FR: check link validity

[TomRoche]

Mostly I use restview to view a reST file (always composed in emacs) before I upload it to some repo or wiki. Occasionally I use it to send HTML email, mostly to Gmail users. I don't normally send HTML mail, but when I do, it's like this:

1. render a file (e.g., mail.rst) until it looks the way I want
2. generate HTML (by C-u in the browser, then cut'n'paste into another file (e.g., mail.html)
3. send mail.html from emacs (yes, I'm one of "those people" :-)
   • as a MIME attachment, but inline (so it doesn't look like an attachment).
   • BCCing me, so I know what it looks like. Usually I'll send myself a test post first, so I can be sure of how Gmail renders it ... which may not be the same as restview+firefox renders it.

And, just now, I found a way that Gmail renders better--or at least, more user-friendly-ly--than restview. In reST, I have a tendency to create one-off links like

`foo, and bar <http://foo/bar>`_

until I create the same link several times, whereupon I create a reference, like

.. _foo, and bar: http://foo/bar

and reuse that. Unfortunately, sometimes I forget to remove the angle brackets, and so create a legal-but-useless reference like

.. _foo, and bar: <http://foo/bar>

restview will cheerfully render that, so I thought I was OK until

1. I generated HTML to a file.
2. I sent a test email (to myself on Gmail) using the inline-attached HTML.
3. I opened the email in Gmail.

I then noticed that several links did not display! Whereupon I checked my reST file, and noticed the problem.

So if there's a way to do a simple syntax check of URIs before rendering, and to throw an error on bad URIs, this would probably help users.

[mgedmin]

Sounds like a good idea.

Maybe it should be a docutils feature though? Would you like to contact them?

[mgedmin]

BTW I'm not against adding a check for this in restview itself, if upstream proves unresponsive.

I'm not likely to find the time to work on it in the near future, so a pull request would be welcome. I already had code for processing all the links in the document in c444140 (grep for checkLinkSchemes), but removed it when I switched to readme in de44c77. I'm not saying that's the best approach -- maybe overriding some method in SyntaxHighlightingHTMLTranslator would be easier?