EPG123 Bug Reports

An evolving, supported alternative to Rovi
Forum rules
★ Download the latest EPG123 here: https://garyan2.github.io/ <> Setup guide here: https://garyan2.github.io/install.html
stuartm

Posts: 721
Joined: Mon Nov 05, 2012 8:05 pm
Location: Longmont, CO

HTPC Specs: Show details

#501

Post by stuartm » Tue Dec 27, 2016 7:28 pm

I think something similar has been reported before but I couldn't find the post. I run my guide update pretty late at night. During time off for the holidays I happened to be watching TV late one night during the time EPG123 normally runs its update. After I had finished watching TV and shut down WMC the next time I tried to watch something (This is using HDHomerun Prime cable tuners) I got the error "No tuners available". I had to reboot my PC to get the tuners back again. So it seems that running the EPG123 update while tuners are in use will leave the tuners in a lockied or unusable state.

stuartm

Posts: 721
Joined: Mon Nov 05, 2012 8:05 pm
Location: Longmont, CO

HTPC Specs: Show details

#502

Post by stuartm » Tue Jan 17, 2017 9:59 pm

I have noticed recently that there seem to be some surprising missing movie thumbnail art from my EPG123 Guide. e.g. recent movies like Gladiator and Friends with Benefits. EPG123 generally does a really good job of finding movie art thumbnails. e.g. Old obscure movies and Spanish language usually work so I wonder what would cause missing art for fairly mainstream movies?

User avatar
garyan2

Posts: 7438
Joined: Fri Nov 27, 2015 7:23 pm
Location:

HTPC Specs: Show details

#503

Post by garyan2 » Wed Jan 18, 2017 2:00 am

stuartm wrote:I have noticed recently that there seem to be some surprising missing movie thumbnail art from my EPG123 Guide. e.g. recent movies like Gladiator and Friends with Benefits. EPG123 generally does a really good job of finding movie art thumbnails. e.g. Old obscure movies and Spanish language usually work so I wonder what would cause missing art for fairly mainstream movies?
I'm surprised no one has piped in earlier. This isn't just movies, it is series images as well. There seems to be a problem with the Gracenote feed that doesn't recognized/acknowledge program IDs for image searches so there is nothing for EPG123 to process. There seems to be a date/time when the problem started occurring since all the older stuff still works. You may have noticed that the new series "Ransom", "Emerald City", and "The Mick" don't have images which all started the beginning of this year. There have been others, but can't think of which ones. It is also sports images, but that's just a minor annoyance for me.

I have found another path to getting images, but there is a problem with it. Whereas the method epg123 is using right now, all the images are specific to the program, the other method gives you access to a pool of images related to the program, but quite possibly for a different language. There is no way I can determine what is best.

I'll send another email to rkulagow to see if we've made any progress in getting this corrected.
- Gary
Keeping WMC alive beyond January 2020. https://garyan2.github.io

chunki

Posts: 5
Joined: Sat Feb 04, 2017 2:38 am
Location:

HTPC Specs: Show details

#504

Post by chunki » Tue Feb 07, 2017 3:33 am

I tried to record all the "NBA Basketball" and I chose "Record Series" from the Guide. Checking WMC, there was only one game in the series. I tried different channels, The behavior stayed the same, one game in each "NBA Basketball" series. What can I do to record the "New" "NBA Basketball" as a series?

User avatar
garyan2

Posts: 7438
Joined: Fri Nov 27, 2015 7:23 pm
Location:

HTPC Specs: Show details

#505

Post by garyan2 » Tue Feb 07, 2017 3:45 am

chunki wrote:I tried to record all the "NBA Basketball" and I chose "Record Series" from the Guide. Checking WMC, there was only one game in the series. I tried different channels, The behavior stayed the same, one game in each "NBA Basketball" series. What can I do to record the "New" "NBA Basketball" as a series?
The series ID of each game has a unique number and is not considered part of a series. There is a method to use I that outlined here (http://www.thegreenbutton.tv/forums/vie ... 94#p112594) that will do what you want.

Just to let you, and everyone else know, based on what you just observed, I have changed the series IDs of sport events to their respective titles in the upcoming version. That way, all NBA Basketball games will be considered part of the same series. There is a caveat, however. This will only be for events whose ID from Schedules Direct starts with a "SP". Those that start with an "EP" will be considered a separate series. I believe it will get us all 90% where we want to be.
- Gary
Keeping WMC alive beyond January 2020. https://garyan2.github.io

chunki

Posts: 5
Joined: Sat Feb 04, 2017 2:38 am
Location:

HTPC Specs: Show details

#506

Post by chunki » Tue Feb 07, 2017 4:09 am

Record by Title works. Thanks!

