[PATCH] D65986: Allow setting the VFS to 'real' mode instead of default 'physical'

Fri Aug 9 14:00:43 PDT 2019

jfb marked an inline comment as done.
jfb added inline comments.

================
Comment at: clang/include/clang/Driver/Options.td:2012
+  HelpText<"Use the virtual file system in 'real' mode, or 'physical' mode. In 'real' mode the working directory is linked to the process' working directory. In 'physical' mode it has its own working directory, independent of (but initially equal to) that of the process.">,
+  Values<"real,physical">;
 def imultilib : Separate<["-"], "imultilib">, Group<gfortran_Group>;
----------------
JDevlieghere wrote:
> jfb wrote:
> > sammccall wrote:
> > > Mea culpa: naming these `physical`/`real` doesn't really explain either the implementation difference nor the observable effect.
> > > (I thought this internal name didn't matter much and didn't want to rename `getRealFilesystem`...)
> > > 
> > > Can we avoid propagating these bad names further? Maybe instead something like `-os-workdir` or `-workdir=os|virtual` or similar?
> > Maybe? What do other folks who actually tough VFS think? I'm happy to go either way, I was following what seemed to be the prevalent naming (and parroted the comments on those functions) but I understand that it might not have been naming that you thought would leave those function 🙂
> Personally I like the `-workdir=os|virtual` because  it better conveys the semantic meaning, rather than how it's implemented.
I renamed it. I think it's better to abandon this revision and go with Michael's suggestion, though.

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D65986/new/

https://reviews.llvm.org/D65986