Switching from G+ to Discourse comments on blog

blog

#1

Hey,

I’ve noticed that G+ comments on our Blog does not load anymore, perhaps this is cross-site protection after we switched to HTTPS only. I suggest to switch to Discourse via embedded HTML if that’s feasible for our case:


#2

That sounds like a superb idea! I’ll take a look after CfgMgmtCamp :wink:


#3

Took a while, but the linking is set up and the PR to the website for the Javascript is open - https://github.com/theforeman/theforeman.org/pull/1125. Thanks for the idea!


#4

Thanks for doing it. I think we have enough blogposts converted and we don’t necessary need all of them to have comments. Commenting on two years old blogpost is usually a bit late anyway. Let’s stop it. :slight_smile:


#5

Sorry, I’m not clear on what you want to stop? Are you suggesting to revert the integration?

I think those comments are interesting, actually, especially when you consider that didn’t happen when they were originally posted. It shows that the integration is already doing what we want - raising the profile of the blog and encouraging interaction.

I appreciate there’s been a few days of “blog spam” (although only to mailing list mode users, the integration doesn’t seem to generate notification emails), but with only 104 blog posts total, we’re past “peak blog spam” already. Give it a few days and it’ll calm down.


#6

Looks like clicking on “Show full post” leads to some missing content - take a look at the last blog post for example (Foreman Community Newsletter - June 2018), the RFCs and Plugin news are cut out.


#7

Interesting, that’s the first post that’s gone up using the feed parser instead of the HTML scraper. I wonder if its stripped some H levels… Will look tomorrow


#8

Okay, thanks.


#9

Looks like the URL tripping is happening on older posts too, so it’s not just the feed parser. II’ve asked upstream about what’s happening here (https://meta.discourse.org/t/embedding-discourse-comments-via-javascript/31963/267) - my guess is that it’s something to do with having URLs mixed with bullet-points or some other format-detection weirdness.

I think it’s minor enough that we don’t need to take action about it. Non-newsletter posts seems fine, so that’s only one post per month affected until we figure this out, and the integration is worth it.