On 27 August 2014 17:34, Bert Huijben <b...@qqmail.nl> wrote: > > >> -----Original Message----- >> From: Ivan Zhakov [mailto:i...@visualsvn.com] >> Sent: woensdag 27 augustus 2014 15:29 >> To: Bert Huijben >> Subject: Re: [Patch] R/W lock slowness on Windows (Was: Windows R/W lock >> comment / Reader Writer lock performance on Windows) >> >> Hi Bert! >> >> Did you get any feedback about this from APR developers? What do you >> think about my suggestion that I've posted on apr@ mailing list? > > I didn't get any feedback except yours :( > > 1) > In general apr doesn't promise anything about its output arguments on error > cases, so I'm not sure if we should start look at this specifically here. > (Other platforms don't care and I see at least one that don't set it on all > error cases) > Yes, I know that general practice do not initialize output arguments on error, but I think it worth to keep old behavior because this is very sensitive area (deadlocks and etc).
> 2) > Same reasoning: Other platforms differ here, so I don't see any requirement > to keep Windows strictly as it was. Ok. -- Ivan Zhakov
