Python packages and egg-info directories

不想你离开。 提交于 2019-11-29 20:06:44

The .egg-info directories get only created if --single-version-externally-managed was used to install the egg. "Normally", installing an egg would create a single directory (or zip file), containing both the code and the metadata.

pkg_resources (which is the library that reads the metadata) has a function require which can be used to request a specific version of the package. For "old-style", regular imports, easy_install hacks a .pth file to get the egg directory onto sys.path. For --single-version-externally-managed, this hacking is not necessary, because there will only be a single version installed (by the system's pacakging infrastructure, e.g. rpm or dpkg). The egg-info is still included, for applications that use require (or any of the other pkg_resources binding mechanisms).

If you want to install a package by hard-linking, I recommend to use "setup.py develop". This is a command from setuptools which doesn't actually install the egg, but makes it available site-wide. To do so, it creates an egg-link file so that pkg_resources can find it, and it manipulates a .pth file, so that regular import can find it.

易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!