drop some non-ASCII characters from HISTORY.rst #20
Conversation
This file was using some odd unicode character instead of a normal apostrophe in a couple of places. Replace one occurrence with a normal apostrophe and remove the other (the possessive "its" does not contain an apostrophe).
|
How the heck did that get in there? Anyhows, thanks for spotting it! |
drop some non-ASCII characters from HISTORY.rst
|
No problem. I see you put me in AUTHORS.rst as 'AdamWill' - you can credit me as Adam Williamson, awilliam AT redhat DOT com , but this really seems too trivial to merit a credit to be honest! |
|
I'm pretty fanatical in my belief that all work should be credited properly. :) |
|
@AdamWill thanks for fixing this! @pydanny any chance you could cut a release to PyPI containing this fix? I know it's really trivial and not this project's fault, but the pip bug prevents us from installing your package on any of our servers :( It's the same issue as in https://github.com/kennethreitz/requests/pull/2420 thanks! |
We can't install ocflib on the wheezy servers anymore because of a bug in pip that prevents packages from being installed that have undecodable bytes (via the current locale, which in Puppet is "C"). This was fixed in pydanny/cached-property#20 but no release has been made yet. This can be reverted once a release is made, or when we're fully on jessie (which installs cached-property via a Debian package).
|
I'll get a release out this evening or tomorrow. 😄 |
|
Just release 1.3.0 "Bold Bottle": https://github.com/pydanny/cached-property/releases/tag/1.3.0 |
This reverts commit 23a4377. @pydanny released a new version (1.3.0) of cached-property: pydanny/cached-property#20 (comment) Also closing rt#4160: https://rt.ocf.berkeley.edu/Ticket/Display.html?id=4160
|
@pydanny thanks for the quick release! <3 |
This file was using some odd unicode character instead of a normal apostrophe in a couple of places. Replace one occurrence with a normal apostrophe and remove the other (the possessive "its" does not contain an apostrophe).
This could cause a problem when
setup.pydoes this:In Python 3, when you do
open().read(), a unicode decode occurs. The 'default' encoding used in this case depends on details of system configuration, for instance the locale. In some cases, it can be ASCII - and so when the file contains non-ASCII characters, this causes a traceback. (This was happening in the Fedora buildsystem; presumably our buildbots don't have a UTF-8 locale).We could do this:
but that's not valid in Python 2,
open()does not have theencodingargument there. So at that point I figured it was easier to just find the non-ASCII characters and fix those than try and figure out a fix for reading non-ASCII chars that's safe for both Pythons.