1041185 - (e10s-ime) [meta] IME related issues with e10s

Status: RESOLVED FIXED

(Firefox :: General, defect, P3)

Description

[Jim Mathies [:jimm]]

Tracking IME / e10s related bugs that would prevent us from turning e10s on by default on Nightly.

Comment 1

[:Gijs (he/him)]

(tentatively adding stuff I've seen pass by in Untriaged triage, please tell me off if I'm doing the wrong thing...)

Comment 2

[Chris Peterson [:cpeterson]]

Move old M2 P3 bugs to M4 (because tracking-e10s=m5+ flag doesn't exist yet).

Comment 3

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

bug 975383 must improve performance at composing string because it reduces IPC caused by compositionupdate. Most key type during composition causes two events, compositionupdate and text event. So, count of IPC per key will be reduced to half.

Comment 4

[Jim Mathies [:jimm]]

Removing m4 designation from this tracking bug, we have various individual ime bugs that are included in individual milestones.

Comment 5

[Jim Mathies [:jimm]]

Hey Masayuki, how's our e10s_ime support looking currently? Do we have any open bugs that we should consider blocking e10s rollout on?

Comment 6

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Basically, IME has already been available in e10s mode. Although, if content process is busy, i.e., there are some pending content updates at posting new IME event, IME may behave odd especially in TSF mode on Windows (e.g., suggest list of commit strings at inputting text is located at top-left of the window or screen). This is ugly, so, ideally, we should fix bug 1173694 for reducing the frequency of this bug. I hope to fix this bug (my first priority of e10s work is this for now) but optional for first release.

And bug 1203871 is optional. If we can fix bug 1203871, we can reduce a lot of CPU time at caching focused editor's content information in TabParent (collected by ContentCacheInChild and stored by ContentCacheInParent). Reducing the CPU cost of content cache may help to hide the above issue.

If we have much time, these bugs should be fixed before first release (they will create/change only e10s path, so, not so risky for non-e10s). If our first priority is providing same quality as non-e10s mode, I think that these two bugs should be fixed before first release. But if our priority is releasing e10s as soon as possible, especially, releasing e10s for a few users is good for getting bug reports which we've not found yet, it's okay these 2 bugs to keep open.

Comment 7

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

I think that currently, we don't have any reports around IME due to e10s/Fission even though bug 1173694 has not been fixed yet. So we don't have any big task to support it anymore. Let's close this meta bug.