User avatar
12noon

Posts: 127
Joined: Mon Oct 06, 2014 4:23 pm
Location:

HTPC Specs: Show details

#507

Post by 12noon » Sun Feb 26, 2017 2:06 am

I upgraded to v1.1.0 yesterday. Everything went well. I really appreciate the clear instructions for the transition utility. I'm glad you made it a separate tool. :)

Everything seems to work fine, but I do get these errors. It's not obvious to me what (if anything) I should do about them.

Code: Select all

7:38:35 PM - The MD5 hash of Program EP005178512656 does not match the expected MD5 provided by the Service Schedule.
7:38:35 PM - The MD5 hash of Program EP005178512658 does not match the expected MD5 provided by the Service Schedule.
7:38:35 PM - The MD5 hash of Program EP024336730017 does not match the expected MD5 provided by the Service Schedule.
7:38:35 PM - The MD5 hash of Program EP019152142439 does not match the expected MD5 provided by the Service Schedule.
7:38:35 PM - The MD5 hash of Program EP024689140160 does not match the expected MD5 provided by the Service Schedule.
7:38:35 PM - The MD5 hash of Program EP024689230150 does not match the expected MD5 provided by the Service Schedule.
7:38:36 PM - The MD5 hash of Program EP000006320018 does not match the expected MD5 provided by the Service Schedule.
7:38:46 PM - The MD5 hash of Program EP000007990167 does not match the expected MD5 provided by the Service Schedule.
USA 60005
WOW Chicago Suburbs - Digital
USA-IL58819-X

User avatar
garyan2

Posts: 7438
Joined: Fri Nov 27, 2015 7:23 pm
Location:

HTPC Specs: Show details

#508

Post by garyan2 » Sun Feb 26, 2017 5:32 am

12noon wrote:I upgraded to v1.1.0 yesterday. Everything went well. I really appreciate the clear instructions for the transition utility. I'm glad you made it a separate tool. :)

Everything seems to work fine, but I do get these errors. It's not obvious to me what (if anything) I should do about them.

Code: Select all

7:38:35 PM - The MD5 hash of Program EP005178512656 does not match the expected MD5 provided by the Service Schedule.
7:38:35 PM - The MD5 hash of Program EP005178512658 does not match the expected MD5 provided by the Service Schedule.
7:38:35 PM - The MD5 hash of Program EP024336730017 does not match the expected MD5 provided by the Service Schedule.
7:38:35 PM - The MD5 hash of Program EP019152142439 does not match the expected MD5 provided by the Service Schedule.
7:38:35 PM - The MD5 hash of Program EP024689140160 does not match the expected MD5 provided by the Service Schedule.
7:38:35 PM - The MD5 hash of Program EP024689230150 does not match the expected MD5 provided by the Service Schedule.
7:38:36 PM - The MD5 hash of Program EP000006320018 does not match the expected MD5 provided by the Service Schedule.
7:38:46 PM - The MD5 hash of Program EP000007990167 does not match the expected MD5 provided by the Service Schedule.
Nothing to be done ... this is more an FYI than any error. The discrepancies will be cleared up on next update or ~6 hours later, whichever is latest.
- Gary
Keeping WMC alive beyond January 2020. https://garyan2.github.io

Sea3

Posts: 38
Joined: Wed Aug 07, 2013 4:54 am
Location:

HTPC Specs: Show details

#509

Post by Sea3 » Sun Feb 26, 2017 5:47 pm

I've been seeing the MD5 hash issue for 8 days since updating to v1.1.0 for Comcast Snohomish County - Digital Everett. Seems to be a different number of warnings each day but doesn't seem to be clearing up. Doesn't sound like I should be worried, right?

User avatar
12noon

Posts: 127
Joined: Mon Oct 06, 2014 4:23 pm
Location:

HTPC Specs: Show details

#510

Post by 12noon » Sun Feb 26, 2017 6:25 pm

Nothing to be done ... this is more an FYI than any error. The discrepancies will be cleared up on next update or ~6 hours later, whichever is latest.
Sure enough. It happened twice on Saturday and not Sunday morning. Thanks!
USA 60005
WOW Chicago Suburbs - Digital
USA-IL58819-X

User avatar
garyan2

Posts: 7438
Joined: Fri Nov 27, 2015 7:23 pm
Location:

HTPC Specs: Show details

#511

Post by garyan2 » Sun Feb 26, 2017 6:54 pm

Sea3 wrote:I've been seeing the MD5 hash issue for 8 days since updating to v1.1.0 for Comcast Snohomish County - Digital Everett. Seems to be a different number of warnings each day but doesn't seem to be clearing up. Doesn't sound like I should be worried, right?
I just downloaded all in your lineup and didn't get any errors. This appears to be my failed attempt at reducing these alerts and I didn't account for something. I think you might be in the case where the information in your guide will not be up-to-date on the identified programs. You could clear out your cache and build a new mxf file and the errors should no longer be present.

