IRC channel logs

<Googulator>While waiting for Rust to bootstrap again, I decided to check out Yale Haskell a 2nd time - and noticed something.

<Googulator>The original v2.05 source distribution's README mentions running Yale Haskell under a language named "t" or "T" - although in 2.05 (which is unfortunately the only surviving version I currently know of), it's mentioned to be broken, and the referenced "t-support" directory is missing.

<Googulator>Other sources state that Yale Haskell was originally written in "t, a dialect of Scheme", and then ported to Common Lisp.

<pabs3>it happens I recently archived a site related to t/scheme98 or something like that.... let me see

<Googulator>And the vast majority of Yale Haskell's source code (apart from what's in Haskell) uses the ".scm" extension, and reads very much like Scheme, and not Common Lisp.

<Googulator>There's a layer named "mumble" (a pun on "lisp" I guess...) responsible for converting/adapting(?) the ".scm" code to Common Lisp.

<Googulator>But we already have a fully bootstrapped Scheme implementation - in fact, we have 2 of them (Mes and Guile)!

<pabs3>aha https://mumble.net/~jar/s48/index.html

<pabs3> https://mumble.net/~jar/tproject/index.html

<pabs3>see also https://www.s48.org/ which I didn't archive yet

<Googulator> http://www.cs.cmu.edu/afs/cs/project/ai-repository/ai/lang/scheme/impl/t/src/ has the source code for T version 3.1 - unfortunately it's a compiler that exclusively targets dead architectures

<Googulator>_and_ it's self-hosted too

<Googulator>looks like Lennart Augustsson is still active in the world of Haskell: https://www.youtube.com/watch?v=Zk5SJ79nOnA

<Googulator>(he's the original author of hbc)

<Googulator>this is also apparently him: https://github.com/augustss

<pabs3>the author of MicroHs

<Googulator>seems like he's also the author of hbc and one of the authors of lmlc...

<Googulator>someone from here (who is good at contacting outsiders, i.e. probably not me) should get in touch with him

<Googulator>see if he still has any of the old Haskell source code

<janus>i wrote Lennart an email about HBC in 2022 and i never got a reply

<janus>but i get replies for the issues i post about MicroHs and MicroCabal

<Googulator>archived this just in case: https://archive.org/details/ghc-history.tar

<Googulator>I was able to find some old directory listings of a now dead mirror of the Glasgow university FTP, and unfortunately it seems by the time that FTP died, the oldest Haskell it still had a copy of was 0.29 :(

<Googulator>so it's unlikely that any surviving mirror or archive would have pre-0.26 versions

<Googulator>OTOH for some good news, I was able to retrieve a lot of nhc development history, thanks to an old backup of York University's FTP on archive.org

<Googulator>we have nhc12 v0.8, nhc13 versions 971106, 971219, 980212, 980304, 980319, 980320, 980327, 980501, 980624, 980706 and 980501 - as well as C-transpiled code for the last 4 versions

<Googulator>plus, nhc98 versions 1.12, 1.14, 1.14a, 1.16, 1.18 and 1.20

<Googulator>(1.18 and 1.20, as well as 1.22, can also be found on haskell.org)

<Googulator>actually, we have more nhc98 versions in this archive - 1.0pre1 through 1.0pre19, 1.02, 1.04, 1.06, 1.08 and 1.10 are included too, they're just in a separate directory

<Googulator>plus 1.0pre7 and 1.0pre8 C-transpilation

<Googulator>and a previously lost version of lml / hbc, v0.998.3pp

<oriansj>and we have https://github.com/blynn/compiler.git (which is a haskell subset we can already bootstrap)

<oriansj>perhaps if MicroHs's blob was instead built by that, it might start another chain of interest?

<Guest84>"Hi, I wanted to know where the download-distfiles.sh script is obtaining its sources from. Could you clarify this?

<Googulator>It parses the `sources` files under steps/

<Guest84>where does he get the links from?

<oriansj>Guest84: do you mean where do we find the links to put in the file? or how the links are parsed out of the file?

<Guest84>yes

<oriansj>Guest84: please clarify

<Guest84>I want to know where the source files get the links from

<oriansj>Guest84: we pick upstream sources and use a mirror we have.

<Guest84>I want to automate and in many cases it gives problem that's why I want to get rid of those links any solution will you suggest

<oriansj>Guest84: well if your goal is automation, I suggest making your own mirror of the files and changing the links to point to your local mirror.

<Guest84>՝

<Guest84>And the rootfs.py link?

<aggi>after a year of absence i got back to tccboot for a moment today

<aggi>couldn't reproduce a working version of it last time i tried

<aggi>out of curiosity: anyone ever succeeded reproducing tccboot from _source_ since when it was first released year 2004?

<matrix_bridge><Andrius Štikonas> I don't think anybody did...

<pabs3>Googulator: btw, if you want anything saved to web.archive.org at the original URLs, let me know and I can get them saved via ArchiveBot. also they have a git archiving project (and there is also SWH for that)

<Googulator>Good to know - although the only one of what I'm archiving now with an "original URL" is GHC, which comes from https://gitlab.haskell.org/ghc/ghc.git

<Googulator>I've since also uploaded https://archive.org/details/nhc98-history.tar and https://archive.org/details/york-haskell-archive.tar - but these are repackages by me (nhc98-history is originally a Darcs repo which I converted to Git, while york-haskell-archive is extracted from a much larger FTP dump tarball)

<Googulator>BTW, just hit a potential gold mine: http://web.archive.org/web/19970220060949/http://www.cs.chalmers.se/pub/haskell/

<Googulator>turns out, Chalmers's FTP server also spoke HTTP, just under a different URL

<Googulator>and unlike the FTP version, this one got archived :)

<Googulator>unfortunately no older GHC or HBC releases than what we already have, but quite a few early NHCs :)

<civodul>Googulator: maybe you’ve already seen this, but if not, this may be relevant: https://elephly.net/posts/2017-01-09-bootstrapping-haskell-part-1.html

<Googulator>yes, that's where I started :)

<civodul>excellent :-)

<civodul>i hope you can find the next steps!

<civodul>it seems to be quite a journey

<Googulator>Just found Yale Haskell 2.2 sources! (http://web.archive.org/web/19970728140842/http://haskell.cs.yale.edu:80/haskell/yale/oldStuff/haskell-2.2-source.tar.gz - newer than 2.05, which was already known)

IRC channel logs

2024-06-07.log