You are here

Error generating REX in version 5.1.4

31 posts / 0 new
Last post
davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week
Error generating REX in version 5.1.4

Hi, happy new year

I am testing the new version 5.1.4 and I get an error when generating the REX

In version 5.0.1.1 it works correctly

Thanks for your time

Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

the postprocessor was rewritten since V5.0.1.1
i tried V5.1.3 and same error.
so no doubt the rewrite broke something,sent a message to jan to check for the problem.

WGMaker
Offline
WGMaker's picture
WG++ Team memberDonator
Joined: 12 years
Last seen: 4 hours
Is the support helpful?
support us

Hi .. that problem is caused by a new feature of REX that allows to convert episode numbers format from xmltv_ns to onscreen vice versa. For that it checks if an episode pattern is specified in rex.config and doing that it tries to read the url element in the channel section of the xmltv source file.
Your epgsinrexDEP.xml doesn't have these url elements. That causes the crash.
I have changed the source code such that that won't happen anymore. Use evaluation version 5.1.4.1 of the program on GIT https://github.com/SilentButeo2/webgrabplus-siteinipack/blob/master/eval...

It can also be fixed if you add an empty url element to all channels in the channel section of the xmltv source

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

Thank you very much for the quick solution, REX postprocessing now works correctly.

But, I have encountered another problem in version 5.1.4.1, with the same ini I don't get the same output as with version 5.0.1.1

Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

what ini are you using.
the (?) at the end of the title's in the 5.1.4.1 xml usually means something is broken in the ini.

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

It is an ini modified by me, I think the part that gives problems is this:

* ARREGLO TITULOS CON ":"
if {('temp_1' ~~ "TEMPORADA")|end}
title.modify {remove('subtitle' not "")|: 'subtitle'}
else ('temp_1' not ~ "TEMPORADA")
subtitle.modify {clear}
temp_3.scrub {single||||}
end_if

* TITLEORIGINAL
videoquality.modify {addstart|'temp_3'}
if {('videoquality' ~~ "TEMPORADA")|end}
videoquality.modify {clear}
else ('title' = 'videoquality')
videoquality.modify {clear}
end_if

Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

else ('temp_1' no ~ "TEMPORADA")
should be
else ('temp_1' not~ "TEMPORADA")
i assume u meants not contain?

also the (?) means the index title and the details title dont match.
you should see messages in your webgrab log showing the title for each for every show that has the (?) in the title.

Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

there were issues with using if/else from V5.1 to V5.1.3(i think).
jan said he had fixed this in V5.1.4 but i havnt had time to do any tests to confirm.
if i remember correctly the arguments were being ignored.

Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

you could also try this,it should do the same without the need for if/else

* ARREGLO TITULOS CON ":"
title.modify {remove(['temp_1' ~~ "TEMPORADA"] ['subtitle' not ""] &)|: 'subtitle'}
loop {('temp_1' not ~ "TEMPORADA" max=1)|end}
subtitle.modify {clear}
temp_3.scrub {single||||} ****replace this with your actual scrub,i assume posting in the forums messed it up
end_loop

* TITLEORIGINAL
videoquality.modify {addstart|'temp_3'}
videoquality.modify {clear(['videoquality' ~~ "TEMPORADA"] ['title' = 'videoquality'] \|)}

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

Exactly, I have checked my ini and indeed the new versions do not process the if/else correctly

WGMaker
Offline
WGMaker's picture
WG++ Team memberDonator
Joined: 12 years
Last seen: 4 hours
Is the support helpful?
support us

Hi
if/else and loops only work properly in 5.1.4 if scopes are applied for all lines in the siteini . Also on places where scopes don't seem necessary.
Upload your ini , I will try to fix it . I might also be able to change the source such that it isn't that strict with scopes anymore. I prefer if you attach it to a mail, see http://webgrabplus.com/content/contact

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

.ini file sent

Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

https://github.com/SilentButeo2/webgrabplus-siteinipack/blob/master/eval...

Version 5.1.4.2
fixed : incorrect settings of loopid and ifid if not specified with end
fixed : several bugs that caused if and especially else not to execute correctly

rex bug was fixed in V5.1.4.1
Version 5.1.4.1 evaluation
- Postproces
fix : crash when the xmltv channel section does not have a element

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

Hello, how quickly you found the solution.

It seems that everything works fine in version 5.1.4.2, but there is one parameter when creating the REX that does not work

category lang="es" 'category'
category lang="kodi" 'category(, )

Category "kodi" is not processed.

My REX is so special because I run webgrab on a very limited VPS, I had to simplify all the processes to create an xml with all the information and then with awk I create different xmls for different applications, KODI, tvheadend, PLEX, etc.

