Blurring strictly vertical/horizontal path yields unexpected result

Bug #1422510 reported by Hachmann
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Inkscape
New
Undecided
Unassigned

Bug Description

See attached svg.

1) Make vertical or horizontal path from two nodes using the 'only vertical or horizontal lines drawing' function of Bezier tool or by aligning them (then also more than two nodes give the same result) using the node alignment.
2) Give the path a stroke and a stroke width if it doesn't have them.
3) Move the blur slider.

Result: No path to see any more. If exported, yields a 1px x path length (plus blur radius?) transparent picture.

0.91 from official Ubuntu repos on Linux Mint 17.1

Edit: also looks wrong in browser, but not in my systems file preview gthumb.

Revision history for this message
Hachmann (marenhachmann) wrote :
description: updated
Revision history for this message
su_v (suv-lp) wrote :

This is the expected result (i.e. in accordance with the SVG 1.1 specification) - view the file in a modern web browser or in Squiggle (Batik 1.7)

Proposing to link as duplicate to
- Bug #168943 “Blurring a zero-height or zero-width path”
  https://bugs.launchpad.net/inkscape/+bug/168943

(The filter effects region is based on the geometric dimensions of an element - a horizontal or vertical line has only one dimension and no area (the stroke is a style property, not part of the geometry))

tags: added: filters-svg svg
Revision history for this message
su_v (suv-lp) wrote :

On 2015-02-16 23:09 (+0100), Hachmann wrote:
> Edit: also looks wrong in browser, but not in my systems file
> preview gthumb.

The browser is likely more compliant with the current SVG specification than the image viewer. AFAICT gthumb uses librsvg for SVG images - librsvg was never considered a reference SVG implementation for Inkscape.

Revision history for this message
Hachmann (marenhachmann) wrote :

Thank you suv, didn't know that. I thought I remembered it looked differently in 0.48, but I must have been wrong. I might want to use the answers section instead next time...

Will this be different in the 2.0 spec?
I never noticed that a single stroke path's blur depends on its direction, or the bounding box. (I'd make it depend on it's stroke width, if anything... )
If it doesn't account for the stroke, it shouldn't account for it in any direction... or the other way round... but that's a spec thing then, and not something you could change ;)

Revision history for this message
su_v (suv-lp) wrote :

See also: related discussion (wrt SVG spec) in the comments of bug #1229971

Linking as duplicate to bug #168943 (see comment #2 there).

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Bug attachments

Remote bug watches

Bug watches keep track of this bug in other bug trackers.