***linas_ is now known as linas
<wingo>just landed cps cse, better than the old tree-il cse <wingo>would be nice to have a verilog backend, somehow... <taylanub>Compiling master seems a lot more intensive than stable ... <tupi>wingo: i started to update guile-clutter to 1.12, first step first, learning [the fewest possible things about C of course :)], folowing HACKING... and i have quizz, some i1'll send an email, maybe, some i'd like to ask here, here is 1: it seems that overrides/clutter.defs is not automatically updated, i wonder why, did i miss a step maybe? <mark_weaver>my guess, simply by the name, is that overrides are things where the automatic wrapper gets it wrong, thus overriding what it would have done. so, by definition, that file has to be done by hand. <tupi>civodul: hello. tx for your help the other day, it did work <tupi>mark_weaver: i'm asking because it could be automatically be built, i think <mark_weaver>saying that everything could be done automatically is, I think, equivalent to saying that no overrides are needed. <tupi>mark_weaver: by extracting from include/clutter-1.0/clutter/deprecated/*.h per major, such as CLUTTER_DEPRECATED_IN_1_8 ... <tupi>so i'm curious and like to know why wingo didn't do it <civodul>tupi: does Guile-Clutter use G-Wrap? <tupi>mark_weaver: that's another question [for wingo], why overrides subdirectories 'only', and not both an overrides and a deprecated... <tupi>mark_weaver: just to make sure, i'm saying this because unless i missunderstood [which is why i'm asking], deprecated can be fully automatic. then overrides would _only_ contain manually maintained files [per definition], which i understand is necessary, as explained in HACKING [from the master branch version] ***ijp is now known as the-universe
<mark_weaver>taylanub: if you're using gc-7.4.0 on guile master, then you probably need to either upgrade gc to current git, or set GC_MARKERS=1 ***the-universe is now known as ijp
<taylanub>oh, ok. thought it was related to a change I made at first, but that's unlikely when I think about it *tupi wonder why h2def has been written in python and not in scheme <taylanub>hrm, I still get out-of-memory while compiling assembler.go