IRC channel logs

<aggi>pnut cannot digest steps/tcc-head (instead of steps/tcc-0.9.26) without further treatment, not yet

<aggi>testing here indicates no way around tcc-0.9.26 as initial bootstrap version (with pnut)

<aggi>after all, if pnut can compile tcc, that's an almost complete c-compiler itself

<matrix_bridge><Andrius Štikonas> Yes, i think everybody wants to get rid of tcc-0.9.26 fork. And in the end there might be multiple options

<matrix_bridge><Andrius Štikonas> Not just a single bootstrap chain that one must use

<aggi>currently thinking how to handle these variants (mes-cc|pnut|..); considering a gentoo-overlay style workflow for these individual steps/

<aggi>and the transition from M2 towards tcc is the most tricky, conceptually, because anything which can compile tinycc more or less is a candidate itself to replace tinycc

<aggi>with the main difference, any such thing can/must be spawn from M2 directly to crack open the "self-hosting" chain

<aggi>seems tcc-head applied live-bootstrap patches/ ; except for the pass1.kaem fiwix one and pass2.sh/static-link.patch which i think could be handled differently

<aggi>having a little trouble to rebase/apply the fiwix patch with pass1.keam onto tcc-HEAD

<aggi>hopefully i got the right spot and diff

<aggi>and since it's some simple-patch i don't know yet how such are updated/re-based

<aggi>sigh, got one spawned directly after pnut/tcc-0.9.26: $ /var/tmp/tinyroot/usr/bin/tcc -version

<aggi>tcc version 0.9.28rc (i386 Linux)

<aggi>some of these changes might even help with swapping tcc-0.9.26 with tcc-head-pnut one step earlier

<aggi>in any case, some latest tcc-HEAD can be re-integrated with tiny-bootstrap/live-bootstrap early

<aggi>and so far, steps-pnut/ OVERLAY pays off with compile-time performance while testing significantly

<aggi>for this alone it's a bliss laurenth succeeded with his hackjob (just a few weeks ago i see)

IRC channel logs

2026-03-11.log