On 17/02/13 11:17, Allan McRae wrote: > On 17/02/13 11:08, Andrew Gregory wrote: >> On Sun, 17 Feb 2013 10:32:48 +1000 >> Allan McRae <[email protected]> wrote: >> >>> On 16/02/13 12:02, Andrew Gregory wrote: >>>> alpm_filelist_contains was being used to search for resolved paths, but >>>> searching in the unresolved paths, causing it to miss matches. >>>> >>>> Signed-off-by: Andrew Gregory <[email protected]> >>>> --- >>> >>> Instead of requiring _alpm_filelist_resolve() to be called before >>> calling alpm_filelist_contains(), just add the call in >>> alpm_filelist_contains() itself. If the filelist has already been >>> resolved, it returns immediately anyway. >>> >>> As an aside - this whole situation really sucks! We now are resolving >>> every path... >>> >> >> I considered doing that. The reason I decided not to was that >> alpm_filelist_contains wouldn't be able to easily notify the caller if >> there's an error resolving the files. We aren't checking for those >> errors at the moment, but it would be good to add that at some point. > > But there is no indication that the resolve function needs to be called > first for the filelist_contains function to work. >
Also, I just noticed that alpm_filelist_contains() is public, _alpm_filelist_resolve() is not... This is a bit of a mess. Allan
