| 3 Dec 2025 |
 helle (just a stray cat girl) | * btw, we have investigated the underlying issue and we know EXACTLY the potential starting points, but ooof, it is one hell of a chore | 20:04:24 |
 Puna | https://github.com/NixOS/nixpkgs/pull/460845
Unlike the previous version, all new versions are built entirely from source by using Hugs to interpret the microhs source code.
cpphs was also booted from source in a similar way.
The goal is to allow Nixpkgs users to more easily experiment with using microhs, and perhaps one day boot GHC from source.
dunno if they have the required spoons to push this to the end, but maybe it can get the ball rolling
| 20:12:46 |
| helle (just a stray cat girl) | sadly currently out of spoons to further explore, but will make a note | 20:15:15 |
 dish [Fox/It/She] | In reply to @aleksi:pikaviestin.fi
I see live-bootstrap people have also bootstrapped autotools to prevent relying on autogenerated configure scripts. I don't know if nixpkgs wants or needs that though that is a lot of work I would not want to do atm | 20:31:39 |
| dish [Fox/It/She] | considering you have to bootstrap diff versions of automake/autotools in a hyperspecific way and you have to do like 2 dozen vers to make actual progress | 20:32:25 |
| dish [Fox/It/She] | for now I don't mind pre-generated header files but that may be something we do later | 20:32:47 |
| *  Qyriad cries in rustc bootstrap | 20:33:01 |
| dish [Fox/It/She] | I would like to get the current pt 1 PR I made merged, then your improvements to i686/x64/riscv, then my statics, then migrate bootstrap-files to minimal-bootstrap for supported arches | 20:34:00 |
| dish [Fox/It/She] | last part is the most complicated and one I'd need the most help with since stdenv is... annoying | 20:34:27 |
| helle (just a stray cat girl) | I mean I may just end up maintaining some of these extremely long bootstraps, because I see value in them, but yeah, they are a mess and sloooow | 20:35:18 |
| dish [Fox/It/She] | In reply to @qyriad:katesiria.org
cries in rustc bootstrap oof yeah | 20:36:34 |
| dish [Fox/It/She] | mrustc needs to keep up for rustc bootstrap to be even partially viable | 20:36:55 |
| dish [Fox/It/She] | * mrustc needs to keep up for rustc bootstrap to be even somewhat viable | 20:37:08 |
| dish [Fox/It/She] | In reply to @helle:tacobelllabs.net
I mean I may just end up maintaining some of these extremely long bootstraps, because I see value in them, but yeah, they are a mess and sloooow oh yeah there is 100% value in it | 20:37:26 |
| dish [Fox/It/She] | but atm I would rather have an imperfect bootstrap be used than a perfect bootstrap that never gets merged | 20:37:53 |
| dish [Fox/It/She] | then we can improve it over time | 20:37:59 |
| dish [Fox/It/She] | ~~which would involve a lot of mass rebuilds but if we can do it in big sets and also maybe get a hydra jobset then its more reasonable~~ | 20:38:38 |
| helle (just a stray cat girl) | oh yeah, part of ours is intentionally not going to get merged any time soon over this | 20:39:01 |
 aleksi | Definitely agreed | 21:45:44 |
| dish [Fox/It/She] | question about your patches, is there anything in them that would be useful to improve i686 bootstrap? or is it just improvements for x86_64/rv64? | 21:57:00 |
| dish [Fox/It/She] | just trying to understand the scope of the patchset relative to if I want to take any changes into my existing upgrade branch | 21:57:14 |
 Alyssa Ross | We have runInLinuxVM! | 22:04:11 |
| aleksi | Hmm, yeah there might be little things: - GCC v8 upgraded to v10 (for better RV64 backend, mostly, but can be good for i686)
- bash 5.2 to 5.2.37, which fixes the strtoimax issue
- fixed egrep, the makefile from live-bootstrap is wrong
- fixed strncasecmp in heirloom stubs, the old implementation can segfault if you give it a string literal as an argument, which heirloom does
| 22:04:21 |
