Dear Akira-san -- Akira Kakuto wrote: > It seems that I was wrong: > > By experiments, I found that the renaming, > cachefile.NEW --> cachefile, almost always > fails if executed from the XeTeX binary even > in a user's writable directory. Thank you -- further experiments after installing a further new font have indeed shewn this to be the case. It would also seem that, in my case at least, that I have write access to the default location for FC_CACHEDIR so the effect of explicitly setting that to C:\TeX\Live\2016\FC_cache was to allow the initial creation of a new cache (which of course helped) but did not address the real problem that you have identified with :
#ifdef _WIN32 unlink ((const char *) atomic->file); /* fails always if executed from the XeTeX binary */ #endif > Thus if you find a delay of time, run fc-cache -v. > > If messages > invalid cache file: ... > are shown, run > > fc-cache -v > > once more. Thank you, and noted for future reference. ** Phil. -- Philip Taylor
-------------------------------------------------- Subscriptions, Archive, and List information, etc.: http://tug.org/mailman/listinfo/xetex
