[Announce] modRana: a flexible GPS navigation system - Page 144 - maemo.org

Active Topics

The N900 is now 15 years old! (8)
to Nokia N900 by whayong - 12 hrs, 6 mins ago
Sailfish OS for the Motorola Moto G7 Power (XT1955-5) - (ocean) (4)
to SailfishOS by Maemish - 5 days, 8 hrs ago
more...

Page 144 of 207

144

Thread Tools

whayong	2013-07-28 , 07:00
Posts: 177 \| Thanked: 118 times \| Joined on Nov 2007 @ California	#1431

Might have been asked already, or might be far out there, but can Modrana use Nokia's Ovi maps file for maps? Maps that have been downloaded from the PC suite onto the N900.

Quote & Reply |

The Following 3 Users Say Thank You to whayong For This Useful Post:
MartinK, petur, Wikiwide

Android_808	2013-07-28 , 08:35
Posts: 1,203 \| Thanked: 3,027 times \| Joined on Dec 2010	#1432

Might be a T&Cs conflict.

Quote & Reply |

The Following 3 Users Say Thank You to Android_808 For This Useful Post:
MartinK, whayong, Wikiwide

MartinK	2013-08-13 , 19:48
Posts: 1,548 \| Thanked: 7,510 times \| Joined on Apr 2010 @ Czech Republic	#1433

Originally Posted by petur

This redownload feature reminds me of a bug I saw lately: For some reason modRana failed to download a few tiles while on the road, driving somewhere. On the way back, when we passed the same place again, the failed tiles were still missing, not redownloaded.

I would expect that if a failed tile comes back into view it would try to download it again?

Historically modRana did not check that what id downloaded was actually a tile image. And some tile servers return a HTML error page when they can't return a tile (too low zoomlevel, too many tiles downloaded, generic tileserver error,...).

So it could happen that modRana would download something that was not actually a tile image and then after coming to the area, it will try to show the image and fail - and because the "image" is already downloaded, modRana won't try to download it again.

But quite some time ago I've added a check, that tests downloaded tiles and only saves them if they really are jpg/png/gif images (possible thanks to the magic numbers embedded in header of these image file formats).

If you are using some old tiles, it could be there are still some non-image-tiles present - or I forgot to apply the check in one of the tile download methods.

BTW, using the new redownload feature in the ON or Update modes should overwrite all tiles, including the offending ones, in the affected area.

Originally Posted by Wikiwide

One bug-vulnerability in Python-for-Maemo. When Python application decides that it wishes to use GPS, and Maemo knows that GPS was disabled by user, Python first asks user, with a system dialog, whether user wishes to "Enable positioning?", and then, irrespective of user's decision, force-enables GPS (blinking icon in status-bar; "Searching..." entry in status-menu), even though, according to user preferences, GPS is still disabled.
The only way to disable GPS is to either politely ask the benevolent application to not use GPS (because it apparently somehow disregards user preferences) or kill the malevolent (without an option to turn off GPS) application because it's apparently a loophole in Python-Maemo-GPS interaction (not stealthy, but still...)
Does anybody else have the same bug (Nokia N900, Fremantle, CSSU-Testing)?
Best wishes.

Thanks for reporting the issue, I'll check it out.

Registered as issue 26 on Github for the time being.

For the time being, you can just disable GPS in Options->Location->GPS, but I agree modRana should respect system settings.

Originally Posted by int_ua

Okay, now I'm sure it was Modrana. Sometimes I get a deadlock after leaving Modrana open for something like 20 minutes (autolock time being 5 minutes). It happens one in 10 cases maybe, the phone stays dark no matter what, reacts only with LED and only to charging. Anyone else experiencing this? It started before the last update. Sorry, no time to search the thread at the moment, leaving town.

I remember this happening a long time ago but it has been some time since I've last experienced this. It could be helpful if you could try to ssh the frozen device (if possible) to check RAM and CPU usage and/or set modRana to log to file and extract the file.

One possible cause could be some sort of memory leak, that would make the N900 swap and be unresponsive. Another candidate could be an endless loop somewhere in modRana or some library it uses pinning the CPU.

Or maybe a GPU driver crash (I remember stuff like that happening in the past).

__________________
modRana: a flexible GPS navigation system
Mieru: a flexible manga and comic book reader
Universal Components - a solution for native looking yet component set independent QML appliactions (QtQuick Controls 2 & Silica supported as backends)

Last edited by MartinK; 2013-08-13 at 19:54.

Quote & Reply |

The Following 3 Users Say Thank You to MartinK For This Useful Post:
int_ua, skanky, Wikiwide

MartinK	2013-08-13 , 20:01
Posts: 1,548 \| Thanked: 7,510 times \| Joined on Apr 2010 @ Czech Republic	#1434

Originally Posted by whayong

Might have been asked already, or might be far out there, but can Modrana use Nokia's Ovi maps file for maps? Maps that have been downloaded from the PC suite onto the N900.

Even if I disregard all the issues with re/distributing the map files, they are in some undocumented proprietary format.

And even then, the files contain just the vector data for the given are. You still need a renderer (which modRana does not have yet), that would draw the map layer based on the vector data for the given area.

