Re: [SyncEvolution] [PATCHES] hijack error 404 to 401 when appropriate (2nd take)

On Di, 2011-06-21 at 17:59 +0300, Salvatore Iovene wrote: > Hi Patrick, > > On Tue, 21 Jun 2011 16:39:40 +0200 > Patrick Ohly <patrick.ohly(a)intel.com&gt; wrote: > > > commit 8c55193d34400a2e94089d9fa2e750866c491515 > > > Author: Salvatore Iovene <salvatore.iovene(a)linux.intel.com&gt; > > > > > > NeonCXX: don't trust libneon's escape and unescape > > > functions. > > > > Do you have reason to not trust libneon here? We rely on the "Neon > > does not return NULL" semantic in various places. However, I must > > admit that I don't know whether it applies here. NULL might > > indicate something other than out-of-memory here, like "bad > > input". > > I have had unescape return NULL for "%u" more than once. I thought > that if the unescaping (or escaping) should fail, it's better to > return the original string, because, well, it couldn't indeed be > (un)escaped. Makes sense. Adding such information to the commit message or (better) as source code comments helps reviewers and future readers of the code. That person might even be you... I certainly rely more on source code comments than my own memory to remind me why I have written a piece of code the way I have ;-}