Redraw fails with clipped object patterns

[Ben-hin]

I draw some objects, clip it and transform it to a pattern.

Next I draw a rectangle and fill it with this pattern.

When I move a window (like Fill and Stroke) over the
rectangle, drop it there and move it away again, the
pattern disappears from under the window.

I suspect it doesn't get redrawn, because zooming in or
out makes it reappear.

Tested on WinXP, Inkscape 0.44

[Prokoudine (prokoudine)]

Originator: NO

Reproducible on current SVN. Also, make a pattern out of a clipped object
makes this object invisible. Raising priority.

[Tom Davidson (tjd-mit)]

Confirmed in SVN 16987 (Jan 7, 2008) on Fedora Core 6, with a lot of these errors to the command line:

File display/nr-arena-item.cpp line 306 (?): Assertion item->state & NR_ARENA_ITEM_STATE_BBOX failed
File display/nr-arena-item.cpp line 306 (?): Assertion item->state & NR_ARENA_ITEM_STATE_BBOX failed

See also 2 bugs newly marked as dupes of this one...

[Mark Duncan (eattheapple)]

This bug still exists in 0.46 on Ubuntu 8.10.

[Guillermo Espertino (Gez) (gespertino-gmail)]

Confirmed in 0.46+devel

[Kjohrf (kjohrf)]

I'm seeing this too in 0.46. Hoping for a fix...

[Kjohrf (kjohrf)]

Note that it affects "Export Bitmap" so that can't be used as a workaround (like it does for blur to PDF). :(

[Pablo Trabajos (pajarico)]

I tried your example with the pattern Stripes 1:1 and can't reproduce it on 0.47pre4. It's the pattern you're using very complex, with many nodes? If it is, please attach the file so I can test it.
Can you try a recent version?

[Kjohrf (kjohrf)]

This is still a major problem with 0.47pre4! I ran into it today after downloading that. Same process as described by the original bug reporter here.

[Pablo Trabajos (pajarico)]

I think I didn't read correctly so don't mind about my comment #7. I just tried pre4 and the problem is there.

[Pablo Trabajos (pajarico)]

However, placing a dialog on top of it doesn't cause any glitch. Zooming in and out doesn't fix it either, the redrawing is still partial. Changing opacity DOES force a full render that disappears after a zoom or displacing the canvas.

[Pablo Trabajos (pajarico)]

Testcase.

[Kjohrf (kjohrf)]

Thanks for confirming. Hope this can get fixed in 0.47!

[Kjohrf (kjohrf)]

It didn't get fixed in 0.47. I just hit it again today. Major problem because there is not way, I think, to use this feature and get proper output. Hoping for a 0.47.1 fix release for some of these major issues.

[quazgar (quazgar)]

I just want to note that similar effects also happen with masked objects, the cause probably being the same. So hopefully they will vanisch together.

[su_v (suv-lp)]

This issue affects imported PDF files (specially those that have been edited in Inkscape - round-tripping PDF -> SVG > PDF), and export to EPS as well.

Related:
Bug #643093 “In imported PDF, some objects not rendered / partially rendered”:
<https://bugs.launchpad.net/inkscape/+bug/643093>

[David Millar (davmillar)]

Still experiencing this in 0.48.1 on Windows Vista.

[jonathan kuhl (kuhl-jonathan)]

I have been experiencing this error with 0.48.1 on Windows Vista as well.

Not only was my pattern behaving strange, I was also having all my clipped objects disappear when ever I experimented with a filter (occurs when i undo the effect).

[ScislaC (scislac)]

I just tried with trunk which recently changed to the cairo renderer and can not reproduce.

[ScislaC (scislac)]

I tried against 0.48.x and could reproduce... it is indeed fixed in trunk!

[su_v (suv-lp)]

Reopening, due to regression in recent trunk (10450 ok, 10454 not ok) - please revert the bug status to 'Fix Committed' if you don't agree and think this should be tracked in a new report.

Patterns with clipped objects seem to render as bitmaps and do not scale when zooming in (the pattern fill looks pixellized when zoomed in or even at zoom 1:1 with a PDF file from bug #643093).

See attached comparison of screenshots at 550% with r10450 and r10454 (on Mac OS X 10.5.8 (i386)).

[Tavmjong Bah (tavmjong-free)]

Can't reproduce. I think ~suv is talking about bug 955141 (which I just check in a fix for).

[su_v (suv-lp)]

Tavmjong Bah wrote:
> Can't reproduce. I think ~suv is talking about bug 955141 (…).

Yes, in comment #21 I was referring to the more recent regression (IIRC at that time bug #955141 hadn't been filed yet).

Closing with bug status 'Fix committed' (comment #19-20).