Regards David

Attachments: 
Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

jan would have to confirn this bu i suspect its because of the language attribute.
one of the updates in V5.1.4.x is the ability select between title and title original.
webgrab identifies each buy it language attribute.
the rex documentation only mentions the title about this so i'm not even sure if its supported for other elements.

example
source xml
<title lang="es">Spanish title</title>
<title lang="en">English title</title>

rex config
<title lang="en">'title</title>

only the english title would be in the rex guide.xml,the spanish title would be omitted.
if this convertion holds true for other elements(if they support the lang attribute) then the reason why its omitted is because there no category with the lang="kodi" attribute in your source.xml
i assume the category with lang="es" works because the category exists with this attribute.

Edit:
i did some tests wuith V5.4.2
i dont think the language attribute is support for category element(or its a bug) as no category elements are written if a language attribute is used the rex config.

also there's another bug i found.
if source xml has multiple categories then multiple language attributes are add to rex guide.

source xml
<category lang="en">Movie</category>
<category lang="en">Comedy</category>

rex config
<category>'category(,)'</category>
rex guide.xml
<category lang="en,en"&ht;Movie,Comedy</category&ht;

if i use..
<category lang="en">'category(,)'</category>

nothing is written to rex guide.xml even though the source has category elements with this language attribute.
thats why its either a bug or the lang attribute applies to the title only.

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

I am using REX Version: 3.1.1 which for my use works perfectly.
In tvheadend I use an xml with "es" and "en" as languages for different configurations, and the "kodi" language I use to generate categories separated by "," for KODI to identify them.
I know it's a bit "special" the way I use webgrab.

Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

talked to jan about this and here the reply..

It seems that you cannot set a new the attribute name without adding the /a for the element value . This works
rex config
<category lang="kodi">'category/a(, )'</category>
rex guide result..
<category lang="kodi">Series(es), Drama(es), DramaSeries(en)</category>
But without /a it doesnt. Which more or less is correct because the attribute is used as element selection . So without /a it tries to locate categories with lang kodi, which arent there.

Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

you could also use /a+

rex config
<category lang="kodi">'category/a+(, )'</category>
rex guide result..
<category lang="kodi">Series(lang es), Drama(lang es), DramaSeries(lang en)</category>

/a and /a+ is explained in the docs.

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

OK, I understand the reasoning, with your solution it works.

Now I have another problem, when doing the REX some "desc lang" are lost that in captured xml if they are there.

Sorry to give you so much work ;-)

Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

the /a and /a+ applies to all elements that have a lang attribute,description is included.
title
subtitle
description
category
country
premiere
other that i remember off the top of my head.
the full list is avilable in the xmltv.dtd
https://github.com/XMLTV/xmltv/blob/master/xmltv.dtd

WGMaker
Offline
WGMaker's picture
WG++ Team memberDonator
Joined: 12 years
Last seen: 4 hours
Is the support helpful?
support us

It should work for all supported attributes, not only lang .. like system, type, even role

there is a table in the documentation , chapter 5.1.6.5

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

The problem is that it does not always have the same behavior.
As I said, when doing the REX, information is lost.

Attachments: 
WGMaker
Offline
WGMaker's picture
WG++ Team memberDonator
Joined: 12 years
Last seen: 4 hours
Is the support helpful?
support us

Sorry I don't understand that . What (and where) in the captured xml (is that the 'conrex' one?) is lost . Can you point to a location in that file ?

You already answered that .. will look why ..

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

I attach two images, one made with REX 3.1.1 and the other with REX 3.2.1.

With version 3.2.1 it seems that information is lost, but it does not happen in all the programs, only in some of them.

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

I think I know what happens, version 3.2.1 does not like "|".

Blackbear199
Offline
Blackbear199's picture
WG++ Team memberDonator
Joined: 9 years
Last seen: 13 min

its fine to use the |
appears to be the custom lang="sc",lang="cc" breaking this.
if you use just <desc> results look ok.
jan will have to look into it.

WGMaker
Offline
WGMaker's picture
WG++ Team memberDonator
Joined: 12 years
Last seen: 4 hours
Is the support helpful?
support us

Contrary to what Blackbear found, the | character must not be used ! It messes up the lines in rex.config while parsing.
If you remove all | or replace them with some other character it all seems to work .

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

I understand, so I will continue using REX 3.1.1 which works fine for me.

davidmumar
Offline
davidmumar's picture
Donator
Joined: 4 years
Last seen: 1 week

Hello again

I think the new version of webgrab 5.1.4.3 still has problems with if/else, this command does not execute the first condition

if {('index_category' ~ "sport")|end}
url.modify {addstart|'showicon'}
else
url.modify {addstart|'temp_4'}
end_if

WGMaker
Offline
WGMaker's picture
WG++ Team memberDonator
Joined: 12 years
Last seen: 4 hours
Is the support helpful?
support us

send me the complete ini in which this occurs. Use the wg++ email [email protected] .. thanks Jan

WGMaker
Offline
WGMaker's picture
WG++ Team memberDonator
Joined: 12 years
Last seen: 4 hours
Is the support helpful?
support us

Upgrade to 5.2.2 beta

Log in or register to post comments

Brought to you by Jan van Straaten

Program Development - Jan van Straaten ------- Web design - Francis De Paemeleere
Supported by: servercare.nl