Mailing List Archive: MythTV: Dev

PARTLY SOLVED Re: length mismatch between programinfo

Index | Next | Previous | View Threaded

brad+mydev at templetons

Jan 19, 2009, 9:20 PM

Post #1 of 9 (2051 views)
Permalink

PARTLY SOLVED Re: length mismatch between programinfo

On Mon, Jan 19, 2009 at 08:50:34PM -0800, Brad Templeton wrote:
> On Mon, Jan 19, 2009 at 07:18:13PM -0800, Brad DerManouelian wrote:
> > On Jan 19, 2009, at 7:14 PM, Brad Templeton wrote:
> >
> >> On Mon, Jan 19, 2009 at 05:11:10PM -0800, Brad Templeton wrote:
> >>> Just compiled to lastest trunk 19741M, but noticing a few issues.
> >>> The most serious
> >>> is that when the frontend (same machine or different) attempts to
> >>> load the recorded
> >>> list, it gets an error as follows:
> >>>
> >>> 2009-01-19 17:02:52.229 Loading window theme from /usr/local/share/
> >>> mythtv/themes/default/recordings-ui.xml
> >>> length mismatch between programinfo
> >>> 2009-01-19 17:02:52.332 SortedList is Empty
> >>>
> >>
> >> I should add that when this diagnostic is printed, numrecordings is
> >> 150 (which is correct),
> >> NUMPROGRAMLINES is 47, but strList.size() is only 189, not nearly
> >> large enough to
> >> match the test in remoteutil.cpp that is failing.
> >
> > I can't reproduce this using 19710. Everything is working as expected.
> > My memory limit is also 128M. I've got 924 programs. I start getting
> > programs back immediately but it takes a long time for the entire page
> > to completely build.
>
> Well, the numbers indicate that while the query in the backend is returning
> all 150 programs, only 4 rows (4*47+1 = 189) are being sent in the protocol.
>
> I am trying to figure out how to blame this on my own mods to the database,
> which add a "suggested" column to the various recording tables, could be causing
> this. I have removed my patches for this test compile, so it must be the extra
> column in the database itself. I thought myth was robust over there being extra
> columns in some of the tables to allow code to be tested like this...
> (I am loathe to remove my columns quite yet)
>
> I notice that recordedprogram has only 146 entries while recorded has 150, but
> the query is returning the 150 results, it's the converting them to programinfos
> and spitting those out that is somehow aborting.
> _______________________________________________
> mythtv-dev mailing list
> mythtv-dev [at] mythtv
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev

Ok, I found that 3 of the programs in my "recorded" database had values of 127 in either
the "bookmark" field or the "cutlist" field. As I recall in earlier versions of
the program these fields were global and contained other values. Possibly
something didn't handle that transition right?

Looking at the code in mainserver.cpp, I can't see why these values were causing
the creation of the programinfos to abort, but the all rows starting with the first one
to have 127 in the tinyint field were not being iterated on.

The work-around weas to patch those values in the sql database so they were "1",
and now the frontend is able to download the recorded list, and most functions seem
to be operating.

On to another matter: Years ago, I added code to my own copy of Mythtv to support
"suggestions" (a la tivo) generated by my program tvwish. I added a column called
"suggested" to record, oldrecorded and recorded, and arranged for the flag to be
copied to the tables at the right time. I put in code so that recordings that
were a suggestion would thus be tagged so that they could be displayed with a
different font in your theme (I made them smaller and darker, but did not just
put them all at the end of the list as Tivo does.)

It's been a while since I did this, and myth has changed, so I am curious -- is this
the best way to do it today? The suggestions are in different program groups so
that is not the way to identify them. I like the smaller darker font, but a little
icon displayed to the side of the entry, or in the status screen could also be appropriate.
I welcome thoughts.

-------------------

Another problem: In the status box, the icons are in two rows. The bottom row displays
stereo, CC and HDTV, and they are always displayed, even on programs that do not have
the HDTV bit set in the recordedprogram table. This is with GANT.

