bookmark.db

vic1972 escribió:
working
Only had a couple of testing, but it seems now it is fine.
have a try ... and enjoy

Wow, it is an excellent info Thanks
I will try it.

vic1972 escribió:
working
Only had a couple of testing, but it seems now it is fine.
have a try ... and enjoy

So I tested it and it works great

I encountered a small problem only:
Scenario: While I am playing ts file, the timer recording takes place -> the message window with 30 seconds countdown is shown -> the machine goes to TDT mode. The recording is not active yet (It will start after aprox. 30 seconds). If I push RedButton immediately then it works correctly, but if I push the RedButton with delay (eg 35 seconds), ie in the moment when recording is already active then RedButton doesn't work at all. But If I stop recording manually then RedButton works again. I studied PipeManagement Log and there is no record about RedButton in that time so it seems that DdvPlayer doesn't communicate thru stdin/stdout at that time (because the same problem also affects other color buttons). However, I found out the workaround - if the RedButton doesn't work during the active recording then I can push the Home button twice and RedButton starts working again.

It seems to me that it is a problem with lack of the communication with DvdPlayer (at the time of recording) which we couldn't solve it so it maybe mentioned in Harmony manual.

OK,
I will have a loot at that scenario.
I will also add more debugging information in the log, so you do more complete testing.

Far for closing this feature, we still have to do some improvements/investigation:
- The scenario you commented.
- The use of ln -f changes de date/time of original file:
I have to check if we can save the use of touch command, probably is not necessary ...
I also have to check the other variants of ln, including cp.

- The load of the triggers.
Still thinking about overwriting bookmark.db vs loading triggers; and the load in PM.

Hi,
I've tried to reproduce the issue, hit the red button in 35 seconds, that is 5 seconds after the little windows dissapear,
and it worked fine.
Anyway, I will keep testing this. I have added some more logs.

I have also eliminated the touch command, I think it is not necessary. (we will know today
Available for download.

vic1972 escribió:Hi,
I've tried to reproduce the issue, hit the red button in 35 seconds, that is 5 seconds after the little windows dissapear,
and it worked fine.
Anyway, I will keep testing this. I have added some more logs.

I have also eliminated the touch command, I think it is not necessary. (we will know today
Available for download.

I am sorry, maybe I described the problem in a wrong way - the time of 35 second was the example value only. The "core" of problem is that the RedButton doesn't work in time when the recording has already started (so the "REC" appears on the LG display etc.). I have to push the "Home" button twice to make the RedButton working again ...

Question to anyone regarding the use of link command: ln

Why on my LG system, I cannot use (soft link) ln -s in the REC folder ( /tmp/hdd/volumes/HDD1 type ufsd )
and I can use ln -s in the /root folder ... ?

Obviously they are different filesystem, etc.
But the fact is that in REC folder I can only use hw link: ln file1 file2
and not soft links: ln -s

I think I explain it earlier. The /tmp/hdd/root is Linux ext3 filesystem so the soft/hard links a fully supported. But /tmp/hdd/volumes/HDD1 is NTFS filesystem which is mounted by commercial product UFSD ( http://www.paragon-software.com/ntfs/ufsd.html ) which obviously doesn't support the soft links. the NTFS support some type of links but it is used as a "shortcut". In CygWin under Windows you can use soft/hard link, but it is normal text files with link definition and CygWin interpreter read it do the action described in it.

Keltek escribió:I think I explain it earlier. The /tmp/hdd/root is Linux ext3 filesystem so the soft/hard links a fully supported. But /tmp/hdd/volumes/HDD1 is NTFS filesystem which is mounted by commercial product UFSD ( http://www.paragon-software.com/ntfs/ufsd.html ) which obviously doesn't support the soft links. the NTFS support some type of links but it is used as a "shortcut". In CygWin under Windows you can use soft/hard link, but it is normal text files with link definition and CygWin interpreter read it do the action described in it.

Hi Keltek, thanks for your exact explanation. Based on previous empirical tests I anticipated that there would be a such issue there ...

yes , ok, do not panic
Ideal would be the soft link, but anyway, I think that with even using hard link, this Repeat functionality is interesting enought

The only little problem I see using hard link is that the date of the original file is also changed...

vic1972 escribió:yes , ok, do not panic
Ideal would be the soft link, but anyway, I think that with even using hard link, this Repeat functionality is interesting enought

The only little problem I see using hard link is that the date of the original file is also changed...

