Scott Blomquist [Tue, 3 Jul 2012 06:56:21 +0000 (23:56 -0700)]
Fix dimension checking for array initializer expressions
Rodrigo Kumpera [Mon, 2 Jul 2012 20:51:49 +0000 (17:51 -0300)]
Add new gc options to the man page.
Rodrigo Kumpera [Mon, 2 Jul 2012 20:33:15 +0000 (17:33 -0300)]
Add two heap growth knobs: save-target-ratio and default-allowance-ratio.
Rodrigo Kumpera [Mon, 2 Jul 2012 19:55:32 +0000 (16:55 -0300)]
Simply a bit the memgov code. Kill a duplicated header declaration.
Rodrigo Kumpera [Mon, 2 Jul 2012 18:53:41 +0000 (15:53 -0300)]
Fix allowance calculation and introduce a constant in place of a magic number.
* sgen-conf.h: Introduce SGEN_DEFAULT_SAVE_TARGET_RATIO tunnable.
* sgen-memory-governor.c: Fix allowance calculation to use the
right size of the los space.
Rodrigo Kumpera [Mon, 2 Jul 2012 18:07:32 +0000 (15:07 -0300)]
Extract the nursery allowance heuristics to a memgov API.
Rodrigo Kumpera [Mon, 2 Jul 2012 15:32:36 +0000 (12:32 -0300)]
Move the sgen_*_os_memory APIs to memgov.
Rodrigo Kumpera [Mon, 2 Jul 2012 15:32:02 +0000 (12:32 -0300)]
Use new atomis with allocated_heap.
Rodrigo Kumpera [Mon, 2 Jul 2012 15:22:04 +0000 (12:22 -0300)]
Add SGEN_ATOMIC_ADD_P for pointer sized integers.
Rodrigo Kumpera [Fri, 29 Jun 2012 15:25:37 +0000 (12:25 -0300)]
Add some FIXME.
Rodrigo Kumpera [Thu, 21 Jun 2012 13:45:31 +0000 (10:45 -0300)]
Move the heap limits code to the memgov file.
Rodrigo Kumpera [Wed, 20 Jun 2012 18:51:52 +0000 (15:51 -0300)]
Add new memory governor files.
Martin Baulig [Mon, 2 Jul 2012 18:42:21 +0000 (20:42 +0200)]
Fix #5169: Environment.ExpandEnvironmentVariables().
Martin Baulig [Mon, 2 Jul 2012 17:06:43 +0000 (19:06 +0200)]
Re-enable CSharpCodeProviderTest on the Mac.
I just had a look at these and can't reproduce the problem anymore,
these tests are now working fine on my Mac.
This reverts commit
58ae80aa5e28610c26cb7472e58e156b1db6735b.
Martin Baulig [Mon, 2 Jul 2012 16:52:40 +0000 (18:52 +0200)]
WebRequestTest.cs: Add workaround for t-online.de's DNS server.
Martin Baulig [Mon, 2 Jul 2012 16:42:24 +0000 (18:42 +0200)]
DnsTest.cs: Add workaround for t-online.de's DNS server.
Martin Baulig [Mon, 2 Jul 2012 16:31:56 +0000 (18:31 +0200)]
Fix go-mono.com's IP address in DnsTest.cs.
go-mono.com has 50.21.183.68, not 74.208.222.19.
Rodrigo Kumpera [Mon, 2 Jul 2012 14:20:43 +0000 (11:20 -0300)]
Use isfinite instead of finite as newer xcode removed it.
Rodrigo Kumpera [Mon, 2 Jul 2012 14:17:08 +0000 (11:17 -0300)]
Merge branch 'master' of github.com:mono/mono
Rolf Bjarne Kvinge [Mon, 2 Jul 2012 13:42:58 +0000 (15:42 +0200)]
Make mkbundle bundle assemblies a lot faster.
Preallocate the 256 different lines we might write to the assembly file
instead of allocating several strings for every line.
