This is the mail archive of the guile@cygnus.com mailing list for the guile project.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
; 1. The Emacs interface that was recently included in the snapshot ; obviously needs some Emacs-side support. What would that be? Also, ; is there a way to disable the readline support short of activating the ; Emacs support or hacking boot-9.scm? 'Cause I was perfectly ; comfortable in Ilisp, but the above two new things mess it up (would ; be great to check (isatty? (current-output-port)) before turning on ; readline, for example :). Oh dear, is the readline support actually enabled in a guile interpreter? I just wrote a function (readline-repl). You're right, it shouldn't be called from a non-tty. Maybe it shouldn't be called at all... you can just use an init file to do this: cat >interactive-guile <<EOF #!/luser/yokel/bin/guile -s !# (use-modules (lang rx) (lang ctax reader)) (readline-repl) EOF ; 2. There are some things Guile needs. Just off the top of my head, we ; need a better module system. I remember discussions about it on this ; list. I'm sure some of the maintainers has a preliminary design or a ; sketch lying somewhere, accumulating dust till the said maintainer has ; more time (fat chance of that, it seems!). I'd probably take a stab ; at it if I were sure I'm not making some terrible design errors. I'm ; sure seeing the design or whatever that was aggried apon by the ; maintainers would help a lot, as I'm clearly not seeing the whole ; picture (don't know a thing about dynamic linking, for instance). If ; I actually implement something, the maintainers are always free to ; reject it. So why not publicize the design? That's a good idea; but I don't think anyone ever did decide on a module system. For myself, I wrote a few hacks that lived on top of the current module system that did things like renamed imports, absolute names, and a syntax I liked better. Nobody was very interested... What do you think the module system needs? I'll start with my wishlist, along with how I'd do it with the current module system: * Autoloading (use-modules) * Modules as objects (current-module returns the current one, how do I look up somebody else's?) * Easy explicit lookups (eg: ((@ ice-9 session apropos) "apropos")) * Individual renamed imports ( (define session-apropos (@ ice-9 session apropos)) ) * Renamed imports from a whole module (eg: (define sess (lambda (var) (module-lookup (ice-9 session) var))) ... ((sess apropos) "apropos") Good syntax for all of the above would be nice. ; 3. My two little patches were ignored (q.scm queue exhaustion fix and ; r5rs multiple values). Am I such a bad netizen, or is it just that ; nobody cares about these things? No, I didn't see anything wrong with your posts... what worked for me, though, was to stick them in incoming on red-bean, then notify the (very busy) maintainers about them. They eventually got around to putting the patches in. ; 4. Yup, I know, nobody's got the time... Yeah... but the maintainers have put together a good package (thank you!). I really can't fault them for being busy. Andrew aarchiba@undergrad.math.uwaterloo.ca