- Inkscape Devel - Inkscape Mailman

Updated W3C tests for Inkscape
by Tavmjong Bah 08 Feb '08

08 Feb '08

Hi, I've updated the W3C tests for Inkscape on my website. You can see the current status (for the development version) at: http://tavmjong.free.fr/INKSCAPE/W3C_SVG/htmlInkscapeHarness/full-index.html To make it easier to see the results relevant to Inkscape, I've removed most tests with animation, links, or other SVG interactive features. Note, that Inkscape passes a few tests interactively that it fails when used from the command line. These tests all involve the feImage filter with an external image. Tav

1 0

grids, guides, snapping UI
by bulia byak 07 Feb '08

07 Feb '08

Hi guys, overall you did a terrific job on these features, but I have several comments/questions on the UI: Doc prefs: - "snap guides while dragging" option does not seem to work - no snapping to grid, other guides or objects - can we disable the remove button when there's nothing to remove in Grids? - can we give grid names that at least reflect their type, e.g. AxGrid2222 or RectGrid2222 instead of grid2222? - why two checkboxes, "enabled" and "visible"? what's the difference? It's not clear. I propose to rename them to "Snapping enabled" and "Grid visible". Also if I uncheck "enabled", visible does not work - why? if that's intentional, please just gray out the rest of the grid options when "enabled" is unchecked. - why there's no option for hex grid to show dots? - Snap tab: for what snaps/nodes, the tooltip lists what they may snap to, but here, more important is another question - what are nodes? only path nodes, or special points in shapes? what about gradient handles and text baselines - are they "nodes" too? I think this needs to be clarified right here. What they snap to is clear from the rest of options below. - same about what snaps/bboxes: what they snap _to_ is not important here, it's clear from the rest of options anyway, but that it only works in selector is important. Oops, no, wait; I just noticed that bboxes don't snap to paths/nodes, and nodes don't snap to bboxes. In that case this should be mentioned in the tooltips of course (e.g. "snap to bounding boxes of other objects but not to their paths" and vice versa). - "Snap at specified distance" is misleading; better something like "Snaps only when closer than:" - "Miscellaneous" is a bad label - doesn't this option belong in "what snaps"? Maybe as a suboption under "nodes" - am I right in assuming the rotation center is considered a "node" for snapping? - in any case, "Also snap the rotation center of an object when snapping nodes or guides" - why are guides mentioned here? how can I snap rotation center of a guide? - finally, I think we should rearrange the tabs again :) Tooltips that include "see previous tab" are a sign that something is very awkward here. As I understand the options, I would propose dividing them as follows: first tab: "Snapping" - enable snapping - what snaps, including the rotation center option second tab: "Snap targets" - snap to guides - snap to grids - snap to objects - snap to intersections (in this order!) What do you think? -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

4 8

feDisplacementmap fix
by jf barraud 07 Feb '08

07 Feb '08

Hi all! Atm, feDisplacementmap output depends on zoom factor (see here<https://bugs.launchpad.net/inkscape/+bug/182059> ). I had a look at the code to try to fix this, but I need someone to validate the code. Precise meaning of area, unit transforms, and particulraily of the function "enlarge_area" are not perfectly clear to me. The modifs are all in nr-filter-displacement-map.cpp.<http://launchpadlibrarian.net/11804938/nr-filter-displacement-map.cpp> I just changed "scale" with "scalex" and "scaley" defined as: Matrix trans = units.get_matrix_primitiveunits2pb(); double scalex = scale*trans.expansionX(); double scaley = scale*trans.expansionY(); where apparently appropriate (in the render() and enlarge_area() methods), but I would feel more confortable if someone having a better understanding of the code could check if this is correct or if other changes are needed elsewhere... Thanks! jfb.

3 2

Re: [Inkscape-devel] 'feComponentTransfer'
by Felipe Sanches 07 Feb '08

07 Feb '08

I am stuck on this. Can anybody help please? I have updated the patch to current svn (r17273) On Jan 17, 2008 3:55 AM, Felipe Sanches <felipe.sanches@...400...> wrote: > I started coding it. Though it is not working yet. > > How does inkscape knows that my Class handles a specific tag? I mean... > how will i have 4 different tags (<feFuncR>, <feFuncG>, <feFuncB> and > <feFuncA> be handled by my SPFeFuncNode class ?) > > I supose that you are more confortable with this than me, so, if > possible, could you take my patch and complete it? I have spent hours trying > to figure out how to do it and now I wish to focus on implementing the > render method. > > > I would also ask for nicholas to help us to adapt the UI > > Hoppefully we can have this filter implemented before 0.46 release. Then > the only ones that would be out are feImage and feTile > > > On Jan 16, 2008 5:57 PM, Niko Kiirala <niko@...1267...> wrote: > > > Wed, 16 Jan 2008 12:45:15 -0200 > > "Felipe Sanches" <felipe.sanches@...400...> kirjoitti: > > > > > can you help me? I am not sure how do we deal with nested tags . How > > > do I read their attributes? > > > > I belive, the best way to set this up is to create SP-classes for those > > tags, that can be children. Then implement child_added, remove_child and > > order_changed callbacks in the parent class. These callbacks are not > > strictly required, but following modifications is often useful. > > > > With these, in the parent class you can access this_object->children, > > this_object->children->next etc. (this_object here is a pointer to > > object of the parent class type) Then, you may need suitable accessors > > in the cild classes, that give you the information you need. > > > > See sp-fespecularlighting.cpp for an example. The child classes in this > > case are in sp-fespotlight.cpp, sp-fedistantlight.cpp and > > sp-fepointlight.cpp > > > > -- > > Niko Kiirala > > niko@...1267... > > > >

