build, qt: Fix cross-compiling detection on M1-based macOS (aarch64) #22487
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Member
Author
jarolrod
reviewed
Jul 19, 2021
| @@ -0,0 +1,16 @@ | |||
| To cross compile for the x86_64-darwin host on aarch64-darwin, | |||
| passing QMAKE_APPLE_DEVICE_ARCHS=arm64 to the configure script | |||
| does not work for some reasons. | |||
Contributor
There was a problem hiding this comment.
I would prefer if we could document the reason it doesn't work. And any possible information on how/when/what circumstances this patch could be dropped
Member
Author
There was a problem hiding this comment.
I would prefer if we could document the reason it doesn't work.
I failed to track this bug down, unfortunately.
And any possible information on how/when/what circumstances this patch could be dropped
As this is an ad-hoc hack, we should try to drop it while bumping Qt version in the future.
Member
Author
There was a problem hiding this comment.
I think the #23583 is the correct solution for this problem.
jarolrod
reviewed
Jul 19, 2021
Contributor
There was a problem hiding this comment.
Concept ACK, thanks for fixing this. One concern: https://github.com/bitcoin/bitcoin/pull/22487/files#r671973286
Tested on an M1 Machine. Confirming that with this change we can drop the recommendation to install rosseta.
master:
a) native depends:
Configure summary:
Building on: macx-clang (x86_64, CPU features: cx16 mmx sse sse2 sse3 ssse3 sse4.1)
Building for: macx-clang (arm64, CPU features: neon crc32)
Target compiler: clang (Apple) 12.0.5
Configuration: cross_compile largefile neon precompile_header silent release c++11 c++14 c++1z reduce_exports static stl
b) building for x64:
Configure summary:
Build type: macx-clang (x86_64, CPU features: cx16 mmx sse sse2 sse3 ssse3 sse4.1)
Compiler: clang (Apple) 12.0.5
Configuration: sse2 aesni sse3 ssse3 sse4_1 sse4_2 avx avx2 avx512f avx512bw avx512cd avx512dq avx512er avx512ifma avx512pf avx512vbmi avx512vl f16c largefile precompile_header rdrnd shani silent x86SimdAlways release c++11 c++14 c++1z reduce_exports static stl
pr:
a) native depends:
Configure summary:
Build type: macx-clang (arm64, CPU features: neon crc32)
Compiler: clang (Apple) 12.0.5
Configuration: largefile neon precompile_header silent release c++11 c++14 c++1z reduce_exports static stl
b) building for x64:
Configure summary:
Building on: macx-clang (arm64, CPU features: neon crc32)
Building for: macx-clang (x86_64, CPU features: cx16 mmx sse sse2 sse3 ssse3 sse4.1)
Target compiler: clang (Apple) 12.0.5
Configuration: cross_compile sse2 aesni sse3 ssse3 sse4_1 sse4_2 avx avx2 avx512f avx512bw avx512cd avx512dq avx512er avx512ifma avx512pf avx512vbmi avx512vl f16c largefile precompile_header rdrnd shani silent x86SimdAlways release c++11 c++14 c++1z reduce_exports static stl
While here
I noticed the following line in the output when building qt through depends, how much of a concern can this be?:
> main.cpp:5:12: fatal error: 'linux/input.h' file not found
> # include <linux/input.h>
> ^~~~~~~~~~~~~~~
> 1 error generated.
> make[1]: *** [main.o] Error 1
test config.qtbase_gui.tests.evdev FAILED
Member
It's not a concern. Output like that is expected while Qt is running it's feature checks. |
promag
reviewed
Jul 19, 2021
Contributor
promag
left a comment
There was a problem hiding this comment.
Tested ACK 5a55ba9.
However, I tried the following (hardcoded) and it worked fine:
$(package)_config_opts_darwin += -device-option QMAKE_APPLE_DEVICE_ARCHS=x86_64
make -C depends qt_configured HOST=x86_64-apple-darwin18
$(package)_config_opts_darwin += -device-option QMAKE_APPLE_DEVICE_ARCHS=arm64
make -C depends qt_configured
Note that it doesn't use _config_opts_x86_64_darwin or _config_opts_aarch64_darwin.
Member
Author
What is your Qt Configure summary? UPDATE:
The |
fanquake
reviewed
Jul 20, 2021
Member
fanquake
left a comment
There was a problem hiding this comment.
Does configuring with qt, for a x86-64 cross-compile, on an M1, using this PR, currently work?
Member
Author
|
Updated 5a55ba9 -> 8edac79 (pr22487.01 -> pr22487.02):
|
Member
Author
Yes, it does, after the recent update. |
Member
Author
|
Rebased 8edac79 -> d660e80 (pr22487.02 -> pr22487.03) due to the conflict with #22526. |
This was referenced Jul 26, 2021
Contributor
|
The following sections might be updated with supplementary metadata relevant to reviewers and maintainers. ConflictsReviewers, this pull request conflicts with the following ones:
If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first. |
This reverts commit cdb41d5. It is no longer required because of the previous commit.
Member
Author
|
Rebased d660e80 -> 85888b5 (pr22487.03 -> pr22487.04) due to the conflict with #22448. |
Contributor
|
GUIX hashes |
Member
Author
Guix builds: |
This was referenced Nov 13, 2021
Closed
Member
|
I'm closing this for now. In favour of #23583. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
5 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
#20641 made building of the
qtpackage in depends more structured, and, particularly, fixed cross-compiling detection on Intel-based macOS (x86_64).However, a similar issue still present on M1-based macOS (arm64):
Note an unexpected
cross_compile.Note that the build system failed to detect cross compiling.
In both cases building the configured
qtpackage fails without installed Rosetta 2 (see #22402).This PR:
With this PR:
This PR is based on #22506.