ZANG!!! I'm writing my article now so I'll include that little tidbit (and change my config file). Should I point people to get the latest svn code or just tell them to get the 2013 release?

Sure, no sweat! It doesn't seem like this would really affect many users, so definitely better to wait until you are able to work it out correctly. I may take a deeper look at the code myself, and maybe suggest further changes (though it's hard if I can't test the main use-case) -- I'm glad that you have no problem just pointing out when the "fix" is wrong .

I just tried it without login info and I could only get the 550K condensed game for yesterday's Nats v Cards game. Should I try a different game?

Have you set free_condensed back to 0?

Thanks, dovf, I completely missed that.

I'm writing my article and I realized that the mlbhls site doesn't have any downloads available anymore. I'm going back through the older posts here but can someone let me know how I should tell my readers to get it? I still had the code from last year so I didn't realize this issue had arisen.

See the REQUIREMENTS file in mlbviewer -- it seems like you need to download mlbhls directly from the experimental branch on svn. (I have no experience with this, just happened to notice this in the REQUIREMENTS file today...).

That's what I get for having so many text files open at once. Thanks again.

So I did a little bit of snooping and testing and it looks like, except for that one case, librtmp is actually there in the code and working (sort of) as expected. I may have to recompile my mplayer2 to get some of the streams working correctly, but I can see that the code is at least trying to format the command string correctly.

So I think what I'll end up doing is deprecating free_condensed and update the documentation.

I'm not too concerned about the login issue at startup with null user and pass. I'm glad it lets you continue anyway. But a more permanent fix would be to handle it in MLBSession (something like return None if user/pass blank.) A more elegant fix would be to re-implement the DISABLED_FEATURES list and keep non-subscribers away from those features I know won't work like game audio and game video. This way we won't nag their servers with null logins and requests we know will fail.

Hopefully you find the code a little more readable , a bit better organized after the rewrite (except maybe the media parsing which was largely untouched.)

3 is misleading. mlbhls works fine for both basic and premium (but you need the use_wired_web option for basic). rtmpdump also works for basic, as it presumably does also for premium.

Let me rephrase that.

mlbhls is required for nexdef streams
rtmpdump is required for non-nexdef streams and condensed games

Better?

free_condensed has been deprecated (I removed the selection of the 3GP stream and override it to False for good measure after config file is read) and the documentation has been updated with regards to leaving user= and pass= blank for non-subscribers.

I haven't fixed librtmp yet nor have I fixed the login error. But for now, if nothing changes between now and Opening Day, I would consider this a pretty good Opening Day release.

It always gets more testing and more bugs/errors after Spring Training, anyway.

Now I really have to leave for airport but I'll have plenty of time on their wifi to see if Sourceforge has their problems sorted enough to update the download tarball.

Thanks, daftcat! And yes, I certainly don't think that any of the issues I'm raising --- as they are today, unaddressed --- would change this from being a great Opening Day release!

So, for when you have time to get back to this:

1. How come you left any mention at all of free_condensed in the code -- why not just remove it completely?
2. I'm revisiting my suggested fix from yesterday regarding librtmp (below, modified against the latest changes): note that this would only affect users who have use_librtmp=1; given your comment in the docs about librtmp: "this may not be working yet in 2013. If you want it, ask for it...", I would expect that anyone having use_librtmp=1 is already in a gray zone of maybe working, maybe not? BTW, what is the relation between nexdef and librtmp -- are they mutually exclusive, or not necessarily?
3. Do you have any objections to my publishing a mercurial mirror of mlbviewer on bitbucket?

Meanwhile, have a safe flight and nice weekend!

Looks fine to me.

1. Quick fix was to deprecate it (e.g. remove the difference in stream selection) and an added safety was to override it after cfg file is read. The only reason why I didn't remove it completely is because I put it some if/elif/else code that I haven't had a chance to de-tangle.

2. I don't know if I want to put that patch in only because I haven't had the chance to test all the use cases and the bug seems to only affect condensed games. There is code elsewhere (presumably locateMedia() of which locateCondensedMedia() is an early shortcut within that procedure) that handles librtmp.

In other words, your use case is much smaller compared to other cases that can be affected by your patch.

In more blunt terms, the patch can do more harm than the bug.

At this point, I would say if you can use rtmpdump and you cannot justify to me why you would rather not use rtmpdump, wait it out until I have the time to put a proper fix in.

2a. librtmp/rtmpdump and nexdef are mutually exclusive except condensed games. There are no nexdef streams for condensed games. Rtmpdump or librtmp is exclusively used for condensed games unless you want the lower quality 3GP streams.

3. I don't know what this means to have a mercurial mirror on bitbucket. Does it require any extra effort on my part? Could it result in different versions of the code residing on the mirror than what's on sourceforge? Will there be different versioning schemes? Again, like the patch above, it sounds like it could result in more work for me than it's worth for everyone else. It's GPL so you can do with it what you want but I'm not likely to do anything but maintain the sourceforge versions.

If anyone has been having their game interface crashing these last few days, this patch seems to fix it. When the patch is applied, it might apply a line or two off, as I have also patched the file to allow 800kb/s streams which MLB.TV are providing that on live streams this season.

The patch might also come in handy for the all star game.

Attached Files

interface.txt (1.1 KB, 9 views)