Pádraig Brady wrote:
Now the gnulib replacement is only table lookup and some bit manipulation. Ah it also calls locale_charset()! That must be slow on OSX. Indeed :(
--- It may be a red herring, but I seem to remember Mac choosing to go with what is now a non-standard normalization form (NFD), vs. linux, windows and the web going with NFC.
In computing char lengths it might be the case that something is converting characters to a normalized form and maybe the gnu code uses the more standard form, and somewhere text is being forced through 1 or more normalization transforms? It may be a non-issue, but was something that occurred to me as problematic should text between platforms be shared and run through text processors that want a native normalization form.
