IRC channel logs

<aggi>almost forgot, there is another minor piece dependening on perl a little early

<aggi>isohybrid utility, used to create hybrid ISO which can boot from USB flash instead of CDR

<aggi>it's relevant insofar the bootstrapping target here will be bootable ISO filesystem

<aggi>in this regard cdrtools themselves seem a little problematic too

<aggi>at least i got all those ported to tinycc already

<aggi>another aspect, since i've _never_ played with any LISP/Scheme before

<aggi>i remember GNU guile first, that was pulled into as dependency for GNOME/gtk while ago

<aggi>i noticed the compilation time for guile was significant, hence i detangle and removed that

<aggi>meanwhile, of cause, Scheme did caught my interest, since it's a major language to spawn live-bootstrap

<aggi>good news is, LISP/Scheme are standardized (didn't know that)

<aggi>with many implementations available, including GNU mes

<aggi>and initially i noticed TinyScheme

<aggi>for integration into TinyCC/OS, not sure if that was fully capable implementing related standards

<aggi>anyway, all preparational tasks are finished, tinycc/os is deployed into production, runtime testing passed

<aggi>so, i'll fully focus on a complete review of stage0-posix and live-bootstrap next, and re-integrate a minified tinycc/os release variant

<aggi>i'll try to keep the dependency graph as small as possible (no autotools, no perl, no python, no portage)

<aggi>while ago it was Aboriginal Linux project which seeked to implement a smallest linux/busybox/tinycc possible which can re-compile itself

<aggi>and i'll add full bootstrapping as criteria to this

<aggi>stikonas: good to see you're online

<aggi>since all preperational tasks to re-integrate tinycc/os for a full bootstrap i'll begin a full review of stage0-posix and live-bootstrap now

<aggi>i assume both posix-stage0 and live-bootstrap are sufficiently stable and working on latest HEAD to begin with

<aggi>first i'll test the regular procedure, and then fork live-bootstrap into a repository named tiny-bootstrap with a few changes made there

<aggi>stikones: if there was any known problems on latest HEAD of posix-stage0, livebootstrap or any of the other related repositories (M2-planet etc) please let me know in advance

<matrix_bridge><Andrius Štikonas> Not aware of any problems, though didn't look recently

<aggi>have to take a deep breath... it's been many years of nonstop hacking to get tinycc/os production ready

<aggi>the hope i put into this, it can significantly shrink the total lines of code necessary with initial live-bootstrap (or tiny-bootstrap how i'll name the fork of it)

<aggi>to arrive at a fully capable *nix development host

<aggi>Frans isn't online, currenlty i'm into RTFM and listening to various talks on youtube

<aggi>anyone knows if the MES-replacement is ready for integration with live-bootstrap?

<aggi>i was too listening to the talk from rickmasters again ... seems he's gone too

<aggi>and he too mentioned the major compile-time penalty originates from GNU mes / scheme

<aggi>if anything, time wasted with waiting for compile-jobs to finish is what i dislike most often

<aggi>so, if MES-replacement from Frans was ready, i would consider skipping the GNU mes pieces alltogether

<aggi>besides, in comparison to the tinycc-portage fork which isn't published, i'll git-push the tiny-bootstrap system integration from the beginning

<aggi>other than re-confirming stage0-posix-x86 of cause, i'm not yet sure how to proceed with live-bootstrap/GNU-mes and/or MES-replacement

<aggi>rickmasters mentioned the total compilation time is ~13hours, including binutils/autotools/python/perl/gcc4/linux4 etc... and i'm already certain i can shove a few hours from this with the tinycc/linux2 option ready

<aggi>because that will not need any python/perl/coreutils whatever... i think a linux2/busybox/tinycc/yasm bundled into a bootable ISO will do just fine as developent host to kickstart GNU-toolchain with

<aggi>and well, the linux-tcc system got SMP, which probably will help with spawning gcc etc. significantly

<aggi>according to MES-replacement documentation Frans he considers it mission accomplished for x86... good

<aggi>so i'll skip the scheme pieces alltogether, major reason being compile-time performance

IRC channel logs

2026-02-28.log