Bruno Haible <[EMAIL PROTECTED]> wrote:
> Jim Meyering wrote:
>> I did not set errno, which would normally happen for a real failure.
>> I tried, but __errno_location() returns an invalid address.
>> Anyone know how to set errno via gdb?
>
> I normally use
> (gdb) print errno = 22
> since in a si
Jim Meyering wrote:
> I think I'll have to resort to using a moderately large width, yet
> with limited virtual memory, e.g., via ulimit -v 1.
Here is the beginning of a proposal for m4/printf.m4 along these lines.
*** m4/printf.m4.orig 2007-11-03 05:38:54.0 +0100
--- m4/printf.m4
Jim Meyering wrote:
> Unfortunately, you may well have to use xprintf much more generally,
> (or else check all *printf return values) if you care about robustness
> under low-memory conditions.
Yes, it's not only FreeBSD 5 and NetBSD 3 which are affected by this
problem, but also FreeBSD 6, MacOS
Jim Meyering wrote:
> I did not set errno, which would normally happen for a real failure.
> I tried, but __errno_location() returns an invalid address.
> Anyone know how to set errno via gdb?
I normally use
(gdb) print errno = 22
since in a single-threaded situation __errno_location() == &errno
Unfortunately, you may well have to use xprintf much more generally,
(or else check all *printf return values) if you care about robustness
under low-memory conditions.
This is because the printf, fputs, fwrite, etc. functions
(at least glibc-based functions) always allocate memory upon
stream ini
According to Bruno Haible on 10/30/2007 7:23 PM:
>
> Installed:
>
> 2007-10-30 Bruno Haible <[EMAIL PROTECTED]>
>
> * modules/xprintf-posix: New file.
Plus this followup, so that tests will pass on cygwin (xprintf pulls in
xalloc-die, which requires gettext).
From: Eric Blake <[EMAIL P
