| 24 Jun 2025 |
 sielicki | I'm not super committed to finishing it, but I've been toying around lately with replacing the nix lexer and parser with https://lexy.foonathan.net/ lately | 22:21:47 |
| sielicki | * I'm not super committed to finishing it, but I've been toying around lately with replacing the nix lexer and parser with https://lexy.foonathan.net/ | 22:21:53 |
| sielicki | the value here is primarily in parsing speed (see lexy benchmarks: https://lexy.foonathan.net/benchmark_json/ ) but also that it would be simple to have a unified parser between nix and json | 22:23:30 |
| sielicki | * the value here is primarily in parsing speed (see lexy benchmarks: https://lexy.foonathan.net/benchmark_json/ ) but also that it would be simple to have a unified parser between nix and json, and any other parsing that you should want nix to be capable of understanding | 22:24:08 |
| sielicki | it's trivial to write allocator-aware lexy parsers and/or parsers which defer subexpression parsing. ie: imagine a parser which, in its first pass, scans exclusively for path references and import statements, and aggressively schedules that io work before coming back to parse everything else. | 22:30:39 |
| sielicki | en.cppreference.com/w/cpp/execution can also play a role there | 22:31:33 |
| 25 Jun 2025 |
 jade_ | we use a different parser of similar vibes in lix | 05:13:34 |
| sielicki | what's lix? | 05:13:55 |
| jade_ | uhh, the other major nix implementation, https://lix.systems | 05:14:07 |
| jade_ | ours is using the pegtl library | 05:15:01 |
| sielicki | oh very cool, yeah this looks awesome | 05:18:46 |
|  gkak.laₛ changed their display name from gkaklas to gkak.laₛ. | 17:50:38 |
| 26 Jun 2025 |
|  mdaniels5757 joined the room. | 01:29:04 |
| mdaniels5757 | It looks like the security patch for 2.26.3 didn't make it to nixpkgs release-24.11. Should I add the patch for 2.26.3 first, or would merging and manually backporting https://github.com/NixOS/nixpkgs/pull/419839 be easier? @Mic92 | 01:32:51 |
 Mic92 | it would be | 05:33:17 |
| 27 Jun 2025 |
|  R̴̨͕͇͍̞̮̐̅͆̌̀̉̐͋̈́̃̀͒́̎̅̚̚̚͠͝Ĕ̵̡̛͖͖̟̙̫̱͈̘̞̭͍͍͑̌̄͑̓̋̓̀̈̏̈́͊̇͊͆̉͂̏̀̃̚͘͝͝ͅͅD̶̡̢͔̱̖̮͙͉̘̺͓͍̩̮͈͍͗̃̀̏͌͘͜ͅŚ̸̬̭̯̬͙͇͓̬̩̳̤͚͓̤̩̺͉͖̉͛̓̿̎͊̿̆́̐͂̇͌̄̇̓͘ͅͅT̴̞̫̘̝͇͔̟̪̪̦͂̔̎̀̎ͅŎ̷̡̬̹̪͈̭̣͈̭̭͉̦̖̝̘̪͖͔̥̦̘̻̳Ṋ̶̛̫͈̳̘͚̜̔̋͆̅̈́͊̑͊̉̌̈́̾͑̈́̚ͅË̸̡̨̨̛͇̜̖͔͖̻̟̗̠̙͓̘̗̥͉͇̜͑͆͊͑͑̀̓͒͜͝͝ changed their display name from Redstone to R̴̨͕͇͍̞̮̐̅͆̌̀̉̐͋̈́̃̀͒́̎̅̚̚̚͠͝Ĕ̵̡̛͖͖̟̙̫̱͈̘̞̭͍͍͑̌̄͑̓̋̓̀̈̏̈́͊̇͊͆̉͂̏̀̃̚͘͝͝ͅͅD̶̡̢͔̱̖̮͙͉̘̺͓͍̩̮͈͍͗̃̀̏͌͘͜ͅŚ̸̬̭̯̬͙͇͓̬̩̳̤͚͓̤̩̺͉͖̉͛̓̿̎͊̿̆́̐͂̇͌̄̇̓͘ͅͅT̴̞̫̘̝͇͔̟̪̪̦͂̔̎̀̎ͅŎ̷̡̬̹̪͈̭̣͈̭̭͉̦̖̝̘̪͖͔̥̦̘̻̳Ṋ̶̛̫͈̳̘͚̜̔̋͆̅̈́͊̑͊̉̌̈́̾͑̈́̚ͅË̸̡̨̨̛͇̜̖͔͖̻̟̗̠̙͓̘̗̥͉͇̜͑͆͊͑͑̀̓͒͜͝͝. | 00:54:19 |
