[yabsie]

Now, the bad news; while I expect most of the fault here lies with Netflix, and I'm sure whoever's putting in their valuable time to code this thing is facing an uphill struggle, the level of accessibility from with in XBMC is pretty bare. So far all that works for me are the New Arrivals sub-sections. Everything else returns blank, except for Search, which returns a script error that I'm in no mood to start troubleshooting now. I can play titles from within New Arrivals, but afterwards I have to Alt-F4 out or Alt-Tab back, due to Netflix having to run inside a full screen browser windows. That's just a disclosure from my end; maybe someone else will have better luck.

Yup thats where it is, thats the current state of the plugin, IF you can login you can only watch new arrivals, either from "all " or "top 25"

with my hack (above) you can browse by gener but not actually play the movies. so thats something, right? right?

Does no one who uses XBMC care about netflix, my guess is hardly anyone does, if you are using xbmc you are probably watching your own downloaded content and dont really care much about netflix. So like someone said a few days ago, if we want to get this thing working at all we need to do it ourselves.

A small ragtag bunch of plucky programmers band together and build something almost usable! Hi-Ho python AWAY!

[yabsie]

[quote='elmerohueso' pid='1219366' dateline='1350855510']
Known issues. Fixes from PatDavid start on page 100.

Edit reason: Removed enormous quote and expired link.

...
....
...



I dont know python well and I don't know this code at all, so any help would grand.

Since there appears to be no active development on this, I decided to bite the bullet and setup a python environment hooked up the debugger and tried to figure it out.

So it look like the Genre and experimental search are suing the oData APi and the other code is using the older version of the api.

I managed to figure out the proper regular expression to get the ID out of the matcher that was throwing the "CRITICAL" only to discover that the ID's returned by the odata API are different than the ID used by the player. the odata ids are 6 digit alpha numeric and the 'old' IDs are 10 digit numeric, so even though I manage to fill my lists in the genre search I dont have the 'real' ID for movie playback

so sad

So we need to
1) figure out how to map odata IDs to original IDs
2) figure out how to play via an odata id
3) figure out how to take the other meta data (year title etc) and do another search on the old school api to get the old id

option 3 sucks, because if you list 500 movies per odata genre search and did another search for each movie, you only get list a genre 50 times a day and you blow your 2500 api limit and netflix shuts you down

so 3.5 is my new idea...

3.5)
Right now the we read the odata stream, and create a link file for each movie, that link "ID".html and is essentially the player. When you launch file, you launch the player, (actually it contains a html redirect to netflix with the url of the movie with the numberic ID)

When you click on the link in XBMCFLix genre list, it simnply fires off the file in your browser and away you go to movietown. (when it works)

BUT

if we were to change what happens you build the list and click the link. Instead of linking directly off to the movie we take the stored up meta data for the movie , launch a search for the ID in the RESTful api (based on year name etc) , get the old school ID , write the linkfile, and launch that.

We could also save the mappings in a file, so we have them and only have to search one time,

Now to be even more nutso, imagine someone hosts a websevice somewhere in cloudville, that maps odata IDs to numberic IDs we could have the plugin DL the mappings, and when it creates one that it doesnt know yet it it posts it to webservice, so the next guy will have it, The list will start out small but it could filll up quicky. but thats a bit over the top, but a local mapping file would be smart.

So, I have been a xbmc plugin deveeloper for less than 12 hours, anyone who knows more about this wanna work with me and figuring out how to implement #3.5?

[Seymour_Bootay]

Glad to see I am not the only one having the script error and blank file issue. Thank you for posting the sturcture of the file. I am going to try this out tomorrow. I am a noob to xbmc Addon development, but I make my livelihood programming. I would be interested in helping with development of this Addon, just not to sure how much time I will have available between family obligations and work.

[xm41907]

I've installed the script and successfully linked it to my Netflix account, but when I try to play a video, it doesn't bring up Internet Explorer. It just sits there like I didn't even press a button. I've turned on the debug option in the addon's configuration, but I'm not sure where to access the log to upload it here. Any thoughts on why IE won't open? I do have Sliverlight installed, and can play Netflix fine through the browser, it just isn't opening up.

