0003276: Floating point errors

(0018049)
vdemidov (manager)
19-09-2017 07:15

Please, try latest nightly version. And use Debug version for reproduce this bug. After crash attach *.elf file.

(0018051)
dzwager (reporter)
21-09-2017 08:27

I switched to nightly, downloaded a 7z file. I opened it but cannot figure out what to do with the contents. Please explain.

(0018052)
vdemidov (manager)
21-09-2017 10:41

Save all files with replacement to SAS folder and run SASPlanet.Debug.exe

(0018053)
dzwager (reporter)
21-09-2017 13:23

I extracted the files an copied all to the SAS folder. After that nothing worked anymore.

I then tried to uninstall the software but couldn't find it in the list of programs to uninstall. So then I deleted the SAS folder.

I will now try to reinstall.

Sorry I haven't used Windows since 2007. I bought a Windows computer especially for your SAS planet software but this Windows 10 is, well, impossible to use.

(0018081)
dzwager (reporter)
15-10-2017 06:46

When I copy the nightly files to the SASPlanet folder then (1) SASPlanet will not run anymore and (2) the debug.exe also does not run.

When I try to execute any of the exe files in only the nightly file folder then nothing happens.

In both cases I get a message "The file or directory is corrupted and unreadable"

So now SASPlanet still does not support my Garmin GPS18x USB.

Also the Google and Bing satellite images are not displayed properly (very very low resolution only).

(0018082)
zed (manager)
15-10-2017 10:01

Create new empty folder and use 7zip to unpack archive to it. Then run SASPlanet.NonUnicode.Debug.exe, reproduce error, close SAS, find SASPlanet.NonUnicode.Debug.elf file with error log and attach it to this issue.

(0018083)
dzwager (reporter)
15-10-2017 11:11

Thanks the debug file actually runs without the problems I described.

However I did notice the following problems:

1. The GPS positions me about 60 m south of where I really am on the Google sat image.

2. The GPS speed shows unrealistically high values.

(0018084)
dzwager (reporter)
15-10-2017 11:26

In the regular SASPlanet app I just updated the google version number to 745. But I am right away again getting the floating point errors.

SEE ATTACHED PNG file

(0018085)
zed (manager)
15-10-2017 11:37

What about time? It shows correct value?

(0018086)
dzwager (reporter)
15-10-2017 11:49

The debug version shows a local time anywhere between 03:08 and 03:45

The "stable" version shows similar local time.

For both the UTC time shown is 20:08 - 20:45

The local time here (Jakarta) is 18:45

(Sorry my computer clock was set to Singapore before, 1 hour later. I just fixed that)

(0018087)
zed (manager)
15-10-2017 11:52
edited on: 15-10-2017 11:52

Try this exe: https://yadi.sk/d/FRYuORmy3NmG75 It should print more information about error.

Put SASPlanet.3276.exe to folder with nightly build and run. Log will be named as SASPlanet.3276.elf.

(0018088)
dzwager (reporter)
15-10-2017 12:23

This showed same behaviour as the debug version.

UTC time varied between 20:49 and 21:24

Local time varied between 03:49 and 04:24

Actual local time is 19:23

(0018089)
zed (manager)
15-10-2017 12:34

I don't see error in attached log. Do you reproduce it? Can you take screenshot with error message?

(0018097)
dzwager (reporter)
17-10-2017 12:04

Correct, the debug version shows no floating point errors but the "stable" version does.

The debug version does show an incorrect local and UTC time, even though the GPS pinned my location correctly.

(0018098)
zed (manager)
17-10-2017 19:27
edited on: 17-10-2017 19:32

Your first log (SASPlanet.NonUnicode.Debug.elf) has error:



EConvertError: 14722840.71 Julian cannot be represented as a DateTime.

Try to reproduce this error with SASPlanet.3276.exe. I need to know, why this error happened. It maybe because of invalid time value or bug in SunCalc library.

And forget about the "stable" version. It's outdated. Use latest nightly build - it's more stable then last release. If nightly version works fine, without "floating point error", then here is nothing to do - bug was somehow fixed.

Incorrect time is an another issue and you should open a new one ticket.

(0018106)
dzwager (reporter)
23-10-2017 12:38

Sorry for late reply. Yes I will report the time issue as a separate ticket. Thanks for your assistance.

Users who viewed this issue
User List	Anonymous (8948x), vdemidov (20x), zed (23x), dzwager (25x), AlexWhiter (2x), bk99 (2x), Garl (4x), rass (1x), ingener (1x), Alexander (1x)
Total Views	9027
Last View	22-11-2024 02:59

