Issues with the release make target

[ArthurHoaro]

I had 2 issues with the release_archive make target I did not have before. I changed my dev environment, so I'm not sure if it's due to my configuration or not.

htmldoc

error: invalid command 'bdist_wheel'

I had to run pip install wheel before installing mkdoc (https://stackoverflow.com/a/44862371/1484919). I don't really know python tools so I'm not sure what was going on here.

release_zip

The command mkdir -p $(ARCHIVE_PREFIX)/{doc,vendor} while running make wouldn't create the directories properly.

mkdir -p Shaarli/{doc,vendor}
rsync -a doc/html/ Shaarli/doc/html/
rsync: mkdir "<path>/Shaarli/Shaarli/doc/html" failed: No such file or directory (2)
make: *** [Makefile:126: release_zip] Error 11

$ ll Shaarli
total 4,0K
drwxrwxr-x 2 arthur arthur 4,0K juil. 27 12:14 {doc,vendor}

However running the command directly from my zsh, it was working properly. Maybe I'm missing some dev package...

[virtualtam]

error: invalid command 'bdist_wheel'

This usually happens when using a virtualenv with outdated versions of pip and setuptools, I've seen this recently on Debian Stretch hosts with Python 3.5 virtualenvs. An easy workaround is to pin a minimum version of pip to be used by the virtualenv.

mkdir -p $(ARCHIVE_PREFIX)/{doc,vendor}

This could happen if the path expansion {doc,vendor} fails, e.g. when using sh as the SHELL interpreter. Maybe we should enforce Bash usage in the Makefile?

Reference:

• https://www.gnu.org/software/make/manual/html_node/Choosing-the-Shell.html

Surprisingly, this has been set in the v0.8.7 release but doesn't appear in the master branch: https://github.com/shaarli/Shaarli/blob/v0.8.7/Makefile#L14

[ArthurHoaro]

This usually happens when using a virtualenv with outdated versions of pip and setuptools

It looks like installing my distribution's setuptools package fixed the issue.

Surprisingly, this has been set in the v0.8.7 release but doesn't appear in the master branch

This change fixes the issue. :)
But it looks like this set of changes never made it to the master branch (at least partially, I'm not sure): #1156