1369696 - [TSF] ITfMessagePump and ITfKeystrokeMgr shouldn't be QIed from ThreadMgr until first use

Status: RESOLVED FIXED

(Core :: Widget: Win32, defect, P2)

Comment 1

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

https://treeherder.mozilla.org/#/jobs?repo=try&revision=c5172d58b333b895bebe22e3bef1121e576bcbcb

Comment 2

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

https://treeherder.mozilla.org/#/jobs?repo=try&revision=b31c30034bb8a01a08de2260c5ae94d57809dbd3

Comment 3

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

Attachment: Bug 1369696 - part 1: TSFTextStore::sKeystrokeMgr should be QIed from TSFTextStore::sThreadMgr at first use

sKeystrokeMgr is required only when WM_KEYDOWN or WM_KEYUP message are received.  So, we can put off to initialize sKeystrokeMgr until then.

Review commit: https://reviewboard.mozilla.org/r/148818/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/148818/

Comment 4

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

Attachment: Bug 1369696 - part 2: TSFTextStore::sMessagePump should be QIed from TSFTextStore::sThreadMgr at first use

sMessagePump is necessary only when WinUtils::(Get|Peek)Message() retrieves a message from the queue.  Therefore, we can put off to initialize it until then.

Review commit: https://reviewboard.mozilla.org/r/148820/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/148820/

Comment 5

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

Looks like no perf regression:

part1: https://treeherder.mozilla.org/perf.html#/compare?originalProject=try&originalRevision=6f6883c950cb&newProject=try&newRevision=cc2b6782fb3c&framework=1&showOnlyImportant=0
part2: https://treeherder.mozilla.org/perf.html#/compare?originalProject=try&originalRevision=6f6883c950cb&newProject=try&newRevision=7469639e8770&framework=1&showOnlyImportant=0

The latter shows some slowness only on x86 build, but they disappeared checking with additional patch for another bug: https://treeherder.mozilla.org/perf.html#/compare?originalProject=try&originalRevision=6f6883c950cb&newProject=try&newRevision=e908cc391657&framework=1&showOnlyImportant=0

And according to the graph,
https://treeherder.mozilla.org/perf.html#/graphs?highlightedRevisions=6f6883c950cb&highlightedRevisions=7469639e8770&series=%5B%22try%22,%22dfac9bac6ca266e9075c959c581774267e881d26%22,1,1%5D&timerange=86400
https://treeherder.mozilla.org/perf.html#/graphs?highlightedRevisions=6f6883c950cb&highlightedRevisions=7469639e8770&series=%5B%22try%22,%22bf3413f2f08733f70710103dc0827f715b51ce08%22,1,1%5D&timerange=86400

These two reds are caused by too fast result of the base revision.

And according to this graph,
https://treeherder.mozilla.org/perf.html#/graphs?highlightedRevisions=6f6883c950cb&highlightedRevisions=7469639e8770&series=%5B%22try%22,%22456abe18d122f7a723b75aef5ed9c175481735c6%22,1,1%5D&timerange=86400

The orange result of new revision isn't so slow comparing with other results.

So, they just hit random result.

Comment 6

[Makoto Kato [:m_kato]]

Comment on attachment 8877426 [details]
Bug 1369696 - part 1: TSFTextStore::sKeystrokeMgr should be QIed from TSFTextStore::sThreadMgr at first use

https://reviewboard.mozilla.org/r/148818/#review153422

Comment 7

[Makoto Kato [:m_kato]]

Comment on attachment 8877427 [details]
Bug 1369696 - part 2: TSFTextStore::sMessagePump should be QIed from TSFTextStore::sThreadMgr at first use

https://reviewboard.mozilla.org/r/148820/#review153426

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/96172f97bf0c
part 1: TSFTextStore::sKeystrokeMgr should be QIed from TSFTextStore::sThreadMgr at first use r=m_kato
https://hg.mozilla.org/integration/autoland/rev/ec9869a14c67
part 2: TSFTextStore::sMessagePump should be QIed from TSFTextStore::sThreadMgr at first use r=m_kato

Comment 9

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

https://hg.mozilla.org/mozilla-central/rev/96172f97bf0c
https://hg.mozilla.org/mozilla-central/rev/ec9869a14c67