atsushi Eno [Mon, 9 Jun 2014 03:42:00 +0000 (12:42 +0900)]
[sys.xaml] remove extraneous NET_2_1 and use __MOBILE__ to make it build for XA/XiOS.

9 years ago[jit] Align the stack when using the pushless code on x86.
Zoltan Varga [Sun, 8 Jun 2014 21:18:53 +0000 (23:18 +0200)]
[jit] Align the stack when using the pushless code on x86.

9 years agoMerge pull request #1090 from BrianDurham/patch-2
Alex Rønne Petersen [Sun, 8 Jun 2014 19:05:03 +0000 (21:05 +0200)]
Merge pull request #1090 from BrianDurham/patch-2

Update X11Display.cs - change "Check you DISPLAY"

9 years agoMerge pull request #1089 from BrianDurham/patch-1
Alex Rønne Petersen [Sun, 8 Jun 2014 19:04:33 +0000 (21:04 +0200)]
Merge pull request #1089 from BrianDurham/patch-1

Update Graphics.cs - change "Check you DISPLAY"

9 years agoMerge pull request #1091 from BrianDurham/patch-3
Alex Rønne Petersen [Sun, 8 Jun 2014 17:51:56 +0000 (19:51 +0200)]
Merge pull request #1091 from BrianDurham/patch-3

Update XplatUIX11.cs - change "Check you DISPLAY"

9 years agoMerge pull request #1092 from BrianDurham/patch-4
Alex Rønne Petersen [Sun, 8 Jun 2014 17:51:42 +0000 (19:51 +0200)]
Merge pull request #1092 from BrianDurham/patch-4

Update XplatUIX11GTK.cs - change "Check you DISPLAY"

9 years agoUpdate XplatUIX11GTK.cs - change "Check you DISPLAY"
Brian Durham [Sun, 8 Jun 2014 15:24:35 +0000 (10:24 -0500)]
Update XplatUIX11GTK.cs - change "Check you DISPLAY"

Update XplatUIX11GTK.cs - change "Check you DISPLAY" text to "Check your DISPLAY"