Can 'download my posts' output have consistently useful inline images?

Here on Meta, the “download my posts” files have a mostly complete URL in the src attribute of uploaded images, lacking only scheme, something like this:

<img src="//assets-meta-cdck-prod-meta.s3.dualstack.us-west-1.amazonaws.com/original/3X/0/6/0654...

But on other sites, such as the Discourse hosted bbs.nextthing.co the URLs lack even a hostname:

<img src='/uploads/nextthing/original/2X/f/f2cf3de9...

(I notice that the nextthing .csv file has some src="..." and some src='...', while I only see the first form in the .csv file from Meta.)

1 Like

There should be the hostname (or the CDN if enabled) for local uploads… is that from an old post? Does the URL change after you rebuild the HTML of that post?

2 Likes

That was output from the profile button “Download my Posts” from two sites that I am a user on, not an admin.

Both of those downloads were made yesterday. I’ve got such post archives from a number of sites, but not all of them do I have uploaded images on. I have started to consider also archiving the images I have with the posts, and to do that I need to parse out the HTML. It would be helpful if Discourse would use a consistent format.

Is this still an issue?

Yes, a fresh download of Meta and Nextthing posts show both are in the same format.

2 Likes