This will be corrected in version v1.1.1.
- Gary
Keeping WMC alive beyond January 2020. https://garyan2.github.io

dkrom

Posts: 255
Joined: Mon Nov 17, 2014 11:39 pm
Location:

HTPC Specs: Show details

#512

Post by dkrom » Mon Feb 27, 2017 1:33 am

I've been dabbling with the logos and have one in that I've been editing to the correct dimensions. As I was looking in your current release directions, I noticed that you had directions to manually adjust the width of a channel column. This is not a key in my registry - I did a find and could not find anything related to "ChannelCellWidth". I'm thinking it might be added by another program (maybe mychannellogos?).

User avatar
garyan2

Posts: 7438
Joined: Fri Nov 27, 2015 7:23 pm
Location:

HTPC Specs: Show details

#513

Post by garyan2 » Mon Feb 27, 2017 1:58 am

dkrom wrote:I've been dabbling with the logos and have one in that I've been editing to the correct dimensions. As I was looking in your current release directions, I noticed that you had directions to manually adjust the width of a channel column. This is not a key in my registry - I did a find and could not find anything related to "ChannelCellWidth". I'm thinking it might be added by another program (maybe mychannellogos?).
Yah, I noticed that last release I didn't specify what it was to manually add it. You can add the DWORD in the ProgramGuide\ key. I find a value of 175 (0xAF) works best for me.
- Gary
Keeping WMC alive beyond January 2020. https://garyan2.github.io

robnitro

Posts: 35
Joined: Thu May 03, 2012 10:13 pm
Location:

HTPC Specs: Show details

#514

Post by robnitro » Sun Mar 05, 2017 6:36 pm

Hi,
I've been getting repeats showing as new on shows like "Outrageous Acts of Science" (Discovery channel), any idea on how I could set it up to that the first showing as new, or WMC to not record both showings of a new episode? Example, on at 10pm, and 1 am (for pacific), both get recorded! I have show episode season and number enabled, but WMC doesn't care??

Space

Posts: 2838
Joined: Sun Jun 02, 2013 9:44 pm
Location:

HTPC Specs: Show details

#515

Post by Space » Sun Mar 05, 2017 7:39 pm

This may not be a problem with the guide data itself. The Microsoft tool that merges the MXF file data into the database (LoadMXF.exe) has a bug that can cause this to happen.

I usually see it when multiple airings of an episode are loaded in to the guide and they do not yet have episode specific information (ESI) (they are just generic placeholders for that show without any episode details). Then sometime later those episodes get ESI but for some reason WMC does not see each of the airings as the same episode and continues to want to record all of them.

Again, this is not a problem with the source data, it is a problem with how new data is merged in with existing data. The problem can be "fixed" by clearing out your guide (using ObjectStoreRecoveryTask) and then loading the MXF file (so no merging is required). But this is not a real fix and I would not recommend doing this on a normal basis.

The real fix is to eliminate the merging bug in LoadMXF.exe, but waiting for Microsoft to do that would be a long wait. There was a project by someone on the forums here that was making a replacement for this import utility to fix this bug, but I don't know what the status of that project is.

stuartm

Posts: 721
Joined: Mon Nov 05, 2012 8:05 pm
Location: Longmont, CO

HTPC Specs: Show details

#516

Post by stuartm » Thu Mar 09, 2017 9:07 am

I wonder if EPG123 is becoming a victim of it's own success or else SD is having server problems or they have insufficient capacity. I noticed that of my 4 HTPC's only one had
successfully run an EPG123 update since March 5. I am on 1.0.2 currently and sadly the event log had no indication of failure(no EPG123 error entry) but instead of the usual two
records for a successful run 1)Retrieving the data and 2)loading the mxf file, there was only one record that had it starting the download from SD but no indication that that completed successfully.
When I discovered this I ran manual updates on all my HTPC's and noticed the SD access ran extremely slowly. On one of the machines during the update it several times threw the timed out error and retrying with longer timeout message. After a few of those that run finally threw a json error and popped up a window that EPG123 had stopped running. I had to click
close and re-run EPG123 again and it (slowly) finally successfully completed the download and mxf load. So several things. Is SD having server problems or being overloaded?
Does 1.1.3 always log an error event if any failure prevents a successful download/mxf load? (If so I will update posthaste). My several HTPC's all run the update at different times, none closer
thn 1/2 hour apart so I am not hitting any multiple SD login problems I am pretty sure.

