[GRASS5] [bug #4164] (grass) v.clean: uses 1GB RAM, 1 GB swap for a vector with only 628 boundaries

Radim wrote:

It seems to be strange. Did you get correct result from v.buffer???!!!
It maybe that I just introduced another bug.

Radim,

Sorry for late reply. I've been away for some time.

Regarding the issue: indeed, v.buffer's output is not a clean buffer. I
reproduced my orignal v.buffer command:

v.buffer input=x_extr output=x_extr_buff100 type=boundary layer=1 buffer=100
scale=1.0 tolerance=0.01 debug=clean

and in spite specifying debug=clean the output is full of tiny polygons made
of crossing circles and arcs. What's good, their topology looks allright after
inspection in Qgis Grass digit. Even better, that v.buffer used very little
memory - several times less than before; which is great! It didn't even touch
swap!

Summarising: the memory consumption is sane now, yippie!, but buffer is not
created.

Do you need my dataset for your work on this?

Maciek

-------------------------------------------- Managed by Request Tracker

On 3/21/06, Maciek Sieczka via RT <grass-bugs@intevation.de> wrote:

Radim wrote:
Sorry for late reply. I've been away for some time.

Regarding the issue: indeed, v.buffer's output is not a clean buffer. I
reproduced my orignal v.buffer command:

v.buffer input=x_extr output=x_extr_buff100 type=boundary layer=1 buffer=100
scale=1.0 tolerance=0.01 debug=clean

and in spite specifying debug=clean the output is full of tiny polygons made
of crossing circles and arcs. What's good, their topology looks allright after
inspection in Qgis Grass digit.

Thats correct, debug means debug, if you want final output don't use debug.

Radim

Even better, that v.buffer used very little

memory - several times less than before; which is great! It didn't even touch

swap!

Summarising: the memory consumption is sane now, yippie!, but buffer is not

created.

Do you need my dataset for your work on this?

Maciek

-------------------------------------------- Managed by Request Tracker

On 3/22/06, Maciek Sieczka <werchowyna@epf.pl> wrote:

Now, not using "debug=clean": the v.buffer creates _mostly_ good result
when compared to r.buffer at resolution =1. But there are at least 2
places with errors visible in v.buffer behaviour. Please see the
attached screendump, you will spot them easily.

The YELLOW lines are the v.buffer and r.buffer input.

The RED backgroung is r.buffer output (GOOD).

The BLACK lines are v.buffer output, with at least 2 ERRORS visible,
indicated by GREEN circles.

Are those possible to be fixed Radim?

Not now by me. If possible publish your input data so that it can
be used for debugging in future.

How do look those 2 places if run with debug=clean?

Radim

Maciek

--------------------
W polskim Internecie s? setki milion?w stron. My przekazujemy Tobie tylko najlepsze z nich!
http://katalog.panoramainternetu.pl/