[GRASS-dev] X monitors display rasters half resolution off to the left

Hi,

In Spearfish:

$ g.region rast=landuse -a
$ d.mon x0
$ d.rast landuse

There is a black, vertical artifact line on the right edge of the monitor.

$ d.zoom -> out

It dissapears...

$ d.zoom -> in, until cells are visible

With some luck, you will see such artifact line on the right again -
different colors.

$ d.what.rast

Query some left-most non-null cells. If you click their left side, you
get null instead of the value you expect.

Query the left half of a null cell on the right of some right-most
non-null cell. Instead of null, you get the value you'd expect from
that non-null cell.

Seems like the raster is displayed half resolution off to the left,
compared to d.what.rast query results. A recently introduced bug.

This is propably the reason of the bug I reported as "r.to.vect
feature=lines: the output vector is moved half the resolution east off
the input raster"

http://wald.intevation.org/tracker/index.php?func=detail&aid=322&group_id=21&atid=204

Maciek

Hi,
i can confirm this bug.

http://www-pool.math.tu-berlin.de/~soeren/grass/modules/screenshots/GRASS_display_bug.png

soeren

Am Montag, 5. März 2007 20:14 schrieb Maciej Sieczka:

Hi,

In Spearfish:

$ g.region rast=landuse -a
$ d.mon x0
$ d.rast landuse

There is a black, vertical artifact line on the right edge of the monitor.

$ d.zoom -> out

It dissapears...

$ d.zoom -> in, until cells are visible

With some luck, you will see such artifact line on the right again -
different colors.

$ d.what.rast

Query some left-most non-null cells. If you click their left side, you
get null instead of the value you expect.

Query the left half of a null cell on the right of some right-most
non-null cell. Instead of null, you get the value you'd expect from
that non-null cell.

Seems like the raster is displayed half resolution off to the left,
compared to d.what.rast query results. A recently introduced bug.

This is propably the reason of the bug I reported as "r.to.vect
feature=lines: the output vector is moved half the resolution east off
the input raster"

http://wald.intevation.org/tracker/index.php?func=detail&aid=322&group_id=2
1&atid=204

Maciek

_______________________________________________
grass-dev mailing list
grass-dev@grass.itc.it
http://grass.itc.it/mailman/listinfo/grass-dev

Maciej Sieczka wrote:

In Spearfish:

$ g.region rast=landuse -a
$ d.mon x0
$ d.rast landuse

There is a black, vertical artifact line on the right edge of the
monitor.

I see it too, in 6.3cvs but not 6.2.1.

drag the xmon to be twice as wide as it is tall to see the grey band
clearly.

Hamish

On Tue, Mar 06, 2007 at 11:15:35AM +1300, Hamish wrote:

Maciej Sieczka wrote:
> In Spearfish:
>
> $ g.region rast=landuse -a
> $ d.mon x0
> $ d.rast landuse
>
> There is a black, vertical artifact line on the right edge of the
> monitor.

I see it too, in 6.3cvs but not 6.2.1.

drag the xmon to be twice as wide as it is tall to see the grey band
clearly.

Could

http://mpa.itc.it/markus/tmp/admin98_oceans_buggy.png
be related to this (latlong)?
I also tried to export the oceans only with v.overlay, the
result looks strange (wrong polygon fill, too).

Markus

Maciej Sieczka wrote:

In Spearfish:

$ g.region rast=landuse -a
$ d.mon x0
$ d.rast landuse

There is a black, vertical artifact line on the right edge of the monitor.

Okay; I'm fairly sure that this was introduced with my changes to the
display/raster libraries, specifically the addition and use of
R_scaled_raster().

What I'm less of is whether I introduced a new off-by-one bug or fixed
one which was being cancelled out by another bug elsewhere.

In any case, I'll look into it now.

--
Glynn Clements <glynn@gclements.plus.com>

Glynn Clements wrote:

Maciej Sieczka wrote:

> In Spearfish:
>
> $ g.region rast=landuse -a
> $ d.mon x0
> $ d.rast landuse
>
> There is a black, vertical artifact line on the right edge of the monitor.

Okay; I'm fairly sure that this was introduced with my changes to the
display/raster libraries, specifically the addition and use of
R_scaled_raster().

What I'm less of is whether I introduced a new off-by-one bug or fixed
one which was being cancelled out by another bug elsewhere.

The former.

In any case, I'll look into it now.

Fixed in CVS, AFAICT.

--
Glynn Clements <glynn@gclements.plus.com>

Glynn Clements wrote:

Fixed in CVS, AFAICT.

Thank you Glynn! Great job, as usual :).

Maciek