Skip to content

Loading…

Prefered flavor br #106

Merged
merged 88 commits into from

5 participants

@mdale

please review the test file passes:
/mwEmbed/modules/KalturaSupport/tests/FlavorSelector.preferedFlavorBR.qunit.html?runQunitTests=1

open issues, what is the api to access the selected flavor in flash?
the test account needs more webm flavors to highlight the selection functionality on firefox, ( tested with chrome )

closes #81 please review and merge.

Michael Dale and others added some commits
Michael Dale show controls on native player when .controls is enabled. c486a45
Michael Dale removed extra onplay event,
-- was a bug not a feature of the flash kdp.
231f39a
Michael Dale complete remove the old onplay event 351ac8f
Michael Dale refactored ad flow
Scenario 1: No ads in the clip
changeMedia-mediaReady-doPlay-0%........-100%-playerPlayEnd (there
should be no events related to the video after playerPLayEnd)

Scenario 2: Ad in the beginning of the clip
changeMedia-mediaReady-prerollStarted-adStart-adEnd-preSequenceComplete-doPlay-0%....
100%-playerPlayEnd

Scenario 3: Ad in the middle of the clip
changeMedia-mediaReady-doPlay-0%-.....20%-doPause-midRollStarted-adStart-adEnd-midSequenceComplete-doPlay....

Scenario 4: Ad at the end of the clip
changeMedia-mediaReady-doPlay-0%........100%-payerPlayEnd-postRollStarted-adStart-adEnd-postSequenceComplete

Scenario 5: Seeking occurred in the middle of the video play:
changeMedia-mediaReady-doPlay-0%....playerSeekStart-playerSeekEnd-....-100%-playerPlayEnd
1c19afe
Michael Dale fixed revision number 23595b2
Michael Dale added fallbacks for kRunMwDomReady 21a6776
Michael Dale fixed trailing , 5533ff4
Michael Dale fixed ogg check 493c064
Michael Dale fixed css issue for dependent overrides. b935996
Michael Dale revert of 7960a2f use version urls on
mwEmbed loader when behind a CDN we can't push updates without expires
2ddf1eb
Michael Dale bump version to e 917e629
Michael Dale updated to version 11.8 d3d7852
Michael Dale webm also missing container format 97a41ac
Michael Dale fixes container checks for broken entries 0f92da7
Michael Dale bumped to version .9 3b4ea68
Michael Dale don't restore the player if we never played an ad 464600f
Michael Dale bumped revions to 1.6.11.10 ( don't restore if we don't have a ad ) d84109e
Michael Dale removed debug c8b9850
@avivais avivais Fixed error handling when trying to capture thumbnails e88c536
@ranyefet ranyefet added comment ee5fbab
@ranyefet ranyefet Merge branch 'develop' of github.com:kaltura/mwEmbed into develop 53ed41f
@ranyefet ranyefet Added some comments and a little cleanup cd7637e
Michael Dale send correct expire headers for uiConf js service
-- check if in debug mode instead of uiConf cache state
d3974f2
Michael Dale fixed doubleClick add orientation 1a07e31
@mdale

I think your just talking about the direct download here... but I guess with the new direct mapping api ... we could consider putting the kaltura player flash player in the iframe as well...

Yeah i was talking about the direct download.

@mdale

Math.round( Math.random() ); is only 1 or 0 ... is that what you want here?

Opps. I'll fix it

@mdale

were all these embed method test pages tested with the new kWidget.embed ( I assume they worked ) ?

Haven't tested. But they shouldn't be affected. I only deleted the direct download fallback because kWidget.embed already handle that,

@mdale

sounds good.. .its a lot of inline injections .. but could certainly be cleaned up.

@mdale

its not nessesaraly a new embedPlayer type.. .just a missingPlayer type.

It's the "Link player" type for BlackBerry support so we could emulate the JS api. anyway it doesn't really matter how we call it :P

@mdale

looks good

@mdale

I assume this is consistent with a replay for flash kdp where you have mediaProxy.playFrom ... then on replay it starts at zero? ... would be great to have a test file that validates this behavior for both flash and html5 ;)

Not only for reply. also when you seek.
Anyway yes, this is onsistent with the flash kdp.
We do have "mediaPlayFromTo.html" test file that works on both player (or should be)

@mdale

the .apply call requires .apply( context, arrayOfAguments )

seems to be called there just with the arguments not the context object, it would be cleaner than the list of array items. ( all of which may not be active for a given call )

@mdale

.toString() won't work on an undefined or null value.. which is why I like String( var ) a little better if we are not sure about type ahead of time

@mdale

your doing a == comparison against a type string.. its going to .toString() the value anyway ;)

@mdale

thanks for the modernization of the code Ran! ... keep in mind the original was written to support both flash and html5. We still need to better define some sample onPage plugins for doing that.

