Publish to Disqus

[kylewm]

I would love to be able to POSSE replies to Disqus comments threads!

Would require publishers to give Bridgy access to their Disqus account, but then comments would be published as me.

Here's an example of one I did manually

https://kylewm.com/reply/2014/08/28/2/very-interesting-project-i-m-curious
http://facebook.smallpict.com/2014/08/25/postingToWordpressAndFacebook.html#comment-1563847441

This is a little crazy, but it would be especially cool if it could detect these permissions when working with a Tumblr user signed up with "bridgy for blogs" to recognize webmentions from me and use my Disqus account to comment as me instead of as the blog owner. 💥

[snarfed]

both this and #271 are great ideas! i'd love to use them myself. they'd take some work, obviously, but i think they're pretty straightforward.

design wise, i don't think you'd need any new pieces. disqus user pages would show just the Publish UI section of social_user.html. you'd probably need to extract that out into a new template so you can reuse it, but that's easy. wp.com would just add that publish UI to the existing wp.com template.

code wise, we'd obviously need to add disqus to oauth-dropins and a new source type, but we can probably reuse a lot of the disqus api usage in tumblr.py. same w/wp.com and wordpress_rest.py.

exciting ideas!

(re the NASCAR concern, eh, don't worry about it. we'd need to add a new disqus button to the second row on the front page, but that's fine. the incremental UX cost is pretty low. the only question in my mind is whether that would confuse tumblr users, since they use disqus too, but meh. we'll see.)

[kylewm]

first minor roadblock, I'm looking up disqus thread ID using /threads/details and thread=link:URL.

on some sites the thread is identified by the full URL http://facebook.smallpict.com/2014/08/25/postingToWordpressAndFacebook.html

but on tumblr sites, they want us to chop off the slug http://snarfed.tumblr.com/post/84623272717 not http://snarfed.tumblr.com/post/84623272717/stray-cat

one option is just to try the full url and then try without the last path component if that fails...I could certainly construct a website for which that would be a bad choice, but maybe that's not too likely in the real world?

[snarfed]

heh, create_comment() in tumblr.py already does exactly that. :P as for identifying tumblr sites, at least without fetching and scraping them or asking tumblr's API, meh. I'm fine with just doing this automatically for now.

[snarfed]

if anyone's interested in tackling this, i've written up comprehensive instructions on how to add a new silo to bridgy. feel free to jump in!