Bug 41945 - Bitcode exception-related tests failing
Summary: Bitcode exception-related tests failing
Status: CONFIRMED
Alias: None
Product: Runtime
Classification: Mono
Component: JIT (show other bugs)
Version: 4.5.X
Hardware: PC Mac OS
: Normal normal
Target Milestone: Future Cycle (TBD)
Assignee: Alexander Kyte
URL:
Depends on:
Blocks:
 
Reported: 2016-06-17 17:49 UTC by Alexander Kyte
Modified: 2016-10-13 12:22 UTC (History)
4 users (show)

See Also:
Tags: bitcode_mobile_static
Is this bug a regression?: ---
Last known good build:


Attachments

Description Alexander Kyte 2016-06-17 17:49:02 UTC
exception18.exe fails because of stacktrace formatting issues and this is with
https://github.com/mono/mono/pull/3166 applied.

A lot of the threadpool exception tests are failing, likely due to
similar reasons. 
threadpool-exceptions2.exe 
threadpool-exceptions3.exe 
threadpool-exceptions5.exe 
threadpool-exceptions7.exe 

which is timing out.

Then there is exception3.exe which is reporting failure. It appears that this is due to a finalizer not running, so fin +=1 is never ran.

Stackframes-async.2.exe is getting a stack overflow.

generic-stack-traces2.2.exe is getting a Null Reference Exception.

Lastly, we seem to be violating type safety here:





generics-sharing-other-exc.2.exe
=============== generics-sharing-other-exc.2.exe.stdout ===============

=============== EOF ===============
=============== generics-sharing-other-exc.2.exe.stderr ===============

Unhandled Exception:
System.Runtime.CompilerServices.RuntimeWrappedException: An object that does not derive from System.Exception has been wrapped in a RuntimeWrappedException.
  at Gen`1[T].except () <0x103ff1f30 + 0x0007a> in <filename unknown>:0
  at main.Main () <0x103ff2050 + 0x000c6> in <filename unknown>:0
[ERROR] FATAL UNHANDLED EXCEPTION: Nested exception trying to figure out what went wrong

=============== EOF ===============

generic-array-exc.2.exe
=============== generic-array-exc.2.exe.stdout ===============

=============== EOF ===============
=============== generic-array-exc.2.exe.stderr ===============

Unhandled Exception:
System.Runtime.CompilerServices.RuntimeWrappedException: An object that does not derive from System.Exception has been wrapped in a RuntimeWrappedException.
  at Gen`1[T].except () <0x108fe4d90 + 0x0007a> in <filename unknown>:0
  at main.Main () <0x108fe4f00 + 0x000c6> in <filename unknown>:0
[ERROR] FATAL UNHANDLED EXCEPTION: Nested exception trying to figure out what went wrong

=============== EOF ===============
Comment 1 Alexander Kyte 2016-06-17 18:02:24 UTC
Delegate3.cs appears to be failing because the exception thrown by the delegate is allowed to reach the top-level.
Comment 2 Alexander Kyte 2016-06-17 18:03:42 UTC
Delegate2 also appears to rely on exception behavior for control flow. It is failing.
Comment 3 Alexander Kyte 2016-06-20 21:01:34 UTC
It looks like LLVM's inliner is dramatically shortening some of the expected stack traces for some of these tests. generic-stack-traces.exe is an example of a test that cannot pass while we allow LLVM to inline.
Comment 4 Ingo Krabbe 2016-10-13 12:22:25 UTC
All those tests work for me on centos6 with gcc-4.7, but the delegate2 test, that fails with an exception.

Compiling delegate2.cs with Visual Studio throws the same exception with a bit different runtime beaviour, naturally, because thread scheduling will be quite different on a windows system. But the test would fail too, on windows, so I assume, this test should just be removed, shouldn't it?

Note You need to log in before you can comment on or make changes to this bug.