On the other hand, the currently planed vector map rendering architecture consist of two parts - a vector geometry source and the Kothic map renderer using data from this source.

So once that is up and someone reverse-engineered the Nokia maps file format and wrote a vector geometry source for it, they might be of some use. Not very probably though & and would require a lot of effort that could be spent to to improve rendering or open vector data storage & generation.

Quote & Reply |

The Following 7 Users Say Thank You to MartinK For This Useful Post:
cheve, fw190, nokiabot, sixwheeledbeast, skanky, whayong, Wikiwide

petur	2013-08-13 , 22:57
Posts: 650 \| Thanked: 497 times \| Joined on Oct 2008 @ Ghent, Belgium	#1435

Originally Posted by MartinK

Historically modRana did not check that what id downloaded was actually a tile image. And some tile servers return a HTML error page when they can't return a tile (too low zoomlevel, too many tiles downloaded, generic tileserver error,...).

So it could happen that modRana would download something that was not actually a tile image and then after coming to the area, it will try to show the image and fail - and because the "image" is already downloaded, modRana won't try to download it again.

But quite some time ago I've added a check, that tests downloaded tiles and only saves them if they really are jpg/png/gif images (possible thanks to the magic numbers embedded in header of these image file formats).

If you are using some old tiles, it could be there are still some non-image-tiles present - or I forgot to apply the check in one of the tile download methods.

You may remember that I was somehow involved in that history (hit and/or reported that issue).

But that cache is long gone, lost in a reflash....

I hit this one only once, during normal download-as-you-go, where I lost connectivity, the tiles failed to download, and later (with connectivity) didn't get downloaded.

If I ever see it again I'll try to note down exact circumstances or load from commandline.

__________________
Affordable mobile internet in Belgium: Try Mobile Vikings
2 GB, 1000 SMS and 15 euro of talk time for.... 15 euro

Quote & Reply |

The Following 5 Users Say Thank You to petur For This Useful Post:
Estel, MartinK, nokiabot, sixwheeledbeast, Wikiwide

saponga	2013-08-14 , 20:45
Posts: 432 \| Thanked: 917 times \| Joined on Jun 2011	#1436

Originally Posted by Estel

Since one of last updated, ModRana hangs at calculating online routing (needs killing). The following error appears in terminal:

Code:

** dialog accepted **
Clicked, sending route:addressRoute
route: address routing
route: routing from 52.222719,18.251084 to Piekary 16, Poznań
onlineServices: worker starting
Sending message: ml:notification:workInProgressOverlay:enable
worker: routing from 52.222719,18.251084 to Piekary 16, Poznań
Sending message: ms:turnByTurn:start:enabled
tbt: starting worker thread
TBTWorker: started
Sending message: ml:notification:m:use at own risk, watch for cliffs, etc.;3
notification: message: use at own risk, watch for cliffs, etc., timeout: 3.0
notification@dmod: message: use at own risk, watch for cliffs, etc., timeout: 3.0
Xlib: unexpected async reply (sequence 0x25f8)!

/Estel

It's the same here even though i've already updated to 0.42.1.

Quote & Reply |

The Following 2 Users Say Thank You to saponga For This Useful Post:
Estel, Wikiwide

MartinK	2013-08-15 , 16:22
Posts: 1,548 \| Thanked: 7,510 times \| Joined on Apr 2010 @ Czech Republic	#1437

Originally Posted by saponga

It's the same here even though i've already updated to 0.42.1.

Searching for similar error messages seams to pinpoint the cause to using the (non-threadsafe) GTK from multiple threads.

This is quite probably the issue in this case - the route search runs in a thread and might be manipulating the GTK context once a result is returned. Anyway, I'll investigate the relevant code in modRana and will try to do something bout it.

Quote & Reply |

The Following 3 Users Say Thank You to MartinK For This Useful Post:
rotoflex, saponga, Wikiwide

	nokiabot	2013-08-15 , 18:30
	Posts: 1,974 \| Thanked: 1,834 times \| Joined on Mar 2013 @ india	#1438

goog sat maps are giving error i just installed from repos after a reflash

Quote & Reply |

The Following User Says Thank You to nokiabot For This Useful Post:
Wikiwide

petur	2013-08-15 , 20:51
Posts: 650 \| Thanked: 497 times \| Joined on Oct 2008 @ Ghent, Belgium	#1439

Just came up with an idea for better map orientation:

What if you would take into account multiple positions to figure out the heading when moving at slow speed. Currently when halted or moving very slow, the map will rotate even if your heading isn't changing. So for slow speeds maybe watch multiple points to detect heading?

Just an idea... if I didn't write it down here I might forget

__________________
Affordable mobile internet in Belgium: Try Mobile Vikings
2 GB, 1000 SMS and 15 euro of talk time for.... 15 euro

Quote & Reply |

The Following 4 Users Say Thank You to petur For This Useful Post:
int_ua, MartinK, whayong, Wikiwide

brandon	2013-08-17 , 00:35
Posts: 4 \| Thanked: 9 times \| Joined on Aug 2011	#1440

In the work i do professionally, we ignore the heading when the speed in <5kph. Perhaps an idea is to augment the heading at low speeds is to simply calculate it from the 2 (or more) most recent positions (i.e. completely ignore the heading contained in the gps data).