- Inkscape Devel - Inkscape Mailman

Re: [Inkscape-devel] arrowheads
by Bryce Harrington 09 Mar '04

09 Mar '04

On Tue, 9 Mar 2004 jasonmod@...262... wrote: > > The way that the SVG spec wants us to do it is using the 'marker' > > approach. I've found it's kind of a pain to implement but in theory if > > we do it right, the files will work in any compliant SVG editor or > > viewer. Would you be interested in helping me finish implementing it? > > I can give you the run-down on what's left to do to get it working, as > > best as I know, if you'd like? > > Sure. Okay, cool. The way that SVG does markers is via the 'marker-start', 'marker-end', etc. style tags. Unlike most other style tags, like line caps, the markers use uri's that point elsewhere in the document for the definition of what to be drawn. I.e.: "marker-start:url(#mArrow);" The place that the mArrow item is defined is in the <defs></defs> section of the SVG document. If you look on the screenshots page down at the bottom, there's an example I made of some markers, and you can download the svg document by clicking on the screenshot to see how the SVG is laid out. The code in the codebase currently has some buttons that will put the marker-start style tag on the lines, which is fine if mArrow is already present in the <defs> section (i.e., if you manually write it into the doc). So the next step is to figure out how to make it automatically insert it when the user clicks that button and adds an arrowhead to a line in the drawing. Ideally, we should also remove the def when there are no longer any lines using it. The gradients work in a very analogous way, so it's possible that code can be stolen from there. I haven't looked at the gradient code at all so am not sure what's there. Would you like to take a look at it and see if you can understand how it handles adding/removing its defs items, and if the code could be reused in some fashion? If we can get this defs insert capability done, I *think* that'll be enough to get the feature officially included. Beyond that, then things get fun. ;-) Ultimately I think we want some way to let users define their own markers. Bryce

1 0

Re: inset/outset
by bulia byak 09 Mar '04

09 Mar '04

>can't reproduce the bug >i just spotted a mistake in MakeOffset, btw: at line >557, the 'maxPt' should be a 'maxAr' (damn >cut'n'paste) >it may be the cause of your crashes Made that change, but it still crashes. Here's a 100% reproducible case. Load the attached file, select the star only, press ctrl+) twice, ctrl+( twice, ctrl+). At this moment it crashes. _________________________________________________________________ Tired of spam? Get advanced junk mail protection with MSN Premium http://join.msn.com/?pgmarket=en-ca&page=byoa/prem&xAPID=1994&DI=1034&SU=ht…

1 0

inset/outset
by bulia byak 09 Mar '04

09 Mar '04

I reworked the sp_selected_path_do_offset similarly, so that it also can work with multiple selections and preserves z-order and parent. When testing it, I pressed ctrl+) and ctrl+( many times on a star with 32 sharp long tips, and sometimes (not sure when, seemingly at random) it crashed for that shape. Here's the backtrace of one such crash: Program received signal SIGSEGV, Segmentation fault. 0x408e96d0 in mallopt () from /lib/i686/libc.so.6 (gdb) bt #0 0x408e96d0 in mallopt () from /lib/i686/libc.so.6 #1 0x408e7fd2 in realloc () from /lib/i686/libc.so.6 #2 0x081a2e3b in Shape::AddEdge(int, int) (this=0x89f7c50, st=106, en=107) at Shape.cpp:1241 #3 0x081972d6 in Path::DoLeftJoin(Shape*, double, join_typ, NR::Point, NR::Point, NR::Point, double, double, double, int&, int&, int, int, double) (dest=0x89f7c50, width=2, join=join_pointy, pos={_pt = {369.90625, 535.15625}}, prev=Cannot access memory at address 0x6a ) at PathStroke.cpp:591 #4 0x081a8292 in Shape::MakeOffset(Shape*, double, join_typ, double) (this=0x89f7c50, a=0x89b5758, dec=Unhandled dwarf expression opcode ) at nr-point.h:34 #5 0x080d28cc in sp_selected_path_do_offset(bool) (expand=false) at splivarot.cpp:1125 #6 0x080d1c5d in sp_selected_path_inset() () at splivarot.cpp:713 #7 0x081520c7 in sp_action_perform(_SPAction*, void*) (action=0x85cb678, data=0x0) at action.cpp:128 #8 0x080bdc7d in sp_shortcut_invoke(unsigned, _SPView*) (shortcut=50331688, view=0x829d270) at shortcuts.cpp:33 #9 0x08145bf6 in sp_window_key_press (widget=0x869d5d8, event=0x86d2a04) at window.cpp:31 #10 0x4010192c in _gtk_marshal_BOOLEAN__BOXED () from /usr/lib/libgtk-x11-2.0.so.0 #11 0x0869d5d8 in ?? () #12 0x086d2a04 in ?? () _________________________________________________________________ MSN Premium with Virus Guard and Firewall* from McAfee� Security : 2 months FREE* http://join.msn.com/?pgmarket=en-ca&page=byoa/prem&xAPID=1994&DI=1034&SU=ht…

