1462257 - Microsoft Pinyin Chinese IME - can't delete the last letter with backspace

Status: VERIFIED FIXED

(Core :: DOM: UI Events & Focus Handling, defect, P2)

Description

[ershiwo]

Attachment: 123.gif

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:61.0) Gecko/20100101 Firefox/61.0
Build ID: 20180514150347

Steps to reproduce:

Input some Chinese words in text area（or any other place) with Microsoft Pinyin in Windows 10 1803, backspace the incorrect letters.


Actual results:

The last letter won't be deleted and it was in selected state.


Expected results:

There is nothing after delete all incorrect pinyin letters.

Comment 1

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

Odd, I cannot reproduce this bug.

Comment 2

[Makoto Kato [:m_kato]]

This seems to be regression by bug 1456381.  And I can reproduce this on both x86 (VM) and x64 using Microsoft Pinyin.  Last backspace selects that character instead of deleting character.

Comment 3

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

Could you take a look or attach a log of "MOZ_LOG=nsTextStoreWidgets:5,sync"?

Comment 4

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

Ah, after I restart my computer, I reproduce this bug.

Comment 5

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

> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   Locking (TS_LF_READWRITE) >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::OnEndComposition(pComposition=0x0000029476172930), mComposition={ mView=0x0000029476172930, mString="n" }
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::RecordCompositionEndAction(), mComposition={ mView=0x0000029476172930, mString="n" }
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::MaybeDispatchKeyboardEventAsProcessedByIME(), does nothing because not necessary to dispatch keyboard event
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=1, acpEnd=1 (length=0), reverted=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::ContentForTSFRef(): mContentForTSF={ mText="n" (Length()=1), mLastCompositionString="n" (Length()=1), mMinTextModifiedOffset=4294967295 }
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::RecordCompositionEndAction(), succeeded
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::OnEndComposition(), succeeded
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::OnStartComposition(pComposition=0x000002940F02EF30, pfOk=0x000000A0D03FCA10), mComposition.mView=0x0000000000000000
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::RecordCompositionStartAction(aComposition=0x000002940F02EF30, aRange=0x000002940F02EDB8, aPreserveSelection=false), mComposition.mView=0x0000000000000000
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::RecordCompositionStartAction(aComposition=0x000002940F02EF30, aStart=0, aLength=1, aPreserveSelection=false), mComposition.mView=0x0000000000000000
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=1, acpEnd=1 (length=0), reverted=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::ContentForTSFRef(): mContentForTSF={ mText="n" (Length()=1), mLastCompositionString="n" (Length()=1), mMinTextModifiedOffset=4294967295 }
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::MaybeDispatchKeyboardEventAsProcessedByIME(), does nothing because not necessary to dispatch keyboard event
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=1, acpEnd=1 (length=0), reverted=false
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::RecordCompositionStartAction() succeeded: mComposition={ mStart=0, mString.Length()=1, mSelectionForTSF={ acpStart=0, acpEnd=1, style.ase=TS_AE_END, style.fInterimChar=false } }
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::OnStartComposition() succeeded
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::SetText(dwFlags=not-specified, acpStart=0, acpEnd=1, pchText=0x00007FF872DEDD1C "", cch=0, pChange=0x000000A0D03FCA48), mComposition.IsComposing()=true
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SetSelectionInternal(pSelection={ acpStart=0, acpEnd=1, style={ ase=TS_AE_END, fInterimChar=false} }, aDispatchCompositionChangeEvent=false), mComposition.IsComposing()=true
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=0, acpEnd=1 (length=1), reverted=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::MaybeDispatchKeyboardEventAsProcessedByIME(), does nothing because not necessary to dispatch keyboard event
> [Parent 29116: Main Thread]: W/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SetSelectionInternal() Succeeded but did nothing because the selection range isn't changing
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::InsertTextAtSelectionInternal(aInsertStr="", aTextChange=0x000000A0D03FCA48), IsComposing=true
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=0, acpEnd=1 (length=1), reverted=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::ContentForTSFRef(): mContentForTSF={ mText="n" (Length()=1), mLastCompositionString="n" (Length()=1), mMinTextModifiedOffset=4294967295 }
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::MaybeDispatchKeyboardEventAsProcessedByIME(), does nothing because not necessary to dispatch keyboard event
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D270   TSFTextStore::Content::ReplaceTextWith(aStart=0, aLength=1, aReplaceString=""), mComposition={ mStart=0, mString="" }, mLastCompositionString="n", mMinTextModifiedOffset=4294967295, firstDifferentOffset=0
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::InsertTextAtSelectionInternal() succeeded: mWidget=0x000002947639C000, mWidget->Destroyed()=false, aTextChange={ acpStart=0, acpOldEnd=1, acpNewEnd=0 }
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SetText() succeeded: pChange={ acpStart=0, acpOldEnd=1, acpNewEnd=0 }
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::OnUpdateComposition(pComposition=0x000002940F02EF30, pRangeNew=0x0000000000000000), mComposition.mView=0x000002940F02EF30
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::MaybeDispatchKeyboardEventAsProcessedByIME(), does nothing because not necessary to dispatch keyboard event
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::OnUpdateComposition() succeeded but not complete
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::OnEndComposition(pComposition=0x000002940F02EF30), mComposition={ mView=0x000002940F02EF30, mString="" }
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::RecordCompositionEndAction(), mComposition={ mView=0x000002940F02EF30, mString="" }