Issue History
Date Modified	Username	Field	Change
19-09-2017 07:01	dzwager	New Issue
19-09-2017 07:15	vdemidov	Note Added: 0018049
19-09-2017 07:15	vdemidov	Status	new => feedback
21-09-2017 08:27	dzwager	Note Added: 0018051
21-09-2017 08:27	dzwager	Status	feedback => new
21-09-2017 10:41	vdemidov	Note Added: 0018052
21-09-2017 10:41	vdemidov	Status	new => feedback
21-09-2017 13:23	dzwager	Note Added: 0018053
21-09-2017 13:23	dzwager	Status	feedback => new
21-09-2017 13:34	vdemidov	Priority	immediate => low
21-09-2017 13:35	vdemidov	Status	new => feedback
15-10-2017 06:46	dzwager	Note Added: 0018081
15-10-2017 06:46	dzwager	Status	feedback => new
15-10-2017 10:01	zed	Note Added: 0018082
15-10-2017 10:02	zed	Status	new => feedback
15-10-2017 11:09	dzwager	File Added: SASPlanet.NonUnicode.Debug.elf
15-10-2017 11:11	dzwager	Note Added: 0018083
15-10-2017 11:11	dzwager	Status	feedback => new
15-10-2017 11:11	dzwager	File Added: GpsInfo.ini
15-10-2017 11:25	dzwager	File Added: floating point error.png
15-10-2017 11:26	dzwager	Note Added: 0018084
15-10-2017 11:37	zed	Note Added: 0018085
15-10-2017 11:49	dzwager	Note Added: 0018086
15-10-2017 11:52	zed	Note Added: 0018087
15-10-2017 11:52	zed	Note Edited: 0018087	View Revisions
15-10-2017 12:22	dzwager	File Added: SASPlanet.3276.elf
15-10-2017 12:23	dzwager	Note Added: 0018088
15-10-2017 12:34	zed	Note Added: 0018089
17-10-2017 12:04	dzwager	Note Added: 0018097
17-10-2017 19:27	zed	Note Added: 0018098
17-10-2017 19:32	zed	Note Edited: 0018098	View Revisions
23-10-2017 12:38	dzwager	Note Added: 0018106
23-10-2017 20:06	zed	Status	new => closed
23-10-2017 20:06	zed	Assigned To	=> zed
23-10-2017 20:06	zed	Resolution	open => unable to reproduce

Notes
(0018049) vdemidov (manager) 19-09-2017 07:15	Please, try latest nightly version. And use Debug version for reproduce this bug. After crash attach *.elf file.

(0018051) dzwager (reporter) 21-09-2017 08:27	I switched to nightly, downloaded a 7z file. I opened it but cannot figure out what to do with the contents. Please explain.

(0018052) vdemidov (manager) 21-09-2017 10:41	Save all files with replacement to SAS folder and run SASPlanet.Debug.exe

(0018053) dzwager (reporter) 21-09-2017 13:23	I extracted the files an copied all to the SAS folder. After that nothing worked anymore. I then tried to uninstall the software but couldn't find it in the list of programs to uninstall. So then I deleted the SAS folder. I will now try to reinstall. Sorry I haven't used Windows since 2007. I bought a Windows computer especially for your SAS planet software but this Windows 10 is, well, impossible to use.

(0018081) dzwager (reporter) 15-10-2017 06:46	When I copy the nightly files to the SASPlanet folder then (1) SASPlanet will not run anymore and (2) the debug.exe also does not run. When I try to execute any of the exe files in only the nightly file folder then nothing happens. In both cases I get a message "The file or directory is corrupted and unreadable" So now SASPlanet still does not support my Garmin GPS18x USB. Also the Google and Bing satellite images are not displayed properly (very very low resolution only).

(0018082) zed (manager) 15-10-2017 10:01	Create new empty folder and use 7zip to unpack archive to it. Then run SASPlanet.NonUnicode.Debug.exe, reproduce error, close SAS, find SASPlanet.NonUnicode.Debug.elf file with error log and attach it to this issue.

(0018083) dzwager (reporter) 15-10-2017 11:11	Thanks the debug file actually runs without the problems I described. However I did notice the following problems: 1. The GPS positions me about 60 m south of where I really am on the Google sat image. 2. The GPS speed shows unrealistically high values.

(0018084) dzwager (reporter) 15-10-2017 11:26	In the regular SASPlanet app I just updated the google version number to 745. But I am right away again getting the floating point errors. SEE ATTACHED PNG file

(0018085) zed (manager) 15-10-2017 11:37	What about time? It shows correct value?

(0018086) dzwager (reporter) 15-10-2017 11:49	The debug version shows a local time anywhere between 03:08 and 03:45 The "stable" version shows similar local time. For both the UTC time shown is 20:08 - 20:45 The local time here (Jakarta) is 18:45 (Sorry my computer clock was set to Singapore before, 1 hour later. I just fixed that)

(0018087) zed (manager) 15-10-2017 11:52 edited on: 15-10-2017 11:52	Try this exe: https://yadi.sk/d/FRYuORmy3NmG75 It should print more information about error. Put SASPlanet.3276.exe to folder with nightly build and run. Log will be named as SASPlanet.3276.elf.

(0018088) dzwager (reporter) 15-10-2017 12:23	This showed same behaviour as the debug version. UTC time varied between 20:49 and 21:24 Local time varied between 03:49 and 04:24 Actual local time is 19:23

(0018089) zed (manager) 15-10-2017 12:34	I don't see error in attached log. Do you reproduce it? Can you take screenshot with error message?

(0018097) dzwager (reporter) 17-10-2017 12:04	Correct, the debug version shows no floating point errors but the "stable" version does. The debug version does show an incorrect local and UTC time, even though the GPS pinned my location correctly.

(0018098) zed (manager) 17-10-2017 19:27 edited on: 17-10-2017 19:32	Your first log (SASPlanet.NonUnicode.Debug.elf) has error: `EConvertError: 14722840.71 Julian cannot be represented as a DateTime.` Try to reproduce this error with SASPlanet.3276.exe. I need to know, why this error happened. It maybe because of invalid time value or bug in SunCalc library. And forget about the "stable" version. It's outdated. Use latest nightly build - it's more stable then last release. If nightly version works fine, without "floating point error", then here is nothing to do - bug was somehow fixed. Incorrect time is an another issue and you should open a new one ticket.

(0018106) dzwager (reporter) 23-10-2017 12:38	Sorry for late reply. Yes I will report the time issue as a separate ticket. Thanks for your assistance.

Relationships

SASGIS