mono.git
7 years agoMerge pull request #2377 from joelmartinez/docs-multiassembly-extension-fix
Jonathan Pryor [Tue, 31 May 2016 16:01:02 +0000 (12:01 -0400)]
Merge pull request #2377 from joelmartinez/docs-multiassembly-extension-fix

[mdoc] Extension method crash fix, removes -multiassembly

7 years ago[mdoc] Extension method crash fix, removes -multiassembly
Joel Martinez [Wed, 16 Dec 2015 16:34:05 +0000 (11:34 -0500)]
[mdoc] Extension method crash fix, removes -multiassembly
The tests added during the initial development of the -multiassembly argument failed to test the presence of extension methods;
the result was a crash that occurred when there were extension methods present in multiple assemblies.

This patch addresses the issue by putting a guard clause around the code that adds the extension method to the index.
Tests have been modified accordingly to verify this bug and include extension methods

Additionally, the -multiassembly parameter was removed entirely. This behavior is now triggered by the presence of -apistyle,
which is where support for "multiple assemblies" was meant to work anyways. To review, this is to support cross-platform
development, where you have multiple -- *different* -- assemblies for each platform. These assemblies will share a certain
cross-section of APIs that will be *exactly* the same in each assembly (as opposed to multiple versions of literally the same assembly).
As a result, every member will now contain an individual `AssemblyInfo` tag that also contains the name of the assembly.

7 years agoMerge pull request #3078 from akoeplinger/fix-test-exclude-filename
Alexander Köplinger [Tue, 31 May 2016 15:24:11 +0000 (17:24 +0200)]
Merge pull request #3078 from akoeplinger/fix-test-exclude-filename

[build] Make exclusion filename consistent between library and test

7 years agoMerge pull request #3067 from lambdageek/dev/monoerror-stringy
Aleksey Kliger (λgeek) [Tue, 31 May 2016 15:13:53 +0000 (11:13 -0400)]
Merge pull request #3067 from lambdageek/dev/monoerror-stringy

[runtime] Deal with MonoErrors from string_to_utf8_checked changes

7 years agoMerge pull request #3059 from lateralusX/jlorenss/win-x64-dyncall-gsharedvt-support
lateralusX [Tue, 31 May 2016 14:32:40 +0000 (16:32 +0200)]
Merge pull request #3059 from lateralusX/jlorenss/win-x64-dyncall-gsharedvt-support

[aot] Added support for GSHAREDVT and DYNCALL on Windows x64 for full AOT builds.

7 years agoMerge pull request #3076 from akoeplinger/fix-ziptime-offbyone
Alexander Köplinger [Tue, 31 May 2016 14:06:41 +0000 (16:06 +0200)]
Merge pull request #3076 from akoeplinger/fix-ziptime-offbyone

[WindowsBase] Fix off-by-one error in ZipTime

7 years ago[WindowsBase] Fix off-by-one error in ZipTime
Alexander Köplinger [Tue, 31 May 2016 09:59:49 +0000 (11:59 +0200)]
[WindowsBase] Fix off-by-one error in ZipTime

The date/time struct in zip.h defines the month value as

    uInt tm_mon;            /* months since January - [0,11] */

.NET's DateTime object in turn stores months from 1-12. We didn't
take this into account when passing the values from managed->native
so the month ended up being off by one.

We didn't use this codepath before https://github.com/mono/mono/pull/3036
that's why it didn't show up until now.

7 years ago[build] Make exclusion filename consistent between library and test
Alexander Köplinger [Tue, 31 May 2016 11:10:28 +0000 (13:10 +0200)]
[build] Make exclusion filename consistent between library and test

The test assembly used *.dll.excludes before while the library uses *.dll.exclude.sources.
Make this consistent to reduce confusion.

Also remove corlib_cmp.dll.excludes which is from a profile that no longer exists.

