klimek added inline comments.
================
Comment at: lib/Format/Format.cpp:1920-1922
+ std::error_code EC = FS->makeAbsolute(Path);
+ assert(!EC);
+ (void)EC;
----------------
amaiorano wrote:
> klimek wrote:
> > I think if makeAbsolute doesn't work, we will probably want to err out here:
> > if (EC) {
> > llvm::errs() << ... << EC.message << ...
> > return Style;
> > }
> Okay. I based myself on the fact that the old code didn't do error checking
> on fs::make_absolute - in fact most calls to fs::make_absolute in the
> codebase assert on the result. I suppose it's because it's just a string
> manipulation function, and if it fails, the subsequent code that will make
> use of the path will fail with a more useful error.
>
> In any case, I don't mind handling the error here as you propose.
>
I don't think it's overly important in this case. Generally, I like a good
error message more than an assert that gives me no info on what happened...
https://reviews.llvm.org/D27971
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
