This is the mail archive of the
cygwin-developers
mailing list for the Cygwin project.
Re: [RFC] libgfortran dll i/o redirection lossage caused by order-of-termination issue
On Mon, Jul 20, 2009 at 12:44:10PM +0100, Dave Korn wrote:
>Corinna Vinschen wrote:
>> Hi Dave,
>
> *waves*
>
>> your mail to cygwin-talk looks like you made some progress in terms of
>> this problem, is that right? Are you going to propose a patch soon?
>
> Yep, and yep later this week. Want to run full gcc-4.3.3-1 testsuite first,
>but fortran results looked promising:
>
>> === gfortran Summary ===
>>
>> -# of expected passes 24749
>> -# of unexpected failures 218
>> +# of expected passes 24925
>> +# of unexpected failures 42
>> # of expected failures 3
>> # of unsupported tests 142
>> /gnu/gcc/releases/4.3.3-1/gcc4-4.3.3-1/build/gcc/testsuite/gfortran/../../gfortran version 4.3.3 20090124 (beta) 1 (GCC)
>
> That's with the Cygwin DLL simply patched to schedule an atexit call to
>dll_global_dtors directly after the exe's global dtors get run.
>
> (Also, going to fix libffi once and for all (both for 4.3.3 and HEAD), this
>will I hope clean up a ton of java FAILs as a knock-on effect. Am working on
>that first while waiting the three days or so it takes to run the full tests...)
So, you've removed the call from do_exit() then? If so, please also remove the
ES_GLOBAL_DTORS definition.
If this is working, I have no problems with it being checked in, assuming Corinna
agrees.
cgf