1 0

simplify command
by bulia byak 09 Mar '04

09 Mar '04

Fred, I noticed you changed something in the simplify algorithm, and I just reworked the command itself. Now it: - works with multiple selected objects - preserves the z-order position of the objects - preserves the parent (i.e. you can simplify within group) I also changed the hard-coded threshold value of 1.0 that you had. Now that value is: - taken from preferences (options.simplifythreshold) - multiplied by the size of the simplified objects, so you get the same result if you scale the object - accelerated when you call the command is quick sequence. The last one is especially useful IMHO. You press Ctrl+L; some simplification happens, but not enough for you. Without acceleration, you're stuck; you may go to preferences and edit the threshold there but it's a major hassle. Now, you can simply keep pressing Ctrl+L quickly enough (faster than twice a second), and on each press, threshold is increased by 0.5 of the original value. If you call simplify again after some pause, the value is back to its prefs default. As for the algorithm itself, with acceleration it works well enough in my testing, I can always drive a path to a one- or two-node degenerate path and get a good range of intermediate simplification degrees along the way. However, when I end up with a one-node degenerate path which is closed, that path tends to grow quickly and rotate if I try to simplify it more, which is kinda weird. Unclosed paths are OK, they get to a two-node bezier and stop there. Can you do something about this growth&rotation? _________________________________________________________________ MSN Premium with Virus Guard and Firewall* from McAfee� Security : 2 months FREE* http://join.msn.com/?pgmarket=en-ca&page=byoa/prem&xAPID=1994&DI=1034&SU=ht…

1 0

inkscape-devel news gateway
by Bryce Harrington 09 Mar '04

09 Mar '04

For any interested, I notice that gmane has picked up inkscape-devel and can be accessed via news: nntp://news.gmane.org/gmane.comp.graphics.inkscape.devel Bryce

1 0

Re: [Inkscape-devel] Re: freehand strokes bug
by bulia byak 08 Mar '04

08 Mar '04

For now, I committed the round -> rint change so that windows folks can work with latest builds. I tested it and it works. When you decide on what's the best function to be used instead of round, feel free to change. Fred, big thanks for debugging this! > > ok, found the bastard. a math function difference, as > > i suspected. so the correct answer is: > > "the round() at line 377 of Shape.h behaves like a > > ceil() on windows" (don't ask me why) > > it is of prime importance for the rounding done in the > > intersector that round() behaves like a round, ie that > > [0...0.5] maps to 0 and [0.5...1] maps to 1. to fix > > it, the quick way is to use rint() instead (which i > > believe is what i was using), but that may induce > > other incompatibilities with other systems. at any > > rate, i'm not qualified to handle platform-specific > > stuff, so i leave it to you to choose the appropriate > > course of action > >Depending on what you need, there are different ways of achieving this >portably -- also there are some finer points like how negative numbers >should be handled, and whether 0.5 gets rounded up or down... > >I'm CCing Nathan on this since he has better mathematical background to >answer your question than I do. > >-mental > > ><< signature.asc >> _________________________________________________________________ Get business advice and resources to improve your work life, from bCentral. http://special.msn.com/bcentral/loudclear.armx

