IRC channel logs

<Googulator>rekado: did you see my post about the newly found Yale Haskell sources?

<Googulator>rekado it appears the Yale Haskell team was aware of clisp, and actually began to port to it

<Googulator>(an older version of clisp, that is - the included port doesn't work with modern clisp, and I was unable to compile the period-accurate clisp 2.11 sources found @ http://www.cs.cmu.edu/afs/cs/project/ai-repository/ai/lang/lisp/impl/clisp/source/haible/clispsrc.tgz )

<fossy>stikonas, I can reproduce the qemu issue, but i'm not sure why it occurs

<matrix_bridge><Andrius Štikonas> fossy: it's caused specifically by tcc-boot upgrade

<matrix_bridge><Andrius Štikonas> (Even though this is not a compiler used to build fiwix...)

<matrix_bridge><Andrius Štikonas> I need to test more to fins out which commit

<matrix_bridge><Andrius Štikonas> But given that most are riscv specific I have some suspects

<matrix_bridge><Andrius Štikonas> https://gitlab.com/janneke/tinycc/-/commits/mes-0.26.1?ref_type=heads

<matrix_bridge><Andrius Štikonas> Might be https://gitlab.com/janneke/tinycc/-/commit/dd46e018f64cbd67d907ce6ea91923e627206363

<aggi>finally, i could reproduce tccboot, moving back onto slackware11 as buildhost

<aggi>there was several issues... tccboot when compiled itself with tcc got problems/crashes then

<aggi>yet when compiled with gcc3, it's easily reproduced with a tiny change made to Makefile

<aggi>however, i want to avoid gcc for as long as possible in the dependency graph, and want to keep that _optional_ for GNU stuff later

<aggi>this means, currently, tccboot/kernel-bootstrapping of mine remain locked/blocked against gcc, contradicting the whole idea

<aggi>and, something else, the AoT kernel compilation of linux-2.4 as published by seyko2 seems to merely be a sanity check before kernel-sources are stuffed into tccboot

<aggi>to verify compilation linking only, not intending to boot such a kernel-image, and instead use tccboot later

<aggi>i don't see an easy way of debugging/repairing tccboot compiled with tcc itself, hence i had to get back to seeing why the AoT-compiled kernel crashes

<aggi>for this, i got a known-working state re-produced from source, to see how tccboot JIT-kernel image and tcc AoT-kernel image differ

<aggi>and imo, if tccboot cannot be compiled with tcc but relies upon gcc3 early, it's of little to no use in a kernel-bootstrapping scenario

<aggi>that's probably why tccboot was mostly abandoned, and there was no interest in maintaining a linux-2.x with tcc-toolchain for native boot either

<aggi>because, linux-2.5 already made some changes to Kbuild incompatible with tcc, among other stability issues to stabilize this

<aggi>there's two options to proceed: repairing tccboot compiled with i386-tcc, and/or repairing the AoT compiled kernel-crash

<aggi>and both required reproducing tccboot (which was rather trivial after moving from aarch32 to buildhost=x86)

<aggi>i may have to see too how buildhost poisoned buildtarget with the i686-musl-cross-gcc within tccboot

<aggi>then i can avoid the cumbersome slackware11-x86 buildhost setup

<aggi>ACTION departs for a little while

IRC channel logs

2024-06-12.log