On 30.06.2015 17:37, Mattia Rizzolo wrote: >> for the new python-requirements-detector I've got a UnicodeDecodeError in >> the tests against >> Python 3.4 [1] from reproducible builds CI [2]. That usually points to the >> locale set to some >> non-UTF-8 (like "C"), but which isn't the case here? > > Yes, the first build is not done in a unicode (like en_US.utf-8). > Actually the locale is not set at all. > > In fact I can get the very same error in my pbuilder on my notebook. I saw > somewhere (I don't really recall where/when) that something without any locale > setted picked up something near the standard C locale in situation like that. > > The second build (which your package didn't even reached) is run with > LC_ALL=fr_CH.UTF-8 instead.
I've found the environment settings on the CI page, LANG but not LC_ALL is set for the first build to some UTF-8. Could you help me where I've got that wrong: LANG provides the main locale while LC_ALL if set overrides all individual LC_ settings. LC_ALL needs to be set? Dan -- http://qa.debian.org/developer.php?login=debian%40danielstender.com 4096R/DF5182C8 46CB 1CA8 9EA3 B743 7676 1DB9 15E0 9AF4 DF51 82C8 -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
