Active Topics

 



Notices


Reply
Thread Tools
Posts: 189 | Thanked: 143 times | Joined on Nov 2009
#171
This one for example:

https://github.com/smokku/pebble/com...250927a832315f


I found it out because in your code there is no message "Can not decode message data length invalid".

grep -R "Can not decode message data length invalid" * reveals nothing as for master of smokku it does.
 
javispedro's Avatar
Posts: 2,355 | Thanked: 5,249 times | Joined on Jan 2009 @ Barcelona
#172
Originally Posted by MaemoUser View Post
This one for example:

https://github.com/smokku/pebble/com...250927a832315f
That one is certainly merged in, being quite old...
It's probably a brand new crash.

I have rewritten the way packets are read because I was "losing" packets when compared to Android's pebbled, and I suspected it's because they were close to the limit of what can be sent in a single RFCOMM frame (which is around ~330 bytes). In the original version either the entire packet fits in a single ->readAll() or it gets dropped. However, it turned out that only the DATA_LOG messages are large, which are actually mostly useless...

I have commited some changes right now to clean up the frame decoding parts a bit. I doubt they'll fix whatever crash you're seeing. Try to get a stack trace, or look at the logs for any hint.
 
Posts: 189 | Thanked: 143 times | Joined on Nov 2009
#173
But the code is not included if I search for this code part in the js-testing branch. I will try to merge it manually and see if it works afterwards.
 
javispedro's Avatar
Posts: 2,355 | Thanked: 5,249 times | Joined on Jan 2009 @ Barcelona
#174
Originally Posted by MaemoUser View Post
But the code is not included if I search for this code part in the js-testing branch.
Yes it is (read that with Monty Python John Cleese's voice ).

If you search for _that_ specific string you won't find it, because I've rewritten that part. Every other change including the voice related ones are there.

EDIT: If you believe you're getting the "0" packets that commit tried to workaround I'd like to know more information before simply readding that workaround. In my tests the official Android app _does not_ ignore "0" packets, so if the watch is sending them we must be doing something wrong. I would prefer to diagnose that and be accurate to the Android app behavior.

If you are using QtCreator, can you just launch pebbled it in debug mode and pinpoint where in pebbled's source does it crash? And also add the few debug log lines from before that?

Last edited by javispedro; 2014-12-08 at 15:30.
 
Posts: 189 | Thanked: 143 times | Joined on Nov 2009
#175
Ok, will do it when I am at home .
 
javispedro's Avatar
Posts: 2,355 | Thanked: 5,249 times | Joined on Jan 2009 @ Barcelona
#176
I have built a package using QtCreator itself. I have implemented almost everything I wanted to implement, so now only the boring parts remain, like the UI for managing apps . This may take a while, so I decided to put out a package for now.

Note the disclaimers still apply -- I have only been using this for a couple of days, so it may be "problematic". Bugs may cause to e.g. corrupt the apps installed on the watch, in which case you may need to "revert to factory settings". Since I haven't implemented firmware upgrades or anything like that, it should be impossible to actually brick a Pebble.

If you need to return to Smoku's branch ensure you fully uninstall my package first.
 
Posts: 189 | Thanked: 143 times | Joined on Nov 2009
#177
Ok, looks like I have a strange version of Twebble it crashes app and watch immediatly. Weather works without a problem. Which version of Twebble do you use?
 
javispedro's Avatar
Posts: 2,355 | Thanked: 5,249 times | Joined on Jan 2009 @ Barcelona
#178
Originally Posted by MaemoUser View Post
Ok, looks like I have a strange version of Twebble it crashes app and watch immediatly. Weather works without a problem. Which version of Twebble do you use?
Does it happen the moment you try to open it on the watch, even when it is disconnected?
This may happen if it's been misinstalled or corrupted. Try uninstalling it (either using qdbus org.pebbled .... , or an Android/iOS device) and then reinstalling it.

I use "1.1.0", downloaded from http://izqui.me/twebble .
 
Posts: 189 | Thanked: 143 times | Joined on Nov 2009
#179
The problem with Tweeble seems to be solved in the latest git commits . Tweeble now working fine for me.
 

The Following User Says Thank You to MaemoUser For This Useful Post:
javispedro's Avatar
Posts: 2,355 | Thanked: 5,249 times | Joined on Jan 2009 @ Barcelona
#180
Originally Posted by MaemoUser View Post
The problem with Tweeble seems to be solved in the latest git commits . Tweeble now working fine for me.
Thanks, I've updated the package accordingly. It seems like the crash was triggered when the Twitter login page took more than a few milliseconds to respond
 
Reply

Tags
pebble, pebble time, pebbled


 
Forum Jump


All times are GMT. The time now is 13:31.