scratchfoundation/paper.js
1
0
Fork 0
mirror of https://github.com/scratchfoundation/paper.js.git synced 2025-01-10 06:41:59 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions
5843 commits 36 branches 45 tags 42 MiB
Commit graph

458 commits

Author SHA1 Message Date
Jürg Lehni
86b1b74869 Prevent detection of tiny overlaps and streamline addOverlap() code. 2015-10-03 12:55:32 -05:00
Jürg Lehni
2bed611648 Improve Curve#getPart() to directly handle reversed curves and write docs for it. 2015-10-03 12:55:05 -05:00
Jürg Lehni
6fb4b7e3c4 Change the way overlaps are detected in lines. 
We don't really care weather they are actually fully collinear, we only really care about the distances from the beginnings- and end-points of one line from the other, since that proximity will affect results elsewhere.
 2015-10-03 11:46:50 -05:00
Jürg Lehni
2167e458ae Since we're using sorting now, we can add all start- / end-point intersections before finding the ones within the curves. 2015-10-03 10:38:45 -05:00
Jürg Lehni
c0bb6890bd Switch back to CURVETIME_EPSILON now that overlap edge-case appears to be handled. 2015-10-01 20:47:56 -05:00
Jürg Lehni
9b883e5fb6 Introduce new TRIGONOMETRIC_EPSILON with higher precision than GEOMETRIC_EPSILON. 2015-10-01 20:44:27 -05:00
Jürg Lehni
f5012a78e9 Reformat nested ternary operators again. 
This got messed up in a recent refactoring.
 2015-10-01 20:24:47 -05:00
Jürg Lehni
72f9705615 Fix overeager refactoring in c77165be3a 2015-10-01 08:49:26 -05:00
Jürg Lehni
fee3a90329 Fixed leaked globals. 2015-10-01 07:05:00 -05:00
Jürg Lehni
b8c6eb46ad Fix weirdness of Curve#divide() modifying the wrong Curve object. 2015-10-01 06:52:08 -05:00
Jürg Lehni
c77165be3a Fix issue in Curve#divide() that lead to intersection segments being linked up wrongly. 
Relates to #784
 2015-10-01 06:21:17 -05:00
Jürg Lehni
53dd726057 Rename ignoreStraight argument to _setHandles 2015-10-01 05:55:22 -05:00
Jürg Lehni
75a004187e Improve Line.intersect() to more reliably find interesctions at the beginnings / ends. 
Relates to #784
 2015-10-01 04:38:48 -05:00
Jürg Lehni
53ff973f06 Perform the faster check first. 2015-10-01 03:38:35 -05:00
Jürg Lehni
5f706a4a5d Use lower tolerance in bezier clipping code. 
This really should be Numerical.CURVETIME_EPSILON, but I get better results using Numerical.GEOMETRIC_EPSILON. Perhaps Numerical.CURVETIME_EPSILON / 2 is the right value to use though.
 2015-09-30 14:19:40 -05:00
Jürg Lehni
9bcf369e6a Ony split potentially self-intersecting curves if there are actual canditates. 2015-09-30 13:44:51 -05:00
Jürg Lehni
45040abc53 More clean-ups for #773 
- Use local variables instead of array lookups for values used repeatetly, and merge with pre-existing variables.
- Add some more comments and reference to long explaining post in issue.
 2015-09-30 13:40:01 -05:00
Jürg Lehni
1231153553 Start cleaning up code from #773 
- Use Line object isntead of static methods
- Do not rely on Math.sign() as it's not supported on all browsers
- Wrap lines at 80 char width.
 2015-09-30 13:07:55 -05:00
Jürg Lehni
07874034e9 Merge pull request #793 from iconexperience/Fix-for-#773 
Fix for #773
 2015-09-30 11:51:37 -05:00
Jan
ec70fa1806 Fix for #773 
Indroduced more reliable method for finding self intersection on curves.
 2015-09-30 12:19:09 +02:00
Jürg Lehni
317b809fee Only calculate non-parametric bezier curve if values are actually used. 2015-09-26 11:46:54 -05:00
Jürg Lehni
ea3cc63e2e Reformat code a bit. 2015-09-26 07:41:03 -04:00
Jan
3fa810a557 Bugfix for #791 and performance improvement 
Prevent infinite loop for degenerate curve. Additionally only calculate sy if necessary and save one call to `Curve.solveCubic`
 2015-09-26 07:36:12 -04:00
Jürg Lehni
fd927cbe22 Properly solve issues with self-intersecting special case. 
(e.g. shapes resembling the infinity sign)
 2015-09-24 12:49:39 -04:00
Jürg Lehni
cc7e60e51a Revert db1ecdddd5 and fix issue properly this time. 
Hopefully?
 2015-09-23 12:44:00 -04:00
Jürg Lehni
20f950ac65 Implement #isFirst() / #isLast() tests on Segment and Curve. 2015-09-23 12:26:44 -04:00
Jürg Lehni
db1ecdddd5 Fix filtering of locations at ends of curves in addCurveLineIntersections() 
Only occured when the line / curve had to be flipped.
 2015-09-21 16:56:08 -04:00
Jürg Lehni
84bcc537e1 Simplify addCurveLineIntersections() and exclude end points. 2015-09-21 12:13:53 -04:00
Jürg Lehni
e36319b71a Give PathItem#getIntersections() a way to filter found intersections right away. 
And use it in #getCrossings()
 2015-09-21 10:41:59 -04:00
Jürg Lehni
b5af47a7b1 Implement a better approach to calculate Path#clockwise... 
...merging code with Path#area.

