/pool_versions should now show changes to pool descriptions.
More things are logged to /mod_actions: Mods changing or deleting anyone's comments, forum posts or user feedbacks; mass reverting all of a user's tag edits; creating/deleting IP bans.
Fixes
A bug with autocompletion for @mentions not working on comments or when editing forum posts was fixed.
A bug with moderator IP addr search being broken was fixed.
The bug with the edit mode submit button being greyed out was fixed.
/posts/random.json works correctly now.
Posts with a lot of favorites or comments should load a little faster.
Does someone know how the favorite list is created when you want to see who favorited a post? I always thought that the person who favorited a post first is shown as the lowest user and the last user who favorrited a post is shown at the top of that list. For example post #2588951 was favorited by me and I was the second user when I did that. But when I click under "Information" on "Show Favorites", I'm listed somewhere in the upper regions. Is that a bug? If not, then I want to know how that list is generated, because it kinda seems off to me.
Apologies if this has already been mentioned, but for the past week or so, over 90% of the time when I translate a commentary and submit it, the translation won't show until I reload the page. This has occurred on both Max OSX 10.10.5 and the latest iteration of Windows 10, both utilizing their respective most recent versions of Firefox. It's not a site-killer, by any means, but I figured I'd bring it up, just in case.
I think that bug was pushed along when the feature to add commentary from source or another post was added, but yeah it does need to be checked out. Same thing happens to me.
I think it's become impossible to unvote a previously voted post. Before this, if you clicked on "vote up/down", you got the message "you have already voted for this post", and the line changed to "unvote". Now it shows the message, but doesn't change, so you can't unvote.
Listed in profile amount of post you uploaded is bit different from actual amount. Most likely some kind of a bug with counter then you upload several posts at once, like prepare ~10 and hit upload button at same time (notice while doing this). And if you search for user (even self) it will show you right amount (or closer to right at least).
Listed in profile amount of post you uploaded is bit different from actual amount. Most likely some kind of a bug with counter then you upload several posts at once, like prepare ~10 and hit upload button at same time (notice while doing this). And if you search for user (even self) it will show you right amount (or closer to right at least).
Yeah, it's just as you say, when you upload several posts simultaneously the upload counter may not be updated correctly. The Post Changes counter has the same problem, mine is off by over 60,000 edits because of rapid tagscripting and bot editing.
The listing on the /users page doesn't use that counter, the counts there are calculated exactly (except for the Edits column). Incidentally that's why that page is a little bit slow, it's calculating all the counts for every user.
The new DText parser was enabled everywhere. If you notice any dtext formatting oddities, report them.
Ugh, I've been putting off learning the DText Ragel parser because of it's complexity... :p But I guess I'll have to if I want to accurately emulate how it parses (or misparses) URL and Textile links....
Edit:
But, since this is now the main parser, I can finally report on the [nodtext] tag, which seems to need a beginning space in order to work
Working example (has a beginning space)
Text[b]strong text[/b]
You should not see the [B] tags if the above works.
Nonworking example (does not have a beginning space)
[b]strong text[/b]
You should not see the [B] tags if the above works.
Ugh, I've been putting off learning the DText Ragel parser because of it's complexity... :p But I guess I'll have to if I want to accurately emulate how it parses (or misparses) URL and Textile links....
You can try and compile ragel source into a C library, most languages have bridges to using native libraries. That's the entire reason Albert switched to ragel anyway, so that parsing is done on lower level.
Another DText oddity is nested [b] or other styling tags.
1. Working example (2 opening/closing [b] tags):
Some bold and bolder text. Some regular text.
2. Non-working example (2 opening, 1 closing [b] tags):
Some bold and bolder text. Some regular text.
Not sure what the proper behavior should be here...? Should the second [b] just be ignored once the bold option has been set...? If so, then 1. would have a trailing [/b] tag on full display. Still, I'm more in favor of that, as I see no reason why nesting should be allowed... unless of course, the option is ever added for multiple levels of boldness.
Edit:
Created issues 5, 6 and 7 for all of the above mention issues
For the issue Type-kun brought up, upon closer inspection, it was caused by out of order [b] and [i] tags. After fixing the order, the page now renders as it should.
Also, just as a reminder, certain URL parsing still doesn't work quite correctly as demonstrated on http://testbooru.donmai.us/posts/11. However, as discussed on issue #2744, there's not really a good way to handle this.
undefined method `updater_id' for #<UserFeedback:0x007f343070e840> Did you mean? update updated_at app/models/user_feedback.rb:13:in `block in <class:UserFeedback>' app/controllers/user_feedbacks_controller.rb:39:in `update'
undefined method `updater_id' for #<UserFeedback:0x007f343070e840> Did you mean? update updated_at app/models/user_feedback.rb:13:in `block in <class:UserFeedback>' app/controllers/user_feedbacks_controller.rb:39:in `update'
Can't really take care of it right now, will do later unless someone steps in first. It'll probably be better to use link_to_user, too, to get colorized usernames.
I think link_to_user wouldn't work here, HTML in modactions is escaped so you can only use dtext. It could be changed, but you'd have to be careful about escaping user-controlled input in modactions.