1 0

CVS for test cases
by Peter Linnell 08 Mar '04

08 Mar '04

Hi all, >From our IRC chat, we agreed to have a cvs tree for test cases. Well, I have the first one to submit, can someone set that up? This is probably related to ongoing code refactoring, but this svg imports to/from Inkscape 0.37 <> Scribus 1.1.5+ perfectly. I've renamed the file ugly.svg to give you an idea of what happens. :) Peter

2 1

aux toolbars + true poly mode
by John Cliff 08 Mar '04

08 Mar '04

hi folks, I implemented a true poly mode in the star tool, as opposed to the flattened star that was there before (so its still editable once you've created it.) As a part of that I implemented the aux toolbar for the star mode, and once I'd worked it out for one context I figured I might as well do the others. So rect, star and spiral now all have toolbars that expose the functionality from the tool options dialog. They always modify the tool options, and also affect any of the type you currently have selected (regardless of what else is in the selection) this means if you create a star and change the sides spinner the # of sides on your new star will update. Anyway bulias commited it, feel free to poke holes in it, I'm new to C++ and this codebase so there may well be more elegant ways of doing some of it. Thanks to bulia, mental and ishmal for answering all my dub questions on jabber. cheers John __________________________________ Do you Yahoo!? Yahoo! Search - Find what you�re looking for faster http://search.yahoo.com

1 0

arrowheads
by unknown＠example.com 08 Mar '04

08 Mar '04

Hi, I downloaded Inkscape and noticed that it didn't have any way to put arrowheads on strokes; so I added that feature. I have had no contact with the Inkscape group previously, how do I offer the feature for use in the code? I've added to the stroke-style dialog two additional cap types (solid triangle arrow and line-drawn arrow like "->"), and also a head-size numeric selector kind of similar to the existing stroke width selector and a radio button for selecting whether the arrowhead should be on both, left, right, or neither end of the stroke (non-arrowhead ends get butt_straight rather than the specified cap style.) I believe the interface look-and-feel and also the code style of the stuff I've added are basically indistinguishable from what's already there. Some caveats: I haven't delved into libart, so if you set Inkscape to use the libart stroke calls instead of the livarot calls, then it will basically ignore my additions. Also, there seemed to be some disabled WITH_MARKER_GUI code that maybe was intended to do something similar, but didn't seem to do anything when enabled. My additions ignore the existing disabled marker stuff. - jason Jason Modisette http://snork.home.texas.net/jason

2 1

Re: freehand strokes bug
by bulia byak 07 Mar '04

07 Mar '04

> > Do you know any code saying "if something, do not > > draw the stroke"? Since > >not that i know of... but it's hard to fully >understand what inkscape does, there are so many >pieces stitched together > > > the stroke disappears completely, this is likely to > > be such a "draw or not > > draw" switch. I know of one, which you introduced > >the behavior is shape-dependent; you can have a single >cubic bezier patch rendering with one value of the 2nd >control point, and if you move that point slightly, it >won't render anymore. So you're observing that behavior? Great, that gives me hope :) >that's much like the intersector >bugs i was having, so i suppose it's the >ConvertToShape() eating the stroke, either because the >input is incorrect, or because it failed to uncross >the polygon. there are maybe other places where this >can happen, but i don't think they come into play >here. sadly my gdb refuses to see debug symbols in >inkscape.exe... (tho it was compiled with the >appropriate cflags) does anyone know how to tame gdb >on mingw? I don't know :( Anyone? _________________________________________________________________ http://join.msn.com/?pgmarket=en-ca&page=byoa/prem&xAPID=1994&DI=1034&SU=ht…

2 1