7 years ago[System] Fixes build
Marek Safar [Tue, 31 May 2016 11:44:22 +0000 (13:44 +0200)]
[System] Fixes build

7 years agoMerge pull request #3075 from knocte/man_bump
Marek Safar [Tue, 31 May 2016 09:44:36 +0000 (11:44 +0200)]
Merge pull request #3075 from knocte/man_bump

[man] bump man's mono version number too (to 4.5.2)

7 years ago[man] bump man's mono version number too (to 4.5.2)
Andres G. Aragoneses [Tue, 31 May 2016 08:51:23 +0000 (16:51 +0800)]
[man] bump man's mono version number too (to 4.5.2)

Following this version bump:
https://github.com/mono/mono/commit/069cea3712a5bd4959e505260cdc4a0c37861e77

(Also add a note to configure.ac as a hint until
someone comes up with a way to sync man/mono.1's
version automatically.)

7 years ago[System] Keep broken api in mobile profiles only
Marek Safar [Tue, 31 May 2016 08:06:34 +0000 (10:06 +0200)]
[System] Keep broken api in mobile profiles only

7 years ago[WindowsBase] Fixes test assert expectation logic
Marek Safar [Tue, 31 May 2016 08:05:13 +0000 (10:05 +0200)]
[WindowsBase] Fixes test assert expectation logic

7 years ago[System] Remove empty test
Marek Safar [Mon, 30 May 2016 22:42:30 +0000 (00:42 +0200)]
[System] Remove empty test

7 years agoMerge pull request #3063 from marek-safar/rs-Net
Marek Safar [Tue, 31 May 2016 08:00:21 +0000 (10:00 +0200)]
Merge pull request #3063 from marek-safar/rs-Net

[System] WebRequest from referencesource

7 years ago[System] WebProxy from referencesource
Marek Safar [Fri, 27 May 2016 15:57:45 +0000 (17:57 +0200)]
[System] WebProxy from referencesource

7 years ago[ci] Bump timeout in runtime test step
Alexander Köplinger [Tue, 31 May 2016 05:04:27 +0000 (07:04 +0200)]
[ci] Bump timeout in runtime test step

We were seeing timeouts in the runtime step if two builds happen to be scheduled
at the exact same time on the ARM workers on Jenkins, causing increased load and
longer test times. They aborted the runtime step right before it'd finish.

Bump the timeout a bit to avoid this.

7 years ago[bcl] Comment out a win32 pinvoke on the mobile profile, since ios can't have unresol...
Zoltan Varga [Tue, 31 May 2016 03:14:32 +0000 (23:14 -0400)]
[bcl] Comment out a win32 pinvoke on the mobile profile, since ios can't have unresolvable pinvokes.

7 years agoBump version to 4.5.2.
Alex Rønne Petersen [Mon, 30 May 2016 18:14:22 +0000 (20:14 +0200)]
Bump version to 4.5.2.

7 years agoMerge pull request #3056 from BrzVlad/fix-multiple-binprot
Vlad Brezae [Mon, 30 May 2016 16:06:09 +0000 (19:06 +0300)]
Merge pull request #3056 from BrzVlad/fix-multiple-binprot

[sgen] Exclusive write on binary protocol file

7 years agoMerge pull request #3057 from BrzVlad/fix-major-log3
Vlad Brezae [Mon, 30 May 2016 16:04:28 +0000 (19:04 +0300)]
Merge pull request #3057 from BrzVlad/fix-major-log3

[sgen] Refactor collection logging

7 years agoMerge pull request #3071 from adamburgess/fix-ci-links
Alexander Köplinger [Mon, 30 May 2016 10:36:55 +0000 (12:36 +0200)]
Merge pull request #3071 from adamburgess/fix-ci-links

Update README with new CI builds/badges

7 years ago[jit] Emit code to throw method access exceptions instead of throwing them from the...
Zoltan Varga [Mon, 30 May 2016 07:40:56 +0000 (03:40 -0400)]
[jit] Emit code to throw method access exceptions instead of throwing them from the JIT, so the stack trace is correct. Fixes #41280. (#3072)

7 years agoRevert "[jit] Emit code to throw method access exceptions instead of throwing them...
Zoltan Varga [Mon, 30 May 2016 02:25:36 +0000 (22:25 -0400)]
Revert "[jit] Emit code to throw method access exceptions instead of throwing them from the JIT, so the stack trace is correct. Fixes #41280."

