[mapserver-users] Mapserver 5.0.3 on Ubuntu 8.10-AMD64 TrueType font problem

[Martin Koch]

Thanks for your interest,

When I change something in the path to the fontset, I invoke this error

msLoadFontset(): Unable to access file. Error opening
fontset /mapserver/mapdata/font/fonts.txt. (which is the crippled path)

When I change something in the path to the desired font in the fontset I
invoke this error...
msGetLabelSize(): TrueType Font error. Could not find/open font

This made me conclude my paths are correct and/or the font is opened
with the correct paths..

Directing the errors to apache's error.log yields these two lines:

[Wed Apr 15 08:40:15 2009] [error] [client 192.168.34.85] CGI Request 1
on process 31346
[Wed Apr 15 08:40:16 2009] [error] [client 192.168.34.85] Premature end
of script headers: mapserv

from which I conclude the process has started and prematurely ended...

Kind regards, 

Martin

-  
_____________________________
Martin Koch

Teamleider data

Bridgis b.v.
Sint Walburgkerkpad 3
4001 MG Tiel


Op vrijdag 10-04-2009 om 09:55 uur [tijdzone -0400], schreef Daniel
Morissette:
> Martin Koch wrote:
> > 
> > I used the same mapfiles. Everything works fine appart but when
> > I use TrueType fonts for displaying labels I get a '500 Internal Server
> > Error'. A default bitmap-font works well, but not the truetype font's I
> > refer to in my fontset.
> > 
> > I checked all my path's tripple, compared both servers for different
> > right's settings, apache2 settings and installed packages. I reinstalled
> > dependencies as libfreetype6 but still 'no go' with truetype.
> > 
> 
> I'd say check the paths inside your fontset again. I found and fixed a 
> very similar issue in v5.2 earlier this week, and it was caused by the 
> font file not being found. I would not be surprised if the same issue 
> existed in 5.0 as well:
> 
>    http://trac.osgeo.org/mapserver/ticket/2973
> 
> Another hint: set CONFIG MS_ERRORFILE "stderr" in your mapfile and look 
> for any error in the Apache error_log. In the case of the ticket above, 
> the "font not found" error was not trapped properly but it would have 
> showed up in the error_log before the crash (unless stderr buffering 
> would have prevented it from getting to the log file before the crash).
> 
> Daniel