* Show-page: split out some effects
Break apart the large effect for easier tweaking, abiding by the 1-action-per-effect guideline.
This also fixes the dependency list. After checking with Haffa, the exclusion of `pathname` and `hash` from the list is probably to avoid hitting `doResolveUri`. This should no longer be an issue after the split, so no suppression needed.
* Show-page: fix "no content found" flicker
## Issue
1539: "Recently been seeing a flicker of "content not found" when direct loading content or channel links on mobile"
## Cause
`doResolveUri` can only be called after the initial mount (in the effect), so at first render, `isResolvingUri` won't be set yet, so we briefly ended up in the "no content found" path.
## Change
Factor out the logic that determines if we should send `doResolveUri`. This allows us to know that the upcoming effect will be sending `doResolveUri`, so we can hold off displaying "no content found".
* Show-page: don't delay the spinner when resolving
## Reason
The delay causes a blank page + footer, which looks bad.
* Remove claim from channel page when geoblock conditions hits
Ticket: 1100
* Handle geoblocked channel page
## Ticket
1100
## Issue
"If a channel is geoblocked, the entire channel page (even showing titles, etc. would be against the law)"
## Issues
- Not safe to call the parent callback from the render function. It must be called from an effect or
- `onHidden` is only called when blacklisted, but it can also be hidden from other circumstances.
- While those other circumstances doesn't apply for FYP (the first client that relies on `onHidden`, it's incorrect from a code perspective, unless it is renamed to `onBlacklisted`.
## Change
- Move the callback execution into an effect.
- Ensure `onHidden` is for all situations that it is hidden.
Effects cannot be defined after the render function returns early, so it was not possible to add new effects when the derived variables that it depends on are way down below.
1. Hardcoding breaks any image changes done on dev instances or locally.
2. The relative path is bad, and there are few more instances of it. But continue as is for now, will file a ticket to fix this.
3. Fix caching.
- Don't want to show it in Incognito.
- Only show it in when entered from homepage, or in the Help page.
- Record the hash of the viewed announcement and update the wallet with it.
This stores the hash for the last viewed announcement.
The intention is so that the announcement won't re-appear when logging into another device.
However, this does mean that announcements would need to wait until the first sync to decide whether to appear or not, which can be quite a delay.
This requires an accompanying commit in the homepages repo (see "Add announcement support").
Using `raw-loader` to import markdown files as a string works for the app-side, but didn't work well in web-side. Falling back to copying the announcement files to `dist` and reading them via `readFileSync`. Don't really like this -- will return with an alternative.
Moving away from the function version to the object version. Trying to track down why the effect that closes the modal ran when the pathname did not change. My only guess is the unstable hideModal reference due to the function mapping. There is no customization needed, so the object version is preferred anyways.
`doOpenModal` isn't used, so I just removed that.
* PublishFile: fix render function violation
Per doc:
> A React component should not cause side effects in other components during rendering.
Even in own render function (allowed to call), it should be avoided as it could cause infinite loops.
* PublishFile: fix useEffect infinite loop due to bad dependency
## Issue
One of the effects was adding an internal wrapper function as a dependency. As this is a functional component, the wrapper is re-created on every render and would spark the effect. That effect also updates redux (depending on the code path), so we end up in a loop.
## Change 1
Two options to fix the dependency:
1. Just remove the wrappers from the list, since we "know" it is essentially the same function (i.e. it's not function-variable that could point to something else at runtime).
2. Peek into the wrapper and determine what are the actual dependencies (usually props or data derived from props).
Solution 2 is the norm.
Aside: wrappers are usually the root-cause of incorrect dependencies, because they mask away the actual code. Need to always peek into it.
## Change 2
Next, change the dispatch-to-props map from function version to object version so that we have stable references to the actions. The object version is also preferred when we don't need to make any customizations to the actions.
This reverts commit f6f15531d4.
The root-cause was known (8dd0982f), and I don't think it worked because:
1. The issue happened at the web side which didn't have logging.
2. The `[object Promise]` came after the `get`, not before.
`tileUris` is being used as an effect dependency, but it gets re-created on every render (same content, different reference), so the `setUriBuffer` effect ended up in a loop.
* Allow drawer to be resized
- Basically re-writes the drag behavior into the Drawer component instead of using SwipeableDrawer, getting more flexibility of where to stop and what to do
- More improvements like the backdrop effect, animations and window resize behavior
* Fix console errors
* Close drawer on unmount so other pages dont load open
* Allow livestream chat to be resized horizontally
* Fix mobile browser size
- address bar etc could be on the way
* Handle popout chat
* Fix pause with floating player
working example
use my forked package
touchups to presentation
Make progress bar thumbnail sprite pixel perfect
add vtt path
fix progress bar a bit
add a black background
use odysee team package
fix two issues for merge
more analytics + refactor
- passes player with time to start (until we move this api to watchman)
- supports livestream metrics for buffering
- fixes bug with buffering over 10 second period
- less head calls by moving to videojs-events
* review fixes
## Issue
I couldn't find where the i18n went because the "events" refactor is just as confusing as the original -- unrelated things are still lumped together in a file.
Also, factoring based on events isn't useful -- it is features that drive what events are needed, not the other way around. This forces features to register events here, and do other things elsewhere? It will be more intuitive to have a one-file-per-feature structure.
## Change
Use existing frameworks to encapsulate things to manageable units/features:
(1) the React useEffect files (can be used isolate out React features like 'tap-to-mute' handling).
(2) the videojs plugins framework.
It's also possible to move the use of `useClaimListInfiniteScroll` into `claimList` itself, but opted to keep them separate for now so that `claimList` doesn't end up a bloated file like `claimListDiscover`.
