Forum user: blacktrash blacktrash.org

Ah, right, I missed that section on combining streams, sorry.

However, as you say, it does not work for the last clip - and, as, for demonstration purposes, I'm using only one, this is the last ;-)

Still, testing this reveals another bug! The scaling parameter suddenly fails to work.

Updated the test page again; on the whole I'd say, that this is quite broken :-(

I can't find a streams attribute, but I can find the playlist attribute for instream clips.

However, the resulting behaviour with an instream playlist is exactly the same. I've updated the test page accordingly.

Sounds like you're looking for this demo: how to make anchor links play a movie.

Yes, you have to prefix the url (the path inside your upload bucket) with mp4:, mp3:, or flv:. Probably tells the rtmp server what kind of stream to expect.

Why do you still need to hide the url? I don't think anyone has access to the file except those who also have access to the bucket, and it's not a full uri anyway.

P.S. Could you please write your code inside formatting tags, it's very hard to read. There's a formatting help popping up once you start editing a message. Thanks.

According to the docs flowplayer needs at least Flash 9.0.0. That's almost 9.0.115 ;-)

And there's the possibility to serve mp4 with flv fallback. Of course the files would have to be found ;-)

The one disadvantage of Flash: it's quite cpu hungry, and with mp4, depending on your compression settings, even more so.

You say "alongside" the video? Do you know the texts for each clip in the playlist in advance? If yes, you could work with cuepoints and in JavaScript only. This player displays the lyrics of the songs it is playing simultaneously. I'm telling you it's at the bottom of the page (nevermind the terse aesthetics), otherwise gmccomb will tell me again that no one can find the player ;-)

You could do the same with a video.

Looks like it has nothing to do with file suffixes, but that the file simply is not in the specified location:

~$ wget http://fitzvideo.com/pktest/pk_test_choice.html
--2009-11-17 02:08:26--  http://fitzvideo.com/pktest/pk_test_choice.html
Resolving fitzvideo.com (fitzvideo.com)... 66.226.69.155
Connecting to fitzvideo.com (fitzvideo.com)|66.226.69.155|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 553 [text/html]
Saving to: `pk_test_choice.html'

100%[==============================================================================>] 553         --.-K/s   in 0s      

2009-11-17 02:08:26 (14.3 MB/s) - `pk_test_choice.html' saved [553/553]

~$ wget http://fitzvideo.com/pktest/halfsizemovies/choice.flv
--2009-11-17 02:08:34--  http://fitzvideo.com/pktest/halfsizemovies/choice.flv
Resolving fitzvideo.com (fitzvideo.com)... 66.226.69.155
Connecting to fitzvideo.com (fitzvideo.com)|66.226.69.155|:80... connected.
HTTP request sent, awaiting response... 404 Not Found
2009-11-17 02:08:35 ERROR 404: Not Found.

You can see I got your html page but nothing on the url specified for your video.

I can see the video, it overlaps the overlay background on and extends further to the right. You probably have to adjust the css directives for the player container (height, width etc.) and possibly for the overlay background (padding).

I've made a sketchy outline on how you can implement thumbnails in this post.

The css z-index directive must be camelCased when used in JavaScript:

zIndex: 50

Also, setting it to 1 might be too low (screen and/or play button might take that already). Have tried with just leaving it out. And some other properties you are setting explicitly are at their default value: be lazy and just tweak what you actually want to tweak ;-)

I agree that not all api methods work as expected. It's sort of absurd to query the autoPlay attribute manually (not "auto" at all ;-) ). I guess autoPlay only has its intended effect when it is set before loading the player.

The problem with the flexibility of the plugin architecture is that the developers cannot imagine what crazy ideas the users will have ;-) See also this bug I opened, which imho is quite serious, as one is not prone to test every possibility oneself.

Anyway, I was just trying to propose a pragmatic solution to your problem.

Ok, I've tried to clean up my first stab example and make it more general. It's sort of a clip.load function (which might be nice to have):

window.onload = function () {
  var i, b = document.getElementsByTagName("button"),
  pl = [
    {url: "clip0.flv", autoPlay: false},
    {url: "clip1.flv", autoPlay: true}
  ],
  p = $f("player", "flowplayer.swf");

  function condplay() {
    // retrieve playlist index
    var c = pl[parseInt(this.id.substring(1), 10)];
    if (p.getState() > 1) {
      p.stop();
    }
    p.getClip(0).update(c);
    if (c.autoPlay) {
      p.play();
    }
  }

  for (i = 0; i < b.length; i = i + 1) {
    // give button id corresponding to playlist item
    b[i].id = "b" + i;
    b[i].onclick = condplay;
  }
};

<div id="player"></div>

<p>1st clip <button type="button">autoPlay: false</button></p>

<p>2nd clip <button type="button">autoPlay: true</button></p>

Wait! This works:

window.onload = function () {
  var b = document.getElementsByTagName("button"),
  pl = [
    {url: "clip0.flv", autoPlay: false},
    {url: "clip1.flv", autoPlay: true}
  ],
  p = $f("player", "flowplayer.swf", {
    clip: {
      onUpdate: function (c) {
        if (c.autoPlay) {
          this.play();
        }
      }
    }
  });

  function condplay(c) {
    p.setClip(c).getClip(0).update();
  }

  b[0].onclick = function () {
    condplay(pl[0]);
  };

  b[1].onclick = function () {
    condplay(pl[1]);
  };
};

• use clip.update() as trigger event
• the 0 as argument is mandatory for getClip(0), in case the playlist is empty

It works for me when I click on the grey area around the centered white area (Firefox, Safari, Opera, IE6). But, as I said in my bug report, this worked far better in version 1.1.1 of jquery.tools.

Have you tried onBeforeClose? Theoretic rationale: once the overlay is closed the player is not accessible.

Problem is, as you noticed, that different browsers have different load order. Note too, that in some situations Safari (or any other WebKit browser) crashes when unload is called; I filed a bug regarding this issue.