[Xmltv-devel] possibly time for a release?

[Chris Butler]

I received an e-mail the other day from a user in Switzerland, asking about
an update to get a working version of tv_grab_ch_search.. According to the
nightly validator, the new ch_search has been OK for the past couple of
weeks.

Would now be a good time to release with the updated grabber?

[Karl Dietz]

Post by Chris Butler
I received an e-mail the other day from a user in Switzerland, asking about
an update to get a working version of tv_grab_ch_search.. According to the
nightly validator, the new ch_search has been OK for the past couple of
weeks.

Is _ch_search considered done for now and tested?

Post by Chris Butler
Would now be a good time to release with the updated grabber?

There are two some small things that would be nice to get in before a
release, but I don't mind either way.

_fi_sv could be made compatible with elder perl installations (see
FIXME in the code) quite easily, if no one is up for that I can give it
a try this weekend.

_il is missing from the .exe builds, but I don't know the cause. Is it
just an oversight or is it the dependency on iconv? It seems to be only
used to convert windows-1255 to utf-8, that could easily be change to
using Encode instead.

Both tasks are be about small refactorings to avoid one-off-dependencies
for single grabbers and result in a common set of available grabbers
across different platforms.


Regards,
Karl

PS: The MANIFEST / make distclean could use an update, too. :)

[Karl Dietz]

Post by Karl Dietz
_fi_sv could be made compatible with elder perl installations (see
FIXME in the code) quite easily, if no one is up for that I can give it
a try this weekend.

done, output was identical in a quick test.

Post by Karl Dietz
PS: The MANIFEST / make distclean could use an update, too. :)

partially done.

I'd like "make (dist)clean" to remove t_* (used by test_grabbers) and
/tmp/.xmltv (the cachedir) as my usual development cycle is as follows:
$ vi grab/$x/tv_grab_$x
$ make
$ grab/test_grabbers --only $x

and sometimes it's important to clean out the left overs of
test_grabbers to avoid errors.
Any suggestions on how to do that in the right way?

Regards,
Karl

[Robert Eden]

Post by Karl Dietz

Post by Karl Dietz
_fi_sv could be made compatible with elder perl installations (see
FIXME in the code) quite easily, if no one is up for that I can give it
a try this weekend.

done, output was identical in a quick test.

I'm having some problems with _fi_sv under windows

C:\xmltv>tv_grab_fi_sv --config-file grab\fi_sv\test.conf --days=1 --output tv.xml
Will download listings for 9 configured channels

Downloading data for channel group 3
Use of uninitialized value in pattern match (m//) at C:\Perl\bin\tv_grab_fi_sv line 428.
Usage: XML::LibXML::parse_html_string(self, string) at C:\Perl\bin\tv_grab_fi_sv line 437.


I'll try and figure it out later. tv_grab_il is working now and a new alpha-exe has been generated.

Robert

[Robert Eden]

Post by Karl Dietz
_il is missing from the .exe builds, but I don't know the cause. Is it
just an oversight or is it the dependency on iconv? It seems to be only
used to convert windows-1255 to utf-8, that could easily be change to
using Encode instead.

IIRC the exe issue was due to ICONV wanting to include 100+ files or so (about 8 more MB)

I'll try your patch and see what that does.

What do you guys think of shooting for a 11/20 release? I can do it quicker if everyone is ready.

Robert

[Nick Morrott]

Post by Robert Eden

Post by Karl Dietz
_il is missing from the .exe builds, but I don't know the cause. Is it
just an oversight or is it the dependency on iconv? It seems to be only
used to convert windows-1255 to utf-8, that could easily be change to
using Encode instead.

IIRC the exe issue was due to ICONV wanting to include 100+ files or so (about 8 more MB)
I'll try your patch and see what that does.
What do you guys think of shooting for a 11/20 release?  I can do it quicker if everyone is ready.

I have a few bits and pieces I'd like to get added to uk_rt if there's
a release on the horizon. 2010-11-20 is good for me.

Cheers,
Nick

[Robert Eden]

Post by Nick Morrott
I have a few bits and pieces I'd like to get added to uk_rt if there's
a release on the horizon. 2010-11-20 is good for me.

Is everyone ready for a release? The status page looks pretty good.

Robert

[Karl Dietz]

Post by Robert Eden

Post by Nick Morrott
I have a few bits and pieces I'd like to get added to uk_rt if there's
a release on the horizon. 2010-11-20 is good for me.

Is everyone ready for a release? The status page looks pretty good.

I'm ready, just commited some last minute fixups. Holding back on all
other changes until after the release.

Regards,
Karl

[Nick Morrott]

Post by Nick Morrott
I have a few bits and pieces I'd like to get added to uk_rt if there's
a release on the horizon. 2010-11-20 is good for me.

Is everyone ready for a release?  The status page looks pretty good.

I've got some regex optimisations I will hold off applying until after
the release, but things seem to be good generally.

Is _in okay? The validator show that as the only grabber with fatal
errors in the last few runs.

Cheers,
Nick

[Karl Dietz]

Post by Nick Morrott

Post by Robert Eden

Post by Nick Morrott
I have a few bits and pieces I'd like to get added to uk_rt if there's
a release on the horizon. 2010-11-20 is good for me.

Is everyone ready for a release? The status page looks pretty good.

I've got some regex optimisations I will hold off applying until after
the release, but things seem to be good generally.
Is _in okay? The validator show that as the only grabber with fatal
errors in the last few runs.

it's ok in principle ;)


_in was ok when I last took a look. But it seems to be broken right now.
After reenabling the xmltv subscription (every now and then the option
gets reset) validation failed.

The site exports 2 days of xmltv data but current download is still for
2010-11-12 - 2010-11-13 (tested with the lineup DigicableDelhi) but
looks fine otherwise.

The lineup Airtel currently validates perfectly fine but will like not
do so in the nightly tests due to having only today and today+1 in the
data but the tester filtering out today+2 and today+3 due to an unlucky
combination of date math and the time the tests run.

Regards,
Karl

[Robert Eden]

Post by Nick Morrott
Is _in okay? The validator show that as the only grabber with fatal
errors in the last few runs.

I noticed that myself... I don't think there's a problem with the grabber, but one with the data source. I'll contact them.

Robert