apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Branko ─îibej <br...@xbc.nu>
Subject Re: cvs commit: apr/threadproc/win32 proc.c
Date Tue, 02 Jul 2002 23:55:00 GMT
William A. Rowe, Jr. wrote:

> At 05:25 PM 7/2/2002,  wrote:
>>   Modified:    threadproc/win32 proc.c
>>   Log:
>>   Reverting the 1.76 and 1.77 changes, because they didn't work.
>>   The child handles weren't properly inheritable, and redirected command
>>   output got lost in the bit bucket.
> On Which Flavor of Win32?  Tests on XP and 2K indicated this works.

Then the tests aren't good enough. I had problems with Subversion on 
Win2k because the child process output went do /dev/null instead of to 
the supplied handles.

>>   -        if (rv == APR_SUCCESS)
>>   -            apr_file_inherit_set(attr->child_in);
> Of course apr_file_dup[2] handles aren't inheritable, only the 
> child_[in|out|err]
> flavors needed to be.
> If there is a problem, it is NOT in this patch you reverted.  It is 
> probably
> localized to apr_file_inherit_set().  That API didn't exist when the 
> original
> 'make inheritable duplicates' was added.

The first order if business is to get HEAD working again, regardless.

> And you are now passing cloned parent-side handles again to the child
> process which means the parent can't signal the file closed, because
> closing the parent handle doesn't close the handle in the child process.

I'm not sure I understand this. If you want to make the file handle 
inheritable, you must create a duplicate.

Brane ─îibej   <brane@xbc.nu>   http://www.xbc.nu/brane/

View raw message