[Xastir] It crashed..

[Kirk Mefford]

I'm starting to think my computer was just having a bad day or something 
because I can't seem to duplicate the problem.  Xastir closed on me several 
times yesterday but I only caught the error twice.  First the error I 
listed below then shortly after I sent the email.  It had a very similar 
error and had barely finished loading.

I've loaded exactly what I had last night same maps, same zoom level and 
ran all of my log files for the last 3 days and Xastir happily parsed 
everything and is still running fine.  Hopefully it doesn't happen again.

Strange.....very strange.


At 11:58 PM 5/23/2003, Curt Mills, WE7U wrote:
>On Fri, 23 May 2003, Kirk Mefford wrote:
>
> > I had an image loaded with about 7 tiger shapefiles loaded.  It had only
> > been running for a few minutes.  I had just finished changing the
> > properties on a few maps and hit "ok" to reload the maps and Xastir
> > disappeared with the following error message:
> >
> > Xlib: unexpected async reply (sequence 0x336a5b)!
> > X Error of failed request:  BadLength (poly request too large or internal
> > Xlib l
> > ength error)
> >    Major opcode of failed request:   83 (X_ListInstalledColormaps)
> >    Serial number of failed request:   3369564
> >    Current serial number in output stream:   3369563
> >
> > Any ideas what happened?
>
>You pushed the wrong button???
>
>Well, it _tells_ you what happened, that's not obvious?  hi hi
>
>X Errors are kind of cryptic.  Most of the time _WE_ don't even know
>what they mean.  There are several layers to X, the lowest being Xt,
>the next one up being Xlib, and then the layers above that (I think)
>are Motif, Lesstif, or OpenMotif.  I have big thick books here and
>at work on all of these layers.  More than I would want to try to
>lift at one time.
>
>In this case it failed on an Xlib call, the middle layer.  It's likely
>that it died on an XFillPolygon() or XPolygonRegion call, as those are
>what we use to draw polygons.  Those are used within draw_symbols.c,
>maps.c, and rotated.c.
>
>At the moment that's probably as close as I can pin it down.  What I'm
>wondering now is whether it is a received polygon that did it:
>
>1) A weather alert that caused a NOAA Shapefile polygon to get drawn
>2) A received multipoint packet that caused a polygon to get drawn
>
>Perhaps one of your maps has a problem, or we have a problem drawing
>it.  X has a problem drawing things that overflow 16 bits, so if we
>didn't protect ourselves properly everywhere in the code, it could be
>that we tried to draw to a pixel outside our range.
>
>Does it die every time it tries to draw or index one particular map?
>Can you try loading exactly those maps again, one at a time, and then
>perhaps all at once, to see if you can duplicate the problem?
>
>Did you perhaps have logging turned on when it crashed and can save
>that logfile away?  If so, try reloading the logfile to see if
>something in it caused the crash.
>
>Give me a way to duplicate the crash and I can fix it.
>
>Another possibility is that you have a corrupt map.  I had some of
>those but found them pretty quickly when I implemented map indexing.
>
>Curt, WE7U.                             archer at eskimo.com
>http://www.eskimo.com/~archer
>   Lotto:  A tax on people who are bad at math. - unknown
>Windows:  Microsoft's tax on computer illiterates. - WE7U.
>The world DOES revolve around me:  I picked the coordinate system!"