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::MaybeDispatchKeyboardEventAsProcessedByIME(), does nothing because not necessary to dispatch keyboard event
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=0, acpEnd=0 (length=0), reverted=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::ContentForTSFRef(): mContentForTSF={ mText="" (Length()=0), mLastCompositionString="n" (Length()=1), mMinTextModifiedOffset=0 }
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::RecordCompositionEndAction(), succeeded
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::OnEndComposition(), succeeded
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::SetSelection(ulCount=1, pSelection=000000A0D03FCA30 { acpStart=0, acpEnd=0, style={ ase=TS_AE_END, fInterimChar=false } }), mComposition.IsComposing()=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SetSelectionInternal(pSelection={ acpStart=0, acpEnd=0, style={ ase=TS_AE_END, fInterimChar=false} }, aDispatchCompositionChangeEvent=true), mComposition.IsComposing()=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=0, acpEnd=0 (length=0), reverted=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::MaybeDispatchKeyboardEventAsProcessedByIME(), does nothing because not necessary to dispatch keyboard event
> [Parent 29116: Main Thread]: W/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SetSelectionInternal() Succeeded but did nothing because the selection range isn't changing
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SetSelection() succeeded
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::GetSelection(ulIndex=4294967295, ulCount=1, pSelection=0x000000A0D03FD760, pcFetched=0x000000A0D03FD8D8)
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=0, acpEnd=0 (length=0), reverted=false
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetSelection() succeeded, acpStart=0, acpEnd=0 (length=0), reverted=false
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::GetSelection(ulIndex=4294967295, ulCount=1, pSelection=0x000000A0D03FD730, pcFetched=0x000000A0D03FD884)
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=0, acpEnd=0 (length=0), reverted=false
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetSelection() succeeded, acpStart=0, acpEnd=0 (length=0), reverted=false
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::GetActiveView(pvcView=0x000000A0D03FD778)
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetActiveView() succeeded: *pvcView=1
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::GetScreenExt(vcView=1, prc=0x000000A0D03FD7E0)
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetScreenExtInternal()
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetScreenExtInternal() succeeded: aScreenExt={ left=990, top=482, right=1496, bottom=503 }
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetScreenExt() succeeded: *prc={ left=990, top=482, right=1496, bottom=503}
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::GetTextExt(vcView=1, acpStart=0, acpEnd=0, prc=0x000000A0D03FD7F0, pfClipped=0x000000A0D03FD7B0), IsHandlingComposition()=true, mContentForTSF={ MinOffsetOfLayoutChanged()=0, LatestCompositionStartOffset()=0, LatestCompositionEndOffset()=0 }, mComposition= { IsComposing()=false, mStart=0, EndOffset()=0 }, mDeferNotifyingTSF=false, mWaitingQueryLayout=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=0, acpEnd=0 (length=0), reverted=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetTextExt() hacked the queried range for not returning TS_E_NOLAYOUT, newvalues are: acpStart=0, acpEnd=0
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetScreenExtInternal()
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetScreenExtInternal() succeeded: aScreenExt={ left=990, top=482, right=1496, bottom=503 }
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetTextExt() succeeded: *prc={ left=992, top=484, right=1000, bottom=501 }, *pfClipped=false
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   Unlocked (TS_LF_READWRITE) <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::FlushPendingActions() flushing Type::eCompositionEnd={ mData="n" }
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::FlushPendingActions(), dispatching compositioncommit event...
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::FlushPendingActions() flushing Type::eCompositionStart={ mSelectionStart=0, mSelectionLength=1 }, mDestroyed=false
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::FlushPendingActions() dispatching compositionstart event...
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::FlushPendingActions() flushing Type::eCompositionEnd={ mData="" }
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::FlushPendingActions(), dispatching compositioncommit event...
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::MaybeFlushPendingNotifications(), putting off flushing pending notifications due to being dispatching events...
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::RequestLock() succeeded: *phrSession=S_OK
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::RequestLock(dwLockFlags=TS_LF_READ | TS_LF_SYNC, phrSession=0x000000A0D03FE788), mLock=not-specified, mDestroyed=false
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   Locking (TS_LF_READ) >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::GetSelection(ulIndex=4294967295, ulCount=1, pSelection=0x000000A0D03FE250, pcFetched=0x000000A0D03FE408)
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::SelectionForTSFRef(): acpStart=0, acpEnd=0 (length=0), reverted=false
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::GetSelection() succeeded, acpStart=0, acpEnd=0 (length=0), reverted=false
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   Unlocked (TS_LF_READ) <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
> [Parent 29116: Main Thread]: D/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::MaybeFlushPendingNotifications(), putting off flushing pending notifications due to being dispatching events...
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190   TSFTextStore::RequestLock() succeeded: *phrSession=S_OK
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::GetStatus(pdcs=0x000000A0D03FD280)
> [Parent 29116: Main Thread]: I/nsTextStoreWidgets 0x000002940B80D190 TSFTextStore::RequestLock(dwLockFlags=TS_LF_READWRITE, phrSession=0x000000A0D03FD2C0), mLock=not-specified, mDestroyed=false

