Mark Probst [Tue, 26 Jul 2011 13:50:05 +0000 (15:50 +0200)]
[sgen] Make mono_sgen_safe_object_get_size() handle forwarded objects.
This is sometimes called on forwarded objects from the major collector
during finalization. Now it doesn't crash anymore.
Zoltan Varga [Mon, 25 Jul 2011 13:35:52 +0000 (15:35 +0200)]
Fix windows build.
Miguel de Icaza [Mon, 25 Jul 2011 23:08:43 +0000 (19:08 -0400)]
Fix invocation to syslog
Atsushi Eno [Mon, 25 Jul 2011 19:25:00 +0000 (04:25 +0900)]
Fix bug #16: use Owner node instead of Parent. Patch by Martin Potter.
Sebastien Pouliot [Mon, 25 Jul 2011 13:58:14 +0000 (09:58 -0400)]
Fixed decoding of PKCS/CMS message with detached ContentInfo. Pull request from Pablo Ruiz Garcia <pablo.ruiz@gmail.com>
Zoltan Varga [Sun, 24 Jul 2011 18:41:58 +0000 (20:41 +0200)]
Fix a crash in the soft-float decompose code when a branch following an fcompare is optimized away. Fixes #705186.
Zoltan Varga [Tue, 12 Jul 2011 01:18:46 +0000 (03:18 +0200)]
Mark Probst [Sat, 23 Jul 2011 19:29:30 +0000 (21:29 +0200)]
Remove references to mono_pthread_key_for_tls(). Fixes builds.
Sebastien Pouliot [Sat, 23 Jul 2011 19:09:56 +0000 (15:09 -0400)]
add put support in certmgr (git pull request 124 by rabink)
Sebastien Pouliot [Sat, 23 Jul 2011 15:03:44 +0000 (11:03 -0400)]
Expose new ReadLines extension methods in the mobile-based profiles
Atsushi Eno [Fri, 22 Jul 2011 14:20:08 +0000 (07:20 -0700)]
Merge pull request #137 from eisnerd/bug-xpath-reverse-axes-497017
XPath reverse axes bug #497017
Mark Probst [Fri, 22 Jul 2011 09:55:30 +0000 (11:55 +0200)]
TLS is always fast now, so remove NO_TLS_SET_VALUE in domain.c.
Mark Probst [Fri, 22 Jul 2011 09:48:36 +0000 (11:48 +0200)]
Remove io-layer Windows API TLS emulation code.
We should always use native TLS. The Windows TLS emulation adds
an unnecessary layer and even slows us down considerably on some
platforms: The TlsGetValue() code not only does a TLS get, but,
since it calls SetLastError(), also a TLS set, so on platforms
with slow TLS set, TLS get will also be slow.
This will almost certainly break something. I beg forgiveness.
Atsushi Eno [Fri, 22 Jul 2011 05:02:48 +0000 (14:02 +0900)]
Avoid possible duplicate xmlns output. Fixes xsp-based hosting somehow.
David Eisner [Thu, 21 Jul 2011 22:28:32 +0000 (23:28 +0100)]
Removal of XPath debugging code relating to bug #497017, leaving the new test cases
David Eisner [Thu, 21 Jul 2011 22:20:13 +0000 (23:20 +0100)]
XPath test cases and debugging code relating to bug #497017
David Eisner [Thu, 21 Jul 2011 22:19:00 +0000 (23:19 +0100)]
Corrected XPath iterator for ancestor-or-self axes for runaway behaviour
Adding a test case for ancestor-or-self along the lines of the existing one for ancestor showed runaway, memory-eating behaviour. In changing CollectResults to remove the livelock, it seemed simplest to remove the extra yield of the root from MoveNextCore. I was unable to ascertain a good reason for producing this element in this special way compared to the AncestorIterator.
