Okay, compiled 0.48 and I'm seeing the VoidSymbol issue previously mentioned. Jon, you backported 9679 to trunk, will you please look into backporting the other changes (if safe) or otherwise backing out the part which causes the problem? It would be nice to move forward with getting 0.48.1 out the door and this is definitely blocking it.

Cheers, Josh

On Tue, Oct 26, 2010 at 12:47 AM, ~suv <suv-sf@...58...> wrote:

On 26/10/10 06:08, Michael Wybrow wrote:

...

On 25/10/2010, at 5:47 PM, Michael Wybrow wrote:

...

On 21/10/2010, at 8:12 AM, Alexandre Prokoudine wrote:

...

Are there any issues of satble version somebody is working on or planning to work on in the coming days?

I'm still looking at a problem with connectors seemingly caused by recent changes to the ConvexHull code (possibly by differences in 2Geom's implementation?).  Should be fixed shortly.

Okay, the fix for this connector bug is now committed to the trunk. See https://bugs.launchpad.net/bugs/666586

Any chance you could take a look at another regression? Bug #591586 “[connectors] 'z' segments closing a path are ignored” https://bugs.launchpad.net/inkscape/+bug/591586

Due to changes how the closing segment (z) is stored in the path data in 0.48, connectors ignore the closing segment of paths and shapes if it is a straight line (still present in r9854 and the 0.48.x branch).

The only workaround I am aware of is to (convert shapes to path and then) convert the closing straight line segment into a curve.

~suv

---

Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in  U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel

On 26/10/2010, at 6:47 PM, ~suv wrote:

Any chance you could take a look at another regression? Bug #591586 “[connectors] 'z' segments closing a path are ignored” https://bugs.launchpad.net/inkscape/+bug/591586

Due to changes how the closing segment (z) is stored in the path data in 0.48, connectors ignore the closing segment of paths and shapes if it is a straight line (still present in r9854 and the 0.48.x branch).

This appears to be a problem with 2geom's path crossing-detection code. Basically, the crossing detection code seems to be ignoring the final line segment of closed paths. As far as I could understand the 2geom code, this is possibly due to the bounds() function in 2geom/crossing.h using Path::size() which will be equivalent to Path::size_open() rather than Path::size_default() (which calls Path::size_closed() for closed paths).

Could familiar with 2geom's crossing detection code, or Krzysztof (who authored the closing segment changes), please have a look into this problem and find the correct way to fix it? The bug report above shows the symptoms.

All the connector code is doing is finding the crossings between the pathvector for the connector and the attached shape: SPCurve* shape_curve = sp_shape_get_curve(SP_SHAPE(item)); // apply transformations (up to common ancestor) shape_curve->transform(item_transform); const Geom::PathVector& shape_pv = shape_curve->get_pathvector(); Geom::CrossingSet cross = crossings(conn_pv, shape_pv);

In this case shape_pv generated from a rectangle object contains a single closed path with a segment count of 4, but crossings with the final (closing) segment of the path are not detected by 2geom.

Cheers, Michael

On 27/10/2010, at 9:28 PM, <J.B.C.Engelen@...1578...> <J.B.C.Engelen@...2237...8...> wrote:

...

This appears to be a problem with 2geom's path crossing-detection code. Basically, the crossing detection code seems to be ignoring the final line segment of closed paths. [...] Could familiar with 2geom's crossing detection code, or Krzysztof (who authored the closing segment changes), please have a look into this problem and find the correct way to fix it? The bug report above shows the symptoms.

Krzystof did not author the closing segment changes, and is not to blame. It happened when we switched our geometry calculations to lib2geom, so blame me. I'll look into the issue hopefully this evening. It could indeed be as simple as the fix you suggested; thanks for finding that.

No worries, Johan. Not looking to assign any blame, just hoping to bring it to the attention of the relevant people. Thanks for looking into the fix.

Cheers, Michael