On Fri, Oct 26, 2018 at 1:24 PM Dylan Baker <[email protected]> wrote: > > This is bad for a couple of reasons, but the worst is that it gets the > shell involved. When the shell gets involved we can start running into > problems with LANG, namely LANG=C. This is particularly obnoxious for > translation files, since there is a very high likelyhood of running into > unicode in them. If we write it in python through file.write we don't > have this problem as python just shovels bits into a file, and the shell > doesn't know, and thus doesn't care.
Not a comment on the implementation, but rather on the commit message -- The shell couldn't care less about any of this, and is entirely uninvolved in the problem. Python 3's stdout/stderr abstractions are what's upsetting things. Instead of taking bytes (since stdout/stderr, and in fact all files, are byte-oriented things deep down inside), they are taking unicode codepoints. This used to work fine in py2, but then it got "fixed". Cheers, -ilia _______________________________________________ mesa-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/mesa-dev