1 0

improving a bit more the feImage filter
by Felipe Sanches 07 Feb '08

07 Feb '08

this is my updated patch for hadling svgelements referenced by id on the feImage filter primitive. It is not working yet but is much better than the previous patch I sent to the list. Now the URIReference is properly handled and a NRArena/NRArenaItem created. Then I create a pibblock and render the svg element in it. When I call the nr_arena_item_invoke_render method I get this error: ** (inkscape:6874): WARNING **: rendering element File display/nr-arena-item.cpp line 308 (?): Assertion item->state & NR_ARENA_ITEM_STATE_BBOX failed I dont know how to get rid of this one. I am going to travel now and will be back only next week. If anybody wants to give it a try, this patch needs a last touch to get working. (it also adds an extra button to the feImage settings UI: to use an element instead of an image, you must select the desired element and then click on the "SVG Element" button) best wishes, Juca

1 1

Expected LPEs for 0.47
by unknown＠example.com 07 Feb '08

07 Feb '08

Hi all, JF sent me this and I couldn't resist posting it here: http://math.univ-lille1.fr/~barraud/Inkscape/effectPreview.svg I especially like the sketch effect! Cheers, Johan

7 8

pattern-along-path LPE recent additions
by jf barraud 07 Feb '08

07 Feb '08

Hi, Hum, Johan is right. I did not consider this as a new feature, but I was wrong. Sorry--- still learning! :-( I think it is definitely worth thinking about it: the point is that pattern-along-path loose a lot of usability if the offset parameters are not there (at least the 'space' added between copies): several people asked me about the old python script, just to have this functionality (that is what made me think of this as a kind of bug...). For instance, in many (most?) cases (foot prints, stars, any shape...), you don't want the copies of the pattern to touch each other. Moreover, the work arround are not easy... Making the dimensions proportional or not to the size of the object is codewise straightfoward, but Johan is right stressing that the right decisions have to be made here. This requires some time to think about it, and might be worth postponing to 0.47. On the other hand, it would be a pitty to release lpe pattern-along-path without it... I am just not experienced at all in such big projects as inkscape, and I'll devoutly follow guru's advices! In particular, I'll perfectly agree if we decide to remove this. JF.

1 0

Fillet Tool Blueprint Discussion
by Joel Holdsworth 06 Feb '08

06 Feb '08

Hi All! I'm in the process of getting together a blueprint for a fillet tool that I'm drafting: http://wiki.inkscape.org/wiki/index.php/SpecFilletChamfer. I'd love to get some discussion and feedback going. I've got a couple of questions that I'd like to discuss at the bottom, and I'd like to hear people's feedback. Thanks Joel

9 22

Filter effects UI
by Niko Kiirala 06 Feb '08

06 Feb '08

Hi There are a couple of issues with filter effects UI, that I'm concerned about in the upcoming release. First, at some point of development, the blur slider was disabled when there was a more complex filter attached to the object. I believe that this is a good thing, because it's not too clear, what it should do. Nevertheless, right now it causes a crash. I haven't dwelved much into this problem and there may be some bigger underlying problem, that needs fixing. But unless there's a plan on how this slider should work with complex filter already attached to the object, I believe it should be disabled in those cases. (see https://bugs.launchpad.net/inkscape/+bug/186281 ) Second is a problem I fully realized only today: there's no user interface for setting enable-background property. Without this, the only working input images are Source Graphic and Source Alpha. Background Image and Background Alpha only contain data, when there's property 'enable-background:new' set on some parent object of the filtered object. Without this, they will just appear to not work. (see https://bugs.launchpad.net/inkscape/+bug/189341 ) (Fill Paint and Stroke Paint don't work because there's no renderer support yet for them in Inkscape) -- Niko Kiirala niko@...1267...

3 3

pattern-along-path LPE recent additions
by unknown＠example.com 06 Feb '08

06 Feb '08

Hi all, Recently, JF added support of some more of path-along-path extension's parameter in the pattern-along-path LPE. (spacing, and offset parameters) I am not too certain that this change is good so close to release. Especially I am concerned with ambiguity or "wrong" calculation of parameter values. Could someone please have a (somewhat thorough!) look at what the new parameter values mean and that the resulting path is consistent with that meaning? What happens for example, when scaling the original path? I really think the meaning of their values should depend on the "scale width in units of length" checkbox. It is easy to revert (the svn revision changed only one file and reverting it does not break compilation), so we can postpone the decision _a bit_. Thanks a lot, Johan

1 0