bh-whd [Thu, 8 Nov 2012 09:24:53 +0000 (10:24 +0100)]
Update mcs/class/Novell.Directory.Ldap/Novell.Directory.Ldap.Rfc2251/RfcFilter.cs
This fixes bug 8278 (https://bugzilla.xamarin.com/show_bug.cgi?id=8278)
Miguel de Icaza [Wed, 7 Nov 2012 20:15:25 +0000 (15:15 -0500)]
Fix #6349: Do not use KQueue on Darwin, as their kernel behaves differently than BSD
Introduce a new define that determines whether we can use kqueue for the threadpool,
and keep the old check for kqueue, in case we need to use it in the future for something
else.
From Gonzalo's notes:
I could not reproduce the problem at all on FreeBSD, but it happened
immediately on the Mac. And when it happened, I could never kill the
process (tried kill from outside and another thread with
Environment.Exit() but no luck). This led me to believe that it was an
OS problem. Both asynchronous operations work fine when they are the
only asynchronous operation on the socket (I turned the other into
synch). When read and writes are asychronous kqueue() changes the
event we are waiting for to be read/write. That is usually not a
problem, but in this particular case, it looked like 2 separate
threads were trying to do the same thing (one for read and one for
write). I tried a simple spin lock per file-descriptor to no avail. I
also tried a fat lock and again no luck. Then I found this:
http://lists.apple.com/archives/darwin-dev/2011/Jun/msg00016.html and
then remembered that back in 2006, when I first tried a patch to
support kqueue(), Geoff found a similar problem to this one (yeah,
what the fuck, Steve!). The problem has not surfaced earlier because
most people that use sockets know how to do it. Setting up read and
write asynch. calls at the same time is not something that pros do,
but...
Solutions?
1. Provide a way to disable kqueue at runtime for cases like this (can
we set the MONO_DISABLE_AIO environment variable in an iOS mono
program? Just Environment.SetEnvironmentVariable before the first use
of sockets/threadpool should work.
2. Disable kqueue by default. The benefits of things like epoll and
kqueue are negligible, really, unless people are dealing with several
hundreds of sockets (~500 and above) and the default poll based
implementation is ok for most applications (xsp would suffer a lot of
contention under load, but xsp uses asynch ops the right way :).
Rodrigo Kumpera [Tue, 9 Oct 2012 21:25:05 +0000 (17:25 -0400)]
Fix mobile profile build of System.dll
Sebastien Pouliot [Fri, 30 Mar 2012 11:54:53 +0000 (07:54 -0400)]
Fix Aes to behave correctly on MonoTouch wrt CryptoConfig (default creation) and fix it's FeedbackSize (for CFB)
Rodrigo Kumpera [Wed, 7 Mar 2012 20:14:50 +0000 (17:14 -0300)]
Use Function<T>:Identity instead of an inline lambda. Fixes 3627.
Rodrigo Kumpera [Tue, 6 Mar 2012 23:38:11 +0000 (20:38 -0300)]
Add workaround to make LINQ 4 Objects order by work under full-aot.
Sebastien Pouliot [Thu, 23 Feb 2012 20:06:30 +0000 (15:06 -0500)]
Avoid multiple potential full-AOT failures with LINQ Enumerable. Fix bug #3285
Sebastien Pouliot [Sun, 17 Jul 2011 20:35:15 +0000 (16:35 -0400)]
Avoid using PredicateOf<T>.Always since it does not work with full-aot. Fix bug #682757...
Rodrigo Kumpera [Tue, 9 Oct 2012 21:25:41 +0000 (17:25 -0400)]
Fix the mobile profile build of System.Core.dll.
