Ivan Ponomarev
Ivan Ponomarev
It seems like you use a bogus toolchain and (or) cmake. Please try this: ``` mkdir build cd build ${ANDROID_HOME}/cmake/3.6.4111459/bin/cmake \ -Wno-dev \ -DCMAKE_SYSTEM_NAME=Android \ -DANDROID_PLATFORM=android-21 \ -DANDROID_ABI=arm64-v8a \ -DANDROID_NDK=${ANDROID_NDK_ROOT}...
It seems like you tried to build it for desktop operating system without a proper Android toolchain. Please try something like that: ``` mkdir build cd build ${ANDROID_HOME}/cmake/3.6.4111459/bin/cmake \ -Wno-dev...
Looks like a problem was in SSH submodules remote paths. I switched all submodule repositories to HTTPS, please try again.
Now there is only one way to do this: to make a fork and fix it in code. Please take a look at ndcrash_dump_write_line function.
Script is successfully run on my machine with MacOS X Sierra, Android NDK r16b, all architectures. But i ran only "full" build and missed a check for "lite" build. Unfortunately...
> How we can change sysroot only at compile time? It's quite tricky and, i think, it doesn't make sense at all. Adding -isystem flag with sysroot/include path is absolutely...
The code that uses addr2line to get a source file and line is broken: See https://github.com/boostorg/stacktrace/blob/develop/include/boost/stacktrace/detail/addr2line_impls.hpp#L209 We pass **addr_** field as an argument to addr2line executable. **addr_** is a virtual...
> how did it make it into the Boost release? :-( I don't know exactly. But I suspect that it's happened because boost stacktrace uses GNU libbacktrace from gcc to...
atos may help https://www.unix.com/man-page/osx/1/atos/
> https://prnt.sc/t63ve8 Is it a screenshot of APK file contents? I suspect no. In order to find a reason of this problem please inspect contents of output APK file, make...