On Wed, May 30, 2012 at 11:59 PM, Alan Coopersmith <alan.coopersmith@xxxxxxxxxx> wrote: > Sorry, I haven't been paying as much attention to this as I should, but > I did notice this thread and wonder about this fix. Please see git log. that should explains all what you want to know. > Isn't this backwards? fontconfig/fontconfig.h has #define FcFalse 0, but at > least the Solaris scandir man page says Doh. right. just a typo. > If this pointer is NULL, then all the directory entries are included. > > which suggests a simpler and more efficient fix of: Good point. > It also looks like you could save a whole bunch of malloc calls by either using > a fixed size buffer of PATH_MAX, or saving the f pointer and realloc'ing it to > be larger if needed, instead of malloc and free every time, such as: True. that makes sense. > Both of those are just suggestions which I have not tested (or even tried to > compile). Thanks for catching this up and kindly your suggestion :) > > -- > -Alan Coopersmith- alan.coopersmith@xxxxxxxxxx > Oracle Solaris Engineering - http://blogs.oracle.com/alanc -- Akira TAGOH _______________________________________________ Fontconfig mailing list Fontconfig@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/fontconfig
