Jump to content

Malformed Gpx Files

sawdust

By sawdust
in Website

Followers 1

Recommended Posts

sawdust

+sawdust

Posted
Posted

I have been using GPX Sonar for quite some time. And have not had any major problems. I run the same querries each week and have been doing so for months now. All of a sudden GPX Sonar will not open one of the files. I get an error message that says the GPX file is malformed. The other 3 querries still work fine. I contacted Fabein, (The GPX Sonar Author) and he told me that the problem is not with the program, it is with the Querry. He suggested I allert Jeremy and go from there. After a couple of tries to Jeremy with no response, I thought I would ask all of you and see if anyone can help. Thanks in advance. Sawdust92

Link to comment
robertlipe

robertlipe

Posted
Posted

There are simple tests that can be run on a PQ (well, simple to geek types) that will tell you if the fault is of the producer or the validator.

 

If it is a recurring problem (i.e. after you follow Hemlocks suggestion) I can help analyze the file and cast my scepter of judgement upon it.

Link to comment
sawdust

+sawdust

Posted
  • Author
Posted
There are simple tests that can be run on a PQ (well, simple to geek types) that will tell you if the fault is of the producer or the validator.

 

If it is a recurring problem (i.e. after you follow Hemlocks suggestion) I can help analyze the file and cast my scepter of judgement upon it.

I have had this same problem for two weeks with the querry. I ran a new one in the same area under a different name and it had the same problem. The querry is centered on zip code 76010, 500 caches, any type, That I do not own, and that I have not found, and is active. I hope this is enough information for you to work with. Sawdust92

Link to comment
robertlipe

robertlipe

Posted
Posted

Well, only you can get your 500 that you haven't found. I just pulled the 500 closest active to that zip (obviously, a different set than you're looking at) and it validates fine.

 

(robertl) rjloud:/home/robertl/src/geo/mgpx

$ SAX2Count blah

blah: 1443 ms (23917 elems, 11509 attrs, 232358 spaces, 1551180 chars)

(robertl) rjloud:/home/robertl/src/geo/mgpx

$ gpsbabel -i gpx -f blah -o csv -F - | wc -l

500

 

Someone needs to analyze your exact PQ or you need to hand it to a validating parser. (I'm guessing if you had one and knew what to do with it, you'd have already done so. :-) I've sent you my email address so we can study this.

Link to comment
as77

+as77

Posted
Posted

Does GSAK open it?

Also, to see what the problem exactly is, you can do this: rename the gpx file to something.xml and then open it in Internet Explorer. It will show you the contents of the file and it will show an error message at the point where it encounters the error (this will be at the bottom of the page because it doesn't display anything further).

Link to comment
robertlipe

robertlipe

Posted
Posted

I've analyzed the bad PQ's from sawdust92. At first, I was thinking it was a problem with the MTA or MUA on his end, but since he has identical malformation in one of the two affected caches in two PQ's, I'm not so sure.

 

Log ID 3702211 in both PQs contains

         <Groundspeak:text encoded="False">1600. Took nothing but left Edgar Alpha Bug TB which next needs to move to a cache with a name that begins w       ithH.groundspeaktext

which has a clearly malformed entry. That's consistent in both. I

 

Log ID 4093462 contains 

         <Groundspeak:text encoded="False">Me and my new adventurer found this one after Celestial Gardens.  He found the stages 1, 2, 3 and I went in for the kill.  Easy to recall stages.  Going back from cache a lady informed us we were walkin in her back yard.[V] She had a hugh dog at her side stating that he runs that area.  There were no fences in that area, one does not know where park ends and her back yard begins. Just got o            utinahurry.Thanksforthefun.groundspeaktext

 

which seems a variation of the same.

 

If I request just those caches in a PQ, I don't see the damage. So I don't know if the PQ generator stumbled on these, his MTA/MUA consistently mangels them or just what...

Link to comment
sawdust

+sawdust

Posted
  • Author
Posted

Jeremy, I have never had any problems with this PQ before. Why now? Nothing has changed with my PC. All of my other PQ's work. I have even set up another PQ in the nearby area and it will not work either. I have 2 different PDA's and the same problem happens on both of them. How can this be fixed?

Link to comment
sawdust

+sawdust

Posted
  • Author
Posted
I've analyzed the bad PQ's from sawdust92. At first, I was thinking it was a problem with the MTA or MUA on his end, but since he has identical malformation in one of the two affected caches in two PQ's, I'm not so sure.

 

Log ID 3702211 in both PQs contains

         <Groundspeak:text encoded="False">1600. Took nothing but left Edgar Alpha Bug TB which next needs to move to a cache with a name that begins w       ithH.groundspeaktext

which has a clearly malformed entry. That's consistent in both. I

 

Log ID 4093462 contains 

         <Groundspeak:text encoded="False">Me and my new adventurer found this one after Celestial Gardens.  He found the stages 1, 2, 3 and I went in for the kill.  Easy to recall stages.  Going back from cache a lady informed us we were walkin in her back yard.[V] She had a hugh dog at her side stating that he runs that area.  There were no fences in that area, one does not know where park ends and her back yard begins. Just got o            utinahurry.Thanksforthefun.groundspeaktext

 

which seems a variation of the same.

