Okay, it's not urgent task so that sounds better waiting for your work done. then revisit this afterwards. thanks for your comment! On Tue, Sep 25, 2012 at 11:35 AM, Behdad Esfahbod <behdad@xxxxxxxxxx> wrote: > On 09/24/2012 10:28 PM, Akira TAGOH wrote: >> So as the developer, I personally failed to find any useful. current >> proposal can do things with same or near costs to current >> implementation. so FcPatternGet*() might be replaced by them in the >> future perhaps. >> >> I may be missing something else. if I'm wrong, please correct me. > > You're not missing anything. Indeed, with my proposed solution, there will be > some extra work happening during the lookups. I don't think we should be > concerned by that though, if we don't know of any usecases where that would be > a measurable performance issue. > > Plus, in the thread-safety work I'm currently doing, I'm replacing with object > lookup hash table with something static and faster, so I'm fairly sure the > overhead will not be a problem in any realistic usecase. > > Moreover, we can add the simple API that addresses the API gap, and always add > an iterator API later if we really see a real performance issue with the > simple API. > > Up to you in the end. > > Cheers, > > behdad -- Akira TAGOH _______________________________________________ Fontconfig mailing list Fontconfig@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/fontconfig
