ZyX, document autoget.py briefly

[marco-oweber]

Can you add a high level overview what that code exactly does? The description can be very short but should contain the source of the repositories.

I imagine something like this:

# autoget.py: queries github and tries to associate github urls with vim script 
numbers by matching script names using some regular expression magic

Looks like one association is wrong: https://github.com/MarcWeber/vim-addon-manager-known-repositories/issues/73

Do you have time to have a quick look at which is likely to be the right fix?

[ZyX_I]

# autoget.py: queries github and tries to associate github urls with vim script 
numbers by matching script names using some regular expression magic

Nothing near this. It downloads archive from www.vim.org and determines a list of file names. Then it checks whether list of file names obtained from the archive is contained in the one in the repository. If it is not, it performs the same check, but for .vim, .py, .pm, .lua and some other files only. Repository name only alters order of checking repository URLs in case there is more then one found.

[ZyX_I]

Add doc string, also add Match.repr

Fixes #75

→ <<cset 2bbb7026acfb5805c9b13a0a1abfa67c7370b90d>>

[marco-oweber]

In the repository ? Which repository URLS are matched against sources from vim.sf.net?

[ZyX_I]

Use argparse to parse arguments. Also adds support for --help

Ref #75

→ <<cset 855f555aabdfdb0785c2dae16b240e662c07743b>>

[ZyX_I]

In the repository ? Which repository URLS are matched against sources from vim.sf.net?

I do not understand. URLs are not matched against sources.

[marco-oweber]

After reading your nice documentation there are no more question. Adding the initial comment lines would have been enough. Now have a look at the original issue: https://github.com/MarcWeber/vim-addon-manager-known-repositories/issues/73. The description contains both: the github repository and a "mirror". Is it better to ask to drop the mirror line or to ignore vim-scripts urls in your code?

[ZyX_I]

Originally my script behaved correctly: it used URL where development goes (though I would expect it to use vim-scripts repo instead: name-based sorting worked, but it worked in reverse putting less likely candidates first). Maybe this is why there are exactly no vim-scripts matches in scm_generated.json.

Now sorting is fixed, but candidate with name match is ignored due to it being in vim-scripts repo.