IRC channel logs

2022-05-02.log

back to list of logs

<oriansj>ok stage0-posix release v1.5.0 is out
***stikonas_ is now known as stikonas
<stikonas>oriansj: thanks!
<oriansj>now to do a quality of life enhancement for stage0-posix
<stikonas>like what?
<stikonas>oh by the way, maybe worth writing some short announcement message to bootstrappable mailing list?
<oriansj>solve the Generate-$arch-answers sorting order problem
<oriansj>but yeah, I'll send that out later tonight
<stikonas>can you not use out own implementation of sha256sum?
<stikonas>which is locale independent
<stikonas>although it's not clear which one to use...
<stikonas>I guess ${ARC}/bin/sha256sum might work
<janneke>hmm, so blood-elf 2.0.0 from mescc-tools 1.3 needs --little-endian, while earlier blood-elf choke on that
<oriansj>janneke: yeah, it is a breaking change needed to support RISC-V
<janneke>hex2 and M1 identify as 1.1.0
<janneke>oriansj: right
<stikonas>that was already present in the last release if I recall correctly
<janneke>that should not be a problem for guix, but it is hard to giving decent configure errors
<oriansj>and the hex2 and M1 problem was fixed in commit: e58fdf98f2b357066bc0dc2189c9514e99bed372
<oriansj>perhaps a 1.3.1 release for mescc-tools to clear that out for you?
<janneke>oh, that would be real nice
<janneke>so, no 0.24 today
*janneke -> zZzzz
<stikonas>I've update live-bootstrap to stage0-posix 1.5.0. (trivial change with no breakages)
<oriansj>stikonas: nice
<stikonas>started testing bauen1's PR now but won't finish today, will finish tomorrow
<oriansj>and stage0-posix refinement is now up
<stikonas>bauen1: have you updated checksums?
<stikonas>I think you forgot to do that in your PR
<oriansj>and release v1.4.0 of mescc-tools is now out
<oriansj>and that was the last release that supports --BigEndian, --LittleEndian, --exec_enable and --BaseAddress flags
<oriansj>as 2 years and 5 major (1 minor) releases (multiple releases warning of the depreciation and how to convert) should be enough for anyone still depending on that behavior to convert
<bauen1>stikonas[m]: no i didn't, not sure how to do that, since not using a tmpfs will fail all the hashes for me right now
<pabs3>TIL the new Hare language has a bootstrap compiler (in C): https://git.sr.ht/~sircmpwn/harec https://harelang.org/
<achaninja>well - I think that is the only compiler atm, but they plan to keep it forever even if they self host
<pabs3>I see
<pabs3>this page mentions it as a bootstrap compiler https://sr.ht/~sircmpwn/hare/sources
<pabs3>so I thought they had a compiler written in Hare too
<achaninja>its not finished
<achaninja>I could be wrong though i guess it changes pretty fast
<achaninja> https://git.sr.ht/~sircmpwn/hare/tree/master/item/cmd/harec
<achaninja>could someone explain the difference between m2-planet and m2-mesoplanet ?
<theruran>achaninja: and write it here? https://bootstrapping.miraheze.org/wiki/Stage0
<theruran>I totally missed about m2-mesoplanet
<theruran>I mean, I've not heard of it
<janneke>oriansj: beautiful, thank you!
<theruran>whoa congrats on the new releases! @stikonas[m] @oriansj
<pabs3>achaninja: an post from ddevault about Hare bootstrap https://lwn.net/Articles/893309/
<pabs3>"Yes, this is a key design choice of the Hare toolchain. We will maintain the C compiler indefinitely, though we do plan on making a second self-hosted compiler available. The language itself will be frozen at 1.0 so the maintenance burden of two compilers is relatively small."
<stikonas[m]>achaninja: M2-Mesoplanet is a fairly new things. It's C preprocessor with more capabilities than in M2-Planet
<stikonas[m]>theruran: new releases are entirely work by oriansj. My only "contribution" is request for new release
<fossy>bauen1: reviewing
<stikonas[m]>fossy: one of us might need to do checksum update
<stikonas[m]>As bauen1 does not seem to have the right environment (see earlier comment in this channel)
<stikonas[m]>I can run checksum update in 10h or so (unless you do it first)