[Seymour_Bootay]

The log to post is xbmc's log located here: %APPDATA%\XBMC\ for windows. For a listing of locations for other OS's and more details on the log please refer to this link http://wiki.xbmc.org/index.php?title=Log_file#Location

[yabsie]

nay-nay-boo-boo! I have working genres on my machine and you dont!

Its magical hacktastic, and tool a bucketload of code changes.

I think we need to officially fork this project, I will look at checking it into Google code or some place like it.

If we are forking we need a new name, I am suggesting "YabsieIsSuperCoolFlix" but ya know, whatever (no I am not 12 : )

But seriously folks I have a working prototype of genres, I am looking at getting a narrow testing release (via a zip ball in this forum ) in a week or two)

Any other bugs anyone hate in this thing?

[yabsie]

Ok I am grumpy again.

So this whole thing is written with specifically requesting json output, and then using a series of complicated regular expressions to tweak out the data. This IMHO is a fragile approach is is probably why it was never really finished.

BUT if we where simply to stick the with the default output of XML, and use the element tree xml parser that comes in python, we could have nice formatted items we can cleanly an reliably parse.

So I am rewriting a lot of the parsing of the movie/Series data, and a lot of the work that I have been doing all weekend is getting tossed, but at least I understand what needs to be done, and I see no reason I cant eventually get a fully working version of this puppy going. Well, I see one reason: TIME Like I said before I have a full time job, another part time coding job and 3 kids, 2 of which play hockey several hours a week, I will do what I can. Sleep is for chumps right?

It may be more than a week before I get it released.

Anyone come up with a good name yet?
Anyone actually reading any of this or am I talking to a wall?

[Jason G]

You're not talking to a wall. I'm watching with interest, and cautiously optimistic, but had previously been so fed up with this plugin that my setup is now switching out of XBMC, into WMC to use their netflix plugin.

[yabsie]

You're not talking to a wall. I'm watching with interest, and cautiously optimistic, but had previously been so fed up with this plugin that my setup is now switching out of XBMC, into WMC to use their netflix plugin.

I went down that road, but WMC plugin only ever offers 75 items, which to me was almost useless I couldn't figure out how to browse by genre so I bailed on it and started hacking away at this

[hiorik]

yabsie if you need help, please ask

[huttdes]

Does no one who uses XBMC care about netflix, my guess is hardly anyone does, if you are using xbmc you are probably watching your own downloaded content and dont really care much about netflix. So like someone said a few days ago, if we want to get this thing working at all we need to do it ourselves.

For me personally it doesn't mean much but my little daughter likes to watch little girl type tv shows and movies on it.... so, well, yeah.... I care :-)

[jasapper]

You're not talking to a wall. I'm watching with interest, and cautiously optimistic, but had previously been so fed up with this plugin that my setup is now switching out of XBMC, into WMC to use their netflix plugin.

+1 for me! Heck I finally got up off my lazy butt and joined the forums so I wouldn't miss any of this promising development. Okay so I was also tired of typing the captcha everytime I fruitlessly searched for anything else for Netflix. And then I found the Amazon Instant plugin which I suspect will be getting alot of use while my patience continues here.

[Sirrico]

It may be more than a week before I get it released.

Anyone come up with a good name yet?
Anyone actually reading any of this or am I talking to a wall?

I'm watching the thread as well; unfortunately my knowledge of any manner of code stops after HTML and a dash of Java, so I'm pretty much useless to development except maybe as a tester. While Netflix integration into XBMC probably doesn't seem like a huge deal, I expect I'm not that only one that finds it frustrating that Hulu, Amazon, and other services integrate basically at the push of a button, and Netflix is the one major provider that just won't make nice with an other wise sleek system. And unless I'm misreading the various archives, there was definitely a time when it did work properly. So more power to you Yabsie, whatever time you feel like sparing is appreciated.

[clizarr]

I am also waiting on a fix for eden on mac OSX. non of the pages are correct for file location etc. perhaps one has to do Show Package Contents.
I hear lots of people building XBMC on Mac Mini but without Netflix I cannot bring this live in my home.
I am willing to try beta code etc. I am not a developer but I am willing to use my hardware to Beta TEST etc.

[xbmcsnapper]

Anyone come up with a good name yet?

How about NetFlux? Seems appropriate given the frequent changes to the API.