#1044 closed defect/bug (Incomplete/Missing infos)
Crash on routing (Sony nav-u92t with 64mb ram)
Reported by: | martin tlustos | Owned by: | KaZeR |
---|---|---|---|
Priority: | major | Milestone: | version 0.5.1 |
Component: | port/wince | Version: | git master |
Severity: | normal | Keywords: | crash, routing |
Cc: |
Description
The first time after deleting navit.log and restarting my device, I can select a route and it will calculate and display it. After a while navit will crash. The next time I restart it, it will not complete to calculate the route, but crash before finishing. With every restart the time until navit crashes get shorter. Starting navit and quickly disable routing in the menu will prevent the crash. Deleting the navit.log file will prolong the time until the next crash happens.
I suspect it has to do with lack of memory. I still have some internal storage memory, and I will try to extend memory with an MS Pro Duo Card and report if there are differences.
Attachments (3)
Change History (16)
comment:1 Changed 11 years ago by tryagain
comment:2 Changed 11 years ago by martin tlustos
O.k. I'll try to help as I can...
The sony nav-u92t is a WinCE based one (an older one, but I don't know exactly what version).
Creating a folder named navit.log helps prevent the log file. (It's getting big pretty fast, but around 99% of the entries is: vehicle_wince:wince_reader_thread:GPS data comes too fast. Have to wait here)
I download navit from svn, so I always use the latest version. I just installed the version from yesterday.
I don't know how big the log file is at the first crash, as I normally drive and can't have a look.
It normally does not crash if I drive around without routing, although I'm not 100% sure, it might just take more time (I don't use it that often any more, because of the crashes).
I'll give it a try navigating next time I drive around and report back.
I'm also attaching the .log file I found today when putting the new version on it.
comment:3 Changed 11 years ago by tryagain
Do you have any news after running navit with log disabled? Does it crash with/without a route when no logging is performed?
Since r5118 WinCE versions should be less noisy in the log. That may help having navit run with logfile enabled but not crashing because of it growing too fast.
comment:4 Changed 11 years ago by martin tlustos
O.k., did some testing. Driving around without routing enabled always worked, so I did not test that...
First, I let it lay around with routing enabled for about 1 hour - no crash.
Then I did three trips, one short (around 30km), two long (~200km). Both went fine, navit did not crash even with routing enabled. That is, as long as I followed the route... I left the route several times, and navit crashed twice, but not the third time.
I also tried a even longer route (~300km), but navit crashed while calculating. But before it had also crashed while calculating the 200km route, so there's still a big improvement.
I'll try to install the new svn and reenable the log to see whether it works better now. I'll report back afterwards.
comment:5 Changed 11 years ago by martin tlustos
O.k., installed it and did a short trip (about 13km in town with routing enabled). Navit crashed twice, once when I left the route and once while I was on route. I'll attach the new log file. I also want to try to enable debug output, but I'll have to wait until I get a Memory Stick Pro Duo Card which a friend will give to me in about a week.
comment:6 Changed 11 years ago by tryagain
Well, logging should be really silenced in r5132, which hopefully will be built tonight. Last builds were even more loud than before because of bug in my commit.
I have some more ideas to test. You may try to reduce memory usage by renaming xpm folder so navit wouldnt find any images and memory requirement would be reduced. If that would help, we could play with images caching so that less memory would be required at expense of CPU load.
comment:7 Changed 11 years ago by martin tlustos
O.k. I downloaded the newest version yesterday and gave it a try. navit.log still isn't silent, but this time it's because I also renamed the xpm directory. Renaming this directory resulted in slightly quicker startup and routing, but nothing essential. It also crashed twice on a 13km trip (and it was much more diffficult to tip on an menu entry of course... ;-). So it seems that image caching changes wouldn't help too much... Anything else I could do/try/help?
comment:8 Changed 11 years ago by tryagain
Please try with r5141 or later. I have found a leak that were activated when you change position with route displayed. That might be one of the main causes of your crashes.
comment:9 Changed 11 years ago by tryagain
btw your navit.log.2.zip seems to be done with quite an old version (at least before r5118).
comment:10 Changed 10 years ago by usul
- Component changed from core to port/wince
- Keywords crash routing added
- Milestone set to version 0.5.1
Do we know anything new here? I never had a crash on my Medion S3857 (but this has more resources) with the SVN.
As stability is in focus of the next hotfix, I will schedule it for 0.5.1
comment:11 Changed 10 years ago by martin tlustos
Well, I did not have any crashes after disabling logging (which I don't need anyway). I'm away from home right now, so can't do any further testing at the moment.
comment:12 Changed 9 years ago by usul
- Resolution set to Incomplete/Missing infos
- Status changed from new to closed
Please give us updates if the crashes appear again. As the trend of this ticket look very good, I pause it for now.
comment:13 Changed 9 years ago by martin tlustos
- Severity set to normal
Yes, I still have crashes. I have navit.log disabled (I created a folder with the same name), but it it takes longer and does not seem to crash quicker every time you restart it... I have to add that with my new cellphone I use navit for android normally, but if you need further testing, I can do it.
Hi!
Having such a resource limited device might help us hunting some memory leaks.
To begin, please answer following questions.
If described workaround doesn't help at all, or only increases the time needed to crash, but not prevents it:
tryagain