| 19 Jan 2026 |
|  @rutger.putter:realiz-it.nl joined the room. | 15:47:47 |
| @rutger.putter:realiz-it.nl left the room. | 16:05:52 |
| 23 Jan 2026 |
 nbp | https://bugzilla.mozilla.org/show_bug.cgi?id=2011902 Just in case people have similar issues, feel free to report back. | 10:25:29 |
| nbp | “In today's Nightly build, we changed the default distribution ID for Mozilla's official Firefox builds from an empty string to "mozilla-official" so that we could differentiate our official builds from Linux distros' repacks or unofficial builds.” | 10:26:42 |
| nbp | * context: “In today's Nightly build, we changed the default distribution ID for Mozilla's official Firefox builds from an empty string to "mozilla-official" so that we could differentiate our official builds from Linux distros' repacks or unofficial builds.” | 10:27:14 |
 hexa | distributionIni =
let
platform = if stdenv.hostPlatform.isDarwin then "Nix on MacOS" else "NixOS";
in
pkgs.writeText "distribution.ini" (
lib.generators.toINI { } {
# Some light branding indicating this build uses our distro preferences
Global = {
id = "nixos";
version = "1.0";
about = "${applicationName} for ${platform}";
};
Preferences = {
# These values are exposed through telemetry
"app.distributor" = "nixos";
"app.distributor.channel" = "nixpkgs";
};
}
);
| 10:47:21 |
| hexa | I think you mean that and what we have should rule out problems originating from that change | 10:48:04 |
 vcunat | Maybe our -bin packages get affected? | 10:48:27 |
| hexa | sure | 10:48:42 |
| hexa | but not sure if that is something we should worry about | 10:49:02 |
| nbp | I will forward the question, is that for binary builds as well? | 10:49:17 |
| hexa | no, source builds only | 10:50:00 |
| hexa | this is part of our buildMozillaMach function | 10:50:13 |
| nbp | Ok, so we should already have experienced the issue, if any, I presume. | 10:50:24 |
| hexa | https://mastransky.wordpress.com/2026/01/23/firefox-linux-in-2025/ | 14:49:02 |
| 27 Jan 2026 |
| nbp | https://bugzilla.mozilla.org/show_bug.cgi?id=2012175 🎉 | 17:12:44 |
| nbp | Do we use ICU from Nixpkgs or from Firefox sources? | 17:13:49 |
| hexa | from nixpkgs | 17:16:17 |
| 28 Jan 2026 |
| hexa | the only maintainable way I see is to patch icu in general | 20:57:11 |
| hexa | tracking these patches i going to be a huge hassle | 21:00:45 |
|  @oskardotglobal:matrix.org left the room. | 21:03:44 |
| hexa | maybe unvendoring isn't worth the hassle in this case | 21:06:18 |
| hexa | vcunat: wdyt | 21:06:44 |
| hexa | https://github.com/mozilla-firefox/firefox/tree/FIREFOX_147_0_2_RELEASE/intl/icu-patches | 21:07:10 |
| vcunat | The (unvendored) library is ~35 MiB. That's quite a nice chunk to share. | 21:08:34 |
| vcunat | But the hassle is annoying, yes. | 21:09:10 |
| vcunat | icu upstream is annoying, e.g. this patch... they haven't applied it, and it's been hanging there for years (!)
https://unicode-org.atlassian.net/browse/ICU-22132 | 21:10:28 |
| hexa | if we correctly applied patches on our unvendored icu, how much sharing would there be left? | 21:12:20 |
| vcunat | I suppose let's vendor it for now. | 21:13:14 |
| vcunat | Like, thunderbird uses its own icu build, though unvendored. | 21:13:55 |