OK, you are right ...

Hello George,
I have changed the link functions, and the touch functions,
as you know these commands handle with files.
My testing are fine, but please, have a testing with your special filenames in czec, just in case.
You can download using the service numeric number.

Also, i have put in .ini at the end two parameters to configure, I think as it is right now it is working fine.

regards

Victor, I realize you start sqlite3 issuing command "/tmp/hdd/root/custom/bin/sqlite3". You should first check this binary is available.
I think you first check this command exist and then perform the greps and then perform the sqlite3. And how did you assure the /tmp/hdd/root is mounted in the moment you issue the command? There can be some delays in some circumstances on different configuration to be this binary available.

Hi, thanks for the tip.
Yes, that is considered.
That execution is done after 30seconds, as i have a thread with this sleep,
then i do this kind of operations that needs all filesystem to be ready.

vic1972 escribió:Hello George,
I have changed the link functions, and the touch functions,
as you know these commands handle with files.
My testing are fine, but please, have a testing with your special filenames in czec, just in case.
You can download using the service numeric number.

Also, i have put in .ini at the end two parameters to configure, I think as it is right now it is working fine.

regards

Hi Victor, thanks a lot! I tested it and it works correctly even with the czech filenames
Btw: Several times I have encountered the problem with the incorrect Power Off (ie "Off" message remained on display so I had to reset the machine) but I think that it is the common problem which is just solving by Keltek & evr ...

Hello George,
I think we might need some checking in the triggers.

Let me explain, yesterday, I was not able to restart any execution in the point where it was interrupted;
neither the Repeat.ts nor normal .ts files.

So, I had a look at the tParameters, and found this very funny:

Código:


sqlite> select * from tParameter;
1|DB_VERSION|1
2|DB_DEVELOP_VERSION|6
3|SETUP_INIT|1
4|RESTORE_DEFAULT_FROM_DB|0
5|0m|��ZM
6|0k|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/Las Tres Mellizas_20110215_0741.ts
7|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/Las Tres Mellizas_20110215_0741.ts|
8|9999m|��ZM
9|9999k|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/ Repeat.ts
11|1m|$�ZM
12|1k|http://192.168.2.119:8200/MediaItems/29.mpg
13|http://192.168.2.119:8200/MediaItems/29.mpg|�
14|2m|=�ZM
15|2k|http://192.168.2.119:8200/MediaItems/30.mpg
16|http://192.168.2.119:8200/MediaItems/30.mpg| 
17|3m|/�ZM
18|3k|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/Bob Esponja_20110212_1538.ts
19|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/Bob Esponja_20110212_1538.ts|
20|4m|��ZM
21|4k|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/ Repeat.ts
22|5m|��ZM
23|5k|http://192.168.2.119:8200/MediaItems/26.mpg
24|http://192.168.2.119:8200/MediaItems/26.mpg| 
25|6m|��ZM
26|6k|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/George De La Jungla_20110215_1925.ts
27|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/George De La Jungla_20110215_1925.ts||
sqlite> .schema


As you can see Repeat.ts appears two time, once in the 9999 and another in the 4 index.

The only different i did yesterday is to play media .ts from the other lg via DLNA,
then you can see the entries like:

23|5k|http://192.168.2.119:8200/MediaItems/26.mpg
Could this be the reason?...


More things regarding the triggers:
*1*
I am not sure, but i think triggers can be optimized using % instead of _
Have a look at the below examples:

sqlite> select * from tParameter where fKey like '_m';
5|0m|��\M
sqlite> select * from tParameter where fKey like '__m';

sqlite> select * from tParameter where fKey like '%m';
5|0m|��\M
8|9999m|


*2*
In one part of the triggers, i see that you calculate the id, makeing a (select) -2
would it be possible to assign a fixed id? like 9999?
I do not remember right now if you already tested or not this possibility ...


Could you please have a look and see if these can be fixed/optimized ..?
Thanks a lot

vic1972 escribió:Hello George,
I think we might need some checking in the triggers.

