On Thu, 6 Jan 2011 18:44:51 +1200 Cedric BAIL <cedric.b...@free.fr> said:
> On Thu, Jan 6, 2011 at 6:33 PM, Mike Blumenkrantz <m...@zentific.com> wrote: > > On Thu, 06 Jan 2011 07:21:18 +0100 > > Sebastian Dransfeld <s...@tango.flipp.net> wrote: > > > >> On 01/05/2011 09:18 PM, Lucas De Marchi wrote: > >> >> + if (errno) > >> > > >> > This is bad... you have to either zero errno before calling open() or > >> > check the return value before. Since open() will not zero errno in case > >> > of succees, doing like you did you're actually carrying whatever value > >> > errno had. > >> > >> Should users always zero errno before checking? Then this needs to be > >> checked throughout efl. I remember a while ago someone actually removed > >> errno = 0 in some code. > >> > >> Sebastian > >> > > Yes, I remember someone removing the errno checks as well. According to the > > man page, however, you must set it to 0 before you check it to receive valid > > results, so I would guess that this is indeed something that should be > > verified throughout the codebase. > > > > Perhaps a good opportunity for a cocci script? :) > > Nah, that's wrong also, errno only mean something if open return a fd > < 0, all other case are not defined. Before checking errno for any > syscall, you need to first check if the syscall failed. kowwekt! :) -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- The Rasterman (Carsten Haitzler) ras...@rasterman.com ------------------------------------------------------------------------------ Learn how Oracle Real Application Clusters (RAC) One Node allows customers to consolidate database storage, standardize their database environment, and, should the need arise, upgrade to a full multi-node Oracle RAC database without downtime or disruption http://p.sf.net/sfu/oracle-sfdevnl _______________________________________________ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel