Feature Request P2
Status Update
Comments
pe...@knego.net <pe...@knego.net> #2
It looks like the culprit for this regression is aosp/3297690. Conrad, could you please take a look and confirm whether this is expected with your changes?
jb...@jbit.net <jb...@jbit.net> #3
Thanks for the triage. I'll take a look!
ch...@google.com <ch...@google.com>
gh...@gmail.com <gh...@gmail.com> #4
Project: platform/frameworks/support
Branch: androidx-main
Author: drchen <
Link:
Fix performance regression caused by rememberPaneExpansionState
Expand for full commit details
Fix performance regression caused by rememberPaneExpansionState
The regression is caused by the introduction of LaunchEffect in the
remember function, which is required to keep the state correctly between
scaffold value changes.
However, when paneExpansionDragHandle is not provided, there's no way
for devs to mutate the default paneExpansionState. In this case, we can
avoid a lot of management overhead by just providing a stub
implementation.
Relnote: Makes pane expansion state null by default
Test: benchmarks and existing instrumentation tests
Bug: 376394520
Change-Id: Ia65f8919ba9aaaa4c611ecb40514ee41d1e8d56f
Files:
- M
compose/material3/adaptive/adaptive-layout/api/current.txt
- M
compose/material3/adaptive/adaptive-layout/api/restricted_current.txt
- M
compose/material3/adaptive/adaptive-layout/src/commonMain/kotlin/androidx/compose/material3/adaptive/layout/ListDetailPaneScaffold.kt
- M
compose/material3/adaptive/adaptive-layout/src/commonMain/kotlin/androidx/compose/material3/adaptive/layout/PaneExpansionState.kt
- M
compose/material3/adaptive/adaptive-layout/src/commonMain/kotlin/androidx/compose/material3/adaptive/layout/SupportingPaneScaffold.kt
- M
compose/material3/adaptive/adaptive-layout/src/commonMain/kotlin/androidx/compose/material3/adaptive/layout/ThreePaneScaffold.kt
- M
compose/material3/adaptive/adaptive-navigation/api/current.txt
- M
compose/material3/adaptive/adaptive-navigation/api/restricted_current.txt
- M
compose/material3/adaptive/adaptive-navigation/src/androidMain/kotlin/androidx/compose/material3/adaptive/navigation/AndroidThreePaneScaffold.android.kt
Hash: 53e5f16e76ae8e1b81468649f4bcc5ec8c06d05f
Date: Tue Nov 12 12:14:14 2024
ik...@google.com <ik...@google.com> #5
How would a RESTful API work? The point of channel API is browser push, and if you wanted to push via HTTP, you already have URLFetch.
gh...@gmail.com <gh...@gmail.com> #6
We have an internal Comet server we wrote that can handle around 7million open connections. For smaller projects, Channel API would be great. Our REST looking apis or nice looking HTTP ones, looks something like this from memory.
http://mycometserver/{ID}/{count} where {count} is what the client thinks they are up to. As a result, the response is the latest count the server says the client should be. If the client and server are the same, the server camps on the response.
To bump the count, there is another API:http://mycometserver/Bump/{ID}/{new count}, which will result in the waiting client to return.
Why couldn't the Channel API, have done that, your POST is similar to our bump. You could have provided this sort of API as well as a the JS version. As a result, I have to sniff traffic to see what your HTTP call looks like.
That being said, I don't need a C# wrapper from you guys, just a nice HTTP api.
To bump the count, there is another API:
Why couldn't the Channel API, have done that, your POST is similar to our bump. You could have provided this sort of API as well as a the JS version. As a result, I have to sniff traffic to see what your HTTP call looks like.
That being said, I don't need a C# wrapper from you guys, just a nice HTTP api.
ma...@gmail.com <ma...@gmail.com> #7
I would love to see this feature too, as I am working on a Silverlight application using App Engine and the Channel API to push data. Taking a detour to JavaScript seems tedious.
wa...@gmail.com <wa...@gmail.com> #8
Add a HUGE +1 for this post as well. Would love to see the Channel API supported by other "channels", get it?
ri...@gmail.com <ri...@gmail.com> #9
An Objective C library would be absolutely fantastic !
th...@gmail.com <th...@gmail.com> #10
Seem like we just need C API...
hr...@gmail.com <hr...@gmail.com> #11
If now channels work client-side by sending http requests, and server-side holding without answer for long time, Google could as well switch everything to Web Sockets when they become more mainstream?
But still I will vote for this. I would very much like to use Channels in my Android app to talk to my Google App Engine cloud.
But still I will vote for this. I would very much like to use Channels in my Android app to talk to my Google App Engine cloud.
ja...@novaleaf.com <ja...@novaleaf.com> #12
some non-webpage channel would be nice, doesn't have to be C# only but if it works in c# others could port the features to other platforms
ur...@gmail.com <ur...@gmail.com> #13
Need this too for Objective-C
vi...@umich.edu <vi...@umich.edu> #14
In response to Comment 13, I recommend you star Issue 7273 (Channel Client for iOS). Google prematurely closed that issue and needs to revisit it.
Description