- Notifications
You must be signed in to change notification settings - Fork1.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to ourterms of service andprivacy statement. We’ll occasionally send you account related emails.
Already on GitHub?Sign in to your account
MPR#7639: configure warnings on Mac OS X High Sierra (10.13)#1371
Conversation
The code in int64align.c is awful in that it reuses a type name |
2c3d1f2 to104031dCompareI have pushed a revised version which renames the type. This seems to operate correctly. (I think we're probably testing |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others.Learn more.
I think this is good. To do it really clean, you'd have to duplicate the detection ofstdint.h and the#ifdef logic fromconfig.h but it's not worth the trouble.
That will do just fine, thanks! Merging... |
MacOS High Sierra balks at the redefinition of `int64_t`.
Cherry-pick: [4.060fa7718] |
MacOS High Sierra balks at the redefinition of `int64_t`.
Another cherry-pick: [4.0523eb28c] |
MacOS High Sierra balks at the redefinition of `int64_t`.
MacOS High Sierra balks at the redefinition of `int64_t`.
It seems that on OS X 10.13 a
typedeffrom a header file comes into scope for theint64align.ctest (used duringconfigure) which definesint64_taslong long. We try to redefine it aslongand an error occurs. (This behaviour is only seen once the XCode tools have been installed.) I don't know why this didn't happen on previous versions---although the previous versions do still define it aslong longin the corresponding header file.Since we are planning to replace the
configureinfrastructure anyway I propose the fix in this pull request. It seems rather unlikely that Intel will introduce architectural constraints on integer alignment in the short term.There was also a semi-related bug in
configurewith a stray backslash.Incidentally, this is not related to MPR#7591 / GPR#1257. The alignment warnings from the linker there are a new introduction which I believe to be due to preparations for a new dynamic loader which forbids misaligned references into the data section. There are some more details here:http://docs.huihoo.com/apple/wwdc/2017/413_app_startup_time_past_present_and_future.pdf. For the moment I don'tthink we need to do anything (any problems with this will show up as linker warnings in any case).
Tested on OS X 10.13 including running the testsuite.
@xavierleroy If you dislike the fix in this pull request, my branch is available for you to push to.