Miguel de Icaza [Mon, 27 Aug 2012 14:36:03 +0000 (10:36 -0400)]
Add some defensive exceptions to WebConnectionStream to try to track down #6329
Miguel de Icaza [Mon, 27 Aug 2012 14:33:43 +0000 (10:33 -0400)]
Add some defensive exceptions to WebConnectionStream to try to track down #6329
Miguel de Icaza [Wed, 15 Aug 2012 20:49:57 +0000 (16:49 -0400)]
mobile/System.dll: Add BindingList<T>
Miguel de Icaza [Tue, 19 Jun 2012 21:14:25 +0000 (17:14 -0400)]
NetworkStream: Bring ReceiveTimeout, SendTimeout, Close to MOBILE profile
Chris Toshok [Mon, 19 Sep 2011 21:44:41 +0000 (14:44 -0700)]
Up the default max persistent connections to 10 for monotouch, and warn when we use more.
Rolf Bjarne Kvinge [Tue, 8 Nov 2011 22:59:20 +0000 (23:59 +0100)]
[Mono.Data.Sqlite] Fix a MonoTouch memory leak
Ensure SQLite3's GCHandle is freed when Close is called instead of Dispose.
Also make SQLiteConnection's Dispose call SQlite3's Dispose to ensure we free
the GCHandle in all cases.
Sebastien Pouliot [Thu, 22 Mar 2012 12:31:09 +0000 (08:31 -0400)]
Redirect, by default, the Console.* and Console.Error.* output to NSLog
Sebastien Pouliot [Wed, 20 Jul 2011 01:39:48 +0000 (21:39 -0400)]
Get better linker results when doing full-aot wrt the cryptographic stack
Sebastien Pouliot [Wed, 20 Jul 2011 22:45:35 +0000 (18:45 -0400)]
Fix default SecurityManager properties to false (they should not be reached unless there's a bug, but they should be right)
Sebastien Pouliot [Thu, 1 Mar 2012 15:14:04 +0000 (10:14 -0500)]
Disable optimization under full AOT. Fix bug #3682
Sebastien Pouliot [Wed, 7 Nov 2012 00:08:30 +0000 (19:08 -0500)]
Use same name/casing for DllImport so we do not duplicate needlessly the extern entries in the metadata
Rodrigo Kumpera [Tue, 6 Nov 2012 23:03:33 +0000 (18:03 -0500)]
Disable the regex compiler under Full AOT.
Rodrigo Kumpera [Tue, 6 Nov 2012 23:03:21 +0000 (18:03 -0500)]
Remove transparent proxy support for monotouch builds.
Rodrigo Kumpera [Tue, 6 Nov 2012 23:00:45 +0000 (18:00 -0500)]
Make System.Guid Full AOT friendly.
* Guid.cs: Make this type more Full AOT friendly by not exposing
the rng type on the field to allow the linker to kill it and by
using a Guid generator that doesn't depend on the Crypto stack.
Rodrigo Kumpera [Tue, 6 Nov 2012 22:57:41 +0000 (17:57 -0500)]
Make the Crypto stack work with Full AOT.
* System.Security.Cryptography.*cs: To make Crypto work with
Full AOT we remove the usage of providers and reflection. We hardcode
the default algorithms for hashing, rng and symmetric cyphers.
Rodrigo Kumpera [Tue, 6 Nov 2012 22:08:53 +0000 (17:08 -0500)]
Don't tag some CAS types as obsolete.
Rodrigo Kumpera [Tue, 6 Nov 2012 21:16:23 +0000 (16:16 -0500)]
Avoid bringing the whole Crypto stack in AssemblyName under FullAOT.
Rodrigo Kumpera [Tue, 6 Nov 2012 21:14:44 +0000 (16:14 -0500)]
Disable code generation for object serialization under Full AOT.
Rodrigo Kumpera [Tue, 6 Nov 2012 21:11:26 +0000 (16:11 -0500)]
Disable more COM code under Full AOT.
Rodrigo Kumpera [Tue, 6 Nov 2012 21:04:50 +0000 (16:04 -0500)]
Remove SRE from Full AOT builds.
Rodrigo Kumpera [Tue, 6 Nov 2012 20:42:27 +0000 (15:42 -0500)]
Remove some COM types under Full AOT.