Thanks

_______________________________________________
mythtv-dev mailing list
mythtv-dev [at] mythtv
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev

kkuphal at gmail

Jan 20, 2009, 6:23 AM

Post #2 of 9 (1929 views)
Permalink

Re: PARTLY SOLVED Re: length mismatch between programinfo [In reply to]

On Mon, Jan 19, 2009 at 11:20 PM, Brad Templeton
<brad+mydev [at] templetons<brad%2Bmydev [at] templetons>
> wrote:

> On Mon, Jan 19, 2009 at 08:50:34PM -0800, Brad Templeton wrote:
> > On Mon, Jan 19, 2009 at 07:18:13PM -0800, Brad DerManouelian wrote:
> > > On Jan 19, 2009, at 7:14 PM, Brad Templeton wrote:
> > >
> > >> On Mon, Jan 19, 2009 at 05:11:10PM -0800, Brad Templeton wrote:
> > >>> Just compiled to lastest trunk 19741M, but noticing a few issues.
> > >>> The most serious
> > >>> is that when the frontend (same machine or different) attempts to
> > >>> load the recorded
> > >>> list, it gets an error as follows:
> > >>>
> > >>> 2009-01-19 17:02:52.229 Loading window theme from /usr/local/share/
> > >>> mythtv/themes/default/recordings-ui.xml
> > >>> length mismatch between programinfo
> > >>> 2009-01-19 17:02:52.332 SortedList is Empty
> > >>>
> > >>
> > >> I should add that when this diagnostic is printed, numrecordings is
> > >> 150 (which is correct),
> > >> NUMPROGRAMLINES is 47, but strList.size() is only 189, not nearly
> > >> large enough to
> > >> match the test in remoteutil.cpp that is failing.
> > >
> > > I can't reproduce this using 19710. Everything is working as expected.
> > > My memory limit is also 128M. I've got 924 programs. I start getting
> > > programs back immediately but it takes a long time for the entire page
> > > to completely build.
> >
> > Well, the numbers indicate that while the query in the backend is
> returning
> > all 150 programs, only 4 rows (4*47+1 = 189) are being sent in the
> protocol.
> >
> > I am trying to figure out how to blame this on my own mods to the
> database,
> > which add a "suggested" column to the various recording tables, could be
> causing
> > this. I have removed my patches for this test compile, so it must be the
> extra
> > column in the database itself. I thought myth was robust over there
> being extra
> > columns in some of the tables to allow code to be tested like this...
> > (I am loathe to remove my columns quite yet)
> >
> > I notice that recordedprogram has only 146 entries while recorded has
> 150, but
> > the query is returning the 150 results, it's the converting them to
> programinfos
> > and spitting those out that is somehow aborting.
> > _______________________________________________
> > mythtv-dev mailing list
> > mythtv-dev [at] mythtv
> > http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
>
>
> Ok, I found that 3 of the programs in my "recorded" database had values of
> 127 in either
> the "bookmark" field or the "cutlist" field. As I recall in earlier
> versions of
> the program these fields were global and contained other values. Possibly
> something didn't handle that transition right?

I'm sorry I didn't see these posts earlier. If you search the archives I
had the same problem. Removing the bookmark fixed this issue for me as
well.

Kevin

brad+mydev at templetons

Jan 20, 2009, 2:16 PM

Post #3 of 9 (1937 views)
Permalink

Re: PARTLY SOLVED Re: length mismatch between programinfo [In reply to]

