Closed belveder79 closed 4 years ago
and more toolchains...
Awesome, thx. Just a quick question: did you ever encounter that running a build for all devices, say armv7, armv7s and arm64, fails, while individual ones work out perfectly?
I ran the “device” toolchain, for which I get a compiler error. But running toolchains individually work out. It is not a memory issue, and I have no idea...
Ruslan Baratov notifications@github.com schrieb am So. 16. Feb. 2020 um 21:50:
Merged #341 https://github.com/ruslo/polly/pull/341 into master.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/ruslo/polly/pull/341?email_source=notifications&email_token=AAA3R4ECU7TQMMIKPJBZ5V3RDGRJFA5CNFSM4KUOFKJKYY3PNVWWK3TUL52HS4DFWZEXG43VMVCXMZLOORHG65DJMZUWGYLUNFXW5KTDN5WW2ZLOORPWSZGOWVHLJ5I#event-3041834229, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAA3R4FAGR7U6BUY7DFZNNDRDGRJFANCNFSM4KUOFKJA .
did you ever encounter that running a build for all devices, say armv7, armv7s and arm64, fails, while individual ones work out perfectly?
Yes, it happens sometimes. Not all CMake code is suitable for a multi architecture build, not all C++ code as well. E.g. try_compile
will run only for a single architecture, and if the results of this operation will be architecture-dependent, you will have errors while compiling the untested architecture. Same problem if you save the results generated by configure_file
C++ source.
Also, note that for a single architecture build we will set CMAKE_SYSTEM_PROCESSOR
, for the multi architecture we can't:
Examples:
Hi, added new toolchain without bitcode. Thx, clemens