[GRASSLIST:1147] Re: More on questions r.in.tiff

[Eric G . Miller]

On Fri, Nov 17, 2000 at 10:11:16AM +0200, juhana.nieminen at helsinki.fi wrote:
> Dear all,
> 
> After vigorous testing on our machines (10 in all) we have come to the
> conclusion that: r.in.tiff is largely sensitive to the amount of
> graphics memory being used at the time. It is not only a problem in
> RH7.0 with xfree 4 but as well in RH6.2 with xfree3.6.  I ran a test
> where parallel Xwindow sessions were used as well as single xwindows
> sessions with  highly graphics memory active programs running - in all
> cases r.in.tiff gave the "core dumped".  In all cases the graphics
> cards had either 16 or 32Mb of memory.  The only configuration where
> r.in.tiff works without a hitch ,is with a  32mb Matrox G400 dual head
> .  It works nicely on all machines when no graphics or GUI  what so
> ever are displayed.
 
Well, like I mentioned, the r.in.tiff in GRASS beta8 has a nasty bug.
It would segfault with truecolor tiffs (which would leave a core file if
your machine is so configured).  I still can't see how it would matter
whether you are using X or not (unless that matters to libtiff).  Sure
it's not an overconsumption of memory?

Anyway, if you want to get the CVS versions of r.in/out.tiff and if you
still have problems, then I'll do more investigation.  Maybe we'll be
making another release before too long.

-- 
Eric G. Miller <egm2 at jps.net>