respec2html is a command line utility that converts a ReSpec source file...

[anssiko]

... to an HTML file. Depends on PhantomJS http://phantomjs.org.

[darobin]

Hi Anssi!

Thanks for this, but I am hesitant to include it because I'm concerned that it's incorrect. I see two problems:

• It waits 1s, but there's no guarantee that that's enough, especially after we add fetching refs from a remote server. (And even without that, on a slow network it could take more.) The correct approach would be to listen for the end-all message on respecEvents (I'm pretty sure that it can be done with PhantomJS).
• It does its own HTML serialisation with outerHTML which I'm unsure is correct in all cases. It would be better to call the internal function that does HTML serialisation — with outerHTML there are quite a few things that could end up being in the tree that don't belong there (e.g. the error dialog, and the menus that we're looking at adding).

Do you want to rework this based on that? If not, I'll get around to it at some point!

[anssiko]

Hi Robin!

Thanks for the review and suggestions. I'm happy to rework this to address your concerns.

[anssiko]

Hi Robin!

Please see the updated pull request for a reworked version. PhantomJS does not seem to provide an API for listening to custom events, so I worked around the issue by polling document.respecDone for doneness. I now use the ReSpec's built-in toHTMLSource() for serialization.

I also did a minor change to legacy.js to make respec2html.js work for local source files if a html5shiv.js is in use. Before the fix, shiv's scheme was set to undefined.

Let me know if you have any further concerns. I validated the reworked respec2html.js with a bunch of specs and with all the spec tested the script produced a 1:1 copy with the manually converted version.

[darobin]

Brilliant, thanks a lot!