A star/polygon with 'Randomized' parameter > 0 changes shape when moved (translated) on canvas.
Steps to reproduce:
1) launch current trunk (default prefs, default new document, locale: en_US)
2) draw a star
3) change parameter 'Randomized:' to e.g. 0.2
4) switch back to select tool
5) move the star (by dragging with the mouse, or with arrow keys)
Expected result:
The star shape is moved to new position, but stays unchanged otherwise.
Actual result:
The star shape is moved to the new position, and the shape gets regenerated (i.e. changes completely).
Based on tests with archived builds on OS X 10.7.5:
- not reproduced with stable 0.48.4, 0.48.5
- not reproduced with 0.48+devel rev <= 12552
- reproduced with 0.48+devel rev >= 12555
this perceived regression was likely introduced with the merge of the GSoC unit-improvement branch in r12554, which changed how transforms for parametrized shapes are optimized (if possibly, transforms get optimized and reflected in the shape parameters instead of stored in a preserved 'transform' attribute):
<http://bazaar.launchpad.net/~inkscape.dev/inkscape/trunk/changes/12555>
Notes:
- non-uniform scaling, rotating or skewing the randomized star prevents the regression from happening (adds a 'matrix()' transformation)
- scaling, or flipping the object does not prevent the regression
Reproduced with Inkscape 0.48+devel r13466 on OS X 10.7.5.
A star/polygon with 'Randomized' parameter > 0 changes shape when moved (translated) on canvas.
Steps to reproduce:
1) launch current trunk (default prefs, default new document, locale: en_US)
2) draw a star
3) change parameter 'Randomized:' to e.g. 0.2
4) switch back to select tool
5) move the star (by dragging with the mouse, or with arrow keys)
Expected result:
The star shape is moved to new position, but stays unchanged otherwise.
Actual result:
The star shape is moved to the new position, and the shape gets regenerated (i.e. changes completely).
Based on tests with archived builds on OS X 10.7.5: bazaar. launchpad. net/~inkscape. dev/inkscape/ trunk/changes/ 12555>
- not reproduced with stable 0.48.4, 0.48.5
- not reproduced with 0.48+devel rev <= 12552
- reproduced with 0.48+devel rev >= 12555
this perceived regression was likely introduced with the merge of the GSoC unit-improvement branch in r12554, which changed how transforms for parametrized shapes are optimized (if possibly, transforms get optimized and reflected in the shape parameters instead of stored in a preserved 'transform' attribute):
<http://
Notes:
- non-uniform scaling, rotating or skewing the randomized star prevents the regression from happening (adds a 'matrix()' transformation)
- scaling, or flipping the object does not prevent the regression
Reproduced with Inkscape 0.48+devel r13466 on OS X 10.7.5.