David Eisner [Thu, 21 Jul 2011 22:07:02 +0000 (23:07 +0100)]
Fixed the XPath built-in position function for bug #497017
According to http://www.w3.org/TR/xpath#predicates, a predicate [n] should be equivalent to [position()=n] for a number expression n. However, we currently have a PredicateIterator intepretation of [n] that has been corrected for the issue outlined in the bug report, leaving the position() function behind with an inappropriate, document-order value for reverse axes. Note that [n] is not treated by parsing or transforming to the same form as [position()=n] before evaluation, valid as that would be.
The discussion in that same section of the recommendation about forward and reverse axes and further remarks in http://www.w3.org/TR/xpath#node-sets make it clear that the predicate [1] has the sense of "nearest" rather than "first in the document" which then, depending on the axis type, may be the same or reverse of the document order. It does not seem to me to be so clear whether the order given by a navigator for a location path expression should mirror this or always give the document order. The .Net implementation seems to always give the document order, as well as predicates having the correct reverse/forward senses, so I have added test cases that explore both.
Zoltan Varga [Thu, 21 Jul 2011 14:33:37 +0000 (16:33 +0200)]
Remove obsolete and broken msvc05 files.
Marek Safar [Thu, 21 Jul 2011 11:33:06 +0000 (12:33 +0100)]
Include context member name in async wrapper to easier debugging
Marek Safar [Thu, 21 Jul 2011 10:03:09 +0000 (11:03 +0100)]
More emit updates to handle await sub-expressions
Atsushi Eno [Wed, 20 Jul 2011 19:57:32 +0000 (04:57 +0900)]
Fix bug #703910 - implement Order support in XmlSerializer.
Marek Safar [Wed, 20 Jul 2011 17:30:07 +0000 (18:30 +0100)]
Type details are no longer need when emitting ldloc
Marek Safar [Wed, 20 Jul 2011 11:31:22 +0000 (12:31 +0100)]
Remove failed approach to await stack handling
Marek Safar [Wed, 20 Jul 2011 10:55:36 +0000 (11:55 +0100)]
Handle await in liften binary expression with compound assignment
Marek Safar [Wed, 20 Jul 2011 09:47:19 +0000 (10:47 +0100)]
Report full method name for IL code mismatch
Sebastien Pouliot [Wed, 20 Jul 2011 17:29:46 +0000 (13:29 -0400)]
Allow IsolatedStorage to work on MOBILE profiles even when CAS types are linked out
Sebastien Pouliot [Wed, 20 Jul 2011 15:39:48 +0000 (11:39 -0400)]
another case where the linker needs help to eliminate CAS
Sebastien Pouliot [Wed, 20 Jul 2011 12:12:49 +0000 (08:12 -0400)]
Ensure the linker can eliminate most CAS-related types in the MOBILE-based profiles
Sebastien Pouliot [Wed, 20 Jul 2011 11:58:29 +0000 (07:58 -0400)]
Ensure the linker can eliminate the *RequestCreator code without extra instructions.
Marek Safar [Wed, 20 Jul 2011 08:03:11 +0000 (09:03 +0100)]
Honour command line checked option during emit phase
Geoff Norton [Wed, 20 Jul 2011 02:33:09 +0000 (22:33 -0400)]
Type matching is fun
Geoff Norton [Wed, 20 Jul 2011 02:28:17 +0000 (22:28 -0400)]
info->signal does not exist on MACH either
Geoff Norton [Wed, 20 Jul 2011 02:27:01 +0000 (22:27 -0400)]
info->stop_count does not exist on MACH
Miguel de Icaza [Tue, 19 Jul 2011 23:35:08 +0000 (19:35 -0400)]
Fix build
Mark Probst [Tue, 19 Jul 2011 16:48:12 +0000 (18:48 +0200)]
New function: mono_valloc_aligned().
Some platforms have APIs for this, so it makes sense to
use them instead of our wasteful approach of allocing more
memory and then freeing off the edges.
Mark Probst [Tue, 19 Jul 2011 15:48:35 +0000 (17:48 +0200)]
[sgen] rehash_roots() must take an int, not a boolean.
