Hi, On Tue 27 Jan 2009 21:55, [email protected] (Ludovic Courtès) writes:
> I think Guile-Lib should check whether [modules] are available at > configure-time and install them iff they are not already available. I'd rather just require a specific version of 1.8.5... Anyway, if you want to do this, once the savannah thing goes through I'll add you to the project :) > Besides, SRFI-35 is broken (see > http://git.savannah.gnu.org/gitweb/?p=skribilo.git;a=blob;f=m4/guile.m4) This kind of thing is what I'd like to avoid. Users won't notice the difference if a newer guile-lib requires the latest release of a specific stable series, since stable series are compatible. >> (2) I removed (math rationalize), as its core functionality, >> `rationalize', has been in Guile since 1.8. > > Maybe it should be left there for compatibility, and just re-export the > procedure? Same for `(container queue)'. (math rationalize) also had unclear licensing, so I'm inclined to keep it out. (container queue) was never documented -- but doing a google code search finds one user... (container async-queue). So I should probably add it back then, sigh... I really don't want to be in the position of being "responsible" for poorly documented modules. > http://permalink.gmane.org/gmane.lisp.guile.guile-lib.devel/454 Pushed out to http://wingolog.org/git/guile-lib.git, still waiting for the asynchronous project start on savannah... Andy -- http://wingolog.org/
