darwin.stdenv: allow patchShebangs during the bootstrap
#242519
Conversation
|
Creating this as a draft PR while x86_64-darwin builds. I feel pretty confident that it will, but I don’t want to waste time on Hydra if it doesn’t. Once the stdenv is built, I’ll set the PR to ready. |
ofborg
bot
added
6.topic: darwin
This fixes pyicu (and any other package that uses `icu-config` instead of the CMake or some other module to get the build flags). What happened here is the bootstrap disables `patchShebangs` to avoid propagating the bootstrap tools to the final stdenv (due to `sh` and `bash` being on the `PATH` from the bootstrap tools). Because of that, the `#!/bin/sh` line in `icu-config` was not updated, causing it to invoke the system bash on Darwin. While that is undesirable in its own right, when the system bash is invoked as `sh`, `echo -n` will print `-n`, resulting in the breakage see in NixOS#241951 (comment). The fix is to build bash earlier in the bootstrap while making sure it is picked up over the one in the bootstrap tools. That allows `patchShebangs` to be enabled during the bootstrap. Any package with scripts that is included in the final stdenv should now have its scripts’ shebang lines properly patched.
reckenrode
force-pushed
the
stdenv-shebangs
branch
from
aad246c
to
856ebe6
Compare
|
Setting this to review. I had to build link bash against the system CoreFoundation to break a dependency cycle on x86_64-darwin. I was able to successfully build a stdenv on both x86_64-darwin and aarch64-darwin, and I was able to build pyicu on aarch64-darwin. I feel pretty confident pyicu will build on x86_64-darwin, but I don’t want to delay longer than necessary. |
|
The failure in |
|
To be clear, Hydra showed the failure the same on both |
|
Ah, I now caught up the confirmation for those questions in another thread: #242202 (comment) |
|
OK, the |
|
Sorry about the delayed response. I was asleep at the time these comments were posted.
If it’s due to icu, then I expect it to be fixed. The
I’m building it now and will report back. |
|
@vcunat Both texlive and R build for me with this commit. |
Description of changes
This fixes pyicu (and any other package that uses
icu-configinstead of the CMake or some other module to get the build flags).What happened here is the bootstrap disables
patchShebangsto avoid propagating the bootstrap tools to the final stdenv (due toshandbashbeing on thePATHfrom the bootstrap tools). Because of that, the#!/bin/shline inicu-configwas not updated, causing it to invoke the system bash on Darwin. While that is undesirable in its own right, when the system bash is invoked assh,echo -nwill print-n, resulting in the breakage see in #241951 (comment).The fix is to build bash earlier in the bootstrap while making sure it is picked up over the one in the bootstrap tools. That allows
patchShebangsto be enabled during the bootstrap. Any package with scripts that is included in the final stdenv should now have its scripts’ shebang lines properly patched.I tested that pyicu builds, and this is the contents of
icu-confignow:Things done
sandbox = trueset innix.conf? (See Nix manual)nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)