***linas__ is now known as linas
<lamefun>why is (define) allowed in (lambda)? there are let and letrec <naboo_>hello, what is xchar type? when i use guile-ncurses i get: expected-type: xchar in error message <stis>yeaaahhh bdw_gc is modded, i'm soooooo awful. <naboo>wingo: for example (normal acs-btee)) ***dsmith-w` is now known as dsmith-work
<stis>Hmm, I now have the gc basics working, yehaaa <stis>civodul: I got a full gc of prolog varibales worrking! <stis>I needed to mod bdw-gc though :-( <stis>Is it possible to make that change into upstream do you think? <civodul>stis: no idea what the change is, so i can't tell :-) <stis>hmm, I will try to make a repo diff <stis>Anyhow I will setup a repo for this so that people an try and look <stis>Then I will make it optional to use this feature in guile-log <stis>meaning at first they need to compile and install a special made bdw-gc <stis>itt is good enough for brave ones :-) <daviid>i've noticed that for opaque types, for which we [have to] provide code in support-* like ClutterMargin for example, have a correspondig goops class <clutter-margin> i.e., but no real goops integration like for other types, like <clutter-actor>, so no (make <clutter-rectangle> #:x ...) but (clutter-rect-init (clutter-rect-zero) 10 10 100 100) ... I wonder if I missed something [I don't think so], and if not, would it be possible to <daviid>'better' get these opaque types 'done' ? <daviid> should it not be: (scm-error 'gruntime-error #f format-string args #f) ? <daviid>not exactly, but I'll find out. there are, apparently, quite a lot of scm-error calls with wrong args, I wonder when scm-error code changed ? <daviid>I found the answer to my 1st quizz, these are subclasses of gvalue-class and can be instanciated using #:value, i.e.: (make <clutter-rect> #:value '(10 10 50 50))