IRC channel logs

2021-12-22.log

back to list of logs

<oriansj> gbrlwck: the logic you are looking for is in the bottom of: boot-0.scm
***nckx is now known as nckxmas
<gbrlwck>oriansj: thanks!
<gbrlwck>fossy: i just opened a pull request for live-bootstrap. i tested it yesterday and it worked
*gbrlwck wonders how mescc can be unhappy about a seemingly trivial c file like lib/mes/eputs.c
<stikonas>gbrlwck: you have offended the linter there, can you split the long line?
<gbrlwck>huh, didn't realize there was a CI script running automatically! nice
<gbrlwck>it is now fixed
<stikonas>gbrlwck: it used to run the whole bootstrap
<stikonas>but eventually live-bootstrap grew too big
<gbrlwck>even cooler
<stikonas>(there is 2h limit on linter runtime)
<stikonas>gbrlwck: what about the other mescc invocation (when building tcc)?
<stikonas>it still has those options you removed in meslibc build
<gbrlwck>ah, i haven't gotten that far yet. but as the use of mescc goes: the -- means ignore the following arguments for the script, initiate an interactive session and use the args for that. so i guess it is (weirdly) unwanted but still working (which confused me before)
<gbrlwck>and i'm not really sure what the intentions behind setting the --no-auto-compile option were. it's useless with MES, but when debugging/running mescc with guile it's kind of a drag (compiled guile is just *so much faster*)
<stikonas>gbrlwck: I'm not sure either. Anyway if it works without it then it's fine. And as for tcc, we can do it later
<gbrlwck>i'm testing it now (with the changes for tcc)
<stikonas>anyway, I've tested that PR, live-bootstrap runs
<gbrlwck>cool
<stikonas>well, if you want to add tcc first, we can do that too
<stikonas>and merge the whole thing
<gbrlwck>should i also update the --LittleEndian/--little-endian stuff?
<stikonas>that's a change for mes-m2, not live-bootstrap
<stikonas>that commit is in upstream mes, juts not in m2 fork
<gbrlwck>ah yes, i thought there needed to be some kaem script patched in live-bootstrap
<stikonas>ok, it's in wip branch https://git.savannah.gnu.org/cgit/mes.git/log/?h=wip
<stikonas>no, that's mescc internally calling mescc-tools
<stikonas>might be worth cherry-picking it onto mes-m2 though
<stikonas>to silence those warnings
<stikonas>but that's another PR to another repo
<gbrlwck>huh, now i get "assert fail: alloc: out of memory" when compiling tcc
<gbrlwck>updated the command line arguments for mes-m2 https://github.com/oriansj/mes-m2/pull/16
<gbrlwck>did i miss something?
<stikonas>gbrlwck: not sure but maybe let's merge the PR without tcc for now
<gbrlwck>yes
<gbrlwck>i am naively giving it another go with MES_MAX_ARENA doubled but i'm not sure that's gonna help
<stikonas>yeah, I'm not sure what's going on
<gbrlwck>i still suspect some bug in mes