Wrong GPX Correlation

[StephaneP]

What steps will reproduce the problem?

1. Load the attached gpx
2. Load the attached picture
3. Ignore the integrated location and try to correlate the picture's position with the gpx

What is the expected result?

You can move the picture on the gpx with the offset.

What happens instead?

The picture's location moves back and forth.

Please provide any additional information below. Attach a screenshot if possible.

Josm seems to read the gpx incorrectly, as the speed indicated on the picture (top left) is wrong.
Warning : The original gnns record is a 5hz nmea.

Other Josm release
tested 8491 OK
tested 9329 NOK

URL:http://josm.openstreetmap.de/svn/trunk
Repository:UUID: 0c6e7542-c601-0410-84e7-c038aed88b3b
Last:Changed Date: 2016-02-03 00:53:09 +0100 (Wed, 03 Feb 2016)
Build-Date:2016-02-03 02:32:49
Revision:9720
Relative:URL: ^/trunk

Identification: JOSM/1.5 (9720 fr) Windows 8.1 64-Bit
Memory Usage: 1145 MB / 5461 MB (170 MB allocated, but free)
Java version: 1.8.0_71-b15, Oracle Corporation, Java HotSpot(TM) 64-Bit Server VM
Dataset consistency test: No problems found

Plugins:
- DirectUpload (31912)
- ImportImagePlugin (31772)
- Mapillary (32040)
- OpeningHoursEditor (31772)
- PicLayer (31895)
- SimplifyArea (31895)
- apache-commons (31895)
- apache-http (31895)
- cadastre-fr (31772)
- continuosDownload (1446070193)
- download_along (31772)
- editgpx (31772)
- ejml (31895)
- geotools (31895)
- jts (31772)
- livegps (31772)
- log4j (31895)
- opendata (31937)
- photo_geotagging (31895)
- photoadjust (32016)
- reverter (32005)
- tag2link (31910)
- todo (29154)
- turnrestrictions (31895)
- utilsplugin2 (32018)

[StephaneP]

gpx

[simon04]

In 9726/josm:

see #12485 - Wrong GPX Correlation, add unit test

Regression from r9383, confusion between s and ms.

[Don-vip]

something remains to do?

[simon04]

StephaneP, would you please confirm that r9726 fixed the issue/s.

[anonymous]

Tested with r9733, and no, it doesn't work correctly.

[StephaneP]

I don't know if it has something related to this bug, but the older release didn't draw the gpx like r9733. Now, there are some strange low speed then high speed segments :

btw : I wrote it was NOK with r9329. I was wrong.

[simon04]

What happened to attachment:gpx_test.gpx​ – the waypoints are in a non-chronological order?

        <time>2016-01-25T06:47:50.600Z</time>
        <time>2016-01-25T06:47:50.200Z</time>
        <time>2016-01-25T06:47:50Z</time>
        <time>2016-01-25T06:47:50.600Z</time>
        <time>2016-01-25T06:47:52.200Z</time>
        <time>2016-01-25T06:47:51.200Z</time>
        <time>2016-01-25T06:47:50.800Z</time>
        <time>2016-01-25T06:47:51.600Z</time>
        <time>2016-01-25T06:47:52.200Z</time>
        <time>2016-01-25T06:47:52.400Z</time>
        <time>2016-01-25T06:47:51.800Z</time>
        <time>2016-01-25T06:47:52.600Z</time>
        <time>2016-01-25T06:47:53.600Z</time>
        <time>2016-01-25T06:47:53.200Z</time>
        <time>2016-01-25T06:47:53.600Z</time>
        <time>2016-01-25T06:47:53.200Z</time>
        <time>2016-01-25T06:47:52.400Z</time>
        <time>2016-01-25T06:47:53.600Z</time>
        <time>2016-01-25T06:47:53.200Z</time>
        <time>2016-01-25T06:47:53.800Z</time>
        <time>2016-01-25T06:47:53.600Z</time>
        <time>2016-01-25T06:47:54.200Z</time>
        <time>2016-01-25T06:47:55.600Z</time>
        <time>2016-01-25T06:47:55.200Z</time>
        <time>2016-01-25T06:47:55Z</time>
        <time>2016-01-25T06:47:54Z</time>

[StephaneP]

Replying to simon04:

What happened to attachment:gpx_test.gpx​ – the waypoints are in a non-chronological order?

I don't know. I opened the nmea file in Josm, converted it to a layer, cut it, then converted it to gpx.

I use directly the nmea file to try to correlate the pictures. I can upload it if you want.

[simon04]

Replying to StephaneP:

I can upload it if you want.

Yes, please!

[StephaneP]

full_nmea_file

[StephaneP]

Replying to simon04:

Yes, please!

Done
btnmeatrack_2016-01-25_06-04-59.zip​

[simon04]

In 9739/josm:

see #12485 - Fix parsing of sub-seconds

Sometimes there are really strange bugs …

[simon04]

In 9740/josm:

see #12485 - Add unit test for nmea reading

[simon04]

r9739 should fix the non-chronological order issue. A new build is available in ≈20 minutes from https://josm.openstreetmap.de/jenkins/job/JOSM/889/jdk=JDK7/

Good that you're tracking with 5Hz, otherwise we this bug would not have been found in ages :)

attachment:2016-02-04-222156_2400x1920_scrot.png​​ shows that the speed colouring should be fixed.

[StephaneP]

Thank you ! I will test r9739 tomorrow.

Yes, with #12486 #12485 #10235 I'm in the millisecond world :)

[StephaneP]

tested with r9742, and the bug isn't here anymore. Thank you Simon !