This reverts commit 04171d2c267d2db54b36877809dca7a890adcfc5.
Revert this as it causes test failures.

7 years agoRevert "[jit] Fix 04171d2c267d2db54b36877809dca7a890adcfc5 so it actually works."
Zoltan Varga [Mon, 30 May 2016 02:25:12 +0000 (22:25 -0400)]
Revert "[jit] Fix 04171d2c267d2db54b36877809dca7a890adcfc5 so it actually works."

This reverts commit 6683c128e4510187b0a6b8a6585095dcd25984c2.
Revert this as it causes test failures.

7 years ago[System.Transactions] Fix the build broken by https://github.com/mono/mono/pull/3061
Alexander Köplinger [Mon, 30 May 2016 00:11:33 +0000 (02:11 +0200)]
[System.Transactions] Fix the build broken by https://github.com/mono/mono/pull/3061

7 years agoMerge pull request #3061 from Thorium/master
Marek Safar [Sun, 29 May 2016 06:22:45 +0000 (08:22 +0200)]
Merge pull request #3061 from Thorium/master

#23050 TransactionScopeAsyncFlowOption

7 years ago[jit] Fix 04171d2c267d2db54b36877809dca7a890adcfc5 so it actually works.
Zoltan Varga [Sun, 29 May 2016 04:24:08 +0000 (00:24 -0400)]
[jit] Fix 04171d2c267d2db54b36877809dca7a890adcfc5 so it actually works.

7 years ago[jit] Emit code to throw method access exceptions instead of throwing them from the...
Zoltan Varga [Sun, 29 May 2016 04:21:58 +0000 (00:21 -0400)]
[jit] Emit code to throw method access exceptions instead of throwing them from the JIT, so the stack trace is correct. Fixes #41280.

7 years agoUpdate README with new CI builds/badges
Adam Burgess [Sat, 28 May 2016 10:03:20 +0000 (20:03 +1000)]
Update README with new CI builds/badges

Adds the debian arm64 build and also changes the amd64/i386 builds to be ubuntu-specific

7 years ago[gsharedvt] Fix the mapping of reference generic instances in gsharedvt wrappers...
Zoltan Varga [Sat, 28 May 2016 05:52:01 +0000 (01:52 -0400)]
[gsharedvt] Fix the mapping of reference generic instances in gsharedvt wrappers, they need to be the same as references.

7 years agoMerge pull request #3002 from lewurm/typespec-caching
Bernhard Urban [Fri, 27 May 2016 23:50:06 +0000 (16:50 -0700)]
Merge pull request #3002 from lewurm/typespec-caching

[aot] use hashtable for caching typespec of classes

7 years ago[sgen] Exclusive write on binary protocol file
Vlad Brezae [Mon, 23 May 2016 18:01:24 +0000 (21:01 +0300)]
[sgen] Exclusive write on binary protocol file

If we have binary protocol enabled and the debugged executable will spawn a new mono instance, both of the processes will write to the same debug file, because they share the environment variable, producing invalid results. Add a pid sufix if output file exists in order to separately obtain binary protocol outputs from all mono instances.

7 years ago[s390x] Fix the build2
Aleksey Kliger [Fri, 27 May 2016 20:19:29 +0000 (16:19 -0400)]
[s390x] Fix the build2

7 years ago[s390x] Fix the build
Aleksey Kliger [Fri, 27 May 2016 20:15:42 +0000 (16:15 -0400)]
[s390x] Fix the build

Broken by b884d17

7 years ago[genproj] Update from Ankit's project
Miguel de Icaza [Fri, 27 May 2016 20:13:17 +0000 (16:13 -0400)]
[genproj] Update from Ankit's project

7 years ago[runtime] Avoid mono_error_raise_exception in icalls and debugging utilities
Aleksey Kliger [Fri, 27 May 2016 19:38:15 +0000 (15:38 -0400)]
[runtime] Avoid mono_error_raise_exception in icalls and debugging utilities

7 years ago[perfcounters] Convert instance name to C string in the icall entrypoint.
Aleksey Kliger [Fri, 27 May 2016 19:37:02 +0000 (15:37 -0400)]
[perfcounters] Convert instance name to C string in the icall entrypoint.