On Tue, Jan 20, 2009 at 08:23:13AM -0600, Kevin Kuphal wrote:
> On Mon, Jan 19, 2009 at 11:20 PM, Brad Templeton
> <brad+mydev [at] templetons<brad%2Bmydev [at] templetons>
> > wrote:
>
> > On Mon, Jan 19, 2009 at 08:50:34PM -0800, Brad Templeton wrote:
> > > On Mon, Jan 19, 2009 at 07:18:13PM -0800, Brad DerManouelian wrote:
> > > > On Jan 19, 2009, at 7:14 PM, Brad Templeton wrote:
> > > >
> > > >> On Mon, Jan 19, 2009 at 05:11:10PM -0800, Brad Templeton wrote:
> > > >>> Just compiled to lastest trunk 19741M, but noticing a few issues.
> > > >>> The most serious
> > > >>> is that when the frontend (same machine or different) attempts to
> > > >>> load the recorded
> > > >>> list, it gets an error as follows:
> > > >>>
> > > >>> 2009-01-19 17:02:52.229 Loading window theme from /usr/local/share/
> > > >>> mythtv/themes/default/recordings-ui.xml
> > > >>> length mismatch between programinfo
> > > >>> 2009-01-19 17:02:52.332 SortedList is Empty
> > > >>>
> > > >>
> > > >> I should add that when this diagnostic is printed, numrecordings is
> > > >> 150 (which is correct),
> > > >> NUMPROGRAMLINES is 47, but strList.size() is only 189, not nearly
> > > >> large enough to
> > > >> match the test in remoteutil.cpp that is failing.
> > > >
> > > > I can't reproduce this using 19710. Everything is working as expected.
> > > > My memory limit is also 128M. I've got 924 programs. I start getting
> > > > programs back immediately but it takes a long time for the entire page
> > > > to completely build.
> > >
> > > Well, the numbers indicate that while the query in the backend is
> > returning
> > > all 150 programs, only 4 rows (4*47+1 = 189) are being sent in the
> > protocol.
> > >
> > > I am trying to figure out how to blame this on my own mods to the
> > database,
> > > which add a "suggested" column to the various recording tables, could be
> > causing
> > > this. I have removed my patches for this test compile, so it must be the
> > extra
> > > column in the database itself. I thought myth was robust over there
> > being extra
> > > columns in some of the tables to allow code to be tested like this...
> > > (I am loathe to remove my columns quite yet)
> > >
> > > I notice that recordedprogram has only 146 entries while recorded has
> > 150, but
> > > the query is returning the 150 results, it's the converting them to
> > programinfos
> > > and spitting those out that is somehow aborting.
> > > _______________________________________________
> > > mythtv-dev mailing list
> > > mythtv-dev [at] mythtv
> > > http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
> >
> >
> > Ok, I found that 3 of the programs in my "recorded" database had values of
> > 127 in either
> > the "bookmark" field or the "cutlist" field. As I recall in earlier
> > versions of
> > the program these fields were global and contained other values. Possibly
> > something didn't handle that transition right?
>
>
> I'm sorry I didn't see these posts earlier. If you search the archives I
> had the same problem. Removing the bookmark fixed this issue for me as
> well.

Is there a bug report? Otherwise others will see this when they upgrade,
and I'll see it again (but know what to do) as there are too many bugs
in trunk right now for me to use it in production and I'm going to
revert. (At least I hope they are bugs, or did they really mean to
change the UI that much?) But the killer bug is that instead of being
better at recording from the 1394 streams, it seems worse. A new kernel
would possibly help, but that's a more one-way change and the new 1394
drivers demand a very recent version, I understand.
_______________________________________________
mythtv-dev mailing list
mythtv-dev [at] mythtv
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev

kkuphal at gmail

Jan 20, 2009, 2:51 PM

Post #4 of 9 (1929 views)
Permalink

Re: PARTLY SOLVED Re: length mismatch between programinfo [In reply to]

On Tue, Jan 20, 2009 at 4:16 PM, Brad Templeton
<brad+mydev [at] templetons<brad%2Bmydev [at] templetons>
> wrote:

>
> Is there a bug report? Otherwise others will see this when they upgrade,
> and I'll see it again (but know what to do) as there are too many bugs
> in trunk right now for me to use it in production and I'm going to
> revert. (At least I hope they are bugs, or did they really mean to
> change the UI that much?) But the killer bug is that instead of being
> better at recording from the 1394 streams, it seems worse. A new kernel
> would possibly help, but that's a more one-way change and the new 1394
> drivers demand a very recent version, I understand.

