[Xastir-dev] Blame Canada for xastir segfaults

[Curt, WE7U]

On Tue, 23 Mar 2004, Owen DeLong wrote:

> Well... I'd say that if we're segfaulting, it's a clear indication that
> something is wrong.  I'm not saying the application shouldn't bail out.  I'm
> saying that there are actions we might be able to take before we bail out
> that could provide the user something more informative and useful than
> the words "Segmentation Fault" and a prompt (or worse if they ran it from
> an X11 launcher).
>
> I think popping up a window indicating that a fatal error occurred, some
> information about where it occurred (we should be able to tell if it
> occurred
> within the shapelib or within Xastir, for example), and a pointer to where
> they can get more useful debugging information would be a good starting
> point.
>
> Providing a convenient way to submit relevant corefiles and traceback
> information to the developer web site would be icing on the cake.

Sure, a handler that pops up something saying "Shapelib caused a
segfault" would be good.  More info would be better.  It has always
bugged me that Shapelib just bails like that.  There are other cases
where libraries can fail, such as if your libtiff/libgeotiff
versions don't match and you open a geotiff file.

Doing this in a manner that doesn't involve a lot of code tweaks may
be a trick.  Would there be a more refined way of doing this than
encasing all calls to external functions inside some macros?

I think this would be a good candidate for the feature request list.
Please add it.  Mention shapelib and geotiff as libs that like to
bail out on us.  I suspect libproj and gdal/ogr may be similar.

--
Curt, WE7U			    archer at eskimo dot com
Arlington, WA, USA		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!"