apparently some test somewhere likes to say hello many times but is speechless today 😂 :nixos: :
"> ERROR: test1 failed! Could not find the word 'hello' 26 times in the sample data."
Today's bug is a bug in `i686-linux` bootstrap binaries in `nixpkgs`.
There `gcc` fails to build as:
> preFixupLibGccPhase
> stat: Value too large for defined data type
Guess what is the cause of the build failure. What has changed in the environment recently?
Spolier: https://trofi.github.io/posts/297-32-bit-file-API-strikes-back.html
@misty @orhun @tshepang awesome. I am about to merge the update for #nixpkgs - https://github.com/NixOS/nixpkgs/pull/252391
🎉
As I got rights on #nixpkgs i also started to get hundreds of notifications every day from that repository.
My current approach here is to just delete notifications as they roll in, if I don't care. From bigger stuff I unsubscribe, like staging stuff I am not involved with for example. Smaller updates I run through nixpkgs-review and review the diff. And I merge them of course.
@musicmatze
Ask that greek guy who many others mix up with Sysiphos about that. And about rice.
My #nixpkgs merge count doubled in the last week. I bet I cannot keep up this rate of change!
Today's bug is a `nixpkgs` bug in the `buildFHSEnv` builder which did not provide enough environment for `ld` wrapper to look up libraries in default `/lib` and `/usr/lib` locations.
As a result an attempt to build some projects (like `gcc` itself) that rely on direct `ld` calls did fail to link.
For comparison `gcc` wrapper did have enough plumbing to Just Work for `/lib` and `/usr/lib`.
Proposed the fix as https://github.com/NixOS/nixpkgs/pull/250500
#nixpkgs disabled direct pushes. After all these years!!!
https://github.com/NixOS/rfcs/pull/156#issuecomment-1677358274
I am almost crying right now.
#CCCamp23 learning: software projects are more eager to get into #nixpkgs rather than getting for example #AlpineLinux packages
#cccamp23 #nixpkgs #alpinelinux