@avivais did all the work. I've just added the getConfig method to grab configurations.

@ranyefet

Not tested... just something i did in a few minutes. I'll continue on Sunday.

remember branches are our friend, since this commit will change a bunch of files once we switch them all over to use this new method.

@mdale

We can now use embedPlayer.getKalutraConfig( 'varName' ); instead of getFlashVars .. I think getKalturaConfig makes it a little more clear whats going on. With get flashvars it may not be clear that it can include uiVars as well.

does this mean we do not display the errors if getFlashvars function is not defined? ... I think it would be preferable to make this check this at the playerControls level and call the same display alert code that we use elsewhere? .. and we should display the alert if we are not able to make the check config check.

Michael Dale and others added some commits
@mdale

doPlay and doStop are valid events.. is there a reason we did not just ~add~ 'playerPlayed', and 'playerPlayEnd', .. instead of replacing doPlay and doStop?

Nu's implementation uses playerPlayed and playerPlayEnd, which are fired only once. For some reason, doPlay is fired twice on KDP

@mdale

ouch .. nice catch.. we should run all our code through js-lent on each commit as well.. I will work with Andrew to set that up.

@mdale

is this a use strict issue or just aesthetic change ? lets try and make a note in the comments when we make changes like this.

I guess it for strict mode.

@mdale

We already have this binding on https://github.com/kaltura/mwEmbed/blob/develop/modules/EmbedPlayer/skins/mw.PlayerControlBuilder.js#L606 ... and I believe we only want to bind this for pseudo full-screen not native fullscreen, since native fullscreen has its own non-overridable binding to exist fullscreen.

Did you find this not working? or does this help address the firefox lack of exit fullscreen event issue? .. Either way we should not have it twice ..if its more robust to apply it all the time ( then lets remove the one on line #L606 )

It didn't work on IE9 so I added that here as well.

Please make a note of this rational as a comment in the code. And Or test if we can remove the duplicate binding and just use the one that works for IE.

@mdale

Please remember to put any sizable commit into a branch instead of directly commit to develop.

The captions selection seems to work on kalturaSupport/test/CaptionsKalturaApi.qunit.html

@mdale

We should better document what return false vs return true means here or consistently return null if we just want to exit the method

@mdale

this actually "fixes" quartiles events for iOS4 ? .. i.e IO4 string casting is messed up ??? ... please make a note in the code, in the form of a comment if that is the case. I would be surprised if that is the case.

Looks like it. Without an explicit casting the comparison fails (Tested on iOS 4.3.1)

Michael Dale added some commits
Michael Dale added play content fallback for failed double click request
-- sometimes CONTENT_PAUSE_REQUESTED is the last event we receive from
the double click plugin. We now give DC 6 seconds to load the ad, else
continue to content playback
953f5dd
Michael Dale also check for "STARTED" event, "LOADED" is not always fired either or
is fired before the "CONTENT_PAUSE_REQUESTED"
49a78b6
Michael Dale increased timeout to 10 seconds..
-- seems there can be playback after 6 or 7
b2a213c
@KalturaGitBot

Live Pull Request Urls

avivais and others added some commits
@avivais avivais Added comment c6a55ec
@avivais avivais Fixed WIDGET_LOADED misfire QC 2030 2f33ba8
Michael Dale adds buffer underrun checks and content resume fallback
-- double click apparently has no built in resume mechanism from a
content pause from buffer underun, we add some fallbacks
-- content resume seems to fails sometimes, added an asynchronous
content resume call
b9c790c
Michael Dale adds buffer underrun checks and content resume fallback
-- double click apparently has no built in resume mechanism from a
content pause from buffer underun, we add some fallbacks
-- content resume seems to fails sometimes, added an asynchronous
content resume call
ee3879a
Michael Dale Revert "adds buffer underrun checks and content resume fallback"
This reverts commit b9c790c.
f9a5b78
Michael Dale Merge branch 'develop' of git@github.com:kaltura/mwEmbed.git into
develop
ea91ee9
@mdale

Note I reverted because this was committed to the wrong branch, should have been committed to the doubleClickNoContentResume branch done here ee3879a

@mdale

We should try to avoid repeat conditional logic ... instead call a method that has that logic... i.e this.triggerWidgetLoaded() ? ... if they are doing the same thing ... or if we can capture the entire playReady event in the same block .. do this.triggerPlayerReady >? .....

ideally we don't have two playerReady's to begin with ;) ... i see that its related to dependency of playerReady being triggered for the changeMedia sequence... maybe we should use a different event for that.. like "setupSourcePlayerDone" or something? ( but that is of course a larger refactor )

@mdale

hmm switched branches inappropriately did not mean to revert the earlier commits restored: b2a213c

@mdale

Avi .. next time please do all re-factoring commits in a branch .. then we can merge once tested.

