SciPy: patch for missing coding lines in some source files

[embray]

As originally reported at ​https://github.com/scipy/scipy/issues/9606 there are a few .py modules in scipy that contain unicode characters in docstrings without supplying a required "coding" line, resulting in a SyntaxError when you try to compile them.

Normally this is not a problem, because when you build scipy with its setup.py it just compiles all .pyc while assuming utf-8 encoding for the source files. If you already have the .pyc file, then, nothing goes wrong because the SyntaxError only arises when the Python compiler tries to read the original source file at runtime.

But if for some reason your files' mtimes get messed up (as apparently happened to me at some point; not sure how) and the .pyc has to be regenerated then you'll get a SyntaxError at import time for these modules.

For now it should suffice to include the following patch:

​https://github.com/scipy/scipy/commit/4a8a35f2c27be9f209cfc6f347e7a57a4a62e693.patch

[embray]

New commits:

​e6ae4f7

Trac #28154: Patch scipy to including missing coding lines on some source files containing unicode.

[vklein]

I guess a package version update is missing : 1.2.0 => 1.2.0.p0.

[vklein]

Apart from that the patch is working and fix the problem.

[embray]

This patch is really only relevant if you're installing scipy from scratch in the first place. If it's already built this affects nothing at runtime so it's not necessary to increase the patch level and force needless reinstalls of scipy.

[vklein]

Ok then.