v.buffer: strange result for lines

[martinl]

Sample data attached: attachment:silnice.txt​

Buffer 1000m - OK (attachment:v_buffer_1000.png​)

v.buffer input=silnice output=silnice1000 distance=1000

Buffer 100m (using GEOS) - surprisingly wrong (attachment:v_buffer_100_geos.png​)

v.buffer input=silnice output=silnice100 distance=100

Buffer 100m (using GRASS) - wrong - not surprisingly (attachment:v_buffer_100_grass.png​)

export GRASS_VECTOR_BUFFER=1
v.buffer input=silnice output=silnice100 distance=100

[martinl]

sample data

[martinl]

Buffer 1000m

[martinl]

Buffer 100m (GRASS)

[martinl]

Buffer 100m (GEOS)

[mmetz]

Replying to martinl:

Sample data attached: attachment:silnice.txt​

Buffer 100m (using GEOS) - surprisingly wrong (attachment:v_buffer_100_geos.png​)

Works for me. Do you have any changes in your local copy?

v.buffer input=silnice output=silnice100 distance=100

Buffer 100m (using GRASS) - wrong - not surprisingly (attachment:v_buffer_100_grass.png​)

export GRASS_VECTOR_BUFFER=1
v.buffer input=silnice output=silnice100 distance=100

Looks better for me, but still with errors.

I'm using GEOS-3.3.6.

Markus M

[martinl]

Replying to mmetz:

Works for me. Do you have any changes in your local copy?

hm, it fails only with original binary map (attachment:silnice.pack​), not with imported ascii map.

[martinl]

original binary map

[mmetz]

Replying to martinl:

Replying to mmetz:

Works for me. Do you have any changes in your local copy?

hm, it fails only with original binary map (attachment:silnice.pack​), not with imported ascii map.

The original binary map is 3D, the imported ascii map is 2D. I guess one of the cleaning tools does not work with 3D output, even though v.buffer creates 2D output. Strange.

Markus M

[martinl]

Replying to mmetz:

The original binary map is 3D, the imported ascii map is 2D. I guess one of the cleaning tools does not work with 3D output, even though v.buffer creates 2D output. Strange.

I overlooked it, good point.

[mmetz]

Should be fixed in r54529.

Markus M

[martinl]

Replying to mmetz:

Should be fixed in r54529.

Confirmed, thanks for super-quick fix.