Tricky behavior... When I remove last 1 character of composition string, MS Pinyin commits composition with the last character. Then, new composition is started and it tries to replace the last character with empty string. We must fail handling the last odd composition.

Comment 6

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

Oh... This is a bug of TextComposition::DispatchCompositionEvent(). It does not dispatch eCompositionChange event from eCompositionCommit event when commit string is empty and it's fired immediately after eCompositionStart.

Comment 7

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

https://treeherder.mozilla.org/#/jobs?repo=try&revision=5b2d37f5e58d0a00c0c3d5533caaef40ef0cf0ab

Comment 8

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

https://treeherder.mozilla.org/#/jobs?repo=try&revision=4b76429fa8bba83db8aa41ffc3cabe28390fe931

Comment 9

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

MS Wubi, MS ChangJie and MS Quick behave same as MS Pinyin.

According to Nightly's telemetry, 55% of Simplified Chinese language users use MS Pinyin or MS Wubi, and 20% of Traditional Chinese language users use MS ChangJie or MS Quick.

Comment 10

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

https://treeherder.mozilla.org/#/jobs?repo=try&revision=74de0c885703d5bc60624566b79d1f32a33bf868

Comment 11

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

Attachment: Bug 1462257 - TextComposition should dispatch eCompositionChange event when eCompositionCommit is being fired immediately after eCompositionStart

TextEditor modifies composition string or selected string when first
eCompositionChange event is received.  However, TextComposition dispatches
eCompositionChange event ("text" event of DOM) only when composition string
becomes non-empty if current composition string is empty.  So, when IME
dispatches only eCompositionStart and eCompositionCommit events for removing
selected text, TextEditor does nothing.  This hacky behavior is used by
MS Pinyin on Windows 10 at least.

For supporting this behavior, we need to make TextComposition dispatch
eCompositionChange event when eCompositionChange(AsIs) event is fired
even before dispatching eCompositionChange event.

Although from point of view of web apps, the hacky composition should be
merged into the previous composition if it's possible but it's out of scope
of this bug.

