PMXBOT Log file Viewer

Help | Karma | Search:

#pypa-dev logs for Saturday the 11th of April, 2015

(Back to #pypa-dev overview) (Back to channel listing) (Animate logs)
[05:07:34] <luzfcb> hello, there is anyone online?
[20:20:31] <wickman> what are the resolution semantics for nested requirements.txt files? e.g. if you have a requirements.txt with a '-r requirements2.txt', should the requirements2.txt inherit the resolver options from requirements.txt, or should it be equivalent to if you did 'pip install requirements2.txt' in isolation?
[20:21:29] <wickman> take for example --no-index and -f http://myrepo.example.com in requirements.txt, should requirements2.txt honor those even though it doesn't specify them?
[20:21:39] <wickman> this is unspecified in http://pip.readthedocs.org/en/latest/reference/pip_install.html#requirements-file-format
[20:27:30] <xafer> currently there is only one finder
[20:27:45] <xafer> which collects all the options found
[20:29:30] <wickman> seems like that's not really the desired behavior, right?
[20:29:56] <wickman> e.g. if some deeply nested requirements.txt does -i and redefines the index or --no-index or the like
[20:32:02] <xafer> yup :)
[20:32:15] <wickman> ok
[20:32:34] <wickman> so if i have an alternate implementation, it seems like i should do what's right and not what pip does? ;-)
[20:33:46] <xafer> it depends of what your goal is :p
[20:36:48] <xafer> but yes the current behavior is not ideal and I'd like to suggest some changes in a near future :)
[21:00:32] <wickman> great, thanks. more on this later.