IRC channel logs

<stikonas_>next4th: should be fairly easy to just try. At least if you don't try kernel bootstrapping we have automated everything. live-bootstrap needs about 3 hours on oldish hardware to go from hex0 to GCC 10

<oriansj>lilyp: I have concerns about: d3c5b3b5c66786fada91d9058667ac15a567060b it appears you just generated the C code from the .vala sources instead of actually bootstrapping the language.

<oriansj>next4th: it appears to the logical continuation of collapseOS by the exact same author. The best FORTH bootstrapping work around; tragic that he effectively does it all on his own.

<stikonas[m]>oriansj: which language is that?

<stikonas[m]>Oh I see it in the logs, it's vala itself

<stikonas[m]>Yeah, it does look like pregenerated c files from valac, even comment says so

<muurkha>duskos is the logical continuation of collapseos?

<oriansj>it appears so

<lilyp>oriansj: that's why I said "partly bootstrapped" – I'm rewriting the generated code to be more readable while being guided by the test suite and the guix.scm that builds vala from it

<lilyp>look at the c files in gee/ to see what I mean: the rewritten sources should be easy enought to understand if you know C

<lilyp>there's also some other generated files that I haven't yet been able to throw out

<lilyp>I think I'll probably throw them out once the core's been translated

<lilyp>that said, many source files are still the pregenerated c from vala itself – thus I wouldn't insert it in the bootstrap chain just yet

<fossy>stikonas[m]: i remember why we don't build a binary package for linux.. it uses too much storage space.. anyways i'm trying to see if we can make it work, otherwise might have to wait until builder-hex0 and non-intramfs in sysa

<stikonas>fossy: oh ok, makes sense...

<stikonas>though it is a bit strange

<stikonas>how big is kernel image

<stikonas>isn't it just a few MiB

<stikonas>fossy: and I've sorted out peparatory work for binutils upgrade now (those packaging issues with get_files, moving autotools, etc)

<stikonas>haven't tried upgrading binutils but that's next

<stikonas>(and then there might be some cleanup work, i.e. removing unused autotools)

<stikonas>so it might end up quite a big PR but should be quite simple to review

<stikonas>fossy: actually last one (automake-1.15) fails to work

<stikonas>fossy: do you think it's worth rebuilding bash first?

<stikonas>or just target older binutils

<stikonas>hmm, maybe I should target slightly older binutils

<stikonas>tcc might struggle to build new ones anyway

IRC channel logs

2023-02-20.log