Convert MonoString to UTF8 early to simplify exception handling.

7 years ago[reflection] Don't raise in {Module,Assembly}Builder icalls
Aleksey Kliger [Fri, 27 May 2016 18:16:24 +0000 (14:16 -0400)]
[reflection] Don't raise in {Module,Assembly}Builder icalls

7 years agoMerge pull request #3037 from mono/system_io_compression_datetime
Alexander Köplinger [Fri, 27 May 2016 18:20:18 +0000 (20:20 +0200)]
Merge pull request #3037 from mono/system_io_compression_datetime

[System.IO.Compression/.Filesystem] Fixed handling of date/time for Zip entries

7 years ago[genproj] Run update-solution-files to track master
Miguel de Icaza [Fri, 27 May 2016 17:52:22 +0000 (13:52 -0400)]
[genproj] Run update-solution-files to track master

7 years agoAdd another build step
Miguel de Icaza [Fri, 27 May 2016 17:27:32 +0000 (13:27 -0400)]
Add another build step

7 years ago[genproj] Do not use \r on the Unix parts of the build hook, add RabbitMQ hooks
Miguel de Icaza [Fri, 27 May 2016 17:24:08 +0000 (13:24 -0400)]
[genproj] Do not use \r on the Unix parts of the build hook, add RabbitMQ hooks

7 years ago[sgen] Refactor collection logging
Vlad Brezae [Wed, 25 May 2016 00:04:40 +0000 (03:04 +0300)]
[sgen] Refactor collection logging

We had the problem that we were logging at sweep finish which can happen anytime. If we happen to suspend a thread, while doing a GC, in the middle of logging we might deadlock if we wait for sweep to finish (which can be blocked waiting for logging).

When logging multiple collections in the same stw pause, we would not log the real results for both collections since we would compute the major sizes at the very end.

We logged collections in sgen_perform_collection, outside the function calls for specific collections, which makes it cumbersome to add additional log entries.

This refactoring solves these issues by having GC code log info at any time. This log entry is enqueued in a list of pending entries. We flush the list of logs, as before, at the end of the stw, when the world is restarted.

7 years ago[file-mmap-posix] Don't use mono_error_raise_exception
Aleksey Kliger [Fri, 27 May 2016 14:38:04 +0000 (10:38 -0400)]
[file-mmap-posix] Don't use mono_error_raise_exception

7 years agoMerge pull request #3060 from directhex/pr-in-version.h
Alexander Köplinger [Fri, 27 May 2016 15:47:54 +0000 (17:47 +0200)]
Merge pull request #3060 from directhex/pr-in-version.h

[jenkins] Include PR number in version.h if PR builder env is set

7 years ago[runtime] Don't raise in mono_assembly_apply_binding
Aleksey Kliger [Thu, 26 May 2016 22:05:07 +0000 (18:05 -0400)]
[runtime] Don't raise in mono_assembly_apply_binding

7 years agoMerge pull request #3050 from lambdageek/dev/monoerror-to_utf8
Aleksey Kliger (λgeek) [Fri, 27 May 2016 14:54:45 +0000 (10:54 -0400)]
Merge pull request #3050 from lambdageek/dev/monoerror-to_utf8

[runtime] Mark mono_string_to_utf8 external only.

7 years agoMerge pull request #3058 from rolfbjarne/git-instead-of-ssh
Alexander Köplinger [Fri, 27 May 2016 14:16:09 +0000 (16:16 +0200)]
Merge pull request #3058 from rolfbjarne/git-instead-of-ssh

[offsets-tool] Use git protocol instead of ssh to allow anonymous access.

7 years agoMerge pull request #3062 from madewokherd/win32threadscompile
Alexander Köplinger [Fri, 27 May 2016 14:06:57 +0000 (16:06 +0200)]
Merge pull request #3062 from madewokherd/win32threadscompile

[win32_threads] Mark an inline function as static.