Review commit: https://reviewboard.mozilla.org/r/256476/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/256476/

Comment 12

[Makoto Kato [:m_kato]]

Comment on attachment 8991557 [details]
Bug 1462257 - TextComposition should dispatch eCompositionChange event when eCompositionCommit is being fired immediately after eCompositionStart

https://reviewboard.mozilla.org/r/256476/#review263610

Comment 13

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/37e818a03a1b
TextComposition should dispatch eCompositionChange event when eCompositionCommit is being fired immediately after eCompositionStart r=m_kato

Comment 14

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

Comment on attachment 8991557 [details]
Bug 1462257 - TextComposition should dispatch eCompositionChange event when eCompositionCommit is being fired immediately after eCompositionStart

Approval Request Comment
[Feature/Bug causing the regression]:
Regression of bug 1456381.

[User impact if declined]:
Chinese language (both Simplified and Traditional) users see odd character when they remove all composition string to cancel current inputting word.

[Is this code covered by automated tests?]:
No.

[Has the fix been verified in Nightly?]:
Not yet, but I confirmed on local build.

[Needs manual test from QE? If yes, steps to reproduce]: 
Yes.
1. Select Microsoft's Chinese IME such as Microsoft Pinyin.
2. Change input mode to Chinese character (perhaps, pressing ShiftLeft key toggles the input mode).
3. Type "ni" into an editor (e.g., <https://w3c.github.io/uievents/tools/key-event-viewer.html>)
4. Type Backspace twice to remove all characters.

Without this fix, you'll see selected "n" in the editor.  If fixed correctly, correctly both "ni" are removed. Note that you may see selected "n" temporarily immediately after press second Backspace. This will be fixed by bug 1475153, but it's not regression. So, it should be ride the train.

[List of other uplifts needed for the feature/fix]:
No.

[Is the change risky?]:
No.

[Why is the change risky/not risky?]:
This starts eCompositionChange event only when "compositionend" follows "compositionstart" without any "compositionupdate". This is really rare case.  And also eCompositionChange event is exposed as DOM "text" event, but this is non-standard and I think that most web apps do not listen to this event. So, most web apps won't realize this behavior change.

If web apps realize this difference, it must be in the case of canceling composition when it receives "compositionupdate" event. However, it won't work well with Chrome (see the comment in testcase). So, most web apps must not do this.

[String changes made/needed]:
No.

Comment 15

[Noemi Erli[:noemi_erli]]

https://hg.mozilla.org/mozilla-central/rev/37e818a03a1b

Comment 16

[Liz Henry (:lizzard) (relman/hg->git project)]

Brindusa, can you or someone on your team verify the fix in nightly? Thanks!

Comment 17

[ershiwo]

(In reply to Liz Henry (:lizzard) (needinfo? me) from comment #16)
> Brindusa, can you or someone on your team verify the fix in nightly? Thanks!

I have test it on this nightly build: http://ftp.mozilla.org/pub/firefox/nightly/2018/07/2018-07-16-11-05-59-oak/
It looks fine.

Comment 18

[ovidiu boca[:Ovidiu]]

I tested this on Windows 10 x64 following the steps from comment 14 and I was able to reproduce the issue on a Nightly build without the fix. I also tested on Nightly 63.0a1(2018-07-18) and the issue is no longer reproducible. 
Based on the above information, I will mark this as verified as fixed on Nightly.

Comment 19

[Brindusa Tot, Desktop QA]

Based on the previous comment, remove the NI.

Comment 20

[Liz Henry (:lizzard) (relman/hg->git project)]

Comment on attachment 8991557 [details]
Bug 1462257 - TextComposition should dispatch eCompositionChange event when eCompositionCommit is being fired immediately after eCompositionStart

Fix for recent regression, verified in Nightly. 
Thanks for adding tests!
Let's uplift for beta 11.

Comment 21

[Andreea Pavel [:apavel]]

https://hg.mozilla.org/releases/mozilla-beta/rev/3743c968feeb

Comment 22

[ovidiu boca[:Ovidiu]]

I tested with FF Beta 62.0b11 on Windows 10 and I can confirm that this issue is fixed.