Hello, On Sat, Jul 18, 2009 at 04:22:36AM +0200, olafbuddenha...@gmx.net wrote: > On Thu, Jul 16, 2009 at 10:41:34PM +0300, Sergiu Ivanov wrote: > > On Fri, Jul 10, 2009 at 04:17:23AM +0200, olafbuddenha...@gmx.net > > wrote: > > > On Sun, Jul 05, 2009 at 07:10:48PM +0300, Sergiu Ivanov wrote: > > > > > I met this style (using intermediate variables) quite often in > > > > unionfs and my understanding is that keeping to such style you > > > > don't clobber the parameters. That is, should fsys_getroot fail, > > > > port will remain unchanged. > > > > > > You are right. However, if it is really important that "port" > > > doesn't get clobbered when an error occurs, this should be > > > explicitely documented in the function comment. If callers don't > > > rely on this behaviour OTOH, I wouldn't bother. > > > > I added the corresponding bit to the comment. IMHO, it is a good > > thing when the function does not clobber its out parameters on errors. > > I wouldn't subscribe to it being a good thing in general. I don't see > how it could avoid programming errors, or otherwise make code more > readable, to always do this. It is useful only in specific cases. > > In this particular case it does indeed seem totally useless: it's not > like the caller might have some other value in there that it would like > to keep using when this call fails...
I see. Thank you for the explanation, I'll keep this in mind. > This is not really an important issue, though. In this case, I think I'll leave it as it is now, both for (probably meaningless) consistency and for saving the time for more important issues. Regards, scolobb