tags 297980 -help tags 297980 +pending thanks Jan Nordholz writes: > Removing the "#ifdef pfa" around the return value check should fix > the problem, although that's likely only mending the symptoms, not > the cause: The code in question has no support whatsoever for the > case that the ioctl()ed FD is already closed, so maybe this should > be detected and handled properly earlier.
I tried that and behold, the bug is fixed! I agree that this fixes the symptoms and not the cause. However, that will be the solution in Debian for the time being. The reasons are (a) to keep the patch minimal, and (b) because the sources in CVS have been refactored and reorganised quite heavily. Of course I will submit the patch to the upstream authors. FYI, the offending #ifdef pfa is still there in CVS, although the file moved from gvd/gnat to common/gnat. Also FYI, the upstream sources now use /dev/ptmx if available, and fall back to a function called allocate_pty_the_old_fashioned_way (sic) if this fails :) Thank you for the invaluable help! -- Ludovic Brenta.
