| 15 Nov 2021 |
 @nixinator:nixos.dev | yeah, but what if doesn't know it knows | 18:31:31 |
| @nixinator:nixos.dev | because it a older version of git. | 18:31:44 |
 balsoft | We don't need to pass it anything | 18:35:28 |
| balsoft | Just let it handle the default branch by itself. | 18:35:36 |
| balsoft | If it knows how to handle non-master branches, let it. If it doesn't, then it would fail anyways. | 18:35:51 |
| 16 Nov 2021 |
 anderscs | Who is working on the nix <-> IPFS integration? I think this would be really cool, but there seems to be several issues. For example the nix store isn't entirely content addressed atm, but it can be turned on. I think this is also an opportunity to optimize the nix store cachability and flexibility. | 15:47:18 |
 [0x4A6F] | Maybe someone in nix#859? | 17:26:36 |
| [0x4A6F] | * Maybe someone in nix#859? | 17:27:25 |
 fzakaria (Old) | Anyone here familiar with LeakSanitizer ?
I am trying to run it on Nix and working through the errors.
I would expect the fix for some is adding missing deletes but its still showing up for some extern global variables. | 18:30:01 |
 andi- | What parts of Nix? There are some expected leaks (within the parser mostly) and it perhaps will be confused by the GC to some degree. | 18:39:34 |
| andi- | I remember building the system daemon with memory sanitiser at some point and that made it terminate almost instantly after startup. | 18:40:09 |
 Moritz Hedtke | In reply to @fzakaria:matrix.org
Anyone here familiar with LeakSanitizer ?
I am trying to run it on Nix and working through the errors.
I would expect the fix for some is adding missing deletes but its still showing up for some extern global variables. See https://matrix.to/#/!VRULIdgoKmKPzJZzjj:nixos.org/$RczAfZSEWi7kX04NYHSdm2q3hzf1iS63CNEGSc6FEpc?via=nixos.org&via=matrix.org&via=nixos.dev the ast is intentinally leaked | 19:32:37 |
| Moritz Hedtke | Around that is also some more probably useful information | 19:33:03 |
|  @mewp:nurupo.pl joined the room. | 20:27:37 |
| @mewp:nurupo.pl | hello, is this the right place to ask how to contribute a particular thing to nix? (not nixos/nixpkgs) | 20:28:53 |
 tomberek | mewp: what are you thinking about? | 20:54:13 |
| @mewp:nurupo.pl | well, I really wanted nix to have haskell's $ operator, so I added one. Now I want to know whether that makes sense to others, and how to go about finding that out | 20:55:00 |
| @mewp:nurupo.pl | I could just make a PR, I don't know whether it's the best place to start, though | 20:55:44 |
| tomberek | PR sounds okay... if it needs to be an RFC i guess people will let you know. Sounds interesting. | 20:56:53 |
| @mewp:nurupo.pl | this is a really simple patch, but it is a language change, so I expect it to be somewhat controversial | 20:57:38 |
| @mewp:nurupo.pl | (just because it does something to the language, even if it doesn't break compatibility etc.) | 20:57:59 |
| tomberek | The rule has been not that current Nix expressions can be parsed by legacy versions of the Nix evaluator; just that the current evaluator can eval old expressions. | 20:59:18 |
| @mewp:nurupo.pl | well, okay then, I posted the PR, we'll see what happens next | 21:20:31 |
 toonn | $ is a little controversial even in Haskell circles. | 21:31:40 |
| @mewp:nurupo.pl | why is that? | 21:32:14 |
 sterni (he/him) | mewp: unlikely this would be added because it is a quite considerable backwards incompatibility | 21:34:46 |
| @mewp:nurupo.pl | how so? | 21:34:57 |
| sterni (he/him) | or at least I see that problem | 21:34:59 |
| sterni (he/him) | well older nix versions wouldn't be able to evaluate any code using $ then | 21:35:17 |
| sterni (he/him) | also there are other consumers of nix code like rnix-parser, hnix which also would need to be adjusted | 21:35:44 |
