Two build fixes for chromium

[kraj]

first commit to fix type is good to go. But I am not sure about second commit, we explicitly changed to use libc++ for native and target portions of chromium build. So I think this patch should no longer be needed.

[schnitzeltony]

But I am not sure about second commit, we explicitly changed to use libc++ for native and target portions of chromium build. So I think this patch should no longer be needed.

Am on very current master on all layers and got error mentioned in additional patch comment. With second patch added, build passes.

[kraj]

But I am not sure about second commit, we explicitly changed to use libc++ for native and target portions of chromium build. So I think this patch should no longer be needed.

Am on very current master on all layers and got error mentioned in additional patch comment. With second patch added, build passes.

master of meta-browser I guess, but which release are you on for oe-core/meta-clang

[schnitzeltony]

All layers are latest master / meta-clang's top commit is 45a3707291a4e8bd2469b0d2bb9cf9e07878ae08 (the latest I can get)

[rwmacleod]

This builds and boots well for me on poky/master.

[kraj]

See https://bugs.chromium.org/p/chromium/issues/detail?id=1189788#c14 and https://bugs.chromium.org/p/chromium/issues/detail?id=1189788#c15. perhaps we should use these options if we are not doing so. Secondly for @schnitzeltony to check if there are any of these options in play when building chromium in failing case

[zboszor]

I took this PR for a ride and I can confirm that it fixes the build issues for me. I haven't tested running chromium-x11 yet but I suspect I won't have any problems with it on x86-64.

[kraj]

@schnitzeltony can you divide this pull into two pieces please ? the typo fix is obvious and we should apply it immediately, I am not too sure about the second fix so lets create a separate pull for that and try to understand why this will still be needed

[schnitzeltony]

Thanks Khem for taking care of first patch.

After recent upgrade to 92.0.4515.131 build fails on other places with same libstdc++ related error as described in second patch.

So I looked around a bit and found in log.do_compile many lines as:

clang-13: warning: argument unused during compilation: '-stdlib=libc++' [-Wunused-command-line-argument]

Maybe we are not doing what we intend to do?

[schnitzeltony]

Interesting:

bitbake -e | grep ^BASE_DEFAULT_DEPS BASE_DEFAULT_DEPS="virtual/aarch64-mortsgna-linux-gcc virtual/aarch64-mortsgna-linux-compilerlibs virtual/libc" BASE_DEFAULT_DEPS:toolchain-clang=" clang-cross-aarch64 virtual/libc libgcc virtual/aarch64-mortsgna-linux-compilerlibs " BASE_DEFAULT_DEPS:toolchain-clang:class-target=" clang-cross-aarch64 virtual/libc libgcc virtual/aarch64-mortsgna-linux-compilerlibs "

looking into clang.bbclas I'd expect at least compiler-rt libcxx. Suggestions?

[schnitzeltony]

Aargh: Forgot chromium-x11:

bitbake -e chromium-x11 | grep ^BASE_DEFAULT_DEPS BASE_DEFAULT_DEPS=" clang-cross-aarch64 virtual/libc compiler-rt libcxx libcxx" BASE_DEFAULT_DEPS:toolchain-clang=" clang-cross-aarch64 virtual/libc compiler-rt libcxx libcxx" BASE_DEFAULT_DEPS:toolchain-clang:class-target=" clang-cross-aarch64 virtual/libc compiler-rt libcxx libcxx"

Look more as expected. So still why all these unused during compilation: '-stdlib=libc++' warnings?

[kraj]

Thanks Khem for taking care of first patch.

After recent upgrade to 92.0.4515.131 build fails on other places with same libstdc++ related error as described in second patch.

So I looked around a bit and found in log.do_compile many lines as:

clang-13: warning: argument unused during compilation: '-stdlib=libc++' [-Wunused-command-line-argument]

Maybe we are not doing what we intend to do?

are you using distcc by any chance ? you could be hitting https://github.com/distcc/distcc/issues/329

[schnitzeltony]

are you using distcc by any chance ? you could be hitting

No - and off topic read mortsgna from back to front...

[kraj]

are you using distcc by any chance ? you could be hitting

No - and off topic read mortsgna from back to front...

OMG !! I thought iti was some brewery name :)