Implement attachment captioning #1719
Conversation
Lymkwi
force-pushed
the
feature/media-captioning
branch
from
ffee5a4 to
a291925
Compare
Starting with Client-Server API v1.10 [0], the `body` field in messages of type `m.image`, `m.audio`, `m.video` and `m.file` can be used as the caption of the attachment. This is fact the way that Nheko rends captions on images, for example. This commit introduces a field in the `UploadHandle`s awaiting upload on the timeline's `InputBar` which holds a caption taken from the input text area. The decision is as follows: - If text bar is empty or full of blanks, send all media with no caption - If the text is an incomplete command, fail - If there are no pending uploads, proceed as done previously (if there is no command recognized send the text, or try and execute the command and if it fails send the text) - If there are pending uploads, only accept uploads if nothing resembling a command name is in the text area. That text becomes the caption for all pending media. Otherwise, try and execute the command, and, if it fails, send it as text. While this workflow for captioning so far is a bit jank, it is the least effort implementation. Links: [0]: https://spec.matrix.org/v1.10/client-server-api/#mimage Signed-off-by: lymkwi <[email protected]>
Lymkwi
force-pushed
the
feature/media-captioning
branch
from
a291925 to
50cdcc0
Compare
|
Note that nheko does not really support rendering long captions on images i.e. it resorts to ellipses for captions that does not fit the image, which is IMO a blocker for this PR. |
My impression as i was implementing it was that the caption display was really meant to show a filename, which is what the body field was used for prior to spec v1.10 Could a solution to that be having a tooltip displaying the caption? That's the only way i can think of for adding the caption without breaking the way the UI renders at the moment |
Well, I would say just display the caption underneath the image... |
2 participants
Starting with Client-Server API v1.10 [0], the
bodyfield in messages of typem.image,m.audio,m.videoandm.filecan be used as the caption of the attachment. This is fact the way that Nheko rends captions on images, for example.This commit introduces a field in the
UploadHandles awaiting upload on the timeline'sInputBarwhich holds a caption taken from the input text area. The decision is as follows:While this workflow for captioning so far is a bit jank, it is the least effort implementation.
Links:
[0]: https://spec.matrix.org/v1.10/client-server-api/#mimage
Do let me know if i missed anything important in the workflow of
InputBar, i tried to get a feeling for what methods had to be called in what order to process the input text but there's always a chance i missed something important.