7 years ago[corlib] Clean up MonoType declaration
Marek Safar [Fri, 27 May 2016 14:00:05 +0000 (16:00 +0200)]
[corlib] Clean up MonoType declaration

7 years ago[System] WebRequest from referencesource
Marek Safar [Wed, 25 May 2016 10:38:50 +0000 (12:38 +0200)]
[System] WebRequest from referencesource

7 years ago[System.IO.Compression.FileSystem] Fixed date time when creating Zip entries from...
Joao Matos [Fri, 27 May 2016 12:03:57 +0000 (13:03 +0100)]
[System.IO.Compression.FileSystem] Fixed date time when creating Zip entries from filesystem files.

7 years ago[System.IO.Compression] Added test for Zip entry timestamp modification.
Joao Matos [Fri, 27 May 2016 12:03:02 +0000 (13:03 +0100)]
[System.IO.Compression] Added test for Zip entry timestamp modification.

7 years agoMade changes suggested by marek.
Tuomas Hietanen [Fri, 27 May 2016 09:40:09 +0000 (10:40 +0100)]
Made changes suggested by marek.

7 years ago[jenkins] Include PR number in version.h if PR builder env is set
Jo Shields [Thu, 26 May 2016 10:49:59 +0000 (11:49 +0100)]
[jenkins] Include PR number in version.h if PR builder env is set

A while ago, we switched our PR builders from using the Github-provided PR branch head (i.e. `pull/merge/XXXX`) to the exact sha1 of the latest commit on the PR (i.e. `abcdefabc`). This change was needed for the performance team, otherwise the changes they wanted to test would be rebased against a new master several times per day, breaking their ability to meaningfully track the changes in a PR.

As a side effect of this, `version.h` went from setting the version from `XXXX/abcdefabc` to `explicit/abcdefabc`, and the Jenkins PR packages use the branch name `explicit` instead of the actual PR number `XXXX`, which is pretty bad from a UX perspective.

This change checks for the `ghprbPullId` environment variable and uses it instead of the actual branch name or `explicit` - it is always set on PR builds.

7 years agoAdded support for GSHAREDVT and DYNCALL on Windows x64 for full AOT builds.
lateralusX [Fri, 27 May 2016 08:22:32 +0000 (10:22 +0200)]
Added support for GSHAREDVT and DYNCALL on Windows x64 for full AOT builds.

Support for GSHAREDVT added for Windows x64. Changes needed
(ones gsharedvt concept was understood, crashes debugged and adjustments located)
where quite small. Primarily around shadow stack used by windows + incorrect handling
of none volatile registers on windows when going over the gsharedvt trampolines.
All GSHAREDVT tests, gshared.exe and aot.tests.exe pass in full AOT builds with above adjustments.

Support for DYNCALL added/enabled for Windows x64. Current AMD64 support was working
for basic scenarios but not enabled on Windows. Identified a problem on AMD64
when structs are passed as input parameters or returned by methods for DYNCALL scenarios
in full AOT:ed builds. This scenario doesn't seem to work for any AMD64 platform for
full AOT and the tests in aot-test.exe handling these scenarios are not run on AMD64.

Adjusted categories on test in aot-tests.exe to reflect what’s really needed by the tests
to run correctly. Also added a new category to some DYNCALL tests that uses struct support
currently not implemented on AMD64, !FULLAOT-AMD64, so at least some of the DYNCALL tests
can be run on AMD64 platforms in full AOT:ed builds.

7 years agoRemove an unused icall.
Zoltan Varga [Fri, 27 May 2016 02:50:23 +0000 (22:50 -0400)]
Remove an unused icall.

7 years agoMerge pull request #3026 from lewurm/is-call-macro
monojenkins [Fri, 27 May 2016 01:25:07 +0000 (02:25 +0100)]
Merge pull request #3026 from lewurm/is-call-macro

[mini] small improvements around `MONO_IS_CALL` macro and `local-prop`

showed up in vtunes when running
```
./mono/mini/mono-sgen --stats -O=all --aot=threads=1,asmonly ~/work/benchmarker/tests/fsharp/lib/bootstrap/4.0/FSharp.Compiler.dll
```

also, I think when `OP_DYN_CALL` was introduced, it wasn't added to the `MONO_IS_CALL` macro. @vargaz, what do you think?

