I'm trying to break string's dependency on unicharutil and other i18n
libraries like uconv. The problem I'm running into is that the string
library (both new and obsolete) provides routines dealing with case,
such as ToUpper(), ToLower() and CaseInsensitiveFindInReadable(). These
routines require the use of the i18n libraries (specifically,
nsICaseConversion) to do figure out what the upper-case of a particular
unicode character is.
What I'd like to do is actually remove case-related unicode APIs from
the string library, and provide them as helper routines that can be
statically linked in as a seperate library for only those consumers
which actually use these routines. The declarations of these routines
would also move out to a seperate header file, like nsUnicharUtils.h.
The static library effectively acts as glue to call related routines in
nsICaseConversion, while retaining the ease-of-use of the current APIs.
Only those consumers who actually need these routines would actually
link against this library.
I've gone through the tree to do this work, and found about a dozen
consumers of such routines for the new string APIs, though I haven't
touched the old APIs.
Comments? See http://bugzilla.mozilla.org/show_bug.cgi?id=100214
Alec
