Branch: refs/heads/master
Home:   https://github.com/D-Programming-Language/druntime

Commit: e777ed1a5edd8312a661867080e9082e7af5b0e3
    https://github.com/D-Programming-Language/druntime/commit/e777ed1a5edd8312a661867080e9082e7af5b0e3
Author: Sean Kelly <sean at invisibleduck.org>
Date:   2011-03-23 (Wed, 23 Mar 2011)

Changed paths:
  M posix.mak
  M src/core/runtime.d
  A src/core/sys/windows/dbghelp.d
  A src/core/sys/windows/stacktrace.d
  M src/core/sys/windows/windows.d
  M win32.mak

Log Message:
-----------
Issue 1001 - print stack trace (in debug mode) when programs die

Okay this is weird.  The auto tester is reporting a failure for this commit with the following error (in Linux):

src/core/sys/windows/stacktrace.d(268): Error: function core.sys.windows.stacktrace.StackTrace.trace synchronized function trace must be a member of a class
make: *** [import/core/sys/windows/stacktrace.di] Error 1

The strange thing is that this code compiles and runs correctly on Windows.  Why the difference?

On Mar 23, 2011, at 12:25 PM, noreply at github.com wrote:

> Branch: refs/heads/master
> Home:   https://github.com/D-Programming-Language/druntime
> 
> Commit: e777ed1a5edd8312a661867080e9082e7af5b0e3
>    https://github.com/D-Programming-Language/druntime/commit/e777ed1a5edd8312a661867080e9082e7af5b0e3
> Author: Sean Kelly <sean at invisibleduck.org>
> Date:   2011-03-23 (Wed, 23 Mar 2011)
> 
> Changed paths:
>  M posix.mak
>  M src/core/runtime.d
>  A src/core/sys/windows/dbghelp.d
>  A src/core/sys/windows/stacktrace.d
>  M src/core/sys/windows/windows.d
>  M win32.mak
> 
> Log Message:
> -----------
> Issue 1001 - print stack trace (in debug mode) when programs die
> 
> 
> _______________________________________________
> D-runtime mailing list
> D-runtime at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/d-runtime

Correction, it compiles and runs correctly on my Windows machine, not the auto-tester Windows machine.  I'll change the code, but it's a bit weird that it worked for me.

On Mar 23, 2011, at 1:53 PM, Sean Kelly wrote:

> Okay this is weird.  The auto tester is reporting a failure for this commit with the following error (in Linux):
> 
> src/core/sys/windows/stacktrace.d(268): Error: function core.sys.windows.stacktrace.StackTrace.trace synchronized function trace must be a member of a class
> make: *** [import/core/sys/windows/stacktrace.di] Error 1
> 
> The strange thing is that this code compiles and runs correctly on Windows.  Why the difference?
> 
> On Mar 23, 2011, at 12:25 PM, noreply at github.com wrote:
> 
>> Branch: refs/heads/master
>> Home:   https://github.com/D-Programming-Language/druntime
>> 
>> Commit: e777ed1a5edd8312a661867080e9082e7af5b0e3
>>   https://github.com/D-Programming-Language/druntime/commit/e777ed1a5edd8312a661867080e9082e7af5b0e3
>> Author: Sean Kelly <sean at invisibleduck.org>
>> Date:   2011-03-23 (Wed, 23 Mar 2011)
>> 
>> Changed paths:
>> M posix.mak
>> M src/core/runtime.d
>> A src/core/sys/windows/dbghelp.d
>> A src/core/sys/windows/stacktrace.d
>> M src/core/sys/windows/windows.d
>> M win32.mak
>> 
>> Log Message:
>> -----------
>> Issue 1001 - print stack trace (in debug mode) when programs die
>> 
>> 
>> _______________________________________________
>> D-runtime mailing list
>> D-runtime at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/d-runtime
> 
> _______________________________________________
> D-runtime mailing list
> D-runtime at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/d-runtime

Is it a problem with the .di file? I've seen very weird stuff when they get out of sync.