Closes #788
 2015-09-18 21:41:54 +02:00
Jürg Lehni
10eafccd1a Implement 0.75 * handle scaling in curve bounds checks. 2015-09-18 21:09:57 +02:00
Jürg Lehni
17dc5eb51a Allow gettings of unstyled bounds on curves without paths. 2015-09-18 20:15:18 +02:00
Jürg Lehni
ae93652b56 Clean up getIntersection() methods. 
Now that they filter the results on the fly.
 2015-09-18 17:31:23 +02:00
Jürg Lehni
9c812335e5 Curve._filterIntersections() is now without a job. 2015-09-17 09:39:22 +02:00
Jürg Lehni
2750c34090 Improve the way intersections are sorted and merged. 
Use a binary search to determine insertion index and compare with neighbours to eliminate doubles.
 2015-09-17 01:03:13 +02:00
Jürg Lehni
30f1441c26 Various boolean code clean-ups. 2015-09-16 18:34:35 +02:00
Jürg Lehni
d0332f843f Renamed Curve#reverse() and Segment#reverse() to #reversed() 
Since they don't modify the object. Also introduce new Segment#reverse(), which does.
 2015-09-16 18:16:48 +02:00
Jürg Lehni
e548391276 Use the zero-epsilon when checking beginnings and ends of curves for overlaps. 
Relates to #786 and #777
 2015-09-16 02:33:56 +02:00
Jürg Lehni
7aef20ae6b Compare intersection points instead of curve time when deciding to merge. 
Use same precision indepenent of curve length.
 2015-09-14 15:18:44 +02:00
Jürg Lehni
e2d2c836e5 Some boolean code clean-up. 2015-09-13 14:19:56 +02:00
Jürg Lehni
ea2ff5ec28 Increase precision in Curve.getParameterOf() 
Usually only requires 0-1 more iteration.
 2015-09-13 13:45:20 +02:00
Jürg Lehni
d84a84c67f Change the way winding contributions are propagated 
The new approach preserves segment sequence.

Relates to #777
 2015-09-13 13:06:01 +02:00
Jürg Lehni
19c9a0e722 Use the correct points on curve2 when checking intersections at beginnings and ends. 2015-09-13 11:52:17 +02:00
Jürg Lehni
d62caf6faa Introduce CURVETIME_EPSILON, to be used when handling curve time parameters. 
Relates to #777
 2015-09-12 22:55:58 +02:00
Jürg Lehni
4f04dae20f Use the correct GEOMETRIC_EPSILON when matching beginnings and ends of curve in Curve.getParameterOf() 
Relates to #777
 2015-09-12 22:26:16 +02:00
Jürg Lehni
a0730756c1 Use correct term for curve parameter renormalization. 2015-09-12 22:13:18 +02:00
Jürg Lehni
085cdd74a2 Use GEOMETRIC_EPSILON when comparing curve start / end points. 
Relates to #777
 2015-09-12 21:56:53 +02:00
Jürg Lehni
3da921a0b0 Improve CompoundPath#reduce() to properly reduce suppaths. 
Relates to #779
 2015-09-12 11:43:41 +02:00
Jürg Lehni
fb5f8c0115 Introduce GEOMETRIC_EPSILON, for isOrthogonal(), isCollinear() and overlap checks. 
Relates to #777
 2015-09-12 10:35:47 +02:00
Jürg Lehni
35f3ac87bf Change checks for startConnected / endConnected to support compound-paths. 
Closes #778
 2015-09-12 10:12:17 +02:00