| 14 Apr 2023 |
|  pbsds joined the room. | 22:19:15 |
| 17 Apr 2023 |
 Growpotkin | sounds fine to me | 14:22:19 |
 infinisil | @room: The next meeting will take place in ~5 minutes - meeting link - live stream - meeting notes | 14:24:10 |
| infinisil | Quick update for here as well: Meetings will be bi-weekly now and on Tuesday instead of Monday, same time though. The first bi-weekly one is in two weeks, 2023-05-02 | 15:30:46 |
| infinisil | Robert Hensing (roberth): Regarding RFC 140, I'd like to discuss Custom Arguments again | 20:10:59 |
| infinisil | We introduced that to make it so that when there's a need for a custom argument to callPackage, one wouldn't have to switch away from using pkgs/unit, instead you could just add a definition to all-packages.nix with that argument. But that's just one possible exception, there's many others, like python3Packages.callPackage, callPackages, etc. | 20:13:28 |
| infinisil | The main problem I have is that a nice property is lost like this: One cannot just look at pkgs/unit/he/hello/package.nix and be sure that that's how pkgs.hello is defined. Because now there's this hidden component of "What if it's overridden with custom arguments in all-packages.nix" | 20:15:55 |
| infinisil | I should probably just make this a comment in the RFC | 20:20:46 |
| infinisil | See my comments here: https://github.com/NixOS/rfcs/pull/140#pullrequestreview-1389030207 | 23:09:12 |
| infinisil | We have at least 3 non-NAT shepherds now thanks to piegames! The steering committee should meet tomorrow and officially announce the shepherd team. Shortly after that (if the shepherd leader hasn't yet) I'll work to find a time to meet with the shepherds to move this along :) | 23:22:56 |
| 18 Apr 2023 |
 Robert Hensing (roberth) | In reply to @infinisil:matrix.org
We introduced that to make it so that when there's a need for a custom argument to callPackage, one wouldn't have to switch away from using pkgs/unit, instead you could just add a definition to all-packages.nix with that argument. But that's just one possible exception, there's many others, like python3Packages.callPackage, callPackages, etc. The difference is those exceptions don't appear and disappear overnight. callPackage arguments do. | 15:39:50 |
| Robert Hensing (roberth) | more detail here https://github.com/NixOS/rfcs/pull/140#discussion_r1170241659 | 15:49:41 |
| Robert Hensing (roberth) | we've already been over this, so I'd like to keep it short | 15:49:58 |
| infinisil | Robert Hensing (roberth): take a look at https://github.com/NixOS/rfcs/pull/140#discussion_r1169340835 | 15:51:58 |
| Robert Hensing (roberth) | infinisil did I miss a negation somewhere? Maybe clarify that callPackage foo { } is only a default that can be overridden in all-packages.nix? | 16:02:12 |
| Robert Hensing (roberth) | or maybe amjoseph missed a negation somewhere | 16:03:49 |
| Robert Hensing (roberth) | discussions about changes to negated exceptions are too complicated for me | 16:04:23 |
 @piegames:matrix.org | Does anybody have the rights to mark some of the old discussions as "resolved"? | 16:09:28 |
| @piegames:matrix.org | infinisil: I must agree with Roberth here that since the last change you are sending mixed signals on the currently discussed point, and I'm unsure as to what's your current position on this. | 16:11:05 |
| @piegames:matrix.org | I think that we all actually agree and there is just a communication issue, but I'm unsure | 16:11:36 |
| infinisil | In reply to @roberthensing:matrix.org
infinisil did I miss a negation somewhere? Maybe clarify that callPackage foo { } is only a default that can be overridden in all-packages.nix? No there's no missing negation. I'm proposing a change to the RFC, I pushed those change to the branch, and am highlighting what changed with the review comments. Previously we had an exception that allowed all-packages.nix to do callPackage foo { /* non-empty */ }, but I'm proposing to remove that exception | 16:43:35 |
| infinisil | In reply to @piegames:matrix.org
Does anybody have the rights to mark some of the old discussions as "resolved"? ✋ | 16:44:02 |
| @piegames:matrix.org | In reply to @infinisil:matrix.org
No there's no missing negation. I'm proposing a change to the RFC, I pushed those change to the branch, and am highlighting what changed with the review comments. Previously we had an exception that allowed all-packages.nix to do callPackage foo { /* non-empty */ }, but I'm proposing to remove that exception How does this qualify as "exception"? Exception to what rule? I think potentially having arguments in a callPackage call is standard? | 17:39:26 |
| infinisil | piegames: It's an exception that allows arguments to package functions to not be pkgs.${name} | 18:42:34 |
| infinisil | And I think that would be a really nice property, for users to just know that from { hello }: that hello points to pkgs.hello and not something else | 18:43:19 |
| infinisil | But I concur that custom arguments are something we need right now, there's no good alternative to it yet | 18:43:35 |
| infinisil | I'll re-add this to the RFC again, I think it's good we discussed this again :) | 18:44:57 |
| 20 Apr 2023 |
|  @federicodschonborn:matrix.org changed their display name from Federico Damián Schonborn to Federico Schonborn. | 01:02:33 |
| 21 Apr 2023 |
 Hari | Redacted or Malformed Event | 01:57:12 |
| Hari | Redacted or Malformed Event | 01:57:26 |