Rodrigo Kumpera [Tue, 6 Nov 2012 19:55:32 +0000 (14:55 -0500)]
Kill silly using directive.
Rodrigo Kumpera [Tue, 6 Nov 2012 15:09:41 +0000 (10:09 -0500)]
Add Full AOT support for System.Linq.Expression.
* System.Core/System.Linq.Expressions/*.cs: To bring LINQ expressions
to Full AOT we disable all code that depends on SRE or dynamic code generation.
Rodrigo Kumpera [Tue, 6 Nov 2012 14:33:07 +0000 (09:33 -0500)]
Rework some bits in Parallel LINQ to work under Full AOT.
Miguel de Icaza [Tue, 1 Nov 2011 18:51:13 +0000 (14:51 -0400)]
Bring Parallel LINQ to the Mobile profile
Zoltan Varga [Tue, 6 Nov 2012 19:21:45 +0000 (20:21 +0100)]
Disable the profilers when using --disable-libraries.
Zoltan Varga [Tue, 6 Nov 2012 19:13:55 +0000 (20:13 +0100)]
Don't build monograph when using --disable-libraries.
Zoltan Varga [Tue, 6 Nov 2012 19:02:02 +0000 (20:02 +0100)]
Don't build monodis when using --disable-libraries.
Rodrigo Kumpera [Mon, 5 Nov 2012 19:13:57 +0000 (14:13 -0500)]
Remove SRE.UnmanagedMarshal dependency from System.Reflection.
* FieldInfo.cs:
* MonoMethod.cs:
* ParameterInfo.cs: Remove silly dep on UnmanagedMarshal to use
MarshalAsAttribute which is what we need in the end.
* icall.c:
* reflection.c: Use the custom attribute type instead of the SRE
one. This allows FULL_AOT_RUNTIME to work with stuff that uses
MarshalAs.
Rodrigo Kumpera [Mon, 5 Nov 2012 19:08:07 +0000 (14:08 -0500)]
Remove FieldInfo::UMarshal, which is dead code.
Rodrigo Kumpera [Mon, 5 Nov 2012 19:02:58 +0000 (14:02 -0500)]
Expose MarshalAsAttribute to the runtime.
Jeffrey Stedfast [Mon, 5 Nov 2012 13:54:11 +0000 (05:54 -0800)]
Merge pull request #488 from roji/email_attachment_fixes
Fixed missing Content-Id email headers
Marek Safar [Mon, 5 Nov 2012 09:39:29 +0000 (01:39 -0800)]
Merge pull request #497 from Shabbafru/ImageListPatches
Fixed clearing of the ImageList
Atsushi Eno [Mon, 5 Nov 2012 09:22:29 +0000 (18:22 +0900)]
Fix JavaScriptSerializer.Deserialize() for dictionary of array of anonymously typed dictionary.
context: https://twitter.com/muo_jp/status/
265373491224801280
Shabbafru [Mon, 5 Nov 2012 08:35:28 +0000 (09:35 +0100)]
fixed clearing of the ImageList
Shay Rojansky [Mon, 5 Nov 2012 01:36:23 +0000 (03:36 +0200)]
Removed debugging prints that slipped through
Sebastien Pouliot [Fri, 2 Nov 2012 21:15:02 +0000 (17:15 -0400)]
Allow setting a custom Profile without using reflection
Marek Safar [Fri, 2 Nov 2012 09:12:41 +0000 (02:12 -0700)]
Merge pull request #492 from pruiz/systx-fixes
Sys.Tx: Avoid uncatched exceptions during Prepare phase, which might crash the whole application.
Rodrigo Kumpera [Thu, 1 Nov 2012 21:43:25 +0000 (17:43 -0400)]
Revert "Merge pull request #464 from gregoryyoung/master"
This commit causes deadlock in the tpool backend in the following way:
thread 1:
socket_io_add locks io_lock
tp_poll_modify waits on new_sem
thread 2:
tp_poll_wait tries to lock io_lock
tp_poll_wait is the responsible to post to new_sem, which it can't do since it's
blocked on io_lock, held by a thread waiting on new_sem.
