Upgraded the hosted instance from alpha 1 to alpha 4.

Alpha Release Changelog

Changes since the first alpha release.

Alpha 2 (internal only)

Cleanup/refactor

Alpha 3

Added Back the CORS Media Upload Proxy

The upload proxy was removed in the early 1.4.x series as it was no longer needed for 0.19.3+ API. However, some instances are either misconfigured or have deliberately restrictive CORS policies around the /pictrs route. To work around this, I re-wrote and re-integrated the CORS proxy handler for image uploads.

Uploading an image will always try a direct connection first. If that fails, it will fall back to the CORS proxy automatically and transparently. If that fails, then you will see an error.

Similarly, the pictrs/image/delete API call was facing similar CORS restrictions. I added an additional CORS handler to proxy those as well. Same as with the upload handler, the delete handler will first attempt a direct call. If that fails, it will try proxying it through the Tesseract server. If that fails, then you will see an error.

Currently, this fallback behavior is “mandatory” in that I do not yet have a user-facing option to disable it. Considering for the first half of Tesseract’s life (and Photon’s for that matter) proxying the upload was the only way it would work, and no one had a problem with it then, I may just leave it as-is.

Alpha 4

Additional Filter Options

Filter Based Upon Community’s NSFW Flag

I added an additional filter option to filter communities based upon their “NSFW” tag. This is useful for several reasons, but the main benefit is that you can filter NSFW communities without filtering NSFW posts in otherwise safe-for-work communities since the “NSFW” tag is used for lots of different purposes. Optionally, you can choose to hide posts to NSFW communities entirely.

NSFW community filtering (not hiding) is also enabled by default now.

Backstory: I had company yesterday and my Lemmy session was open to my .world account from earlier. I absentmindedly refreshed /new right as someone had just done a massive porn dump. The images were blurred, but my friend laughed and was like “did I interrupt you in the middle of something?”. Needless to say, that was embarrassing but could have been worse lol.

I don’t like disabling “NSFW” in the feed since it’s used frivilously as a spoiler tag, trigger warning, a way to attempt to evade spam detection, because the post has a fucking bad word in it, or any other non-NSFW reason under the sun.

In conclusion, to see content posted to NSFW communities without having to click the “reveal” button each time, users must manually disable this filter at Settings -> Filters -> Communities -> Filter NSFW Communities

To Do: Add a server side environment variable option so the admin can disable this. Use case would be if a dedicated NSFW instance is using Tesseract.

Low Score Filters for Posts/Comments are No Longer Disabled if Blind Voting is Enabled

While this breaks with the spirit of the blind voting feature, sometimes things are massively downvoted for a reason (spam, trolling, etc).

Low Score Filters are Enabled by Default

In the spirit of fostering a more pleasant “default” experience, posts and comments with a score of -10 or below are automatically filtered (can still click to reveal them). Note that this is based on the item’s total score (upvotes - downvotes) and not the number of downvotes an item received.

If you want to disable or modify this behavior, you will need to open Settings -> Filters and scroll to the “Filter Low Score Posts” and “Filter Low Score Comments” sections and do one or more of the following:

• Disable each of them if desired
• Opt to completley hide those items
• Modify the score threshold