| R̴̨͕͇͍̞̮̐̅͆̌̀̉̐͋̈́̃̀͒́̎̅̚̚̚͠͝Ĕ̵̡̛͖͖̟̙̫̱͈̘̞̭͍͍͑̌̄͑̓̋̓̀̈̏̈́͊̇͊͆̉͂̏̀̃̚͘͝͝ͅͅD̶̡̢͔̱̖̮͙͉̘̺͓͍̩̮͈͍͗̃̀̏͌͘͜ͅŚ̸̬̭̯̬͙͇͓̬̩̳̤͚͓̤̩̺͉͖̉͛̓̿̎͊̿̆́̐͂̇͌̄̇̓͘ͅͅT̴̞̫̘̝͇͔̟̪̪̦͂̔̎̀̎ͅŎ̷̡̬̹̪͈̭̣͈̭̭͉̦̖̝̘̪͖͔̥̦̘̻̳Ṋ̶̛̫͈̳̘͚̜̔̋͆̅̈́͊̑͊̉̌̈́̾͑̈́̚ͅË̸̡̨̨̛͇̜̖͔͖̻̟̗̠̙͓̘̗̥͉͇̜͑͆͊͑͑̀̓͒͜͝͝ changed their profile picture. | 00:55:42 |
 tomberek | How intrusive would it be? And can it express the various quirks Nix has? | 17:29:23 |
 magic_rb | Would anyone else be interested in git-annex support? Id like to pair with someone, never did any nix work before. It could even maybe be an out of tree plugin? Essentially just needs to scan all annexed symlinks and then resolve them and put them into the store, or refuse eval if some annexed file is unavailable | 17:39:55 |
| 30 Jun 2025 |
|  sinan changed their profile picture. | 02:11:25 |
|  Marc Thieme joined the room. | 11:05:59 |
| Mic92 | how would this be used with nix? | 15:38:37 |
| 1 Jul 2025 |
 fzakaria | oh man | 03:05:34 |
| fzakaria | replacing bison/yacc would be such a big development boost i think; | 03:05:44 |
 djacu | Hey Nix Team
In case you haven't seen the recent post on discourse, the Marketing Team is preparing this year's community survey. I am reaching out to teams to see if there are any questions they would like to add to the survey to better serve the work you all do. More details in the post linked below.
https://discourse.nixos.org/t/community-feedback-requested-2025-nix-community-survey-planning/66155 | 03:27:54 |
 roberth | This would be similar to LFS support, iiuc, which is already implemented | 11:19:06 |
| magic_rb | In reply to @joerg:thalheim.io
how would this be used with nix? Yeah, as robert said, with the current lack of support, if you have a repository with annex you end up with a bunch of dangling symlinks | 11:26:49 |
| magic_rb | In reply to @roberthensing:matrix.org
This would be similar to LFS support, iiuc, which is already implemented Its important to not resolve the symlink but just repoint them to the store. But yes, if base the code on git lfs | 11:27:18 |
| roberth | magic_rb: Ohh the keys can be arbitrary and dependent on the annex backends? Seems like it will be hard to be sure that it's reproducible then - impossible if we want fully general support for it | 11:32:12 |
| magic_rb | In reply to @roberthensing:matrix.org
magic_rb: Ohh the keys can be arbitrary and dependent on the annex backends? Seems like it will be hard to be sure that it's reproducible then - impossible if we want fully general support for it Wdym keys, so the way it works is that there are symlinks checked into git which point into .git/annex/... and the keys there are sha256 hashes of the annexed file content always. So the backend doesnt matter. | 11:39:55 |