7 years ago[sdb] Avoid calling mono_thread_current () in the profiler shutdown callback, sends...
Zoltan Varga [Thu, 26 May 2016 22:55:57 +0000 (18:55 -0400)]
[sdb] Avoid calling mono_thread_current () in the profiler shutdown callback, sends a 0 thread id instead, the managed side ignores it.

7 years ago[mini] avoid checking condition twice in local-prop
Bernhard Urban [Fri, 20 May 2016 18:52:47 +0000 (11:52 -0700)]
[mini] avoid checking condition twice in local-prop

7 years ago[mini] simplify MONO_IS_CALL macro
Bernhard Urban [Fri, 20 May 2016 18:52:44 +0000 (11:52 -0700)]
[mini] simplify MONO_IS_CALL macro

7 years agoMerge System/MonoType.cs into ReferenceSources/RuntimeType.cs, use th… (#3045)
Zoltan Varga [Thu, 26 May 2016 22:33:41 +0000 (18:33 -0400)]
Merge System/MonoType.cs into ReferenceSources/RuntimeType.cs, use th… (#3045)

* Merge System/MonoType.cs into ReferenceSources/RuntimeType.cs, use the latter in the runtime code.

* [sdb] Fix a test which referenced System.MonoType.

7 years ago[win32_threads] Mark an inline function as static.
Vincent Povirk [Thu, 26 May 2016 21:50:42 +0000 (16:50 -0500)]
[win32_threads] Mark an inline function as static.

If the C compiler decides not to inline the function, and it's not declared
static, it'll look for an extern version, and linking will fail.

7 years ago[runtime] Mark mono_string_to_utf8 external only.
Aleksey Kliger [Wed, 25 May 2016 18:01:22 +0000 (14:01 -0400)]
[runtime] Mark mono_string_to_utf8 external only.

Runtime should use mono_string_to_utf8_checked.

7 years agoMerge pull request #3055 from lambdageek/dev/monoerror-cleanup-cleanup
Aleksey Kliger (λgeek) [Thu, 26 May 2016 15:26:33 +0000 (11:26 -0400)]
Merge pull request #3055 from lambdageek/dev/monoerror-cleanup-cleanup

[runtime] Disallow unintended reuse of MonoError after mono_error_cleanup

7 years agoMerge branch 'master' of https://github.com/mono/mono
Tuomas Hietanen [Thu, 26 May 2016 14:07:30 +0000 (15:07 +0100)]
Merge branch 'master' of https://github.com/mono/mono

7 years ago#23050 First attempt to create TransactionScopeAsyncFlowOption on Mono.
Tuomas Hietanen [Thu, 26 May 2016 14:04:25 +0000 (15:04 +0100)]
#23050 First attempt to create TransactionScopeAsyncFlowOption on Mono.

This is a feature that is missing. However I don't have Mono,
so has to be compiled and tested.

7 years agoMerge pull request #3028 from lateralusX/jlorenss/threadpool_warning
Ludovic Henry [Thu, 26 May 2016 13:26:51 +0000 (15:26 +0200)]
Merge pull request #3028 from lateralusX/jlorenss/threadpool_warning

Incorrect warning message on thread pool startup in full AOT:ed Windows build.

7 years agoMerge pull request #3047 from marek-safar/rs-corlib
Marek Safar [Thu, 26 May 2016 09:15:27 +0000 (11:15 +0200)]
Merge pull request #3047 from marek-safar/rs-corlib

[corlib] More System.Runtime.InteropServices from referencesource

7 years ago[offsets-tool] Use git protocol instead of ssh to allow anonymous access.
Rolf Bjarne Kvinge [Thu, 26 May 2016 07:54:40 +0000 (09:54 +0200)]
[offsets-tool] Use git protocol instead of ssh to allow anonymous access.

7 years agoAot runtime noraise3 (#3054)
Zoltan Varga [Thu, 26 May 2016 03:58:36 +0000 (23:58 -0400)]
Aot runtime noraise3 (#3054)

* [aot] Make mono_aot_get_method () external only, use mono_aot_get_method_checked () instead.

* [aot] Avoid raising exceptions in the AOT runtime.

