| 2 Sep 2025 |
 hexa | or rather in fetch-cargo-vendor-util.py | 01:16:14 |
| hexa | * or rather in fetch-cargo-vendor-util.py:267 | 01:16:17 |
| hexa | cc Toma | 01:16:34 |
| hexa | so when is that supposed to be replaced? | 01:19:05 |
| hexa | # cargo config
mkdir -p .cargo
cat $cargoDeps/.cargo/config.toml >> .cargo/config.toml
ln -s $cargoDeps @vendor@
| 01:24:39 |
| hexa | wow, that symlink to ./@vendor@ is something | 01:25:13 |
 Toma | In reply to @hexa:lossy.network
# cargo config
mkdir -p .cargo
cat $cargoDeps/.cargo/config.toml >> .cargo/config.toml
ln -s $cargoDeps @vendor@
Where is this logic? (can't use github search rn)
Seems pretty insane
@vendor@ that is inside pkg.cargoDeps is replaced with an actual name in cargoSetupHook's postUnpack part | 04:55:57 |
| Toma | But it is only replaced in the real /build/.cargo/config.toml, but not the original /build/.cargo/foo-vendor/.cargo/config.toml | 05:04:43 |
| Toma | * But it is only replaced in the real /build/.cargo/config.toml, but not the original /build/.cargo/foo-vendor/.cargo/config.toml (aka. $cargoDepsCopy/.cargo/config.toml) | 05:17:07 |
| hexa | that is in pkgs/by-name/af/affine/package.nix | 11:06:18 |
| hexa | ok, but for some reason the cargo metadata call that mozjs_sys is doing chokes on the leftover @vendor@ | 11:07:31 |
|  QuadRadical (Ping) left the room. | 13:01:40 |
| hexa | ? failed to read root of directory source: /build/servo-0-unstable-2025-09-01-vendor/@vendor@ | 19:17:32 |
| hexa | *
failed to read root of directory source: /build/servo-0-unstable-2025-09-01-vendor/@vendor@
| 19:17:33 |
| hexa | Toma: I mean … it clearly states that is looking in what I assume to be $cargoDeps | 19:18:06 |
| hexa | do you have an idea how we can get that resolved? | 19:18:29 |
|  @aidalgol:tchncs.de joined the room. | 21:54:17 |
| hexa | * ok, but for some reason the cargo metadata call that mozjs_sys is running chokes on the leftover @vendor@ | 22:29:36 |
| 3 Sep 2025 |
|  ghpzin joined the room. | 07:54:42 |
| 4 Sep 2025 |
 wonrax | anyone using crane know why my binary doesn’t include file names and line numbers in backtrace frames (backtrace-rs)? when i build “normally” with cargo on the host, they show up as expected | 05:02:50 |
 Charles | Even in release mode? | 05:11:31 |
| wonrax | yes | 05:18:33 |
| wonrax | crane builds package in release mode by default I think | 05:18:50 |
| wonrax | at least in the craneLib.buildPackage | 05:19:59 |
| Charles | I'm asking because I'm not sure if that info is expected to be included in release builds in the first place | 05:35:10 |
 Alyssa Ross | It is not | 05:59:29 |
| wonrax | i have debug = true set in profile.release in Cargo.toml | 06:00:10 |
| wonrax | i would expect building with nix produce the binary that bahaves the same like running cargo build --release | 06:01:53 |
| wonrax | i would expect building with crane produce the binary that bahaves the same like running `cargo build --release` | 06:04:38 |
| wonrax | btw I created a discussion about this
https://github.com/ipetkov/crane/discussions/892 | 06:06:25 |
