Re: utf8, macos x, and fmt

  • From: Bengt Kleberg <eleberg@xxxxxxxxxxxxxxx>
  • To: wilyfans@xxxxxxxxxxxxx
  • Date: Wed, 14 May 2003 11:31:51 +0200 (MEST)

> Delivered-To: wilyfans@xxxxxxxxxxxxx
> Subject: Re: utf8, macos x, and fmt
> From: "Mark H. Wilkinson" <mhw+wily@xxxxxxxxxxx>
> I've not looked at the code, but if these libraries support the same
> APIs that libplan9c does at the moment wily & sam would probably not
> require any re-writing.

you are correct. the libplan9c api is a subset. there is just more of
everything in russ collection.

> Is russ's libutf related to the one at
> , or are they implementing
> different APIs (one Plan 9 derived and one not)?

both libs have the same utf*()/rune*() functions. alistair also has
lots of unicode*() functions not available from russ.

> libplan9c tries to bring together what bits of the Plan 9 libraries were
> available at the time into one library, so I guess one question is
> whether to split that one library into four or five separate ones.

imho it is preferable to have one lib (and one header file). in this
case it might mean adding all the new functions to libplan9c. suddenly
it seems very attractive to use the existing superset of libraries.
even if they are > 1.

> The other thing 9libs brings is the configure script (which probably
> works out larger than the source of the libraries :-) ). Is that worth
> keeping, or do people look for binary distributions these days?

this is a very good point. i still do not like the autoconf system. and
i still have not built something better. binary distributions might be
a good idea. those who can, will get the source and handle the makefile
edits that are needed. those who can not, will get a binary. presumably
those who can not are using a few, very well know, platforms.

i do not know.


Other related posts: