Bug 56983 – [chromium] Implement preload=none, setPreload hooks to media player

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Victoria Kirst]

[chromium] Add setPreload hooks to Chromium media player

[Victoria Kirst]

Created attachment 86730 [details]
Patch

[Eric Carlson]

Comment on attachment 86730 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=86730&action=review

> Source/WebKit/chromium/ChangeLog:9
> +        setAutobuffer methods. This has been replace with

Typos: methods -> method, and replace -> replaced

[Victoria Kirst]

Created attachment 86851 [details]
Patch

[Eric Carlson]

Comment on attachment 86851 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=86851&action=review

r- for now as I am not sure the details are quite right yet.

> Source/WebKit/chromium/public/WebMediaPlayer.h:103
> +    virtual bool setPreload(Preload) { return false; };

Does this function really  need to return a bool?


> Source/WebKit/chromium/src/WebMediaPlayerClientImpl.cpp:190
> +    return static_cast<unsigned>(MediaPlayer::MetaData);

Don't you want to return m_preload here instead of hard coding MediaPlayer::MetaData?


> Source/WebKit/chromium/src/WebMediaPlayerClientImpl.cpp:216
> +    Frame* frame = static_cast<HTMLMediaElement*>(
> +        m_mediaPlayer->mediaPlayerClient())->document()->frame();

This indentation is odd, I thought at first that a stray tab had snuck in. I am not sure that the line break helps readability.


> Source/WebKit/chromium/src/WebMediaPlayerClientImpl.cpp:259
> +    if (m_delayingLoad)
> +        startDelayedLoad();

If you don't set m_delayingLoad to false before calling startDelayedLoad(), won't it ASSERT if this is called when m_preload == MediaPlayer::None?


> Source/WebKit/chromium/src/WebMediaPlayerClientImpl.cpp:466
> +    if (m_delayingLoad && m_preload != MediaPlayer::None)
> +        startDelayedLoad();

Ditto.

[Victoria Kirst]

Created attachment 87226 [details]
Patch

[Victoria Kirst]

Comment on attachment 86851 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=86851&action=review

Thanks for looking! I uploaded a new patch with the fixes - let me know whatever questions/concerns you still have!

>> Source/WebKit/chromium/public/WebMediaPlayer.h:103
>> +    virtual bool setPreload(Preload) { return false; };
> 
> Does this function really  need to return a bool?

No it doesn't. Fixed!

>> Source/WebKit/chromium/src/WebMediaPlayerClientImpl.cpp:190
>> +    return static_cast<unsigned>(MediaPlayer::MetaData);
> 
> Don't you want to return m_preload here instead of hard coding MediaPlayer::MetaData?

Yes, thanks! :) Fixed!

>> Source/WebKit/chromium/src/WebMediaPlayerClientImpl.cpp:216
>> +        m_mediaPlayer->mediaPlayerClient())->document()->frame();
> 
> This indentation is odd, I thought at first that a stray tab had snuck in. I am not sure that the line break helps readability.

OK! Changed to all on one line.

>> Source/WebKit/chromium/src/WebMediaPlayerClientImpl.cpp:259
>> +        startDelayedLoad();
> 
> If you don't set m_delayingLoad to false before calling startDelayedLoad(), won't it ASSERT if this is called when m_preload == MediaPlayer::None?

It shouldn't - the assert is for m_delayingLoad, not !m_delayingLoad (i.e. m_delayingLoad must be true whenever you first call startDelayedLoad).

[WebKit Commit Bot]

Comment on attachment 87226 [details]
Patch

Rejecting attachment 87226 [details] from commit-queue.

Failed to run "['./Tools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '--bot-id=cr-jail-3', 'apply-..." exit_code: 2

Last 500 characters of output:
s).
patching file Source/WebKit/chromium/src/WebMediaPlayerClientImpl.h
Hunk #1 FAILED at 76.
Hunk #2 succeeded at 88 (offset 1 line).
Hunk #3 succeeded at 108 (offset 1 line).
Hunk #4 succeeded at 131 (offset 1 line).
Hunk #5 succeeded at 144 (offset 1 line).
1 out of 5 hunks FAILED -- saving rejects to file Source/WebKit/chromium/src/WebMediaPlayerClientImpl.h.rej

Failed to run "[u'/mnt/git/webkit-commit-queue/Tools/Scripts/svn-apply', u'--reviewer', u'Eric Carlson', u'--for..." exit_code: 1

Full output: http://queues.webkit.org/results/8273869

[Victoria Kirst]

Since there were merge conflicts, I'll ask scherkus or someone to land by hand!

[Eric Seidel (no email)]

Attachment 87226 [details] was posted by a committer and has review+, assigning to Victoria Kirst for commit.

[Andrew Scherkus]

blah I landed this a while back:
http://trac.webkit.org/changeset/82641

apologies (again!) for not updating this bug

[Andrew Scherkus]

Comment on attachment 87226 [details]
Patch

Obsoleting patch since it was landed a while ago.