With compression mkbundle is about twice as fast now (since compression
speed hasn't changed), without compression it's 5-10x faster, depending
on the number of assemblies to embed.
Rodrigo Kumpera [Mon, 2 Jul 2012 13:21:18 +0000 (10:21 -0300)]
Merge branch 'master' of github.com:mono/mono
Mark Probst [Fri, 29 Jun 2012 17:28:19 +0000 (19:28 +0200)]
[sgen] Make heavy statistics compile again.
Marek Safar [Mon, 2 Jul 2012 10:51:06 +0000 (11:51 +0100)]
Update corcompare
Marek Safar [Mon, 2 Jul 2012 10:50:39 +0000 (11:50 +0100)]
Add missing attribute classes
Marek Safar [Mon, 2 Jul 2012 07:37:18 +0000 (00:37 -0700)]
Merge pull request #362 from JamesB7/system-security
Fix System.Security.AccessControl.MutexAccessRule's constructor.
Marek Safar [Mon, 2 Jul 2012 06:54:22 +0000 (23:54 -0700)]
Merge pull request #367 from robwilkens/work1
Fixes a problem that was preventing mono from compiling
Rob Wilkens [Mon, 2 Jul 2012 02:02:21 +0000 (22:02 -0400)]
Fixes a problem that was preventing mono from compiling
Somehow the compiler became more strict, and a not-implemented class is now
causing mono to fail to build.
The attached patch doesn't really fix anything besides letting it compile.
I was nice enough to throw a NotImplementedException() on the default
constructor.
The compile time error was:
System.IO.MemoryMappedFiles/MemoryMappedFileSecurity.cs(35,15): error CS0122:
`System.Security.AccessControl.ObjectSecurity
<System.IO.MemoryMappedFiles.MemoryMappedFileRights>.ObjectSecurity()' is
inaccessible due to its protection level
Feel free to reject this patch if I am the only one seeing this problem (on
Linux x86_64 build) after most recent git clone of mono/master.
James Bellinger [Sun, 1 Jul 2012 04:59:14 +0000 (00:59 -0400)]
Added a unit test for the bug this is fixing -- previously the constructor just crashed, always.
So, this constructs with valid MutexAccessRights, and also makes sure that the rights are correct.
James Bellinger [Sat, 30 Jun 2012 11:16:56 +0000 (07:16 -0400)]
MutexAccessRule... previously entirely broken as it passed 0 for accessMask.
Now, accessMask matches the MutexRules passed in. The MutexRules enum is a bitmask
corresponding to Win32 ACE accessMasks and the two match.
I verified that accessMask corresponds directly to the MutexRules passed in the
constructor on Win32 using Reflection. It stands to reason, and it turns out to be
the case. System.Security.AccessControl is still mostly broken but this at least
corrects MutexAccessRule.
Marek Safar [Sun, 1 Jul 2012 16:05:58 +0000 (09:05 -0700)]
Merge pull request #363 from JamesB7/system-security2
Implemented the .NET 4.0 generic versions of AccessRule and AuditRule.
James Bellinger [Sun, 1 Jul 2012 15:06:23 +0000 (11:06 -0400)]
Added unit test.
Also added constructors to ObjectSecurity<T>.
It's supposed to have them (they just pass to NativeObjectSecurity),
but the only user in Mono is MemoryMappedFileSecurity so this is
just for completeness's sake. Also, this lets the test depend on neither
System.Core nor internal methods.
Marek Safar [Sun, 1 Jul 2012 13:26:06 +0000 (06:26 -0700)]
Merge pull request #365 from JamesB7/system-security3
AuthorizationRule now accepts NTAccount like it should. Provided a unit test.
James Bellinger [Sun, 1 Jul 2012 13:17:05 +0000 (09:17 -0400)]
AuthorizationRule now accepts NTAccount like it should. Provided a unit test.
Many derived classes of AuthorizationRule have string overloads which pass
'new NTAccount (identity)'. The exception docs state "can be cast as a
SecurityIdentifier" not "is a SecurityIdentifier". Anyway, this fixes all the
string constructors in System.Security.AccessControl.
