IRC channel logs

2021-04-11.log

back to list of logs

<OriansJ`>I must say the README has one heck of a delta
<OriansJ`>So the next release of stage0-posix will unfortunately have to build the x86 binary for mes-m2 (until I properly port it to M2libc)
<Hagfish>does it make sense for me to ask how M2libc compares in size to something like musl?
<fossy>OriansJ`: what do you mean, build the x86 binary for mes-m2?
<fossy>cant --bootstrap-mode just be used?
<stikonas>fossy: should I use one word per line in import-gnulib?
<stikonas>or just normal page width?
<fossy>stikonas: i prefer one word per line, i think that is the most readable
<fossy>like how you did for tar
<stikonas>ok
<stikonas>well, both were pasted from the comments in generated file
<stikonas>one was split, another was one liner
<fossy>oh, right
<fossy>i think if you split it manually it would be better tho
<stikonas>yeah, I split it and updated PR
<fossy>:)
<stikonas>ok, I'll remove pylint since in the end I actually used default file
<stikonas>I added those ignore comments inline
<fossy>alright that sounds good
<fossy>i'm moving back a version to perl 5.10 cause 5.12 seems to be kinda buggy
<stikonas>ok
<fossy>even with ./Configure
<stikonas>I briefly looked at building makeinfo but it needs newer perl
<fossy>ah ok
<fossy>hopefully 5.10 is more simple
<stikonas>yeah, if not 5.8 would definitely work
<fossy>5.12 had almost double the number of .o's than 5.6 for libperl
<stikonas>I did some initial work on 5.8 before
<stikonas>but then we discovered those pregenerated files
<stikonas>so went with that chain starting from 5.0
<stikonas>fossy: the other PR updatede too (although, it's on top of findutils branch)
<stikonas>so includes that commit too
<fossy>oh ok
<stikonas>ok, pylint passes without that pylintcfg
<stikonas>fossy: by the way, when are you building perl, after 5.6 of after gcc?\
<OriansJ`>Hagfish: yes it is certainly a reasonable question. It is smaller but it should grow to the size required to build TCC with mescc.
<OriansJ`>fossy: well if you remember M2-Planet builds require asm blocks to do system calls and such things are unique for each architecture. So mes-m2 is using the asm blocks that janneke did and they are x86 specific. So to make a mes-m2 binary that runs on AArch64 or armv7l, I'll have to provide those bits of functionality on those architectures as well.
<OriansJ`>The path forward would be to implement that functionality in M2libc
<fossy><stikonas> fossy: by the way, when are you building perl, after 5.6 of after gcc?\
<fossy>After gcc
<fossy>OriansJ`: I see
<stikonas>fossy: so are findutils and rootfs.py port ready for merge, or do I need to make more fixes?
<fossy>stikonas: just merged before seeing lol
<stikonas[m]>Thanks
<fossy>Perl is close 2 done
***ChanServ sets mode: +o rekado_
***Noisytoot is now known as Noisytoot__
***Noisytoot__ is now known as Noisytoot
***stikonas_ is now known as stikonas
<stikonas>gforce_de1977: I'm now testing a possible fix for https://github.com/fosslinux/live-bootstrap/issues/92
<stikonas>although, I can't run it as many times as you did
<bauen1>stikonas: can you push your fixes somewhere ?
<stikonas>sure, but completely untested atm
<stikonas>bauen1: https://github.com/stikonas/live-bootstrap/tree/autoconf.as
<stikonas>still building tcc now
<stikonas>bauen1: so I think timestamp of autoconf.in will not be greated than that of autoconf.as
<stikonas>so it should build fine
<stikonas>and we need to regenerate it in any case
<bauen1>stikonas: that's great then
<bauen1>stikonas: not sure if there are other dependencies specified in the makefile
<stikonas>hmm
<stikonas>bauen1: ok, it actually doesn't build at autoconf-2.65
<stikonas>probably autom4te 2.61 is too old
<stikonas>I wonder if dependencies are solved in autoconf 2.65
<gforce_de1977> stikonas: no problem, i will start tomorrow a new massrun, so we have results 24h later
<stikonas>no idea about the other failures right now
<stikonas>but if we can fix at least this one, that would still be good
<gforce_de1977>stikonas: it does not matter for now, it is just a computer which has time and electricity 8-)))
<gforce_de1977>so step by step
<stikonas>by the way, you might need to do some minor adjustments to your script
<stikonas>since we removed rootfs.sh now
<stikonas>./rootfs.py is mostly compatible, but command line flags now have dashes in front, e.g. ./rootfs.py --chroot (or -c)
<gforce_de1977>ok
***nckx is now known as jorts
<stikonas>bauen1: so unfortunately autoconf that patch doesn't help...
<stikonas>although, it's still good to merge, but I should remove "Fixes: #92" from description...
<stikonas>(I'll possibly convert it to simply rm bin/autoconf.in file)
<stikonas>hmm, I think the correct workaroud would be to first run make -C lib