Let me explain, yesterday, I was not able to restart any execution in the point where it was interrupted;
neither the Repeat.ts nor normal .ts files.

So, I had a look at the tParameters, and found this very funny:

Código:


sqlite> select * from tParameter;
1|DB_VERSION|1
2|DB_DEVELOP_VERSION|6
3|SETUP_INIT|1
4|RESTORE_DEFAULT_FROM_DB|0
5|0m|��ZM
6|0k|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/Las Tres Mellizas_20110215_0741.ts
7|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/Las Tres Mellizas_20110215_0741.ts|
8|9999m|��ZM
9|9999k|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/ Repeat.ts
11|1m|$�ZM
12|1k|http://192.168.2.119:8200/MediaItems/29.mpg
13|http://192.168.2.119:8200/MediaItems/29.mpg|�
14|2m|=�ZM
15|2k|http://192.168.2.119:8200/MediaItems/30.mpg
16|http://192.168.2.119:8200/MediaItems/30.mpg| 
17|3m|/�ZM
18|3k|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/Bob Esponja_20110212_1538.ts
19|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/Bob Esponja_20110212_1538.ts|
20|4m|��ZM
21|4k|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/ Repeat.ts
22|5m|��ZM
23|5k|http://192.168.2.119:8200/MediaItems/26.mpg
24|http://192.168.2.119:8200/MediaItems/26.mpg| 
25|6m|��ZM
26|6k|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/George De La Jungla_20110215_1925.ts
27|fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/George De La Jungla_20110215_1925.ts||
sqlite> .schema


As you can see Repeat.ts appears two time, once in the 9999 and another in the 4 index.

The only different i did yesterday is to play media .ts from the other lg via DLNA,
then you can see the entries like:

23|5k|http://192.168.2.119:8200/MediaItems/26.mpg
Could this be the reason?...


More things regarding the triggers:
*1*
I am not sure, but i think triggers can be optimized using % instead of _
Have a look at the below examples:

sqlite> select * from tParameter where fKey like '_m';
5|0m|��\M
sqlite> select * from tParameter where fKey like '__m';

sqlite> select * from tParameter where fKey like '%m';
5|0m|��\M
8|9999m|


*2*
In one part of the triggers, i see that you calculate the id, makeing a (select) -2
would it be possible to assign a fixed id? like 9999?
I do not remember right now if you already tested or not this possibility ...


Could you please have a look and see if these can be fixed/optimized ..?
Thanks a lot

Hi Victor,

it's seems very strange, because the record of Repeat file's 3rd row is missing, actually:
1. The record with id = 10 doesn't exist at all
2. The record with id = 22 belongs to another file with filenumber of "5"
So none of the records mentioned above contains Repeat file fullname in text form (ie "fileset:///tmp/ramfs/volumes/HDD ( Internal )/REC/ Repeat.ts"). Well, there are 2 records with Repeat file fullname in BLOB form (id = 9 & id = 21) but itsn't enough for correct working

As no part of triggers can delete db records it seems that it was caused by DvdPlayer itself. Maybe it's connected with using DLNA - I don't know actually because I have never used the DLNA so far. So I will try to test it ...

About optimalizations:

*1*
Well, the goal is to check if the fKey contains values from "0m" to "255m". Unfortunately there would not be possible to use regular expressions in sqlite, so we are limited to use sql wildcards only. In case of using "%m" pattern there will be also matched the record with fKey = '9999m' which we don't want to, actually. So I avoided this situation by using '_m' or '__m' or '___m' to match the values with maximum 3 characters before the 'm' letter.

*2*
You are right! Using the relative id calculation (ie id value - 2) is the "relict" of earlier trigger versions which didn't use the fixed file number "9999". So instead of

Código:

... (SELECT id FROM tParameter WHERE fKey = (SELECT fValue FROM tHarmony WHERE fKey = 'RecentFilenameText')) - 2 ...

you can use this:

Código:

... (SELECT id FROM tParameter WHERE fKey = (SELECT fValue FROM tHarmony WHERE fKey = 'RecentFilenameNumber') || 'm') ...

Hope it helps ...