<littlebobeep>Irvise_: What can be done with OpenGL bindings to s7?
<littlebobeep>stikonas[m]: Where can I read about the licensing issue with heirlooms-tools? Why does it require musl?
<littlebobeep>oriansj: Can you explain why --BigEndian, --LittleEndian, --exec_enable and --BaseAddress flags are being deprecated please?
<janneke>oriansj: there's something funny with ${foo:-<variable expansion here>} in kaem 1.4.0
<janneke>$ echo 'foo=${foo:-${bar}}; echo ${foo}' | foo=foo kaem --strict -f /dev/stdin
<janneke>foo}
<janneke>$ echo 'foo=${foo:-${bar}; echo ${foo}' | foo=foo kaem --strict -f /dev/stdin
<janneke>foo
<Irvise>littlebobeep: what do you mean about OpenGL in s7? AFAIK, s7 is mostly meant to disappear in your C/C compatible src as an extension language. It can be used as a Scheme interpreter too.
<Irvise>I suppose you can just create Scheme symbols that have some OpenGL meaning and then drive OpenGL from there. That would be just like any other Scheme implementation basically.
<Irvise>Maybe I am missing something :)
<stikonas>littlebobeep: those options (--LittleEndian) were deprecated long time ago (in favour of --little-endian), not now
<stikonas>littlebobeep: and heirloom-tools stuff was described in this channel many times, it's just a license incompatibility with meslibc (CDDL and GPLv3 are incompatible)
<stikonas>but CDDL and MIT is fine
<theruran>janneke's announcement on the mailing list says Mes 0.24 is the work of 4 people over a year. so congrats to everyone!
<theruran>I can't get preForth/seedForth to build because I don't have 32-bit toolchain. but the 2020 FOSDEM talk says it supports AMD64 so I am wondering what I am missing
<vagrantc>oooh, mes 0.24 ... oh the fun
<theruran>the source is GPLv3 so I was wondering about porting it to M1-macro?
<stikonas>theruran: what do you want to port to M1-macro?
<stikonas>seedForth?
<theruran> https://github.com/uho/preForth
<theruran>yeah, the Makefile only builds i386 for some reason. it fails to link on my system
<theruran> https://archive.fosdem.org/2020/schedule/event/forth_new_synthesis/
<stikonas>I don't have fasm to build it either...
<theruran>yeah, it's in Guix but not Portage
*janneke has finally released mes-0.24, phew!
<muurkha>congratulations janneke!
<civodul>yay! \o/
<janneke>thanks muurkha, civodul
<theruran>stikonas: kinda strange it relies on GForth but I guess this is a convenience during development
<stikonas>yes, congratulations to janneke too
<stikonas>janneke: which M2-Planet version have you tested with 0.24?
<stikonas>the older one?
<janneke>stikonas: 1.9.0; see https://git.savannah.gnu.org/cgit/guix.git/tree/gnu/packages/commencement.scm?h=wip-full-source-bootstrap#n309
<stikonas>ok, I'll try the latest one, hopefully it will work too
<stikonas>1.10.0 has a fair amount of changes compared to 1.9.0 but hopefully we didn't break anything much
<janneke>mes is still(!?) using --bootstrap-mode, so it should probably work
<stikonas>bootstrap-mode is no longer necessary
<stikonas>but we can deal with it in the next release
<janneke>it is/was for mes
<stikonas>well, it could be easily patched
<stikonas>that oriansj's fork would have necessary patches
<janneke>possibly, i haven't looked into that at all
<janneke>ah right, i should probably check that
<stikonas>I'll try to pick up any useful stuff from there
<janneke>ah thanks!
<stikonas>see kaem files in https://github.com/oriansj/mes-m2
<janneke>ok
<stikonas>(or I can try to send some updates upstream later, just not right now)
<janneke>that's okay, we can align here
*janneke is going for some r&r anyway
<Hagfish>well deserved :)
<theruran>oriansj: I am still having trouble getting the stage2 Forth to do anything interesting. hello world?
<plasma41>janneke: Congratulations on the Mes 0.24 release!
<theruran>oh hecc I got it. using S"
<janneke>plasma41: thanks!
<plasma41>janneke: Don't forget to tag the new release in git
*janneke pushes tags
<janneke>thanks for the headsup, plasma41
<plasma41>New tag successfully fetched!