7 years ago[reflection] Reinitialize MonoError after mono_error_convert_to_exception
Aleksey Kliger [Wed, 25 May 2016 23:27:39 +0000 (19:27 -0400)]
[reflection] Reinitialize MonoError after mono_error_convert_to_exception

7 years ago[mini] Use an inner MonoError in instance testing in first pass EH
Aleksey Kliger [Wed, 25 May 2016 23:23:16 +0000 (19:23 -0400)]
[mini] Use an inner MonoError in instance testing in first pass EH

7 years ago[runtime] Disallow unintentional reuse of MonoError after mono_error_cleanup
Aleksey Kliger [Wed, 25 May 2016 22:25:37 +0000 (18:25 -0400)]
[runtime] Disallow unintentional reuse of MonoError after mono_error_cleanup

7 years ago[reflection] Cleanup uses of mono_error_cleanup
Aleksey Kliger [Wed, 25 May 2016 22:09:16 +0000 (18:09 -0400)]
[reflection] Cleanup uses of mono_error_cleanup

MonoError must be explicitly reinitialized if it is used again after
mono_error_cleanup.  But it is better to introduce an explicit inner
MonoError when possible.

7 years ago[llvmonly] Avoid calling mono_error_raise_exception () to throw exceptions during...
Zoltan Varga [Wed, 25 May 2016 20:59:57 +0000 (16:59 -0400)]
[llvmonly] Avoid calling mono_error_raise_exception () to throw exceptions during method initialization, call mono_llvm_throw_exception () directly.

7 years agoMerge pull request #3051 from mono/revert-3043-aot-runtime-noraise
Alexander Köplinger [Wed, 25 May 2016 19:50:30 +0000 (21:50 +0200)]
Merge pull request #3051 from mono/revert-3043-aot-runtime-noraise

Revert "[aot] Avoid raising exceptions in the AOT runtime."

7 years agoRevert "[aot] Avoid raising exceptions in the AOT runtime."
Alexander Köplinger [Wed, 25 May 2016 19:48:39 +0000 (21:48 +0200)]
Revert "[aot] Avoid raising exceptions in the AOT runtime."

7 years agoMerge pull request #2996 from UCIS/patch-7
Sebastien Pouliot [Wed, 25 May 2016 19:11:22 +0000 (15:11 -0400)]
Merge pull request #2996 from UCIS/patch-7

Generate RSA keys using public exponent 65537 (was 17)

7 years agoMerge pull request #3036 from tritao/windowsbase_zip_datetime
Alexander Köplinger [Wed, 25 May 2016 17:28:55 +0000 (19:28 +0200)]
Merge pull request #3036 from tritao/windowsbase_zip_datetime

[WindowsBase] Fixed System.IO.Packaging creating Zip archive with wrong date time entries

7 years agoMerge pull request #3016 from lewurm/small-arm-cleanup
Bernhard Urban [Wed, 25 May 2016 16:56:39 +0000 (09:56 -0700)]
Merge pull request #3016 from lewurm/small-arm-cleanup

[mini-arm] remove useless case for iphone_abi

7 years ago[aot] use hashtable for caching typespec of classes
Bernhard Urban [Tue, 17 May 2016 17:43:25 +0000 (10:43 -0700)]
[aot] use hashtable for caching typespec of classes

I had a look in vtunes for

```
$ mono/mini/mono-sgen --stats -O=all --aot=threads=1,asmonly ~/work/benchmarker/tests/fsharp/lib/bootstrap/4.0/FSharp.Compiler.dll
```

and that method showed up as hot.

With this change I get an improvement from ~27.4s down to ~24.3s on my
machine.