This reverts commit
11286da0ac2e2bab7b2d8ab04b9f6a4da4e12131, reversing
changes made to
94bfdf21e6f06424ad96980a27c35c99ba42b1d3.
Rodrigo Kumpera [Thu, 1 Nov 2012 20:23:50 +0000 (16:23 -0400)]
When inflating a (M)VAR make sure we're not using a broken type.
*class.c (inflate_generic_type): When inflating (M)VAR we must
check if the replaced type is not void or typedbyref as both can
later lead to all sorts of nasty crashes or memory corruption.
Mark Probst [Thu, 1 Nov 2012 19:13:11 +0000 (20:13 +0100)]
[sgen] MONO_GC_DEBUG option for disabling the managed allocator.
This can be used to make all object allocations show up in the dtrace probe.
Jonathan Pryor [Thu, 1 Nov 2012 17:59:30 +0000 (13:59 -0400)]
[mdoc] Update XSD @id, @tool for docs team.
Xamarin docs team has requested the following changes to make life
easier and increase tooling support:
//para/@id:
Similar in concept/purpose to the HTML @id attribute.
//para/@tool:
Contains the name of a tool which generated the <para/>.
//exception/list:
Allow lists within <exception/> blocks.
Marek Safar [Thu, 1 Nov 2012 17:27:28 +0000 (10:27 -0700)]
Merge pull request #490 from knocte/master
[System.Configuration] IntegerValidator's maxValue was 0 even if not used
Zoltan Varga [Thu, 1 Nov 2012 17:12:25 +0000 (18:12 +0100)]
Disable LLVM+AOT for methods with clauses on ARM, because some tests still fail.
Marek Safar [Thu, 1 Nov 2012 16:54:29 +0000 (17:54 +0100)]
Implement SqlConnection.DbProviderFactory. Fixes #7936
Alex Rønne Petersen [Thu, 1 Nov 2012 16:09:21 +0000 (09:09 -0700)]
Merge pull request #491 from mono/fix_keyvaluepair
[corlib] Remove the DebuggerDisplay attribute from KeyValuePair
Alan McGovern [Thu, 1 Nov 2012 15:47:18 +0000 (15:47 +0000)]
[corlib] Remove the DebuggerDisplay attribute from KeyValuePair and DictionaryEntry
This attribute gives us a horrible debugging experience when the
'Value' of the KeyValuePair is null. By using this attribute, the
user is told that the KeyValuePair itself is null when they inspect
something like 'new KeyValuePair<string, string> (null, null)'.
This is very confusing.
We should just let the ToString method be invoked, which returns a
nicely formatted string. This is also the behaviour observed when
debugging using Visual Studio.
Zoltan Varga [Thu, 1 Nov 2012 15:15:41 +0000 (16:15 +0100)]
Avoid throwing exceptions from WeakReference.IsAlive after finalization when the underlying GCHandle is already finalized. Fixes #8072.
Zoltan Varga [Thu, 1 Nov 2012 14:41:20 +0000 (15:41 +0100)]
Properly inflate arrays of generic parameters in resolve_object (). Fixes #7964.
Zoltan Varga [Thu, 1 Nov 2012 04:29:24 +0000 (05:29 +0100)]
Make the AOT statistic output more dense.
Andres G. Aragoneses [Wed, 31 Oct 2012 23:33:20 +0000 (23:33 +0000)]
[System.Configuration] IntegerValidator's maxValue was 0 even if not used
When using IntegerValidator, if only MinValue is used, max value should
not be set by default to 0. Fix and unit tests included.
This fixes BNC#654721 (https://bugzilla.novell.com/show_bug.cgi?id=654721)
and therefore allows Enyim.Memcached to be used with Mono.
Rodrigo Kumpera [Wed, 31 Oct 2012 21:52:10 +0000 (17:52 -0400)]
Make gc maps optional and disabled by default for AOT.
