Closed svenstaro closed 9 years ago
Thanks. Apparently Github failed to notify me about this PR :/
For various reasons I reworked your patches. The result is in commit ac31f03644cf4b38c45d1b4c525f984cd279e066. I don't want to use io.open in python2, because the _setup stuff is supposed to work with older versions as well (also, it's subtree'd now).
Can you give the current HEAD a shot? If it works properly, I'll use the same changes for the other projects (like rcssmin), too.
Works for me. The latin-1 does confuse me a bit, though.
Thanks.
Those latin-1 parts modifiy epydoc output. This is just the failsafe way to do it.
Looks resolved to me then!
Great. Thanks again.
Can you make a release for this so people on PyPI get it too? Last release is also quite long ago.
Yes, as soon as I have some free cycles. The last release happened so long ago, because nothing changed so far :-)
You have UTF-8 data that you read during configuration but it might be read as ASCII on systems that use LC_ALL="C". This is the case with configuration scripts such as ansible. My patch is a compatible way of opening UTF-8 as UTF-8 at all times. And yes, this is an actual error for us otherwise.