Thanks for the prompt and accurate response
I see now and understand what you say about _m and %m, cristal clear.
By the time of your answer, it seems you missed yesterday watching a movie for this ...
All users thanks your time.
I will fully test the new triggers and report. thanks again.

fixed actual version:
http://dl.dropbox.com/u/684543/varios/ms450h/Firmwares/work/triggers_SpaceRepeat2.sql

Hello George,
could you please have a look at this bookmark.db,
user pc3 says it is not re-starting well any file.
Probably you can comunicate wiht him for further details.
file is in:
http://dl.dropbox.com/u/684543/varios/ms450h/RePlay/bookmark-dvdplayer.db
and, but i guess is the same:
http://dl.dropbox.com/u/684543/varios/ms450h/RePlay/bookmark-ramfs.db

Thanks very much

vic1972 escribió:Hello George,
could you please have a look at this bookmark.db,
user pc3 says it is not re-starting well any file.
Probably you can comunicate wiht him for further details.
file is in:
http://dl.dropbox.com/u/684543/varios/ms450h/RePlay/bookmark-dvdplayer.db
and, but i guess is the same:
http://dl.dropbox.com/u/684543/varios/ms450h/RePlay/bookmark-ramfs.db

Thanks very much

Hi Victor and pc3,
I checked the db file and it seems that it is a quite old version, because used triggers contain clause

Código:

 "WHERE fKey = (SELECT fValue FROM tHarmony WHERE fKey = 'RecentFilenameText')) - 2

instead of

Código:

 "WHERE fKey = (SELECT fValue FROM tHarmony WHERE fKey = 'RecentFilenameNumber') || 'm')

which we changed recently. Another issue is that some db records are missing. So I am not sure which version of PM the colleague called pc3 has got on its machine but it seems that it could be a similar problem as yours before the triggers fixing (see https://ms450.forosactivos.net/t567p160-bookmarkdb#5868).

Dear pc3, is it acceptable for you to delete both bookmark.db files (ie to lost saved resume playing times) and let the PM "recover" the bookmark.db file?

very good, thanks george.

yes, it seems he is using old version. i think the best for him, is to use last version of the firmware, and then test again.

pc3 ese tema ya fue resuelto, por eso salio una nueva revision.
Favor de instalar la ultima version disponible (en el primer mensaje del foro de testers, gracias)

Thanks George and Victor. I erased both bookmark.db files and reinstalled the latest version and now everything works perfectly.

Good job, good afternoon and good luck.

ok people
time to tune a bit the application.

My actual db:
http://dl.dropbox.com/u/684543/varios/ms450h/RePlay/bookmark.db_13April2011

I have noted now, that after more than a month working fine,
now red button does not detect the last execution point.

and so, it start execution always from the begining.
I do not know what can be the reason, probably it is only me
Is red button working find for you people?

vic1972 escribió:ok people
time to tune a bit the application.

My actual db:
http://dl.dropbox.com/u/684543/varios/ms450h/RePlay/bookmark.db_13April2011

I have noted now, that after more than a month working fine,
now red button does not detect the last execution point.

and so, it start execution always from the begining.
I do not know what can be the reason, probably it is only me
Is red button working find for you people?

Hi Victor,
as I analysed your bookmark.db, the problem is that the 3rd row of record set (ie with path to Repeat.ts) is missing (see that record with id = 19 is missing). Sometimes I noticed that some records are actually missed away "without reason" (eg records no. 13, 79, 101, 123) so I think that is caused by not so well working sqlite client embedded in DvdPlayer
I tried to repair your bookmark.db, I hope that it will be working OK now: www.rysi.duha.cz/lg/bookmarkdb_13April2011.db

ahhh hi
thanks for quick response.
yes, well, no problem for repearing,
what will be nice a kind of auto repear functionality.
best regards and many thanks

vic1972 escribió:ahhh hi
thanks for quick response.
yes, well, no problem for repearing,
what will be nice a kind of auto repear functionality.
best regards and many thanks

Well, auto repairing has been already implemented at the beginning (see Autorecovery* triggers). In fact, I also used it for your bookmark.db So it works perfectly outside the LG (eg in SQLite Manager), unfortunately it doesn't work reliably in environment of DvdPlayer's embedded sqlite and I don't know why, actually Maybe it is connected with manner of processing the triggers in such environment, who knows ...