[PATCH] D68418: [test] Remove locale dependency for mri-utf8.test

[Thomas Preud'homme via Phabricator via llvm-commits]

thopre added a comment.

In D68418#1694539 <https://reviews.llvm.org/D68418#1694539>, @thopre wrote:

> In D68418#1694439 <https://reviews.llvm.org/D68418#1694439>, @gbreynoo wrote:
>
> > You are correct that the locale is required to pass on linux. I had some trouble with this test as the behaviour of python in this area differs between linux / windows and python 2 / python 3. For example this fix appears to be fine for linux, however Windows with python 2 fails:
> >
> >   Traceback (most recent call last):
> >     File "<string>", line 1, in <module>
> >   IOError: [Errno 2] No such file or directory: '\xc2\xa3.txt'
> >
> >
> > I do not like the reliance on the locale however the upstream buildbots all appear to have it installed. Maybe the test should be split into a windows test and a linux test?
>
>
> Mmmh, I need a Windows system to try more then. I guess the current code will output the pound sign into whatever encoding Windows use (which I guess is not UTF-8 but then how does the echo few lines above creates the file correct). Thanks for reverting the commit and sorry for the breakage.


So it's lit that processes the redirection and I guess the file will be read as a UTF-8, decoded into unicode and then on Windows this will output a filename in UTF-16.

Does llvm-ar espects UTF-8 specifically in archive members or can it be anything (e.g. UTF-16)?


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D68418/new/

https://reviews.llvm.org/D68418