Opened 19 years ago

Closed 18 years ago

#1264 closed defect (fixed)

Crash with Scalebar

Reported by: boily@… Owned by: sdlime
Priority: high Milestone:
Component: MapScript Version: 4.4
Severity: critical Keywords:
Cc:

Description

Hi,
after ours of test I've delimited the crash on LABEL in the map file.
When I comment out the LABEL section, it never crash.
LABEL
  TYPE TRUETYPE
  FONT myTime # configured by the FONSET TAG
  BUFFER 3
  COLOR 255 0 0
  POSITION CC
  SIZE 10
END
The same mapserver site is running perfectly on Linux mpaserver 4.1.
Thanks a lot
Roger

Attachments (3)

mapserverCrash.zip (956.0 KB ) - added by boily@… 19 years ago.
Files reproducing crash (cf Readme.txt)
file.map (5.5 KB ) - added by boily@… 19 years ago.
Entire .map file
Bug1264_data.zip (759.1 KB ) - added by assefa 19 years ago.
windows zip file of new lotti data

Download all attachments as: .zip

Change History (19)

comment:1 by boily@…, 19 years ago

Summary: Crash vs FontCrash vs LABEL

by boily@…, 19 years ago

Attachment: mapserverCrash.zip added

Files reproducing crash (cf Readme.txt)

comment:2 by boily@…, 19 years ago

In the Readme.txt I've added the two scale level
at witch it works and it crashes.

comment:3 by boily@…, 19 years ago

Type of shape file is Polygon (not Line) but it
still crash anyway

comment:4 by boily@…, 19 years ago

Summary: Crash vs LABELCrash with Scalebar
After many other tests, I've narrowed the problem to the scalebar
and reference map. Uncommenting scalebar object in the .map file
makes mapserver to crash without any apparently reason. 
Here's the .map scalebar config:
SCALEBAR
  IMAGECOLOR 241 239 226
  LABEL
    COLOR 0 0 0
    SIZE SMALL
  END
  SIZE 350 5
  COLOR 255 255 255
  BACKGROUNDCOLOR 0 0 0
  OUTLINECOLOR 0 0 0
  UNITS meters
  INTERVALS 10
  STATUS ON
END

comment:5 by sdlime, 19 years ago

What does the whole mapfile look like? What calls to MapServer cause the crash? 
Do you have a URL, script or something? Where'd you get the win32 build?

Steve

by boily@…, 19 years ago

Attachment: file.map added

Entire .map file

comment:6 by boily@…, 19 years ago

Hi Steve,
I have attached the entire .map file.
I have found that without the reference and scalebar objects it never crashes.
I have send in the first attachement the files to reproduce the crash. (but I
think this is not enough since there is no reference and scalebar i it)
I have no external link to the demo, sorry.
I've installed the 4.4.1 for win32 from mapserver site.
Thanks for help.
Roger

comment:7 by sdlime, 19 years ago

What does a crash look like? Server 500 error, segfault or what? How are you 
triggering a crash, via CGI or mapscript? Do you have sample URLs?

You might want to take this to the MapServer user mailing list too.

Steve

comment:8 by boily@…, 19 years ago

>What does a crash look like? Server 500 error, segfault or what? 
The crash on win32 just shows the normal crash windows of the mapserver.exe
and the Apache Logs (where I redirects MS_ERRORFILE) only shows a "Premature end
of script headers: mapserver.exe". Nothing else.

>How are you 
>triggering a crash, via CGI or mapscript? 
It's via CGI.
If I comment REFERENCE and SCALEBAR in the .map file, everything goes perfectly.
The problem is with the creation of those images. See previous comments for
configuation of scalebar in the .map file.

>Do you have sample URLs?
Unfortunetly no on windows. This application works perfectly on linux 
at http://sittel.gestin.it/index_sittel.php?c=procida

>You might want to take this to the MapServer user mailing list too
Done, with no success.
Thanks for help.

comment:9 by sdlime, 19 years ago

Cc: mapserver-bugs@… added
I can't really test Windows-only bugs so I've CC'd the DMSolutions guys in 
hopes they might have to look too.

Steve

by assefa, 19 years ago

Attachment: Bug1264_data.zip added

windows zip file of new lotti data

comment:10 by assefa, 19 years ago

Hi There,

 I just tried the test data/url and It did effectivly crashed on windows. Not 
sure what the exact reason is but the crash happened when reading the 
shapefile. Maybe the data or a certain record (could be record 6935 or one 
record below or above) is corrupted somehow but I did not investigate more. I 
used ogr2ogr to recreate the data and I have attached it here. It works fine 
with this new data

comment:11 by boily@…, 19 years ago

Hi,
thanks for your help.
Effectively I've tried your new shape and it works perfectly with the scalebar,
but still crash with the reference map.
I really do not understand why. I'm trying to make more tests to find it out.
Since I'm not able to recreate the crash only with the test data/url I posted
previously and I receive the same crash error with reference map, I should find
out a way to reproduce the error or find out myelf where is the bug in my entire
application (java).

3 questions:
1.Did you rebuilt the shape file only with the command 
>ogr2ogr  -f "ESRI Shapefile" lotti2.shp lotti.shp< ?
2. How did you trace out that the crash was when querying the shape file at
record 6935?
3. What is the best way (since error logs do not give enough informations) to
find out my bug?
Thanks again, Roger

comment:12 by assefa, 19 years ago

roger,

 1.Did you rebuilt the shape file only with the command 
>ogr2ogr  -f "ESRI Shapefile" lotti2.shp lotti.shp< ?

  Yes.

2. How did you trace out that the crash was when querying the shape file at
record 6935?

 I was in the MS VC++ debugger. If the application is built in debug mode when
Mapserver crashes, I can open the debugger and see the line where it crashed. 
It was trying to access that record at the time.

3. What is the best way (since error logs do not give enough informations) to
find out my bug?

 Either use a debugger or prepare me a url like last time for me to reproduce 
the crash easiliy. Can you reproduce it with a simple data/map file ?


 Note that since the last bug was related to the data, I would look in that 
direction first. I might be wrong but the scalebar and legend might not be in 
fault here.
  
 

comment:13 by sdlime, 18 years ago

Does anyone know if this is still an issue these days? I'd like to clear out
these older bugs if possible.

Steve

comment:14 by assefa, 18 years ago

I would say close it and if there are problemes, Roger can open a new bug. It
has been a while since this bug had no activities.

comment:15 by sdlime, 18 years ago

Resolution: fixed
Status: newclosed

comment:16 by sdlime, 18 years ago

Closing for now, marking as fixed although I don't know that's the case...

Steve
Note: See TracTickets for help on using tickets.