| 17 Mar 2025 |
 John Ericson | emily: wait is that new in this PR? | 16:15:18 |
 emily | no, I reported it here before | 16:16:34 |
| emily | the complicated splicing stuff that I don't understand why it's there is broken | 16:16:47 |
| John Ericson | Eelco: So the reason it is not a CanonPath is the symlink target and "location the source tree is for symlink resolution purposes" is an OS path | 16:17:02 |
| John Ericson | even more confusing, in the case of remote file systems it might be an OS path of another operating system | 16:17:33 |
| John Ericson | I am wary of canonicalizing because we don't want a situation where the symlink points outside the source tree, but then another symlink points back inside, to allow the symlink to be resolved | 16:18:33 |
| John Ericson | rather as soon as any intermediate encountered during symlink resolution leaves the source tree, it's game over | 16:19:30 |
| John Ericson | Eelco: Also happy with https://github.com/NixOS/nix/pull/12658 now? | 17:18:53 |
| emily | (I was just mentioning it wrt bumping nixVersions.latest) | 17:36:36 |
| emily | (and that more fixes will need backporting later) | 17:36:45 |
 aidetechbot | oh woops sorry | 22:57:37 |
| 18 Mar 2025 |
|  sinan changed their profile picture. | 06:43:31 |
 trofi | Any chance to merge https://github.com/NixOS/nix/pull/12105? it's 3 months old. | 07:27:11 |
 Robert Hensing (roberth) | It seems that an assertion is failing; see comment | 12:40:39 |
|  BigBabyThor joined the room. | 14:48:57 |
|  Schromp joined the room. | 15:56:47 |
| John Ericson | Las: I feel like we should merge my commits prior to my last one splitting the files, and then do your commits | 18:17:50 |
| John Ericson | Robert Hensing (roberth): have any opinions on that? | 18:18:01 |
 Las | In reply to @Ericson2314:matrix.org
Las: I feel like we should merge my commits prior to my last one splitting the files, and then do your commits From your PR or the restricted store one or? | 18:19:00 |
| John Ericson | Las: from the big one (though yes that includes restricted store) | 18:19:17 |
| Las | My PR has the benefit of being simple, but it’s also simpler to rebase possibly | 18:19:39 |
| Las | Maybe smaller PR then with your changes? Which would include what! | 18:19:57 |
| Las | * | 18:22:20 |
| John Ericson | Las: basically I am thinking my PR isn't churn tooo bad until the last commit | 18:30:58 |
| John Ericson | and I want to understand what state moves back and forth with my split classes | 18:31:20 |
| John Ericson | when you move the stuff to lambdas | 18:31:25 |
| Las | In reply to @Ericson2314:matrix.org
Las: basically I am thinking my PR isn't churn tooo bad until the last commit I’m honestly not sure I’m a fan of the prepareBuild thing. It seems like a change that moves from a local optimum to a worse directionally momentarily and then leaves it there. | 18:33:51 |
| Las | Is it really that much more work to just take it out of the class and return of struct of what it does? | 18:34:10 |
| John Ericson | it's not that I think anything is wrong with your PR, it is just the modularity is supposed to make it easier to understand these changes | 18:34:26 |
| John Ericson | the prepare/unprepare is supposed to be temporary | 18:35:06 |
