WPS tokenizer: REPORT PROBLEMS HERE

[NicolasP]

The WPS loading and displaying system has been changed. Read the commit message or the tracker entry for more information about the changes. There may be some problems left, although a lot of testing has been done already.

If you encounter a problem like a WPS not being displayed correctly or causing a crash, please report it here after having made sure it hasn't been mentioned before. If a particular WPS causes the problem, please provide a link to download it or attach the relevant files. Also, don't forget to mention which target is affected by the problem.

[donutman25]

Where do i find instruct on how to use this new wps system?

[NicolasP]

You use it exactly like the old one. It works like the previous system but has improvements. This topic is for problem reporting purposes. By problems I mean broken tags, broken WPSs or unexpected crashes.

[Llorean]

Basically, everything should be exactly the same. If it's not, let us know.

[robin0800]

I assume this breaks all the WPS patches was it tested with any of them?

[Llorean]

It was made available in the tracker. Whether they tested against it or adapted to it is their problem and has no place in this thread. Official builds only.

[pilot000]

I found a "minor" problem. I use my wps on a RTC-modded H140(see http://www.rockbox.org/twiki/bin/view/Main/WpsIriverH100#RemoMartinelli).
This line has a problem with the different delay times:

Code: [Select]

%ac%?Ia<%t3%xdn %Ia |%t0.4%xda.Please Wait.>;%ac%?Ia<%t3%xdo %Id |%t0.4%xda..Please Wait..>;%ac%?Ia<%t3%xdp %It |%t0.4%xda...Please Wait...>With the newest build (13020), the display changes all 0.4 sec., even the next track info is available, but it should change only all 3 seconds.

[NicolasP]

I found a "minor" problem. I use my wps on a RTC-modded H140(see http://www.rockbox.org/twiki/bin/view/Main/WpsIriverH100#RemoMartinelli).
This line has a problem with the different delay times:

Code: [Select]

%ac%?Ia<%t3%xdn %Ia |%t0.4%xda.Please Wait.>;%ac%?Ia<%t3%xdo %Id |%t0.4%xda..Please Wait..>;%ac%?Ia<%t3%xdp %It |%t0.4%xda...Please Wait...>With the newest build (13020), the display changes all 0.4 sec., even the next track info is available, but it should change only all 3 seconds.

This should be fixed with r13026. Thanks for the report

[AiZ]

Salut Nicolas,

Using r13027, I've some problems with sublines. I'm using a (slightly) modified darkgeek theme on my Nano and for testing purpose, I 've added a line.
If I input A;B;C;D in my wps, ouput is A;B;C;D
If I input %iaA;%itB;%fbC, output is (ID3 artist displayed Ok)A;(ID3 Track Title displayed Ok)B;(File Bitrate displayed Ok)C
If I input %ia;%it;%fb, output is OK.

So, it looks like that what it's just before the semi-colon must be a WPS tag to allow sublines working correctly. If it's ordinary text, cycling is broken.

Or perhaps I'm completely wrong. Can you test this, please ?

A pluche,


        AiZ

[NicolasP]

Salut Nicolas,

Using r13027, I've some problems with sublines. I'm using a (slightly) modified darkgeek theme on my Nano and for testing purpose, I 've added a line.
If I input A;B;C;D in my wps, ouput is A;B;C;D
If I input %iaA;%itB;%fbC, output is (ID3 artist displayed Ok)A;(ID3 Track Title displayed Ok)B;(File Bitrate displayed Ok)C
If I input %ia;%it;%fb, output is OK.

So, it looks like that what it's just before the semi-colon must be a WPS tag to allow sublines working correctly. If it's ordinary text, cycling is broken.

Or perhaps I'm completely wrong. Can you test this, please ?

A pluche,


        AiZ

You were right. It should be fixed with r13034.
Merci

[elborak]

So, if I understand correctly, adding a new WPS tag with the new tokenizer would consist of:

• Add a new enumeration to wps_token_type in gwps.h.
• Add whatever entries are appropriate to all_tags in wps_parser.c.
• Add a case for the new enum to the switch in dump_wps_tokens in wps_debug.c.
• Add a case for the new enum to the switch in get_tag in gwps-common.c.

(Note that this ignores any complications for RTC tags and doesn't address where to actually get the data to support the new tag, e.g. reading a previously unsupported ID3 tag.)

Do I have this right?  As an experiment, I updated http://www.rockbox.org/tracker/task/4961 to the current SVN and it seems to work fine.

[NicolasP]

So, if I understand correctly, adding a new WPS tag with the new tokenizer would consist of:

• Add a new enumeration to wps_token_type in gwps.h.
• Add whatever entries are appropriate to all_tags in wps_parser.c.
• Add a case for the new enum to the switch in dump_wps_tokens in wps_debug.c.
• Add a case for the new enum to the switch in get_tag in gwps-common.c.

(Note that this ignores any complications for RTC tags and doesn't address where to actually get the data to support the new tag, e.g. reading a previously unsupported ID3 tag.)

Do I have this right?  As an experiment, I updated http://www.rockbox.org/tracker/task/4961 to the current SVN and it seems to work fine.

You got it right. I looked at the WPS part of the patch and it seems fine to me

[roolku]

I don't know if you saw my comment on IRC:

[15:50] Nico_P: what do you mean by "strictly positive tiemout" ? does this mean %t0 is not possible anymore? This is a very useful feature (even documented with an example on CustomWPS) and should definitely stay.

There needs to be a way to turn sublines of conditionally and %t0 is what WPSs are using at the moment.

[Llorean]

It has been said that the goal is to keep all the old WPS functionality. Have you tried it yet? If it doesn't work, it's a problem, but it sounds like you haven't tested yet.

[NicolasP]

I don't know if you saw my comment on IRC:

[15:50] Nico_P: what do you mean by "strictly positive tiemout" ? does this mean %t0 is not possible anymore? This is a very useful feature (even documented with an example on CustomWPS) and should definitely stay.

There needs to be a way to turn sublines of conditionally and %t0 is what WPSs are using at the moment.

Should be fixed with r13047. This new version should work exactly like the old code.