7 years ago[aot] Avoid raising exceptions in the AOT runtime. (#3043)
Zoltan Varga [Wed, 25 May 2016 16:35:39 +0000 (12:35 -0400)]
[aot] Avoid raising exceptions in the AOT runtime. (#3043)

7 years ago[WindowsBase] Fix for the new .rels file when it's shorter than the old. A fix for...
Marek Safar [Wed, 25 May 2016 15:01:34 +0000 (17:01 +0200)]
[WindowsBase] Fix for the new .rels file when it's shorter than the old. A fix for #41203

7 years agoMerge pull request #3038 from lambdageek/dev/monoerror-threads
Aleksey Kliger (λgeek) [Wed, 25 May 2016 14:36:20 +0000 (10:36 -0400)]
Merge pull request #3038 from lambdageek/dev/monoerror-threads

[threadpool-ms] Assert if queuing work items fails in the IO selector

7 years agoMerge pull request #3042 from lambdageek/dev/monoerror-upgrade_remote_class
Aleksey Kliger (λgeek) [Wed, 25 May 2016 14:12:25 +0000 (10:12 -0400)]
Merge pull request #3042 from lambdageek/dev/monoerror-upgrade_remote_class

[remoting] Pass MonoError to mono_upgrade_remote_class

7 years agoMerge pull request #3044 from lambdageek/dev/monoerror-llvm
Aleksey Kliger (λgeek) [Wed, 25 May 2016 13:56:53 +0000 (09:56 -0400)]
Merge pull request #3044 from lambdageek/dev/monoerror-llvm

[llvm] Convert MonoError to LLVM ctx failure in process_call

7 years agoMerge pull request #3046 from redbrain/cfnetworking-null-reference
Marek Safar [Wed, 25 May 2016 13:24:41 +0000 (15:24 +0200)]
Merge pull request #3046 from redbrain/cfnetworking-null-reference

ContentStream can be null

7 years agoContentStream can be null if no bytes were available and is closed will cause null...
Philip Herron [Wed, 25 May 2016 13:18:59 +0000 (14:18 +0100)]
ContentStream can be null if no bytes were available and is closed will cause null-exception

7 years ago[corlib] More System.Runtime.InteropServices from referencesource
Marek Safar [Wed, 25 May 2016 12:59:06 +0000 (14:59 +0200)]
[corlib] More System.Runtime.InteropServices from referencesource

7 years ago[WindowsBase] Deal with deleted package parts. A patch from #41203
Marek Safar [Wed, 25 May 2016 09:01:30 +0000 (11:01 +0200)]
[WindowsBase] Deal with deleted package parts. A patch from #41203

7 years ago[genproj] Handle resources properly, add helper scripts to bootstrap
Miguel de Icaza [Tue, 24 May 2016 21:43:21 +0000 (17:43 -0400)]
[genproj] Handle resources properly, add helper scripts to bootstrap

7 years ago[amd64] Implement the general rgctx fetch trampoline.
Zoltan Varga [Tue, 24 May 2016 20:37:58 +0000 (16:37 -0400)]
[amd64] Implement the general rgctx fetch trampoline.

7 years ago[llvm] Convert MonoError to LLVM ctx failure in process_call
Aleksey Kliger [Tue, 24 May 2016 20:19:27 +0000 (16:19 -0400)]
[llvm] Convert MonoError to LLVM ctx failure in process_call

7 years ago[llvmonly] Assert gsharedvt wrapper compile always succeeds.
Aleksey Kliger [Tue, 24 May 2016 19:20:30 +0000 (15:20 -0400)]
[llvmonly] Assert gsharedvt wrapper compile always succeeds.

7 years ago[jit] Optimize ldnull+unbox.any pairs, mcs generates it for code like f = null where...
Zoltan Varga [Tue, 24 May 2016 20:08:11 +0000 (16:08 -0400)]
[jit] Optimize ldnull+unbox.any pairs, mcs generates it for code like f = null where f is a field of a generic type with an open type.

7 years ago[amd64] Add back an assert removed by the previous change.
Zoltan Varga [Tue, 24 May 2016 19:35:25 +0000 (15:35 -0400)]
[amd64] Add back an assert removed by the previous change.

7 years ago[amd64] Handle genericinst return types in gsharedvt.
Zoltan Varga [Tue, 24 May 2016 19:30:29 +0000 (15:30 -0400)]
[amd64] Handle genericinst return types in gsharedvt.

7 years ago[remoting] Pass MonoError to mono_upgrade_remote_class
Aleksey Kliger [Tue, 24 May 2016 17:36:39 +0000 (13:36 -0400)]
[remoting] Pass MonoError to mono_upgrade_remote_class