Mail Archives: djgpp-workers/1998/04/01/03:31:44
On Wed, 1 Apr 1998, Martin Str|mberg wrote:
> Good work, Eli!
Thanks.
> Your last proposed patch regarding djtar corrects the problems I've
> stumbled into.
Great!
DJ, do you need me to resend you the final patch (cc:'ed to workers a
few days ago), you do you have it already?
> kant:~> grep NOSPC /DOS/d/tmp/djlib/djlib202/src/utils/djtar/*.c
> /DOS/d/tmp/djlib/djlib202/src/utils/djtar/epunzip.c: errno = ENOSPC;
> /DOS/d/tmp/djlib/djlib202/src/utils/djtar/untar.c: errno = ENOSPC;
> /DOS/d/tmp/djlib/djlib202/src/utils/djtar/untar.c: if (write_errno == ENOSPC) /* target disk full: quit early */
This is because whoever wrote those parts didn't trust the DJGPP
library ;-). (Or maybe it's a left-over from v1.x or early v2.0
days.) Current versions of `write' set `errno' to ENOSPC
automatically.
> And if you examine the output of djtar you see a printout about NOSPC
> scrolling past, when the disk becomes full. Alas, it scrolls past
> very fast because of a huge amount of EBADF or something similar.
> Perhaps there's another bug lurking somewhere?
What is the case where you see EBADF errors? Is that with the
original djtar, before the patches in the last alpha release? If not,
I'd like to know hoe to reproduce that problem.
- Raw text -