| 22 Feb 2025 |
 alexfmpe | I guess that's usable | 22:59:21 |
| alexfmpe | I did think about shoving 'jailbreak' in passthru, but that only lets me read, not write right? | 23:00:28 |
| alexfmpe | doCheck works because the one I can set is the same mkDerivation will read | 23:00:49 |
| alexfmpe | passthru just 'mirrors' | 23:00:59 |
| alexfmpe | There's no pass-by-reference | 23:01:10 |
| alexfmpe | In reply to @alexfmpe:matrix.org
I guess that's usable As in, few false positivirs | 23:01:54 |
| alexfmpe | * As in, few false positives | 23:02:02 |
 Alex | Well you can override mkDerivation to completely ignore doJailbreak too. It's not entirely clear to me what you're trying to do. | 23:02:13 |
| alexfmpe | In reply to @alexfmpe:matrix.org
As in, few false positives Vanishingly few if I just sed out the jailbreak-cabal line | 23:02:44 |
| alexfmpe | In reply to @alex:tunstall.xyz
Well you can override mkDerivation to completely ignore doJailbreak too. It's not entirely clear to me what you're trying to do. Brute force discovery of needless jailbreaks | 23:03:07 |
| alexfmpe | Find all jailbreaked ones, unjailbreak, attempt build of all, report the ones that worked | 23:03:38 |
| alexfmpe | Technically it can't report the ones whose deps failed, but doing it one at a time takes an eternity of eval | 23:04:23 |
| alexfmpe | * Technically it can't report the ones whose deps failed to build due to that change, but doing it one at a time takes an eternity of eval | 23:04:42 |
| alexfmpe | In reply to @alex:tunstall.xyz
Well you can override mkDerivation to completely ignore doJailbreak too. It's not entirely clear to me what you're trying to do. I guess I could do this for a one off, but the idea is having a fire and forget script one can run whenever | 23:06:50 |
| alexfmpe | Rather than some patch to mkDerivation or generic-builder | 23:07:52 |
| alexfmpe | This is why I said 'from the outside' | 23:08:08 |
| alexfmpe | eh, I guess you could still have a .patch file and apply it from the script | 23:08:32 |
| alexfmpe | Think I can hack a 'getter' and 'setter' for jailbreak by doing invasive surgery to postPatch | 23:08:53 |
| alexfmpe | That lets me use the same overall shape I did for dontCheck hunting | 23:09:11 |
| Alex | In reply to @alexfmpe:matrix.org
Rather than some patch to mkDerivation or generic-builder No patching is needed, it can be done like you'd override any other package in the Haskell set. | 23:53:12 |
| alexfmpe | not sure I follow | 23:54:08 |
| alexfmpe | how would I unset doJailbreak by overriding mkDerivation | 23:55:00 |
| alexfmpe | btw this is how I ended up detecting: v: lib.hasInfix "jailbreak-cabal" v.postPatch; | 23:55:39 |
| alexfmpe | stringly hack, but a decent first approximation | 23:56:07 |
| alexfmpe | also, I only want to do this for the ones for which it is set, otherwise mass rebuild | 23:57:16 |
| 23 Feb 2025 |
| alexfmpe | oh I see what you mean, override the jailbreak argument itself | 00:09:10 |
| alexfmpe | ok that's working | 00:26:13 |
| alexfmpe | get = v: lib.hasInfix "jailbreak-cabal" v.postPatch;
set = x: v: v.overrideAttrs(_: _: { jailbreak = x; });
reset = set false;
| 00:26:23 |
| alexfmpe | ok so it's building...511 packages, ouch
300 of which are amazonka ones | 00:28:58 |
| alexfmpe | disabled their common jailbreak for this one-off run, but probably adding an explicit exception for them to the script given the volume | 00:35:29 |