 

If I request just those caches in a PQ, I don't see the damage. So I don't know if the PQ generator stumbled on these, his MTA/MUA consistently mangels them or just what...

If these are clearly malformed entries, can't one of you guys at Geocaching.com just tweek it, Problem gone? A helpfull person took my PQ and fixed it and then sent it back to me and it works fine now. Would be nice if the home office could do the same thing. Thanks for looking into this. Sawdust92

Link to comment
sawdust

+sawdust

Posted
  • Author
Posted (edited)
Did you get them in a zip file?

No, I did not get them in a zip file. I have always got them regular. Please explain to me how them being non-zipped could be a problem. I get 4 PQ's every week, none of them zipped. Three of them are still working fine. If the zip factor was the problem dosen't it stand to reason that all of the PQ's would be messing up? Sawdust92

Edited by sawdust92
Link to comment
sawdust

+sawdust

Posted
  • Author
Posted
When was the bad GPX file generated? Last weekend? The PQ machine ran out of disk space last weekend and stopped delivering PQs. Maybe this one was the last one generated the moment the disk filled up. :P

I had one generated today and it messed up. The problem has been traced to two particular cache pages. Robert Lipe ran a PQ from his machine and it was messed up for him as well. The caches are GCHJY1 and GC6D53. Robert's earlier post in this thread tells where the error is. Now hopefully it is just a matter of getting someone who can fix it to actually do the fix. Sawdust92

Link to comment
fizzymagic

+fizzymagic

Posted
Posted
Those documents are validated before they are emailed. Why don't you resend the query zipped and see if that fixes it?

Are you saying that the GPX generator is now emitting legal XML? And that it is getting validated? If true, that is wonderful, since it was not the last time I checked, about 6 months ago.

 

Or is the XML still the MS-specific version of XML that encodes entities illegally?

Link to comment
Jeremy

Jeremy

Posted
Posted

What I'm saying is that the objects are validating it, but yes, we found the illegal characters a while back and removed them as a workaround for characters that MSFT in their infinite wisdom designated valid.

 

However, don't mix that former (now fixed) issue with this current (and unable to recreate) issue.

Link to comment
robertlipe

robertlipe

Posted
Posted

To be clear, I have absolutely no problem placing blame when I can authoritatively do so. (Perhaps this is why The House seems to treat me with ill regard.)

 

In this case, I was able to see problems on specific log entries in a specific cache (one that appeared in two PQ's from the same user) and was able to identify and edit away the offenders but I was unable to observe it in manually generated GPX files for that specific cache or in a PQ or a manual download. So there's a big cloud of doubt for me whether the PQ is generated in a naughty way or whetther the users mail transport agent/mail user agent (MTA/MUA) is mangling the data in some obscure way.

 

At one point I (the chief GPSBabel-head) was getting hammered by receivers of defective PQ's from geocaching.com and that's not the case now. While the conspiracy theorist in me would like to head off problems before they escalate (and have developed skills to analyze this as a result of previous hammering) I can't say that I've heard of this problem lately. When I heard "malformed GPX" before, it was spread across the world and many times a week instead of isolated to one user.

 

Right now, I don't know where the problem is. Well, I know that it's on line XXXX of the PQ's that dude received, but I can't point my finger on which side bozoed. All I can say is the files he sent me were consistently mangled; which side did the mangling is a job for Quincy. (That's "CSI" for the under-30 crowd here. )

 

 

After a year of being spanked by the last round of malformed GPX, I'll personally thank the annoited admins to _not_ manually "fix" offending entries when chasing this kind of problem. Yes, it applies pain killer to the user having this specific problem, but it's no help to the 297 people behind him in line when you can't duplicate the problem becuase it was manuall tweaked before it could be analyzed. Please don't be an enabler.

 

Fizzymagic, it was late winter/early spring that the PQ generator was fixed to not mock XML. I suspect life is better than the last time you looked, but perhaps my "fix broken Groundspeak XML" patch is still in place.

Link to comment
sawdust

+sawdust

Posted
  • Author
Posted
If you know the ID of the Pocket Query, post it here (the filename). I'll look on the server itself to check the copy before it was transmitted into the ether.

44186.GPX Jeremy, it came to me bad again today. I am going to try to get it zipped and see if that does anything different. But all 4 of my other PQ's came in good order. Sawdust92

Link to comment
sawdust

+sawdust

Posted
  • Author
Posted (edited)
I received the file and it seems ok. Of course it was sent zipped. I'll have it sent to you zipped and see if that works.

Jeremy, I just recieved the zipped file. It works!! I will go into my PQ page and change all of them to be zipped. A BIG THANKS to all of you who helped to get me going again! Special Thanks to you Jeremy for making time to help! Sawdust92

Edited by sawdust92
Link to comment
Guest
This topic is now closed to further replies.
Followers 1
Go to topic listing
×
×
  • Create New...