@mdale

test

diffalot and others added some commits
@KalturaGitBot

Updated pull request 106

@mdale mdale merged commit 73dfde4 into develop
@mdale

Unfortunate "merge issue" ... this was removed in 7419dba 3 days earlier. Yet this merge looks like it was either not-re-based or not updated with the most recent code before it was pushed to the branch we were tagging from :( This results in recurrence of the .find() error that 7419dba addressed :(

@mdale

@ranyefet did we ever address the questions in the commit. Disable alerts should be supported both as a flashvar and as a uiConf var .. i.e grabbed from playerConfig .. and this check should probably be in the showAlert function .in mw.PlayerControlBuilder .. not mw.EmbedPlayer

@einatr einatr added a commit that referenced this pull request
@einatr einatr new HLS plugin built from 2.32 branch (fixes for FEC-3582, FEC-3415 a…
…nd PLAT-2016) includes PRs #106, #107, #108, #109 and #118
682cb59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Commits on Apr 30, 2012
  1. show controls on native player when .controls is enabled.

    Michael Dale committed
  2. removed extra onplay event,

    Michael Dale committed
    -- was a bug not a feature of the flash kdp.
  3. complete remove the old onplay event

    Michael Dale committed
  4. refactored ad flow

    Michael Dale committed
    Scenario 1: No ads in the clip
    changeMedia-mediaReady-doPlay-0%........-100%-playerPlayEnd (there
    should be no events related to the video after playerPLayEnd)
    
    Scenario 2: Ad in the beginning of the clip
    changeMedia-mediaReady-prerollStarted-adStart-adEnd-preSequenceComplete-doPlay-0%....
    100%-playerPlayEnd
    
    Scenario 3: Ad in the middle of the clip
    changeMedia-mediaReady-doPlay-0%-.....20%-doPause-midRollStarted-adStart-adEnd-midSequenceComplete-doPlay....
    
    Scenario 4: Ad at the end of the clip
    changeMedia-mediaReady-doPlay-0%........100%-payerPlayEnd-postRollStarted-adStart-adEnd-postSequenceComplete
    
    Scenario 5: Seeking occurred in the middle of the video play:
    changeMedia-mediaReady-doPlay-0%....playerSeekStart-playerSeekEnd-....-100%-playerPlayEnd
  5. fixed revision number

    Michael Dale committed
  6. added fallbacks for kRunMwDomReady

    Michael Dale committed
Commits on May 1, 2012
  1. fixed trailing ,

    Michael Dale committed
  2. fixed ogg check

    Michael Dale committed
  3. fixed css issue for dependent overrides.

    Michael Dale committed
  4. revert of 7960a2f use version urls on

    Michael Dale committed
    mwEmbed loader when behind a CDN we can't push updates without expires
  5. bump version to e

    Michael Dale committed
  6. updated to version 11.8

    Michael Dale committed
  7. webm also missing container format

    Michael Dale committed
  8. fixes container checks for broken entries

    Michael Dale committed
  9. bumped to version .9

    Michael Dale committed
  10. don't restore the player if we never played an ad

    Michael Dale committed
  11. removed debug

    Michael Dale committed
  12. @avivais
  13. @ranyefet

    added comment

    ranyefet committed
  14. @ranyefet
  15. @ranyefet
  16. send correct expire headers for uiConf js service

    Michael Dale committed
    -- check if in debug mode instead of uiConf cache state
  17. fixed doubleClick add orientation

    Michael Dale committed
  18. Merge pull request #99 from kaltura/adSequenceSync

    Michael Dale committed
    Ad sequence sync Ran tested VAST, more testing will be done in develop.
  19. restored version tag to 1.6.12alpha

    Michael Dale committed
Commits on May 2, 2012
  1. @ranyefet
  2. @ranyefet

    unbind if we have bind name

    ranyefet committed
  3. @ranyefet
  4. @ranyefet
Commits on May 3, 2012
  1. define resultObject prior to try catch call

    Michael Dale committed
  2. explicitly check for null result object

    Michael Dale committed
  3. bumped version to 1.6.11.11

    Michael Dale committed
  4. Merge branch 'develop' into ResultObjectRecurseFix

    Michael Dale committed
    Conflicts:
    	includes/DefaultSettings.php
  5. Merge pull request #100 from kaltura/ResultObjectRecurseFix

