Add switch to disable progressive support

[osagie98]

1. Adds a runtime switch "--disable_progressive_playback" that will prevent progressive media playback. When set, Cobalt will reject CanPlayType() queries for progressive formats, or will report an error before a ProgressiveDemuxer is constructed.
2. Adds an H5vcc setting "DisableProgressivePlayback". The web app can call window.h5vcc.settings.set("Media.DisableProgressivePlayback", true) to disable progressive playback, and window.h5vcc.settings.set("Media.DisableProgressivePlayback", false) to reenable it.
3. Rearranges methods in CanPlayTypeHandlerStarboard to match the declaration order in CanPlayTypeHandler.

b/382791540

[jasonzhangxx]

My understanding of the "runtime switch" is to enable/disable progressive support via h5vcc, but the pr uses command line switch. @xiaomings to correct me if that's not correct.

[borongc]

We may consider to use h5vcc switch --> base_feature --> enable/disable media feature, as base_feature is the way Chromium uses to enable/disable feature. (cc @yell0wd0g)

We should define how we are going to add switch for Cobalt.

[osagie98]

I can add an H5vcc switch to this PR

[osagie98]

The H5vcc switch is in. @borongc are you suggesting we use BASE_FEATURE for 25.lts?

[borongc]

I got confused with this PR #4547.

For this PR for C25, we should use h5vcc flag for this.

[xiaomings]

Maybe we don't need this switch in C26, given that we are going to experiment on C25 first. We can keep progressive support in C26 (without the disable switch), and remove the support completely when the C25 metrics is in?

[osagie98]

That should be fine - the C26 change is small and can be easily merged anytime if we need to

[osagie98]

cc @yell0wd0g shall we add the progressive switch in C25 first in case we decide to remove support entirely in C26?