Skip to content
This repository has been archived by the owner on Aug 8, 2023. It is now read-only.

[build] Update CircleCI jobs to Xcode 10.0 #13184

Merged
merged 3 commits into from
Nov 2, 2018

Conversation

friedbunny
Copy link
Contributor

/cc @mapbox/maps-ios

@friedbunny friedbunny added iOS Mapbox Maps SDK for iOS build macOS Mapbox Maps SDK for macOS Node.js node-mapbox-gl-native labels Oct 24, 2018
@friedbunny friedbunny self-assigned this Oct 24, 2018
@friedbunny friedbunny requested a review from julianrex October 24, 2018 20:21
@friedbunny friedbunny force-pushed the fb-xcode-10-will-be-the-death-of-me branch from 9ba9e0c to 017c3c1 Compare October 26, 2018 18:06
@friedbunny
Copy link
Contributor Author

friedbunny commented Oct 26, 2018

Tests are exceeding even the 10 second timeout on CI:

https://circleci.com/gh/mapbox/mapbox-gl-native/183935

MBCameraTransitionTests
  testFlyToCameraInDelegateMethod, Asynchronous wait failed: Exceeded timeout of 10 seconds, with unfulfilled expectations: "Map view should finish loading style.”.

https://circleci.com/gh/mapbox/mapbox-gl-native/184371

MBGLIntegrationTests
  testOpenGLLayerDoesNotLeakWhenMapViewDeallocs, Asynchronous wait failed: Exceeded timeout of 10 seconds, with unfulfilled expectations: "Map view should be rendered".
  /Users/distiller/project/platform/ios/Integration Tests/MGLMapViewIntegrationTest.m:143

    [super waitForExpectations:expectations timeout:seconds];
    [timer invalidate];

@friedbunny
Copy link
Contributor Author

@julianrex do you see any alternative to indefinitely bumping timeouts? Flappy tests are (to my mind) worse than no tests, so I’d vote for disabling timing-dependent tests entirely on CI if we can’t resolve this issue.

@julianrex
Copy link
Contributor

Agreed - let's disable until we can come up with a longer term solution, that doesn't require us to constantly tweak fragile settings!

@friedbunny
Copy link
Contributor Author

Per chat with @julianrex we’re going to stop running the Integration Tests target on CircleCI for now.

@friedbunny friedbunny force-pushed the fb-xcode-10-will-be-the-death-of-me branch from 8dd886c to e26a032 Compare November 2, 2018 17:37
@friedbunny
Copy link
Contributor Author

Rebased and ready for review.

@friedbunny
Copy link
Contributor Author

Some persistent, strange, and seemingly unrelated failures with two of the gcc jobs (node-gcc6-debug, linux-gcc5-debug-coverage) after rebasing onto master at 3b9618e:

[116/379] Building CXX object CMakeFiles/mbgl-core.dir/src/mbgl/renderer/layers/render_symbol_layer.cpp.o
FAILED: CMakeFiles/mbgl-core.dir/src/mbgl/renderer/layers/render_symbol_layer.cpp.o 
/src/build/linux-x86_64/Debug/launch-cxx /usr/lib/ccache/g++-6   -DMBGL_USE_GLES2=1 -DMBGL_WITH_EGL=1 -DRAPIDJSON_HAS_STDSTRING=1 -D_GLIBCXX_USE_CXX11_ABI=1 -I../../../include -I../../../src -I../../../mason_packages/linux-x86_64/swiftshader/2018-05-31/include -I../../../platform/default -I../../../platform/linux -I../../../mason_packages/linux-x86_64/libpng/1.6.25/include/libpng16 -I../../../mason_packages/linux-x86_64/libjpeg-turbo/1.5.0/include -isystem ../../../vendor/earcut.hpp/include -isystem ../../../vendor/expected/include -isystem ../../../vendor/geojson-vt-cpp/include -isystem ../../../vendor/kdbush.hpp/include -isystem ../../../vendor/shelf-pack-cpp/include -isystem ../../../vendor/supercluster.hpp/include -isystem ../../../vendor/unique_resource -isystem ../../../vendor/wagyu/include -isystem ../../../vendor/boost/include -isystem ../../../vendor/geojson.hpp/include -isystem ../../../vendor/geometry.hpp/include -isystem ../../../vendor/icu/include -isystem ../../../vendor/polylabel/include -isystem ../../../vendor/protozero/include -isystem ../../../vendor/rapidjson/include -isystem ../../../vendor/variant/include -isystem ../../../vendor/vector-tile/include -isystem ../../../vendor/nunicode/include -Werror -ftemplate-depth=1024 -Wall -Wextra -Wshadow -Wnon-virtual-dtor -Wno-variadic-macros -Wno-unknown-pragmas -fext-numeric-literals -g -fPIC -fvisibility=hidden -fvisibility-inlines-hidden   -DCI_BUILD=1 -std=c++14 -MMD -MT CMakeFiles/mbgl-core.dir/src/mbgl/renderer/layers/render_symbol_layer.cpp.o -MF CMakeFiles/mbgl-core.dir/src/mbgl/renderer/layers/render_symbol_layer.cpp.o.d -o CMakeFiles/mbgl-core.dir/src/mbgl/renderer/layers/render_symbol_layer.cpp.o -c ../../../src/mbgl/renderer/layers/render_symbol_layer.cpp
g++-6: internal compiler error: Killed (program cc1plus)
Please submit a full bug report,
with preprocessed source if appropriate.
See <file:///usr/share/doc/gcc-6/README.Bugs> for instructions.
[117/379] Building CXX object CMakeFiles/mbgl-core.dir/src/mbgl/renderer/sources/render_geojson_source.cpp.o
[118/379] Building CXX object CMakeFiles/mbgl-core.dir/src/mbgl/renderer/layers/render_line_layer.cpp.o
[119/379] Building CXX object CMakeFiles/mbgl-core.dir/src/mbgl/renderer/sources/render_image_source.cpp.o
ninja: build stopped: subcommand failed.
Makefile:367: recipe for target 'node-all' failed
make: *** [node-all] Error 1
Exited with code 2

/cc @tmpsantos

Copy link
Contributor

@julianrex julianrex left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Changes here look good (no idea what's happening with the build failures though)

@friedbunny friedbunny force-pushed the fb-xcode-10-will-be-the-death-of-me branch from e26a032 to 947af30 Compare November 2, 2018 21:49
@friedbunny friedbunny merged commit 947af30 into master Nov 2, 2018
@friedbunny friedbunny deleted the fb-xcode-10-will-be-the-death-of-me branch November 2, 2018 22:27
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
build iOS Mapbox Maps SDK for iOS macOS Mapbox Maps SDK for macOS Node.js node-mapbox-gl-native
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Build ios-integration-test flaky for ios-debug-xcode10
2 participants