Mark Probst [Tue, 19 Jul 2011 12:51:38 +0000 (14:51 +0200)]
[sgen] Use g_usleep() instead of usleep().
Mark Probst [Tue, 19 Jul 2011 11:57:55 +0000 (13:57 +0200)]
[sgen] Include config.h in all of SGen's compilation units.
Jonathan Pryor [Tue, 19 Jul 2011 16:01:08 +0000 (12:01 -0400)]
Add System.Data.Services.Client to monotouch and monodroid profiles.
Sebastien Pouliot [Tue, 19 Jul 2011 14:34:31 +0000 (10:34 -0400)]
remove unrequired using clauses
Sebastien Pouliot [Thu, 14 Jul 2011 20:40:41 +0000 (16:40 -0400)]
avoid potential InvalidCastException when we already deal with bad/non-present extensions (reduce error message when -E:all is used)
Marek Safar [Tue, 19 Jul 2011 13:34:50 +0000 (14:34 +0100)]
Fix resolving referenced extension methods used in nested namespace
Marek Safar [Tue, 19 Jul 2011 07:55:52 +0000 (08:55 +0100)]
Emit valid code for nullable mutators used as an expression
Marek Safar [Fri, 15 Jul 2011 15:08:36 +0000 (16:08 +0100)]
Await support for more expressions
Atsushi Eno [Tue, 19 Jul 2011 00:05:27 +0000 (09:05 +0900)]
fix build regarding ContentType.
Rodrigo Kumpera [Mon, 18 Jul 2011 18:38:35 +0000 (15:38 -0300)]
Fix the altstack check so it works under osx.
Rodrigo Kumpera [Mon, 18 Jul 2011 18:42:45 +0000 (15:42 -0300)]
Do some portability adjustments so altstack works on osx.
* mini-exceptions.c (mono_setup_altstack): Use stack_t instead
of struct sigaltstack. Don't set ss_flags as OSX doesn't like it.
* mini-exceptions.c (mono_free_altstack): Use stack_t instead
of struct sigaltstack.
Rodrigo Kumpera [Mon, 18 Jul 2011 18:13:16 +0000 (15:13 -0300)]
On OSX do SIGBUS on altstack as well since it's used for *0.
Rodrigo Kumpera [Mon, 18 Jul 2011 18:09:15 +0000 (15:09 -0300)]
Handle altstack exceptions in a more portable way on x86.
* exceptions-x86.c (mono_arch_handle_altstack_exception): Push a
MonoContext to the stack instead of an ucontext. It's smaller
and more portable since we don't have to bother with internal
pointers.
Rodrigo Kumpera [Sun, 17 Jul 2011 16:42:34 +0000 (13:42 -0300)]
Fix stack overflow code to work under darwin-x86.
* mini-x86.h: Respect MINSIGSTKSZ under OSX.
* mini-amd64.h: Ditto.
Rodrigo Kumpera [Sun, 17 Jul 2011 01:32:10 +0000 (22:32 -0300)]
Fix mono_thread_get_stack_bounds thread bounds.
Mark Probst [Mon, 18 Jul 2011 17:59:45 +0000 (19:59 +0200)]
[sgen] Use ptrdiff_t instead of ssize_t.
Mark Probst [Mon, 18 Jul 2011 17:58:11 +0000 (19:58 +0200)]
[sgen] Move all signal stuff into sgen-os-posix.c.
This might make Mach stop working or uncompilable.
Mark Probst [Mon, 18 Jul 2011 15:27:13 +0000 (17:27 +0200)]
[gc] Don't do unaligned loads in mono_gc_memmove ().
This caused problems on S390x, at least.
Atsushi Eno [Mon, 18 Jul 2011 17:28:59 +0000 (02:28 +0900)]
support action parameter in SOAP 1.2 content type. Close pull request #133.