Sebastien Pouliot [Wed, 31 Oct 2012 20:04:44 +0000 (16:04 -0400)]
Allow an existing 4.0 assembly to fill in for a 2.0 reference
Sebastien Pouliot [Wed, 31 Oct 2012 20:03:06 +0000 (16:03 -0400)]
Avoid processing (and creating) empty collections
Neale Ferguson [Wed, 31 Oct 2012 16:13:00 +0000 (12:13 -0400)]
Implement GetObjectData
Neale Ferguson [Thu, 11 Oct 2012 00:21:09 +0000 (20:21 -0400)]
Increase flushinput timeout to 1 second as 1/10th second is sometimes not enough
Marek Safar [Wed, 31 Oct 2012 10:24:33 +0000 (03:24 -0700)]
Merge pull request #489 from alesliehughes/master
Pathces for AxHosts, RegistrationServices and Marshal
Alistair Leslie-Hughes [Wed, 24 Oct 2012 05:36:32 +0000 (16:36 +1100)]
Implement Marshal.GenerateProgIdForType
Alistair Leslie-Hughes [Wed, 24 Oct 2012 05:35:16 +0000 (16:35 +1100)]
Implement RegistrationServices.GetManagedCategoryGuid
Alistair Leslie-Hughes [Wed, 24 Oct 2012 05:28:21 +0000 (16:28 +1100)]
Add AxHost.OcxState support
Sebastien Pouliot [Tue, 30 Oct 2012 20:36:39 +0000 (16:36 -0400)]
Allow overriding MarkMethod (without resolving the reference multiple times)
Marek Safar [Tue, 30 Oct 2012 15:56:16 +0000 (16:56 +0100)]
Correctly encode nested typeof type inside namespace.
Marek Safar [Tue, 30 Oct 2012 14:57:22 +0000 (15:57 +0100)]
Don't set sequential type attribute for async containers
Marek Safar [Tue, 30 Oct 2012 12:40:51 +0000 (13:40 +0100)]
Move corlib path check after nostdlib
Sebastien Pouliot [Mon, 29 Oct 2012 23:07:48 +0000 (19:07 -0400)]
Separate namespace and name arguments for Inherits to avoid memory allocations
Rodrigo Kumpera [Mon, 29 Oct 2012 21:10:28 +0000 (17:10 -0400)]
Propertly handle the layout of generic valuetypes. Fixes bxc #7574.
class.c (mono_class_layout_fields): All valuetype checks must account to
the fact that for generic instances the byval_arg/this_arg type will be
of a generic instance and not of a valuetype. The right check is for
MonoClass::valuetype. Fixes bxc #7574.
Marek Safar [Mon, 29 Oct 2012 17:01:42 +0000 (18:01 +0100)]
Add new test
Marek Safar [Mon, 29 Oct 2012 16:49:43 +0000 (09:49 -0700)]
Merge pull request #485 from mtausig/master
Bugfix for DirectoryInfo.GetDirectories()
mtausig [Mon, 29 Oct 2012 15:50:35 +0000 (16:50 +0100)]
Fix: Removed wrong increment
Marek Safar [Mon, 29 Oct 2012 15:34:16 +0000 (16:34 +0100)]
Add missing 3rd System reference
Martin Baulig [Mon, 29 Oct 2012 14:47:49 +0000 (15:47 +0100)]
[#7637][Web]: This is supposed to fail prior to .NET 4.0.