I believe this is the original that references the bookmark value. I
chalked it up to some DB error I encountered on my system since it was so
far removed from that revision. Probably something left over from a while
back.

http://svn.mythtv.org/trac/ticket/2293

Kevin

brad+mydev at templetons

Jan 20, 2009, 3:06 PM

Post #5 of 9 (1915 views)
Permalink

Re: PARTLY SOLVED Re: length mismatch between programinfo [In reply to]

On Tue, Jan 20, 2009 at 04:51:49PM -0600, Kevin Kuphal wrote:
> On Tue, Jan 20, 2009 at 4:16 PM, Brad Templeton
> <brad+mydev [at] templetons<brad%2Bmydev [at] templetons>
> > wrote:
>
> >
> > Is there a bug report? Otherwise others will see this when they upgrade,
> > and I'll see it again (but know what to do) as there are too many bugs
> > in trunk right now for me to use it in production and I'm going to
> > revert. (At least I hope they are bugs, or did they really mean to
> > change the UI that much?) But the killer bug is that instead of being
> > better at recording from the 1394 streams, it seems worse. A new kernel
> > would possibly help, but that's a more one-way change and the new 1394
> > drivers demand a very recent version, I understand.
>
>
> I believe this is the original that references the bookmark value. I
> chalked it up to some DB error I encountered on my system since it was so
> far removed from that revision. Probably something left over from a while
> back.
>
> http://svn.mythtv.org/trac/ticket/2293
>

I don't see a link to create an account so I can re-open this one.
Possibly the truncation took place back earlier while I was running
15102, and only started have an bad effect on the more recent
version.

Though looking at the code, I can't see why having 127 in that field causes
the creation of the program list to silently abort, which is its own
bug.
_______________________________________________
mythtv-dev mailing list
mythtv-dev [at] mythtv
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev

kkuphal at gmail

Jan 20, 2009, 4:59 PM

Post #6 of 9 (1921 views)
Permalink

Re: PARTLY SOLVED Re: length mismatch between programinfo [In reply to]

On Tue, Jan 20, 2009 at 5:06 PM, Brad Templeton
<brad+mydev [at] templetons<brad%2Bmydev [at] templetons>
> wrote:

> On Tue, Jan 20, 2009 at 04:51:49PM -0600, Kevin Kuphal wrote:
> > On Tue, Jan 20, 2009 at 4:16 PM, Brad Templeton
> > <brad+mydev [at] templetons <brad%2Bmydev [at] templetons><
> brad%2Bmydev [at] templetons <brad%252Bmydev [at] templetons>>
> > > wrote:
> >
> > >
> > > Is there a bug report? Otherwise others will see this when they
> upgrade,
> > > and I'll see it again (but know what to do) as there are too many bugs
> > > in trunk right now for me to use it in production and I'm going to
> > > revert. (At least I hope they are bugs, or did they really mean to
> > > change the UI that much?) But the killer bug is that instead of being
> > > better at recording from the 1394 streams, it seems worse. A new
> kernel
> > > would possibly help, but that's a more one-way change and the new 1394
> > > drivers demand a very recent version, I understand.
> >
> >
> > I believe this is the original that references the bookmark value. I
> > chalked it up to some DB error I encountered on my system since it was so
> > far removed from that revision. Probably something left over from a
> while
> > back.
> >
> > http://svn.mythtv.org/trac/ticket/2293
> >
>
> I don't see a link to create an account so I can re-open this one.
> Possibly the truncation took place back earlier while I was running
> 15102, and only started have an bad effect on the more recent
> version.
>
>
> Though looking at the code, I can't see why having 127 in that field causes
> the creation of the program list to silently abort, which is its own
> bug.

I couldn't find it either. It seemed to happen at the database level when I
was looking which was even stranger. The query.next wouldn't advance
through the results IIRC

Kevin

skd5aner at gmail

Oct 31, 2009, 4:51 PM

Post #7 of 9 (1371 views)
Permalink

Re: PARTLY SOLVED Re: length mismatch between programinfo [In reply to]

On Tue, Jan 20, 2009 at 6:51 PM, Kevin Kuphal <kkuphal [at] gmail> wrote:
> On Tue, Jan 20, 2009 at 4:16 PM, Brad Templeton <brad+mydev [at] templetons>
> wrote:
>>
>> Is there a bug report? �Otherwise others will see this when they upgrade,
>> and I'll see it again (but know what to do) as there are too many bugs
>> in trunk right now for me to use it in production and I'm going to
>> revert. � (At least I hope they are bugs, or did they really mean to
>> change the UI that much?) �But the killer bug is that instead of being
>> better at recording from the 1394 streams, it seems worse. �A new kernel
>> would possibly help, but that's a more one-way change and the new 1394
>> drivers demand a very recent version, I understand.
>
> I believe this is the original that references the bookmark value.� I
> chalked it up to some DB error I encountered on my system since it was so
> far removed from that revision.� Probably something left over from a while
> back.
>
> http://svn.mythtv.org/trac/ticket/2293
>
> Kevin
>