Also, I added checking for inheritanceFlags and propagationFlags parameters.
While building the unit test, I found accessMask == 0 throws ArgumentException,
not ArgumentOutOfRangeException, on MS.NET 4.0 (contrary to docs). So,
I've made this match.
Zoltan Varga [Sat, 30 Jun 2012 23:30:02 +0000 (01:30 +0200)]
Fix altstack on osx/amd64. Fixes #5933.
Alan McGovern [Sat, 30 Jun 2012 22:55:38 +0000 (18:55 -0400)]
[Xbuild] Use OrdinalIgnoreCase not InvariantCultureIgnoreCase
It's chewing up a fair bit of CPU time and buys us nothing. We only care
about ordinal ignore case for these scenarios.
James Bellinger [Sat, 30 Jun 2012 17:21:46 +0000 (13:21 -0400)]
Implemented the .NET 4.0 generic versions of AccessRule and AuditRule.
Details worth being aware of, that aren't well documented in MS.NET:
(1) The Rights property is just a casted AccessMask.
I've checked this by testing with MS.NET -- if you use AccessRule<Test>
with an arbitrary struct Test { }, MS.NET has an InvalidCastException
on the Rights getter if you create with ObjectSecurity's factory method
(which is passed an int), and in the constructor if you use AccessRule's
constructor.
(2) The string constructor overloads call 'new NTAccount(identity)'.
What these methods do for AccessRule is not documented by MS.NET,
but this *is* documented for SemaphoreAccessRule, and my testing
with C# shows AccessRule does the same thing. So, I've implemented
this behavior to match.
I will submit a separate patch afterwards to make AuthorizationRule's
accept NTAccount, which is required for the string overloads to work.
Contrary to the MS.NET *documentation* (which Mono follows), AuthorizationRule
actually accepts NTAccount as well as (the documented) SecurityIdentifier.
I verified this by deriving a class from AuthorizationRule.
Martin Baulig [Sat, 30 Jun 2012 04:40:54 +0000 (06:40 +0200)]
Fix UdpClientTest.CloseInReceive() on Mac OS.
On my Mac, somethig is already listening on port 50000; use
a random one instead.
Martin Baulig [Sat, 30 Jun 2012 04:10:09 +0000 (06:10 +0200)]
Disable failing tests on the Mac; filed bug #5934 about it.
Martin Baulig [Sat, 30 Jun 2012 02:43:34 +0000 (04:43 +0200)]
Improve SocketTests.EndConnect().
Rename to BogusEndConnect() and accept a few more error codes.
Martin Baulig [Sat, 30 Jun 2012 01:21:00 +0000 (03:21 +0200)]
SocketTest.ConnectMultiple() for Mac OS.
Rodrigo Kumpera [Fri, 29 Jun 2012 21:27:18 +0000 (18:27 -0300)]
Add test for pinned gc handle.
Rodrigo Kumpera [Fri, 29 Jun 2012 21:25:39 +0000 (18:25 -0300)]
Fix a pair of bugs in the gchandle code.
* gc.c: Make sure pinned handles actually pin on sgen.
* gc.c: Use gc_memmove and gc_bzero when dealing with memory
that has managed references.
Miguel de Icaza [Fri, 29 Jun 2012 19:11:55 +0000 (15:11 -0400)]
Revert the fix for #5747 from
8b5f71ef60146590193298cc870bf72e14b47950, it broke XSP monitoring of the file system
Zoltan Varga [Fri, 29 Jun 2012 18:39:00 +0000 (20:39 +0200)]
Remove empty log files when using test-wrench.
Zoltan Varga [Fri, 29 Jun 2012 18:09:55 +0000 (20:09 +0200)]
Remove some assembly from mkbundle which doesn't work on osx 64 bit.
Rodrigo Kumpera [Fri, 29 Jun 2012 15:46:39 +0000 (12:46 -0300)]
Fix race condition in static's inialization of mono_message_init.
