On Fri, 2011-08-26 at 21:36:31 +0200, Raphael Hertzog wrote: > On Fri, 26 Aug 2011, Guillem Jover wrote: > > +static bool > > +alternative_path_needs_update(const char *linkname, const char *filename)
> IMO you could just as well drop it entirely since you call > xreadlink(linkname, false) (i.e. it can't error out, but it will return > NULL in case of failure). Actually I wanted the common semantics for an x* function. I've now refactored xreadlink to not take a boolean argument. And found out another instance where the code was expecting xreadlink never to return on failure. Included on my next push. > > + linktarget = xreadlink(linkname, false); > > There's a real possibility that linktarget is NULL when linkname doesn't > exist. That's one of the usual case where can_replace_path() will return > true and where path_needs_update() will thus be called. Right, the patch was bogus regarding missing files, “make check” found about it too though. > That's the only problem I could spot. I reimplemented it in a cleaner way. thanks, guillem -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