Atsushi Eno [Sat, 16 Jul 2011 13:48:41 +0000 (06:48 -0700)]
Fix bug 704813: just ignore members of type IEnumerable`1 without Add().
Rodrigo Kumpera [Fri, 15 Jul 2011 20:54:38 +0000 (17:54 -0300)]
If a thread abort is fatal, print the exception before dying.
Rodrigo Kumpera [Fri, 15 Jul 2011 20:49:41 +0000 (17:49 -0300)]
Improve stacktraces when facing a runtime invoke wrapper.
* mini-exceptions.c (mono_handle_exception_internal_first_pass):
Don't trim the trace list if we find a filter clause from a wrapper.
This makes stacktraces across MRI's work.
Extract trace trimming into a macro to reduce code duplication.
Marek Safar [Fri, 15 Jul 2011 14:53:17 +0000 (15:53 +0100)]
Await support for more expressions
Marek Safar [Thu, 14 Jul 2011 12:22:37 +0000 (13:22 +0100)]
Add await support for more expressions
Marek Safar [Thu, 14 Jul 2011 08:16:16 +0000 (09:16 +0100)]
Use array index instead of an iterator when checking interfaces
Mark Probst [Wed, 13 Jul 2011 17:07:52 +0000 (19:07 +0200)]
[sgen] Option for an allowance limit.
The metric for deciding when to trigger a major collection assumes
that there's unlimited memory, so it does, especially since it's
probably buggy and certainly not very smart, sometimes allow too much
allocation to happen. This new option is a way to tell SGen where to
draw the line. Note that memory will still be allocated, it's just
that once the limit is overstepped, major collections happen more
frequently to keep memory usage down.
Marek Safar [Wed, 13 Jul 2011 16:25:06 +0000 (17:25 +0100)]
Implement await for new expressions
Marek Safar [Wed, 13 Jul 2011 16:00:22 +0000 (17:00 +0100)]
Implement async inside a method with MVARs and contraints
Marek Safar [Wed, 13 Jul 2011 08:39:18 +0000 (09:39 +0100)]
Await for array access expression
Marek Safar [Tue, 12 Jul 2011 12:45:59 +0000 (13:45 +0100)]
Return from async method has to also skip following code block
Rodrigo Kumpera [Tue, 12 Jul 2011 14:01:16 +0000 (11:01 -0300)]
Avoid a .cctor on the initialization path.
Rodrigo Kumpera [Tue, 12 Jul 2011 14:01:10 +0000 (11:01 -0300)]
Add shared_perfcounters option to enable minimal.
Mark Probst [Tue, 12 Jul 2011 15:37:32 +0000 (17:37 +0200)]
[gc] Fix a bug in atomic memmove.
Marek Safar [Tue, 12 Jul 2011 08:26:47 +0000 (09:26 +0100)]
Modify few emits to deal with await expression
Rodrigo Kumpera [Tue, 12 Jul 2011 04:32:25 +0000 (01:32 -0300)]
Do DISABLE_PORTABILITY more agressively.
Zoltan Varga [Tue, 12 Jul 2011 00:49:42 +0000 (02:49 +0200)]
Store precise wrapper info for native-to-managed wrappers, needed by AOT.
Zoltan Varga [Tue, 12 Jul 2011 00:08:29 +0000 (02:08 +0200)]
Avoid register allocation problems when a try clause is nested inside a handler block.
Zoltan Varga [Mon, 11 Jul 2011 21:16:40 +0000 (23:16 +0200)]
Get rid of the 'ut_e_' symbols in AOT, use the normal 'ut_' symbols for them.
Zoltan Varga [Mon, 11 Jul 2011 15:10:04 +0000 (17:10 +0200)]
Fix the length of r4const on arm.
Mark Probst [Fri, 8 Jul 2011 19:47:00 +0000 (21:47 +0200)]
[sgen] Debug option for printing the heap usage and minor collection allowance.
Zoltan Varga [Sun, 10 Jul 2011 22:23:36 +0000 (00:23 +0200)]
Fix caching in mono_marshal_get_array_address ().
