See original comment at: https://bugzilla.mozilla.org/show_bug.cgi?id=1160601#c9 WebIDE editor on Mac is unusable when scrolling with the trackpad. Easy to reproduce by loading even a small JS file in the WebIDE editor and scrolling using two fingers on the trackpad. Video where I am attempting to scroll for the entire duration: https://www.youtube.com/watch?v=z6d0402uWPU I turned off APZ and the bug still occurs. See also bug 1239615 which is about APZ. I'm filing this separately since I'm seeing with and without APZ. Feel free to dupe it if you want.

I believe this related to APZ and CodeMirror in general, not something WebIDE is doing. Moving to APZ for now.

(In reply to J. Ryan Stinnett [:jryans] (use ni?) from comment #1) > I believe this related to APZ and CodeMirror in general, not something > WebIDE is doing. Moving to APZ for now. Per comment 0 this happens with APZ disabled as well, so unless you have some justification for your belief, this is not really acceptable. For the moment I'm willing to mark this dependent on bug 1239615 and see if fixing that helps here, but if it doesn't I'm going to throw it back to you.

(In reply to Kartikaya Gupta (email:kats@mozilla.com) from comment #2) > (In reply to J. Ryan Stinnett [:jryans] (use ni?) from comment #1) > > I believe this related to APZ and CodeMirror in general, not something > > WebIDE is doing. Moving to APZ for now. > > Per comment 0 this happens with APZ disabled as well, so unless you have > some justification for your belief, this is not really acceptable. For the > moment I'm willing to mark this dependent on bug 1239615 and see if fixing > that helps here, but if it doesn't I'm going to throw it back to you. Makes sense. I just can't think of anything that would be specific to WebIDE in particular, so hopefully fixing bug 1239615 fixes this too.

Making title less scary.

For this bug, which seems to be both APZ on or off, let's see if a regression window can help.

I'm unable to reproduce this bug locally. Dietrich, please see if you can find the regression window.

I'm not going to be able to do that anytime soon. Can confirm is still unusably bad in Nightly, not magically fixed yet.

(In reply to Dietrich Ayala (:dietrich) from comment #7) > I'm not going to be able to do that anytime soon. Well then, can you provide any more information that will enabled me to reproduce this locally? Unless I am able to a regression window will be impossible.

Hm, I'm on the 12" Macbook. I just reproduced instantly on Dev Edition, Nightly, and Firefox. Maybe hardware?

(In reply to Dietrich Ayala (:dietrich) from comment #7) > I'm not going to be able to do that anytime soon. Can confirm is still > unusably bad in Nightly, not magically fixed yet. Dietrich, could you check in Release (44) as well? I can try to find a regression, but I am starting to think this has been around longer than we guessed. Just want to see if you get a different result. For me, it seems to be present in Release as well.

Yep, see comment #9 :) Interestingly, it is not as bad in 44... I'm not seeing the screen whiteout effect that I see when scrolling up and down quickly in Nightly.

Could you try the patch from Bug 1243589 ? It completely fixes the issue logged in Bug 1239615. I quickly tried WebIDE with it, and it seems to improve the situation as well.

Please retest on the next Nightly; bug 1243589 should fix this or at least improve it significantly. Thanks!

It's WAY better. The interaction is smooth now - the motion of the scroll is exactly matching my physical scrolling movement. However, visually it is still not great, it's skipping bits, looks like. I'll try and make a movie showing it.

(In reply to Dietrich Ayala (:dietrich) from comment #14) > It's WAY better. The interaction is smooth now - the motion of the scroll is > exactly matching my physical scrolling movement. However, visually it is > still not great, it's skipping bits, looks like. I'll try and make a movie > showing it. If by "skipping bits" you mean it looks like checkerboarding (i.e. parts of the code window are blank for a while but fill in later) then it might be https://github.com/codemirror/CodeMirror/issues/3564 (which bug 1232491 would help with). Still a video would help confirm we're talking about the same thing.

https://www.youtube.com/watch?v=ThXBlhYhR_E The latter part of the video is where the effect is visible - looks like a flashing on both the top and bottom of the scrolling area.

A bit hard to tell but that looks like a combination of APZ checkerboarding and what I described in comment 15. I'll mark this fixed by bug 1243589 since the jank issue is fixed, and file a follow-up to track the checkerboarding.