glorp

Posts: 369
Joined: Sun Sep 23, 2012 2:54 pm
Location:

HTPC Specs: Show details

#517

Post by glorp » Thu Mar 09, 2017 4:27 pm

You'll probably find a NetFramework unhandled exception error (1026) in your Windows logs that traces through the Newtonsoft JSON dll. Because it's a NetworkFramework crash, EPG123 terminates and never gets a chance to log anything in my case so there's nothing in its log. It's crashing for some reason probably due to slow response of the servers from what I've seen reported the last few days. I've had several of these over the last year, almost always on a weekend night update. Re-running at a later time manually always works but I've just kind of starting ignoring them because I know they are outside my control. Usually the next days' run works for me.

Gary here's the error trace I got on the last one on 3/5/2017:

Code: Select all

Application: epg123.exe
Framework Version: v4.0.30319
Description: The process was terminated due to an unhandled exception.
Exception Info: Newtonsoft.Json.JsonReaderException
Stack:
   at Newtonsoft.Json.Serialization.JsonSerializerInternalReader.Deserialize(Newtonsoft.Json.JsonReader, System.Type, Boolean)
   at Newtonsoft.Json.JsonSerializer.DeserializeInternal(Newtonsoft.Json.JsonReader, System.Type)
   at Newtonsoft.Json.JsonConvert.DeserializeObject(System.String, System.Type, Newtonsoft.Json.JsonSerializerSettings)
   at Newtonsoft.Json.JsonConvert.DeserializeObject[[System.__Canon, mscorlib, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089]](System.String, Newtonsoft.Json.JsonSerializerSettings)
   at epg123.sdApi.SdApi.sdGetRequestResponse(METHODS, System.String, System.Object, Boolean)
   at epg123.sdApi.SdApi.sdGetScheduleListings(epg123.sdApi.sdScheduleRequest[])
   at epg123.sdJson2mxf.getMd5ScheduleEntries(System.String[], Int32)
   at epg123.sdJson2mxf.getAllScheduleEntryMd5s(Int32)
   at epg123.sdJson2mxf..ctor(epg123.epgConfig)
   at epg123.Program.Main(System.String[])

rkulagow

Posts: 246
Joined: Sun Jul 19, 2015 1:04 am
Location: Schedules Direct

HTPC Specs: Show details

#518

Post by rkulagow » Thu Mar 09, 2017 4:45 pm

stuartm wrote:or else SD is having server problems or they have insufficient capacity.
I've added servers to the pool at AWS, so if you're still having issues please let me know.

stuartm

Posts: 721
Joined: Mon Nov 05, 2012 8:05 pm
Location: Longmont, CO

HTPC Specs: Show details

#519

Post by stuartm » Thu Mar 09, 2017 7:41 pm

glorp wrote:You'll probably find a NetFramework unhandled exception error (1026) in your Windows logs that traces through the Newtonsoft JSON dll. Because it's a NetworkFramework crash, EPG123 terminates and never gets a chance to log anything in my case so there's nothing in its log. It's crashing for some reason probably due to slow response of the servers from what I've seen reported the last few days. I've had several of these over the last year, almost always on a weekend night update. Re-running at a later time manually always works but I've just kind of starting ignoring them because I know they are outside my control. Usually the next days' run works for me.
Gary, Is there a way you can detect this has happened and log an error?

User avatar
garyan2

Posts: 7438
Joined: Fri Nov 27, 2015 7:23 pm
Location:

HTPC Specs: Show details

#520

Post by garyan2 » Thu Mar 09, 2017 7:52 pm

stuartm wrote:Gary, Is there a way you can detect this has happened and log an error?
That's the baffling part. It should be logging something as it is right now. Everything in my code for traffic to/from SD is in a try/catch and should at least log something. I even have code such that even if the main program throws an exception, that before epg123.exe is released from memory it will log an error.

From the above, I take it this is started from the Newtonsoft dll, so I'll start looking down that path. We are using an older version.
- Gary
Keeping WMC alive beyond January 2020. https://garyan2.github.io

Post Reply