* object.c (mono_message_init): Make sure the static field
we test fo check for initialization is the last to init or
we risk an abort due to others been null.
Marek Safar [Fri, 29 Jun 2012 10:30:47 +0000 (11:30 +0100)]
Remove all ifdefs
Marek Safar [Fri, 29 Jun 2012 10:01:44 +0000 (11:01 +0100)]
Use correct iterator return type for XPathEvaluate. Fixes #5902
Zoltan Varga [Thu, 28 Jun 2012 23:59:25 +0000 (01:59 +0200)]
Fix the test-messages test.
Zoltan Varga [Thu, 28 Jun 2012 23:35:06 +0000 (01:35 +0200)]
Fix warnings.
Gonzalo Paniagua Javier [Thu, 28 Jun 2012 19:48:57 +0000 (15:48 -0400)]
Attempt to fix #1553
This should prevent a popular exception to happen again.
Martin Baulig [Thu, 28 Jun 2012 18:41:50 +0000 (20:41 +0200)]
Enable the multicast socket tests on the Mac.
Martin Baulig [Thu, 28 Jun 2012 17:43:19 +0000 (19:43 +0200)]
Fix #5504: UdpClient.JoinMulticastGroup().
* configure.in: Check for if_nametoindex().
* socket-io.c (ves_icall_System_Net_Sockets_Socket_SetSocketOption_internal):
Mac OS doesn't allow interface = 0; use the interface id from the first
non-loopback interface.
* UdpClientTest.cs: Enable JoinMulticastGroup1_IPv6,
JoinMulticastGroup2_IPv6 and JoinMulticastGroup3_IPv6 on the Mac.
Martin Baulig [Thu, 28 Jun 2012 17:34:45 +0000 (19:34 +0200)]
socket-io: Fix endianess in ipaddress_to_struct_in6_addr.
Marek Safar [Thu, 28 Jun 2012 17:14:13 +0000 (18:14 +0100)]
Add new test
Marek Safar [Thu, 28 Jun 2012 08:52:26 +0000 (09:52 +0100)]
Remove NET_2_0
Rodrigo Kumpera [Thu, 28 Jun 2012 15:38:11 +0000 (12:38 -0300)]
Fix posix build.
Rodrigo Kumpera [Wed, 27 Jun 2012 22:57:35 +0000 (19:57 -0300)]
Make the thread abort code multi-thread safe.
Rodrigo Kumpera [Wed, 27 Jun 2012 20:36:54 +0000 (17:36 -0300)]
With Socket:Dispose from using Thread.Abort to less agressive alternative.
* Socket_2_1.cs: When closing a socket we need to abort any
pending syscalls to that socket. For most syscalls the kernel
does it for us. But accept/connect on linux requires extra
love.
We previously used Thread.Abort to implement that, but this
is a pretty radical solution with very complicated ramifications.
So instead now we use a new icall cancel_blocking_socket_operation
that makes sure we only abort the syscall and do nothing else.
This reduces the trouble surface a lot given how tricky
Thread.Abort is.
Rodrigo Kumpera [Wed, 27 Jun 2012 20:22:18 +0000 (17:22 -0300)]
Async close on mac raises ECONNABORTED for connect().
Rodrigo Kumpera [Wed, 27 Jun 2012 20:20:29 +0000 (17:20 -0300)]
Add new cancel_blocking_socket_operation icall to Socket.
* socket-io.c: Add cancel_blocking_socket_operation icall
to Socket so we can abort socket syscalls on other threads
when we decide to close a socket.
Rodrigo Kumpera [Wed, 27 Jun 2012 20:18:37 +0000 (17:18 -0300)]
Respect the new signal ignore flag.
* mini-posix.c: If ignore_next_signal is set, we ignore the
current USR1. This is used to abort syscalls.
Rodrigo Kumpera [Wed, 27 Jun 2012 20:12:12 +0000 (17:12 -0300)]
New field InternalThread::ignore_next_signal. Bump corlib version due to layout change.
