Assume gcc is at least 6.4, the oldest xtoolchain in the ports tree.
Assume clang is at least 6, which was in 11.2-RELEASE. Drop conditions
for older compilers.
Details
- tinderbox with in-tree toolchain
- amd64 buildworld/buildkernel with amd64-gcc xtoolchain (6.4.0)
- amd64 buildworld/buildkernel on 11.2-RELEASE
(There were unrelated failures in 1 and 2.)
Diff Detail
- Lint
Lint Passed - Unit
No Test Coverage - Build Status
Buildable 31019 Build 28725: arc lint + arc unit
Event Timeline
share/mk/bsd.sys.mk | ||
---|---|---|
74 | I just realized I can completely remove these three conditions. |
share/mk/bsd.sys.mk | ||
---|---|---|
223 | I'm glad you asked. I don't know. Are there other compilers capable of building FreeBSD? Intel released a FreeBSD version of their 2016 compiler. I intend to try building FreeBSD with it (in my copious spare time). |
Follow on work may include assuming c++11 or c++14, but this is a great start and should be committed as is, module units from the review.
share/mk/bsd.sys.mk | ||
---|---|---|
142 | We should note that we should clean up warnings produced with these flags. They were originally there as a quick hack to enable gcc5/6 | |
223 | We don't need it. Intel compiler support needs to be based on someone using it successfully. Changes to support it are likely more extensive than this one line... | |
stand/efi/boot1/Makefile | ||
46 | Do we still need this? | |
stand/efi/loader/Makefile | ||
36 | Ditto |
lib/libthr/Makefile | ||
---|---|---|
32–33 | Shouldn't this -Wno option be removed instead ? I do not even see this warning option (either with or without no-) in gcc 10.1 manual. |
lib/libthr/Makefile | ||
---|---|---|
32–33 | The test was not GCC-before-4.3 so I think the patch is what's intended? |
share/mk/bsd.sys.mk | ||
---|---|---|
90 | Can't this be removed since we only support armv6 and armv7 now? If so, that probably should be a separate commit though. |