Fixes my previous commit
1ccb01e.
mtausig [Mon, 29 Oct 2012 12:29:24 +0000 (13:29 +0100)]
Removed null check of parameter (done in Directory.GetDirectories)
Changed foreach loop to for loop
Added some comments
Marek Safar [Mon, 29 Oct 2012 12:09:22 +0000 (13:09 +0100)]
Label inside implicit switch block is accessible from any switch case. Fixes #8036
Marek Safar [Mon, 29 Oct 2012 10:23:03 +0000 (11:23 +0100)]
Reject incomplete conversion unary user operators. Fixes #7935.
mtausig [Mon, 29 Oct 2012 10:06:41 +0000 (11:06 +0100)]
Added Testcase for DirectoryInfo.GetDirectories with parameter
SearchOption.AlLDirectories
mtausig [Mon, 29 Oct 2012 10:06:14 +0000 (11:06 +0100)]
Removed extra-brace
Shay Rojansky [Mon, 29 Oct 2012 00:26:39 +0000 (02:26 +0200)]
Fixed missing Content-Id email headers
When specifying a Content-Id header on Attachments and AlternateViews (but not LinkedResources) the header was not rendered.
This fixes bug #2279.
Also refactored header generation methods.
Pablo Ruiz Garcia [Sat, 27 Oct 2012 20:58:18 +0000 (22:58 +0200)]
Sys.Tx: Avoid uncatched exceptions during Prepare phase, which might crash the whole application. Also try to align internal exception handling flow/logic to the one found in MS.NET.
Zoltan Varga [Sat, 27 Oct 2012 01:10:26 +0000 (03:10 +0200)]
Disable the build of libmini.la when libraries are disabled.
Zoltan Varga [Sat, 27 Oct 2012 01:03:51 +0000 (03:03 +0200)]
Add a --disable-libraries command line option to disable the build of libmono.
Martin Baulig [Fri, 26 Oct 2012 23:15:21 +0000 (01:15 +0200)]
[#7637][Web]: Support empty write streams.
Martin Baulig [Fri, 26 Oct 2012 22:50:10 +0000 (00:50 +0200)]
[6329][Web]: Fixed abort race condition in WebConnection.ReadDone().
ReadDone() now passes the 'WebConnectionData' instance to GetResponse()
and 'readState' has been moved into 'WebConnectionData'.
Abort() can now reset 'Data' while ReadDone() still uses the old
instance, but learns about the abort request from 'data.ReadState'.
Rodrigo Kumpera [Fri, 26 Oct 2012 21:15:48 +0000 (17:15 -0400)]
Handle external protocols correctly in sdb.
* debugger-agent.c: Handle deferred attach in sdb correctly when
an external protocol is used.
Marek Safar [Fri, 26 Oct 2012 20:26:43 +0000 (13:26 -0700)]
Merge pull request #484 from roji/transactions_pspe
Transaction now has limited support for PromotableSinglePhaseEnlistment
Marek Safar [Fri, 26 Oct 2012 20:24:03 +0000 (13:24 -0700)]
Merge pull request #486 from fvalette/master
Fix System.Media.SoundPlayer on Linux
Zoltan Varga [Fri, 26 Oct 2012 20:13:27 +0000 (22:13 +0200)]
Disable LLVM for methods with clauses on darwin, current LLVM versions don't support it.
Zoltan Varga [Fri, 26 Oct 2012 18:10:38 +0000 (20:10 +0200)]
Align the trampoline stack size on MIPS.
Zoltan Varga [Fri, 26 Oct 2012 02:12:17 +0000 (04:12 +0200)]
Clear sb->cached_str in mono_string_utf8_to_builder () so if the same stringbuilder is passed to two pinvoke calls, the modifications made by the second are visible in the result of sb.ToString (). Fixes #7980.
Zoltan Varga [Fri, 26 Oct 2012 00:52:05 +0000 (02:52 +0200)]
Quiet some automake warnings.
Martin Baulig [Thu, 25 Oct 2012 22:57:25 +0000 (00:57 +0200)]
Emit WebAsyncResult's callback on same thread on sync completion.
This fixes System.Net.Http.HttpClientTest.Send_Content_Put ().
Martin Baulig [Thu, 25 Oct 2012 19:48:17 +0000 (21:48 +0200)]
Fix TaskFactory::FromAsync for methods that completed synchronously.
Don't call InnerInvoke() a 2nd time if the 'Begin' method that
completed synchronously already invoked the callback.