Rodrigo Kumpera [Wed, 27 Jun 2012 19:55:18 +0000 (16:55 -0300)]
Add mono_thread_info_abort_socket_syscall_for_close to the mono-threads API.
* mono-threads.c: New functin mono_thread_info_abort_socket_syscall_for_close
specially dedicated to our most special creature - linux.
Say you want to close a socket and make sure a thread making a syscall on it
will have this be aborted, call this method.
Duncan Mak [Wed, 27 Jun 2012 20:49:41 +0000 (16:49 -0400)]
Bump version number to 2.11.3.
Marek Safar [Wed, 27 Jun 2012 20:21:31 +0000 (13:21 -0700)]
Merge pull request #358 from adbre/iss3333
Do not support DST if no adjustment rules are given
Adam Brengesjö [Wed, 27 Jun 2012 18:26:20 +0000 (20:26 +0200)]
TimeZoneInfo: disable DST if no adjustment rules are given.
Adam Brengesjö [Wed, 27 Jun 2012 18:24:13 +0000 (20:24 +0200)]
Unit tests for issue #3333
Marek Safar [Wed, 27 Jun 2012 16:44:29 +0000 (09:44 -0700)]
Merge pull request #357 from Suremaker/master
MonoTests.System.Xml.XmlSchemaDatatypeTests: corrected expected/actual parameters position in Assert.AreEqual method calls.
Wojciech Kotlarski [Wed, 27 Jun 2012 16:04:59 +0000 (17:04 +0100)]
MonoTests.System.Xml: corrected expected/actual parameters position in Assert.AreEqual method.
Tests were written first in .NET environment using Assert.That method which has different order of expected/actual parameters.
Marek Safar [Wed, 27 Jun 2012 16:04:33 +0000 (09:04 -0700)]
Merge pull request #356 from Suremaker/master
MonoTests.System.Xml.XmlSchemaDatatypeTests ChangeType_FromDateTimeTest / ChangeType_ToDateTimeTest correction
Wojciech Kotlarski [Wed, 27 Jun 2012 15:44:10 +0000 (16:44 +0100)]
corrected tests to not use UtcNow
Marek Safar [Wed, 27 Jun 2012 15:40:34 +0000 (08:40 -0700)]
Merge pull request #355 from Suremaker/master
System.XML: added null arguments validation for targetType and namespaceResolver in System.Xml.Schema.XmlSchemaDatatype.ChangeType method
Wojciech Kotlarski [Wed, 27 Jun 2012 15:31:20 +0000 (16:31 +0100)]
System.XML: added null arguments validation for targetType and namespaceResolver in System.Xml.Schema.XmlSchemaDatatype.ChangeType methods
Atsushi Eno [Wed, 27 Jun 2012 14:59:10 +0000 (07:59 -0700)]
Merge pull request #354 from Suremaker/master
Partial implementation of System.Xml.Schema.XmlSchemaDatatype.ChangeType
Wojciech Kotlarski [Wed, 27 Jun 2012 14:50:29 +0000 (15:50 +0100)]
System.XML: Partial implementation of XmlSchemaDatatype::ChangeType
This is not complete but it meets our current validation needs
(better than NIE). The missing stuff is still marked as
[MonoTODO] and [Cattegory("NotWorking")].
Wojciech Kotlarski [Wed, 27 Jun 2012 11:15:30 +0000 (12:15 +0100)]
System.XML: Remove more NET_2_0 directives
Wojciech Kotlarski [Wed, 27 Jun 2012 11:07:49 +0000 (12:07 +0100)]
System.XML: Remove deprecated NET_2_0 preprocessor conditionals
Marek Safar [Wed, 27 Jun 2012 10:15:50 +0000 (11:15 +0100)]
Better error message
Marek Safar [Wed, 27 Jun 2012 07:39:25 +0000 (08:39 +0100)]
ifdef 4.0 using
Martin Baulig [Wed, 27 Jun 2012 04:36:12 +0000 (06:36 +0200)]
Enable all tests on the Mac.
