autogen: fix patchelf invocation to remove generic build path from rpath #106758
Conversation
siscia
changed the title
ofborg
bot
added
10.rebuild-darwin: 0
10.rebuild-linux: 501+
10.rebuild-linux: 5001+
labels
|
LGTM. this resolved an issue I had with cross compiling to armv7 on a WSL+Debian+Nix build platform |
|
Had the same problem and this fixed it. I think we, who have to compile our systems (non-standard nix dir location) without sandboxing (work permission restrictions; no I can't use proot or the like; big brother doesn't allow user namespace in kernel) are, and forever will be, the red-headed step-children. |
|
/rebase staging |
There was a problem hiding this comment.
Thanks! This fixes the linked issue for me. Agree with Sandro that this should be rebased onto staging due to the quantity of linux rebuilds. Otherwise LGTM.
I think we, who have to compile our systems [...] are, and forever will be, the red-headed step-children.
While I certainly wouldn't expect first-class support, I think there are enough of us out there that if we stick up for each other and nip these issues in the bud, we'll have a pretty good user experience. From what I've seen, issues like this are rare and fairly easy to patch around.
| @@ -79,7 +79,7 @@ stdenv.mkDerivation rec { | |||
| '' + stdenv.lib.optionalString (!stdenv.hostPlatform.isDarwin) '' | |||
| # remove /build/** from RPATHs | |||
There was a problem hiding this comment.
| # remove /build/** from RPATHs | |
| # remove build directory (/build/**, or /tmp/nix-build-**) from RPATHs |
might make it a little more clear why this change was needed
|
Guys please don't expect me to work on this. If somebody want to take charge of this PR, please go ahead with any means, it just won't be me. |
|
Yeah no worries, thanks for sharing this patch. Continued in #126557 |
Refers to #106165
Motivation for this change
Things done
sandboxinnix.confon non-NixOS linux)nix-shell -p nixpkgs-review --run "nixpkgs-review wip"./result/bin/)nix path-info -Sbefore and after)