site_path_helper() produces URLs without trailing slashes about blogofile_blog

[douglatornell]

So, this is really a blogofile issue, not a blogofile_blog one, correct? Though, of course, you need a site, which comes from a plugin like blogofile_blog, to demonstrate the issue.

[brandon-rhodes]

Yes, the issue is with the function itself in blogofile/util.py — either we could add a clause to it like the one it already has for assuring that the links it generates are preceded with slashes that also checks for a terminal slash, or we could add the clause behind a test for a new option "append_slash_to_site_paths" or something so that people would have to make a deliberate change to get the new behavior. (And thanks, by the way, for taking over blogofile — it has been great to watch all the activity on github each time I log in, because of your work!)

[douglatornell]

So, just adding a trailing slash would break this use case:

<script src="${bf.util.site_path_helper('js/libs/modernizr-1.7.min.js')}"></script>

That breakage could be avoided by checking for a file extension at the end of the path and only adding the trailing slash if there wasn't one, but that feels like it could be fraught with other exceptions.

Add to the above that "explicit is better than implicit" and we arrive at a new function signature:

site_path_helper(*parts, trailing_slash=False)

It turns out that a signature like that is valid in Python 3, but not in Python 2. Who knew? Certainly not me. To suppport both version the signature has to be:

site_path_helper(*parts, **kwargs)

and the function has to do the kwargs-processing dance.

This may be the first time I've hit one of the syntactic improvements in Python 3 that has really made me regret having to support Python 2 as well. The Python 3 version is just plain clearer and more elegant. Fodder for your code beauty talk?