| 17 Apr 2025 |
 Mic92 | John Ericson: I suppose it depends on the someone restacking pull requests once something is merged manually? | 18:07:42 |
 John Ericson | Mic92: just was told it is a client-side-only thing, so yes I think you are right | 18:08:22 |
| Mic92 | I don't think anyone is opposed if you would use that. It can probably still be combined with the CI statuses from dpulls | 18:10:51 |
| 18 Apr 2025 |
|  void joined the room. | 02:16:31 |
|  Winter joined the room. | 08:57:54 |
| Winter | anyone know if the docs to build locally w/ Meson are still accurate? the prefix env stuff doesn’t seem to work properly — --prefix gets properly passed to meson setup but ninja’s install phase is still trying to install into $PREFIX | 08:58:53 |
 allrealmsoflife | In reply to @winter:catgirl.cloud
anyone know if the docs to build locally w/ Meson are still accurate? the prefix env stuff doesn’t seem to work properly — --prefix gets properly passed to meson setup but ninja’s install phase is still trying to install into $PREFIX Are you trying to build Nix from source? Which docs are you following, the official ones building from source dont say much, and last time I failed building Nix from source too. | 09:10:44 |
 trofi | New json-3.12.0 exposed a bug in how nix deserializes integers: https://github.com/NixOS/nix/issues/13046 | 15:54:09 |
| John Ericson | thanks trofi | 16:59:23 |
| John Ericson | I didn't know we had a $PREFIX defined? | 16:59:54 |
| Winter | In reply to @Ericson2314:matrix.org
I didn't know we had a $PREFIX defined? in the default nix-shell $PREFIX points to nix-util’s $out ;) | 17:00:16 |
| John Ericson | Winter: oh hmm that might be an artifact of how we made the dev shell? | 17:01:27 |
| Winter | yeah. but i guess that begs the question: has nobody tested the dev shell/does nobody use it? | 17:02:23 |
| Winter | like should a shell hook be added to just unset $prefix/$out? that seems simplest but shrug | 17:02:53 |
| Winter | or dontSetPrefix=true ig | 17:03:09 |
| John Ericson | Winter: I don't think we do much installing from the dev shell | 19:21:05 |
| Winter | fair, guess you just use the bins | 19:21:24 |
| John Ericson | what made you want to do that? | 19:21:24 |
| John Ericson | (though I do recall runnning installPhase and it worked with nix develop) | 19:21:40 |
| John Ericson | trofi: gotta say the nlohmann people don't seem super clear about this | 19:22:19 |
| Winter | nothing in particular | 19:22:33 |
| Winter | was just linearly following the docs | 19:22:33 |
| John Ericson | e.g. is a positive integer always parsed as an unsigned number? | 19:22:39 |
| John Ericson | do you need to try both just to be sure? | 19:22:44 |
| John Ericson | maybe the docs need to be changed? | 19:23:07 |
| John Ericson | before you had to install for the install tests | 19:23:12 |
| John Ericson | with since Meson you don't need t | 19:23:18 |
| John Ericson | * with since Meson you don't need to | 19:23:19 |
| Winter | In reply to @Ericson2314:matrix.org
maybe the docs need to be changed? might be best to remove the install section, yeah | 19:32:59 |
| trofi | Yeah. AFAIU the https://github.com/nlohmann/json/issues/4475 as get_ref() can expose mutable references it's unsafe to return signed reference to unsigned underlying type (breaks -fstrict-aliasing assumptions). It was allowed by accident. And AFAIU whether to use signed or unsigned internally is dependent on actual json data. 0 is unsigned while -1 is signed. | 19:46:48 |
