PMXBOT Log file Viewer

Help | Karma | Search:

#pypa logs for Saturday the 1st of October, 2016

(Back to #pypa overview) (Back to channel listing) (Animate logs)
[15:48:20] <ricoalpha> hi team, got refered here from the Python group…
[15:48:32] <ricoalpha> I have multiple packages checked in one repository (repo1)
[15:48:33] <ricoalpha> repo1/packageA/setup.py
[15:48:34] <ricoalpha> repo1/packageB/setup.py
[15:48:36] <ricoalpha> repo1/packageC/setup.py
[15:48:37] <ricoalpha> I am trying to install a specific package (package B) by calling following command:
[15:48:39] <ricoalpha> pip install -e git+ssh://git@github.com/useranm/repo1.git#egg=packageB
[15:48:40] <ricoalpha> I am getting following error message:
[15:48:41] <ricoalpha> Command "python setup.py egg_info" failed with error code 1 in /Users/username/.Envs/repo1/src/packageB/
[15:48:42] <ricoalpha> what am I doing wrong?
[15:49:40] <dstufft> ricoalpha: think you need to add &subdirectory=packageB to the end of that
[15:50:09] <ricoalpha> ok, what does the #egg refer to in this case?
[16:47:20] <xafer> ricoalpha: it helps pip know what package each url corresponds to
[19:48:39] <ricoalpha> how do you manage development of packages e.g. (pip install -e ….) vs repo
[19:48:58] <nedbat> ricoalpha: what does that mean?
[19:49:50] <ricoalpha> lets say you want a new virtualenv…what do you do? do you do a 1) git clone from a URL 2) install -r requirements.txt
[19:50:10] <ricoalpha> or do you just call pip -r URL…?
[19:50:55] <nedbat> ricoalpha: i don't think "pip -r URL" works.
[19:51:09] <nedbat> ricoalpha: for development, you clone the repo, make a virtualenv, and install -r
[19:52:08] <ricoalpha> do you add a “ -e . “ in the requirements.txt ?
[19:52:21] <ricoalpha> to install the cloned repo in editable mode?
[19:54:06] <ricoalpha> not sure how to manage 2 local packages (repo) vs development vs production
[19:54:21] <ricoalpha> how to specify these in the requirements.txt file
[19:54:41] <jessamynsmith> ricoalpha: If I'm using a package not on pypi, yeah I install from url in requirements.txt
[19:55:08] <ricoalpha> and your local repos?
[19:55:09] <nedbat> ricoalpha: the reason to install the product code itself is to run tests, and often the test runner can handle that for you
[19:56:58] <ricoalpha> because when I add something like “-e .” and “-e ../repo1 “ pip freeze generates something like: -e git+ssh://git@github.com/username/project1.git@562ba58a6b98a29fe0914428c097f#egg=repo1&subdirectory=repo1
[20:03:22] <nedbat> ricoalpha: you don't have to use pip freeze to keep your requirements up to date. You can edit the file by hand
[20:03:58] <ricoalpha> is there anything I missing? I just cannot figure out what is the best practice…
[20:08:17] <ricoalpha> http://www.marinamele.com/2014/02/django-best-practices-i-different.html
[20:08:26] <ricoalpha> requirements/{common.txt,dev.txt,prod.txt,test.txt}
[20:08:43] <ricoalpha> anyone following this approach?
[20:13:48] <nedbat> ricoalpha: yes, that's common to have requirements based on the environment.
[22:21:39] <[Tritium]> ricoalpha: I have seen almost a matrix of requirements files by production environment (dev, stage, prod, test, etc), python version, and library version.
[22:22:52] <[Tritium]> granted the * versions were submatracies of test, not dev, stage, or prod
[22:22:56] <ricoalpha> it could get really complicated….
[22:23:02] <[Tritium]> tox
[22:23:50] <[Tritium]> (and theres a tool to compile your requirements files too, but ive seen it used mostly when people are running tox)
[22:24:04] <ricoalpha> I found something interesting here: https://caremad.io/posts/2013/07/setup-vs-requirement/
[22:24:08] <ricoalpha> --index-url https://pypi.python.org/simple/
[22:24:08] <ricoalpha> -e https://github.com/foo/bar.git#egg=bar
[22:24:09] <ricoalpha> -e .
[22:33:51] <bennofs> Is there any way to detect inside a setup.py file whether the current build is using wheels or not?
[22:34:37] <bennofs> because it seems that data_files behaves differently under wheels than under normal pip install
[22:52:39] <[Tritium]> is there a plan on what to use to populate the changelogs on warehouse?