Status Update
Comments
bl...@google.com <bl...@google.com> #2
To be clear, when you say jump pages, do you mean recyclerview is scrolling down, or is viewpager scrolling across multiple fragments?
A repro / sample project demonstrating the issue would definitely be helpful here.
in...@google.com <in...@google.com> #3
You also mentioned alph09 in the title, but alpha07 in the comment - could you specify which version of viewpager and paging you are using?
to...@google.com <to...@google.com> #4
Ah sorry I see that you have used both here, still could you specify which version of viewpager you are using as well?
ma...@google.com <ma...@google.com> #5
View Pager 2 : - implementation "androidx.viewpager2:viewpager2:1.0.0"
Paging 3 : - implementation "androidx.paging:paging-runtime-ktx:3.0.0-alpha09"
Example
Attached Video Explanation
When i click slowly on previous button it swipe single page, but if i click quickly multiple times on previous button it start scrolling more than we click and it's not stopping until we click on viewpager.
to...@google.com <to...@google.com> #6
ap...@google.com <ap...@google.com> #7
Haven't forgotten this - this is on my queue, just that my queue is quite long :)
ap...@google.com <ap...@google.com> #8
ma...@google.com <ma...@google.com>
ap...@google.com <ap...@google.com> #9
This looks like a weird race scenario where the smooth scroll behavior in RV / ViewPager is attempting to do an animated scroll to an absolute position, but Paging is able to PREPEND items to the front of the list faster than the animation can scroll there.
I noticed that you're attempting to enablePlaceholders to mitigate this by telling Paging the total item count, but you never pass a value for placeholdersBefore / After in the initial result on REFRESH. This will prevent Paging understanding how many placeholder items there are in the list. This is probably something Paging should check for, but the proper solution to this ticket should probably be some kind of relative scrolling behavior on RecyclerView or ViewPager, however I expect boundary cases to be quite annoying to deal with there.
Description
When building [ is not defined automatically by
libcxx
], I noticed that_GNU_SOURCE
armv7m-cros-eabi-clang++
.Some comparisons (where
empty.cc
is an empty file):_GNU_SOURCE
is defined:_GNU_SOURCE
is not defined:It looks the baremetal variants don't define it. I guess the assumption is that baremetal targets won't use
libc
?