Martin Baulig [Wed, 27 Jun 2012 02:19:21 +0000 (04:19 +0200)]
Implement Dns.GetHostEntry(string.Empty) on Mac OS.
* configure.in: Add check for getifaddrs().
* socket-io.c: Implement get_local_ips() using getifaddrs()
if SIOCGIFCONF is not available.
* DnsTest.cs: Add GetHostEmpty_StringEmpty().
* UdpClientTest.cs(): Enable Available(), BeginReceive() and BeginSend()
on the Mac.
Martin Baulig [Wed, 27 Jun 2012 00:22:10 +0000 (02:22 +0200)]
Fix UdpClientTest.UdpClientBroadcastTest().
You must not use Send() with an endpoint after using either the
hostname constructor or calling Connect(); see the MSDN docs.
Martin Baulig [Wed, 27 Jun 2012 00:08:03 +0000 (02:08 +0200)]
Fix UdpClientTest.Constructor6(): UDP port 0 seems to be invalid.
Marek Safar [Tue, 26 Jun 2012 21:58:26 +0000 (22:58 +0100)]
Don't register namespace of predefined (optinal) types
Marek Safar [Tue, 26 Jun 2012 21:43:44 +0000 (22:43 +0100)]
Remove wrong usings
Martin Baulig [Tue, 26 Jun 2012 19:52:23 +0000 (21:52 +0200)]
UdpClientTest.Constructor5 now works on the Mac.
Martin Baulig [Tue, 26 Jun 2012 19:39:00 +0000 (21:39 +0200)]
create_object_from_sockaddr(): map ::ffff:0:0 to :: (bug #5502).
::ffff:0:0 is the IPv4 mapped address 0.0.0.0, translate this to ::
(IPAddress.IPv6Any).
Martin Baulig [Tue, 26 Jun 2012 16:52:07 +0000 (18:52 +0200)]
Fix race condition in TimerTest.TestChange().
Set the first interval to 10 milleseconds, not just one. Then read
the bucket.count again after changing the interval, the callback may
fire a few times while we execute the Assert() and t.Change() callls.
Martin Baulig [Tue, 26 Jun 2012 16:33:46 +0000 (18:33 +0200)]
Fix DateTime.ToFileTimeUtf (bug #5697).
Marek Safar [Tue, 26 Jun 2012 12:57:55 +0000 (13:57 +0100)]
Bump DoubleWaitTest timeout
Marek Safar [Tue, 26 Jun 2012 12:25:09 +0000 (13:25 +0100)]
Fix System.Core reference in VS projects
Marek Safar [Tue, 26 Jun 2012 12:24:54 +0000 (13:24 +0100)]
Fix few warnings
Marek Safar [Tue, 26 Jun 2012 11:15:51 +0000 (12:15 +0100)]
Update VS project files
Marek Safar [Tue, 26 Jun 2012 10:38:44 +0000 (11:38 +0100)]
Remove static file index from location and fix few hidden xml parser bugs
Zoltan Varga [Tue, 26 Jun 2012 08:02:25 +0000 (01:02 -0700)]
Merge pull request #353 from robwilkens/bug2014
For Bug 2014: map FcntlCommand to local system command for fcntl_lock
Miguel de Icaza [Tue, 26 Jun 2012 03:57:09 +0000 (00:57 -0300)]
Use Ordinal string comparison here.
Martin Baulig [Tue, 26 Jun 2012 03:56:15 +0000 (05:56 +0200)]
Merge branch 'master' of github.com:mono/mono
Miguel de Icaza [Tue, 26 Jun 2012 03:54:37 +0000 (20:54 -0700)]
Merge pull request #334 from robwilkens/bug5747
Bug 5747 (MacOSX)Fix: KeventWatcher didn't return subdir names in filenames
Martin Baulig [Tue, 26 Jun 2012 03:53:51 +0000 (05:53 +0200)]
Include quotation marks in cookie value.