[vcpkg] Enhancement android triplet#21847
[vcpkg] Enhancement android triplet#21847holmesconan wants to merge 38 commits intomicrosoft:masterfrom holmesconan:enhancement-android-triplet
Conversation
|
See also #21831 |
|
Too many PRs on |
|
In any case, I don't want to repeat PR comments. |
I'm building those successfully with an older version branched off #15605 Actually the complete gdal stack + some qt based libs with precompiled qt libraries. |
I understand this ("branched") as: |
|
I didn't specifically patch anything for cmake base ports. The cumbersome ones were always the others. If I had to guess I'd say they are fine already but never put this on test. |
|
The build or install log files should reflect the android API level and the chosen C++ library. |
|
Here's a geos build log produced with #21831 , default community triplet |
…CTED_CMAKE_(C/CXX)_FLAGS_* instead
Sorry for that. Since @m-kuhn did not mention my issue, I don't know he/she has already made a PR. But after testint that PR, I thought this one should be the primary solution to work with... |
I have read the |
IIUC this is independent from the rest of the work in here. |
| if(VCPKG_DETECTED_${cmake_var}) | ||
| get_filename_component(cmd_PATH "${VCPKG_DETECTED_${cmake_var}}" DIRECTORY) | ||
| get_filename_component(cmd_PROGRAM "${VCPKG_DETECTED_${cmake_var}}" NAME) | ||
| set(ENV{${env_name}} "${cmd_PROGRAM}") |
There was a problem hiding this comment.
Why not set ENV{${env_name}} to the full path?
There was a problem hiding this comment.
It is just some kind of OCD problem, :-(. Just make the log looks good instead of run well.
There was a problem hiding this comment.
IMO changing PATH invites side effects. Program name lookup after updating the path will behave different. Apart from the PATH becoming longer and longer. And this isn't even visible in most logs.
@dg0yt @holmescn @JackBoosY Given the recent increased communication in this PR I wonder if I should close mine. I am happy to help here too. Opinions? |
|
Can you please merge to master and resolve the file conflict? |
|
@jwtowner the new PR won't append the ANDROID_NATIVE_API_LEVEL when it is there, could you please test it? |
I can confirm it works. Thanks. |
|
Successfully built with r21e |
|
@holmescn , FYI, when trying to build using this PR against Android NDK r23 (23.1.7779620), the following error is thrown while configuring libiconv: |
|
Ping @holmescn for reply. |
@nairaner Beside the overlay port, you need to comment |
|
I need to see resolutions for the comments I made before this is marked info:reviewed; to be clear, that means:
|
|
Ping for response. |
|
Ping @holmescn again. |
|
#25009 implements a core part of this PR and answers the question about the OSX exception raised here https://github.com/microsoft/vcpkg/pull/21847/files?w=1#r776056939 while #25201 tackles the root problem of the OSX exception. |
|
Convert this PR to draft since there is no progress. |
vcpkg_configure_makefor Android newest NDK.$ANDROID_PLATFORMset in$VCPKG_ROOT/scripts/toolchains/android.cmakeby parse the Android NDK revision file.What does your PR fix?
Fixes New Android NDK support request #21345 [android] Android NDK r23 and clang support #20844 [icu:arm-android] build failure #18941
Which triplets are supported/not supported? Have you updated the CI baseline?
android, No
Does your PR follow the maintainer guide?
Yes