Details
Diff Detail
- Repository
- rL LLVM
Event Timeline
Shouldn't the fact that the make file exists be stored in a class variable and then the class will still call CleanMakefile() if the make file didn't exist?
I'm honestly not sure. This seems to be the way Sean liked it based on the other thread.
But I'm confused about why any of this would even be in class methods or class variables to begin with, as opposed to instance variables. If you have multiple inline tests, one would be affected by the run of another.
+Sean Callanan <scallanan@apple.com> in case he wants to weigh in. There
seems to be a difference of opinion on whether this is ok to go in as-is,
so hopefully we can get it figured out.
Bump. I'm going to be out all of next week and I don't want to leave with
any pending CLs if at all possible. Is this ok to go in as is?
Yea, and that's how the patch is implemented now. But I think what Greg is
saying is that he thought it should check for an existing Makefile first.
If it exists, make a copy of it, then generate the makefile, then after
everything is done restore the old Makefile. Do you think that's
important? You wrote most of the lldbinline stuff it looks like, so
ultimately if you're ok with it then I am too, but just want to make sure
we're all on the same page.
Going to put this in today then if there are no further issues. This gives
me time to revert today if there are still issues / disagreements, and if
anything comes up next week feel free to revert then since I will be out
all week.