Re: strerror_r and POSIX

[Eric Blake]

On 05/21/2011 11:45 AM, Bruno Haible wrote:
> What about the platforms that don't return ERANGE when they truncate?
> According to <http://sourceware.org/bugzilla/show_bug.cgi?id=12204> these
> are AIX 6.1 and OSF/1 5.1. Are we sure that we override the system's
> strerror_r() on these platforms?

I plan to work around that, by checking whether strlen(msg) < buflen -
1, and if not, use a second iteration with a larger buffer to prove
whether truncation occurred.  It's certainly on my list of things to fix
before declaring strerror_r complete.

>> The other is whether strerror_r may modify buf even when returning
>> ERANGE or EINVAL.  Right now, POSIX does not forbid this, and in fact,
>> it even recommends it in the case of EINVAL.  I took care in
>> http://austingroupbugs.net/view.php?id=398 to also add wording
>> recommending that buf is also modified in the case of ERANGE to always
>> be a NUL-terminated string (except for 0-size buf).  And it is this
>> issue that glibc is _not_ doing (hence glibc bug 12782)
> 
> OK, the ticket in glibc is now fixed. So even if the Austin group rejects
> your proposal, gnulib will be implementing a GNU extension.

Yep, and since Uli fixed __xpg_strerror_r in glibc.git, I now have a
reason to reject older glibc __xpg_strerror_r versions if they don't do
everything nicely.

-- 
Eric Blake   address@hidden    +1-801-349-2682
Libvirt virtualization library http://libvirt.org

signature.asc
Description: OpenPGP digital signature