Conversation
|
I haven't been able to build this myself yet due to a build error, but from looking at the video the tags look a bit too large/prominent. I'd try smaller font size and a bit less padding on them, and it'd be good to avoid the different widths. I'm not quite sure what the options are there until we get AdwWrapLayout, but maybe we could do something like the Mastodon web UI and only show 4 by default, and the rest in a popover or something, so we don't have to deal with multiple rows for now? |
(you are aware of flowbox' width situation, where column items have the same width which is the biggest item's. So the widths will be different when there are multiple rows)
The video is a bit quick but the implementation is the same as mastodon-web. It shows the first 3 and then I click the 'Show 5 More' button which causes them all to appear. No opinion on popover or even dialog Here's some different styles: |
|
|
Should I port it over? I have many use cases for WrapBox that are currently using FlowBox here |
|
Sure, it would be good to get some in-the-wild usage of the widget early on! And it seems like an obvious choice for this particular thing. |
|
After spending some time porting it, I'm not sure if I want to continue. We could use LabelWithWidgets in the meantime however: Screencast.From.2024-10-09.09-33-29.mp4since it's not really made just for custom emojis, but for any widget in general edit: (obv, the video above doesn't feature the styling the others have) |
|
That works for now (with more space between the elements, haha) |
|
Performance is terrible, I'll stick to flowbox for now 🤷 :( |
This one looks good to me!
If you encounter any issues with WrapBox, now would be a great time to provide feedback :) |
Nothing to do with WrapBox itself! Porting it to Vala is quite the challenge due to some incompatibilities between C and Vala and ripping it out of libadwaita as is requires copying over a lot of other things, so it's probably best to wait for 1.7 |
|
Ah yeah, that makes sense 👍 |





fix: #880
I tried out different ways to do this and honestly this is probably the best.
The way libxml2 traverses through a doc is unpredictable (aka, when it reaches an anchor, you can't exactly get the content, you have to go through it and usually the
#character and the hashtag name are in different elements). That makes it impossible to detect that an anchor is a hashtag quickly and before parsing and appending it. Same thing for detecting if the last paragraph only has hashtags.The structure also depends on the backend.
While we now have a predictable structure, the same issue as above is still present. It becomes too difficult to reliably detect if the last paragraph is only made of hashtag anchors.
Regex. (Anchors are predictable as it's being ran on the converted HTML)
Screencast.from.2024-10-07.06-43-17.mp4