Rodrigo Kumpera [Sat, 9 Jul 2011 16:10:44 +0000 (13:10 -0300)]
Fix signal chaining to handle a stupid previous handler.
Rodrigo Kumpera [Fri, 8 Jul 2011 23:44:21 +0000 (20:44 -0300)]
Disable the exception thread under iOS since it's neither needed nor working.
Atsushi Eno [Fri, 8 Jul 2011 21:50:58 +0000 (06:50 +0900)]
Full equivalence comparison is unnecessary for MessageEncoder ContentType.
based on https://github.com/mono/mono/pull/130
Zoltan Varga [Fri, 8 Jul 2011 12:37:44 +0000 (14:37 +0200)]
Resurrect the AOT profiler.
Marek Safar [Fri, 8 Jul 2011 07:14:17 +0000 (00:14 -0700)]
Merge pull request #131 from txdv/master
Encoding constants should be in lower case.
Andrius Bentkus [Fri, 8 Jul 2011 04:32:57 +0000 (06:32 +0200)]
Fix upper case encoding constants.
name.ToLowerInvariant ().Replace ('-', '_') should ensure that they
are all lowercase.
Jonathan Pobst [Thu, 7 Jul 2011 21:52:46 +0000 (16:52 -0500)]
'async' is becoming a keyword. Fix variables named async.
Jonathan Pobst [Thu, 7 Jul 2011 21:50:49 +0000 (16:50 -0500)]
Handle DBNull's.
Marek Safar [Thu, 7 Jul 2011 15:34:28 +0000 (16:34 +0100)]
Initial skeleton of secondary approach of await stack handling
Marek Safar [Tue, 5 Jul 2011 16:29:42 +0000 (17:29 +0100)]
Remove debug comment
Jeffrey Stedfast [Wed, 6 Jul 2011 22:21:29 +0000 (18:21 -0400)]
Reverted previous commit, struct sizes vary way too much between platforms
Jeffrey Stedfast [Wed, 6 Jul 2011 22:06:14 +0000 (18:06 -0400)]
Use a struct utsname when calling uname() instead of an 8K buffer
Jeffrey Stedfast [Wed, 6 Jul 2011 17:32:05 +0000 (13:32 -0400)]
Add System/Platform.cs to *System*.dll.sources
Rodrigo Kumpera [Wed, 6 Jul 2011 17:01:15 +0000 (10:01 -0700)]
Merge pull request #128 from paszczi/master
Fix for bug #701733
Jeffrey Stedfast [Wed, 6 Jul 2011 16:42:24 +0000 (12:42 -0400)]
Make internal Platform class static
Jeffrey Stedfast [Wed, 6 Jul 2011 16:06:12 +0000 (12:06 -0400)]
Load the correct type for MacOSX (as opposed to iOS)
Zoltan Varga [Wed, 6 Jul 2011 14:17:09 +0000 (16:17 +0200)]
Do some work towards making ARM fp mode runtime configurable, still far from working.
Zoltan Varga [Wed, 6 Jul 2011 14:16:16 +0000 (16:16 +0200)]
Prefix ARM FPA codegen macros with 'FPA'.
Zoltan Varga [Wed, 6 Jul 2011 13:37:32 +0000 (15:37 +0200)]
Avoid calling add_method () twice.
Zoltan Varga [Wed, 6 Jul 2011 02:12:59 +0000 (04:12 +0200)]
Add an ExportSymbol property to MonoPInvokeCallbackAttribute to instruct the AOT compiler to generate a global symbol for the native-to-managed wrapper.
Jeffrey Stedfast [Tue, 5 Jul 2011 21:53:54 +0000 (17:53 -0400)]
Oops, we need Platform.cs too
Jeffrey Stedfast [Tue, 5 Jul 2011 21:52:25 +0000 (17:52 -0400)]
Implemented a shared way to do MacOSX detection in System.dll