Revision as of 16:33, 31 July 2014

Known Issues with Character Encoding

• Need to verify that we can dump perl strings at XMLTV::Writer and it will do the right thing with regard to escaping anything outside $encoding into XML entities.
• tv_grab_hr tv_grab_no_gfeed tv_grab_se_swedb fix commited upstream
• #1910245 should add a test for HTML entities in the generated XML. (hint ´ is invalid XML!)

Issues with Time Zones

• tv_grab_dk_dr DST issues
• tv_grab_il DST issues
• tv_grab_it DST issues
• tv_grab_pt_meo DST issues
• tv_grab_uk_bleb DST issues, floating start>stop leads to wrong date calculation and time offsets

Memory Leaks??

• #2612996 tv_grab_na_dtv dumps core on windows as .exe, grows quite as perl

Cleanup List

Feel free to take anything from the list

• #1880681 the bug was solved (not a bug), the suggestion for Supplementary Files is turning one can of worms into another...
• tv_grab_huro has no maintainer?
  • #2748362 site changes: holes in the collected programs
  • #2837668 site changes: unexpected hash references
  • #2858285 close it, was "no channels found", the grabber does not fail completely anymore (see status)
  • #2910015 close it, was "no programs on channel 1", test_grabbers tests with that channel succesful

Cleanup SourceForge Project

• remove group/status/category examples from the tracker (might check for other unused stuff while there)

Check for Breakage caused by LWP::Simple

• silent uncompression
• silent code page conversion
• silent proxy handling

Maybe it's best to move most uses over to our own Get_nice.

Potential Data Sources

Candidates for wrapping into Static File Grabbers

• _cz_arcao: XMLTV export from arcao.com. Provides explicit time offsets.
• _dk_ontv: XMLTV export from ontv.dk.
• _eu_phazer: XMLTV service from tvprofil.net aka Phazer XMLTV Service. Notice that they provide timestamps in their local time as floating time which is intepreted as UTC...
• _fr_kazer: XMLTV service from kazer.org. done
• _it_ambrosa: XMLTV export from ambrosa.net. Explicit about non-commercial use only.
• _ru_teleguide: XMLTV export from teleguide.info. Provides explicit time offsets.

Configuration API

• [1] possible extensions/clarifications from a consumers POV
• list in the supplementary files mapping DVB/ATSC id to grabber/channel. Then let --list-channels & co. enrich channel list with related ids. MythTV's Browser Based Setup might be a consumer for this to allow automatic mapping of channels in the guide to channels on the video source.

Data Sinks

• Check [2] and see if all are mentioned here

Best Practices

Consumers of XMLTV Data

• be prepared that xmltv ids really might be similar to FQDN (255 characters max.) the longest I've seen in the wild is 69 characters (_es_laguiatv)

Random Pieces of Information

You might receive the same transport stream on multiple frequencies

TS 101 211 - DVB Guidelines on implementation and usage of Service Information (SI)

NOTE 1: The cell_id cannot be used to identify a service. The combination of service_id and original_network_id 
remains a unique identification of a service.

It is recommended to make all receivable multiplexes with the same transport_stream_id but with different
cell_ids available to the user, and only when a service (not a transport stream) is available through multiple
multiplexes to select a preferred multiplex based on e.g. reception quality.

Any reference resolution from a transport_stream_id or a service_id (e.g. from a linkage_descriptor
transport_stream_id/service_id pair) to a multiplex/frequency requires consideration to handle the potential multiplicity

Note that in networks deploying the service_availability_descriptor, the unique identification of a transport stream by
the tuple (transport_stream_id, original_network_id), can often be sensibly replaced by identification through the triplet
(transport_stream_id, original_network_id, cell_id).

Proper TV Metadata Schema Bits and Pieces (Hi TVBrainz)

• Some series have multiple sets of episode titles per locale, usually its one set per broadcasting company
  • the Exes http://www.fernsehserien.de/the-exes/episodenguide/staffel-1/16232
• Some series have a different title per season
  • Elephant Princess http://de.wikipedia.org/wiki/Elephant_Princess
• Some series have alternate titles
  • The Killing http://de.wikipedia.org/wiki/Kommissarin_Lund_%E2%80%93_Das_Verbrechen this series also uses roman or arabic numbers in addition to the title as season specific series title
• Usually the episode title is unique per series, but some series have multiple episodes with the same title
  • Lindenstraße http://de.wikipedia.org/wiki/Lindenstra%C3%9Fe/Episodenliste#Mehrfache_Folgentitel
• some series started on radio and continued on tv
  • Die Hesselbachs http://de.wikipedia.org/wiki/Die_Hesselbachs
• some series got rebranded
  • Pusteblume -> Löwenzahn http://en.wikipedia.org/wiki/L%C3%B6wenzahn
• many series don't do seasons
  • Eisenbahnromantik http://www.swr.de/eisenbahn-romantik/ (lots of repeats on bags of tv stations and lots of full episodes on youtube, good for testing VOD integration due to no regional restrictions http://www.youtube.com/user/Eisenbahnromantik )
• some episodes belong to multiple series
  • it is common for documentary brands to buy unrelated documentary movies and mini series that are run under a bigger brand
  • - example of the two episodes that belong to two series goes here -
• some series have a different order per country/station
  • Wickie und die starken Männer http://forums.thetvdb.com/viewtopic.php?f=41&t=18059
• some series contain the same episode multiple times / with multiple episode numbers
  • Eisenbahnromantik http://www.swr.de/eisenbahn-romantik/

The set of titles (series/season/episode) should be marked as being a true alternate or just a typo/search hint. Some titles are international (used for all languages without localization) titles - showing a poster of the international is better then defaulting to a specific locale.