Resurecting an old thread, but wanted to let anyone who might stumble
across this to know I've submitted a new ticket:
http://svn.mythtv.org/trac/ticket/7467

I ran into this problem when upgrading from .21-fixes to .22-fixes
(RC2). Turns out I had one recording from 2005 with a value of 127 in
the bookmark column. It rendered my watch recordings list from
populating at all. The workaround (as mentioned) is to manually edit
the value to 1 or 0.

Thanks!
Matt
_______________________________________________
mythtv-dev mailing list
mythtv-dev [at] mythtv
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev

torbjorn.jansson at mbox200

Nov 7, 2009, 7:52 AM

Post #8 of 9 (1292 views)
Permalink

Re: PARTLY SOLVED Re: length mismatch between programinfo [In reply to]

> -----Original Message-----
> From: mythtv-dev-bounces [at] mythtv [mailto:mythtv-dev-bounces [at] mythtv]
On
> Behalf Of Matt S.
> Sent: Sunday, November 01, 2009 12:52 AM
> To: Development of mythtv
> Subject: Re: [mythtv] PARTLY SOLVED Re: length mismatch between
programinfo
>
> On Tue, Jan 20, 2009 at 6:51 PM, Kevin Kuphal <kkuphal [at] gmail> wrote:
> > On Tue, Jan 20, 2009 at 4:16 PM, Brad Templeton
<brad+mydev [at] templetons>
> > wrote:
> >>
> >> Is there a bug report? �Otherwise others will see this when they
upgrade,
> >> and I'll see it again (but know what to do) as there are too many bugs
> >> in trunk right now for me to use it in production and I'm going to
> >> revert. � (At least I hope they are bugs, or did they really mean to
> >> change the UI that much?) �But the killer bug is that instead of being
> >> better at recording from the 1394 streams, it seems worse. �A new
kernel
> >> would possibly help, but that's a more one-way change and the new 1394
> >> drivers demand a very recent version, I understand.
> >
> > I believe this is the original that references the bookmark value.� I
> > chalked it up to some DB error I encountered on my system since it was
so
> > far removed from that revision.� Probably something left over from a
while
> > back.
> >
> > http://svn.mythtv.org/trac/ticket/2293
> >
> > Kevin
> >
>
> Resurecting an old thread, but wanted to let anyone who might stumble
> across this to know I've submitted a new ticket:
> http://svn.mythtv.org/trac/ticket/7467
>
> I ran into this problem when upgrading from .21-fixes to .22-fixes
> (RC2). Turns out I had one recording from 2005 with a value of 127 in
> the bookmark column. It rendered my watch recordings list from
> populating at all. The workaround (as mentioned) is to manually edit
> the value to 1 or 0.

I ran into this issue too.
Would be nice with that one line fix to be added to the schema update code.
It may save other old users the trouble.

_______________________________________________
mythtv-dev mailing list
mythtv-dev [at] mythtv
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev

mtdean at thirdcontact

Nov 7, 2009, 8:07 AM

Post #9 of 9 (1292 views)
Permalink

Re: PARTLY SOLVED Re: length mismatch between programinfo [In reply to]

On 11/07/2009 10:52 AM, Torbj�rn Jansson wrote:
>> Resurecting an old thread, but wanted to let anyone who might stumble
>> across this to know I've submitted a new ticket:
>> http://svn.mythtv.org/trac/ticket/7467
>>
>> I ran into this problem when upgrading from .21-fixes to .22-fixes
>> (RC2). Turns out I had one recording from 2005 with a value of 127 in
>> the bookmark column. It rendered my watch recordings list from
>> populating at all. The workaround (as mentioned) is to manually edit
>> the value to 1 or 0.
>>
> I ran into this issue too.
> Would be nice with that one line fix to be added to the schema update code.
> It may save other old users the trouble.

http://svn.mythtv.org/trac/ticket/7467#comment:2

Mike
_______________________________________________
mythtv-dev mailing list
mythtv-dev [at] mythtv
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads