| 28 Jan 2026 |
 dish [Fox/It/She] | but that is coming from my perspective so... grain of salt | 22:49:04 |
 whispers [& it/fae] | same, but :/ | 22:49:36 |
 Atemu | One issue that is not clear to me is sustanability | 22:54:20 |
| Atemu | Is the bootstrap compiler still getting updates? | 22:54:45 |
| Atemu | Or does the chain grow indefinitely? | 22:55:10 |
| Atemu | Because that would not scale well w.r.t. rebuilds | 22:55:37 |
| Atemu | This is the reason why we don't have rust source bootstrap either | 22:57:14 |
| dish [Fox/It/She] | In reply to @atemu12:matrix.org
Is the bootstrap compiler still getting updates? which one? go's bootstrap compiler is gccgo, so whatever that supports is supported by the bootstrap. Currently, with go the bootstrap chain would get 1 version longer per year, as go only increments their version reqs every other minor release, and each release is 6 months apart. Afaik there aren't any plans for them to change that or shrink the chain at a later time, besides gccgo actually getting developed to support later golang features to shrink our bootstrap chain | 23:07:56 |
| dish [Fox/It/She] | so to answer your question, the bootstrap chain does grow infinitely. The good thing wrt rust is that go is much faster to build, so we don't have to worry about compile times as much, and the chain grows much slower. But it will grow unmanagable at some point, and there's the thorny details of new architectures etc | 23:09:06 |
| dish [Fox/It/She] | the chain in my PR is gccgo -> 1.17 -> 1.20 -> 1.22 -> 1.24/25 | 23:10:32 |
| dish [Fox/It/She] | once 1.26 is released, 1.24 becomes part of its bootstrap chain | 23:10:46 |
| dish [Fox/It/She] | * once 1.26 is released, 1.24 becomes part of its bootstrap chain because of how go's bootstrap process works | 23:13:20 |
| dish [Fox/It/She] | https://go.dev/doc/install/source#go14 | 23:13:44 |
| dish [Fox/It/She] | * upstream docs are at https://go.dev/doc/install/source#go14 | 23:13:53 |
| Atemu | These details should be documented in the PR; this is critical information for deciding on this change! | 23:15:08 |
| Atemu | Also how long exactly are we talking? If it'd only scale to, like, an hour for the whole bootstrap in the next 10 years (given current velocity), that'd be a very different proposition to it scaling to multiple hours | 23:17:37 |
| dish [Fox/It/She] | checking compile time rn | 23:18:22 |
| dish [Fox/It/She] | uhh | 23:30:10 |
| dish [Fox/It/She] | hmm | 23:30:11 |
| dish [Fox/It/She] | im worried that the minimal-bootstrap PR just broke staging | 23:30:21 |
| dish [Fox/It/She] | 
Download 28-18-30-34.png | 23:30:51 |
| whispers [& it/fae] | fwiw, stdenv itself did manage to build fine on x86_64-linux https://hydra.nixos.org/build/320736662 | 23:31:43 |
| dish [Fox/It/She] | any ideas why my build is broken | 23:33:41 |
| dish [Fox/It/She] | I just rebased my branch to latest staging, will push to the above PR | 23:33:57 |
| whispers [& it/fae] | did a git fetch upstream refs/pull/435345/merge:go-bootstrap && jj new go-bootstrap && nom-build -A go in nixpkgs and gccgo is in the process of configuring just fine for me? | 23:39:49 |
| dish [Fox/It/She] | okay this does seem to be exclusive to my homelab for... reasons :think | 23:37:17 |
| dish [Fox/It/She] | * okay this does seem to be exclusive to my homelab for... reasons 🤔 | 23:37:23 |
| dish [Fox/It/She] | hmmmmmmm | 23:41:36 |
| dish [Fox/It/She] | very strange, checking my homelab | 23:41:42 |
| 29 Jan 2026 |
 rosssmyth | Wrote my thoughts about block on freebsd support in the Go PR | 00:41:52 |