On 23 March 2011 22:00, Sean Kelly <sean at invisibleduck.org> wrote:
> Correction, it compiles and runs correctly on my Windows machine, not the auto-tester Windows machine. ?I'll change the code, but it's a bit weird that it worked for me.
>
> On Mar 23, 2011, at 1:53 PM, Sean Kelly wrote:
>
>> Okay this is weird. ?The auto tester is reporting a failure for this commit with the following error (in Linux):
>>
>> src/core/sys/windows/stacktrace.d(268): Error: function core.sys.windows.stacktrace.StackTrace.trace synchronized function trace must be a member of a class
>> make: *** [import/core/sys/windows/stacktrace.di] Error 1
>>
>> The strange thing is that this code compiles and runs correctly on Windows. ?Why the difference?
>>
>> On Mar 23, 2011, at 12:25 PM, noreply at github.com wrote:
>>
>>> Branch: refs/heads/master
>>> Home: ? https://github.com/D-Programming-Language/druntime
>>>
>>> Commit: e777ed1a5edd8312a661867080e9082e7af5b0e3
>>> ? https://github.com/D-Programming-Language/druntime/commit/e777ed1a5edd8312a661867080e9082e7af5b0e3
>>> Author: Sean Kelly <sean at invisibleduck.org>
>>> Date: ? 2011-03-23 (Wed, 23 Mar 2011)
>>>
>>> Changed paths:
>>> M posix.mak
>>> M src/core/runtime.d
>>> A src/core/sys/windows/dbghelp.d
>>> A src/core/sys/windows/stacktrace.d
>>> M src/core/sys/windows/windows.d
>>> M win32.mak
>>>
>>> Log Message:
>>> -----------
>>> Issue 1001 - print stack trace (in debug mode) when programs die
>>>
>>>
>>> _______________________________________________
>>> D-runtime mailing list
>>> D-runtime at puremagic.com
>>> http://lists.puremagic.com/mailman/listinfo/d-runtime
>>
>> _______________________________________________
>> D-runtime mailing list
>> D-runtime at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/d-runtime
>
> _______________________________________________
> D-runtime mailing list
> D-runtime at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/d-runtime
>

The basic issue is that I had a static synchronized class function.  For whatever reason, my version of DMD on Windows didn't complain, but everything else did (including the Windows test machine).  I can't explain why it was okay when I built it.  Perhaps I forgot to test that change before checking in my code.  I could have sworn that I didn't change anything between testing and committing.  Oh well.

I had thought that a static synchronized function would simply synchronize on the classinfo instance, but I can see why that may not be desirable as an in-built feature.  I'm doing this manually now instead.

On Mar 23, 2011, at 2:05 PM, Don Clugston wrote:

> Is it a problem with the .di file? I've seen very weird stuff when they get out of sync.
> 
> On 23 March 2011 22:00, Sean Kelly <sean at invisibleduck.org> wrote:
>> Correction, it compiles and runs correctly on my Windows machine, not the auto-tester Windows machine.  I'll change the code, but it's a bit weird that it worked for me.
>> 
>> On Mar 23, 2011, at 1:53 PM, Sean Kelly wrote:
>> 
>>> Okay this is weird.  The auto tester is reporting a failure for this commit with the following error (in Linux):
>>> 
>>> src/core/sys/windows/stacktrace.d(268): Error: function core.sys.windows.stacktrace.StackTrace.trace synchronized function trace must be a member of a class
>>> make: *** [import/core/sys/windows/stacktrace.di] Error 1
>>> 
>>> The strange thing is that this code compiles and runs correctly on Windows.  Why the difference?
>>> 
>>> On Mar 23, 2011, at 12:25 PM, noreply at github.com wrote:
>>> 
>>>> Branch: refs/heads/master
>>>> Home:   https://github.com/D-Programming-Language/druntime
>>>> 
>>>> Commit: e777ed1a5edd8312a661867080e9082e7af5b0e3
>>>>   https://github.com/D-Programming-Language/druntime/commit/e777ed1a5edd8312a661867080e9082e7af5b0e3
>>>> Author: Sean Kelly <sean at invisibleduck.org>
>>>> Date:   2011-03-23 (Wed, 23 Mar 2011)
>>>> 
>>>> Changed paths:
>>>> M posix.mak
>>>> M src/core/runtime.d
>>>> A src/core/sys/windows/dbghelp.d
>>>> A src/core/sys/windows/stacktrace.d
>>>> M src/core/sys/windows/windows.d
>>>> M win32.mak
>>>> 
>>>> Log Message:
>>>> -----------
>>>> Issue 1001 - print stack trace (in debug mode) when programs die
>>>> 
>>>> 
>>>> _______________________________________________
>>>> D-runtime mailing list
>>>> D-runtime at puremagic.com
>>>> http://lists.puremagic.com/mailman/listinfo/d-runtime
>>> 
>>> _______________________________________________
>>> D-runtime mailing list
>>> D-runtime at puremagic.com
>>> http://lists.puremagic.com/mailman/listinfo/d-runtime
>> 
>> _______________________________________________
>> D-runtime mailing list
>> D-runtime at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/d-runtime
>> 
> _______________________________________________
> D-runtime mailing list
> D-runtime at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/d-runtime