- User Since
- Aug 15 2014, 7:40 AM (319 w, 5 d)
Mon, Sep 28
Fri, Sep 25
Thanks, @bruno. Modifying the existing logic is risky. If we have to do that, I think a new patch will be better. The purpose of this patch is to solve an ABI bug.
Tue, Sep 22
This looks pretty useful, thanks for adding this Volodymyr. LGTM.
Mon, Sep 21
Hi Richard. Thanks for the quick feedback and for the testcase!
I'm taking this over from Hongtao (w/ his blessing :)
Tue, Sep 1
Hi Volodymyr, this looks pretty interesting. Can you share screenshots or examples of how the output looks like?
Aug 31 2020
Yes the wrapper is definitely problematic. I'm checking how the diagnostic itself is handled correctly and start from there when I have time. For now would it be better to separate this into multiple patches and get the diagnostic improvement in first?
Aug 28 2020
Hi Andrew, thanks for improving this. I think this makes sense: dependents can choose to not use modules without having to trigger the build system to rebuild all dependencies. Can you add a simple testcase to prove the point of the change?
Aug 18 2020
Hi Wei, this looks handy! Minor stylish comments below.
Nice catch! LGTM
Neat unit test! LGTM (one minor suggestion below).
Hi Zixu, thanks for working on improving this.
LGTM as is, minor suggestion below.
Jul 14 2020
Apr 21 2020
Mar 9 2020
I also want to understand why you want this change. I'm not convinced this is the behavior we want and it's not like we encourage people to use header maps in the first place. I'd be fine with such behavior under a CC1 flag though. Falling back to the filesystem is something that pertains more to a VFS kinda thing than header maps.
Mar 3 2020
Mar 2 2020
Hi Michael, thanks for improving this.
Feb 27 2020
Feb 24 2020
Feb 19 2020
Thanks for working on this.
Feb 13 2020
Address Richard's review.
Thanks for taking a look Richard, comments inline.
Jan 31 2020
- Update patch after Volodymyr's review.
- Refactor bits done as part of rG90f58eaeff5f, update it to remove that part.
- Why are you adding ODR hash support for RecordDecl and not TagDecl? We already have support for EnumDecl, so TagDecl seems like a good candidate to cover both. Honestly, I don't know if it is possible or a good idea but it looks plausible enough to consider.
Jan 22 2020
Jan 10 2020
Updated the patch to refactor mode code.
Jan 9 2020
Jan 8 2020
Remove some FIXMEs that are now done.
Change the approach: handle type merging for tag types using the ODRHash mechanism