    Michael Dale committed
    Result object recurse fix
  6. build the captions menu after captions are ready

    Michael Dale committed
    left align frewheel restore interface
  7. removed duplicate break;

    Michael Dale committed
  8. fixed _this ref in timed text setup callback

    Michael Dale committed
  9. freewheel should call updateUiForAdPlayback directly

    Michael Dale committed
  10. bumped version to 1.6.11.12

    Michael Dale committed
  11. opps, one more _this

    Michael Dale committed
  12. @ranyefet

    merge GoogleAnalytics fixes

    ranyefet committed
  13. fixed convertTTML2HTML method

    Michael Dale committed
  14. added empty player kWidget test

    Michael Dale committed
  15. @ranyefet
  16. removed debug line

    Michael Dale committed
  17. comment clean up

    Michael Dale committed
  18. @ranyefet
Commits on May 4, 2012
  1. improved code documentation for kWidget

    Michael Dale committed
    -- made embed calls more uniform in calling arguments 
    -- also replaced some instances of kEmbedSettings with 'settings'
  2. Merge pull request #103 from kaltura/deploy_vp2k

    Michael Dale committed
    Deploy vp2k, includes updates to freewheel plugin for new ad timeline and minor fixes to close captions
Commits on May 5, 2012
  1. close #96 IE9 should not leadWithHTML5, and instead use flash.

    Michael Dale committed
    -- if you want html5 with IE9, use the forceMobileHTML5 flag.
Commits on May 6, 2012
  1. @avivais
  2. @avivais

    Merge branch 'develop' of https://github.com/kaltura/mwEmbed into dev…

    avivais committed
    …elop
    
    Conflicts:
    	modules/TimedText/mw.TimedText.js
  3. @ranyefet
  4. @avivais
  5. @ranyefet
Commits on May 7, 2012
  1. @avivais
  2. @avivais
  3. @avivais
  4. added play content fallback for failed double click request

    Michael Dale committed
    -- sometimes CONTENT_PAUSE_REQUESTED is the last event we receive from
    the double click plugin. We now give DC 6 seconds to load the ad, else
    continue to content playback
  5. also check for "STARTED" event, "LOADED" is not always fired either or

    Michael Dale committed
    is fired before the "CONTENT_PAUSE_REQUESTED"
  6. increased timeout to 10 seconds..

    Michael Dale committed
    -- seems there can be playback after 6 or 7
Commits on May 8, 2012
  1. @avivais

    Added comment

    avivais committed
  2. @avivais
  3. adds buffer underrun checks and content resume fallback

    Michael Dale committed
    -- double click apparently has no built in resume mechanism from a
    content pause from buffer underun, we add some fallbacks
    -- content resume seems to fails sometimes, added an asynchronous
    content resume call
  4. adds buffer underrun checks and content resume fallback

    Michael Dale committed
    -- double click apparently has no built in resume mechanism from a
    content pause from buffer underun, we add some fallbacks
    -- content resume seems to fails sometimes, added an asynchronous
    content resume call
  5. Revert "adds buffer underrun checks and content resume fallback"

    Michael Dale committed
    This reverts commit b9c790c.
  6. Merge branch 'develop' of git@github.com:kaltura/mwEmbed.git into

    Michael Dale committed
    develop
Commits on May 9, 2012
  1. set source of player on init

    Michael Dale committed
    -- fixes iPad native controls
    -- adds test for disabling iPhoneShowHTMLPlayScreen
  2. updated test name disable show html play screen

    Michael Dale committed
  3. @ranyefet

    Merge pull request #111 from kaltura/iPhoneShowHTMLPlayScreen

    ranyefet committed
    IPhone show html play screen
  4. @avivais
  5. @avivais
  6. @ranyefet
  7. @avivais
  8. Merge pull request #110 from kaltura/doubleClickNoContentResume

    Michael Dale committed
    Double click no content resume, fixes no content resume issue for double click ad plugin on some ads.
  9. typo fix

    Michael Dale committed
Commits on May 10, 2012
  1. @avivais

    Added uiConf

    avivais committed
  2. @avivais
  3. follwo up to 95694bc include tag in

    Michael Dale committed
    text area
Commits on May 11, 2012
  1. @diffalot
Commits on May 12, 2012
  1. Merge branch 'develop' into preferedFlavorBR

    Michael Dale committed
    ( ensures updated inject file for automated testing )
Something went wrong with that request. Please try again.