March 2008 - Inkscape Devel - Inkscape Mailman

ill2svg.pl
by Alexandre Prokoudine 27 Mar '08

27 Mar '08

Hi, We have at least one bug on old AI files importer, namely LP145145. In the past I already suggested moving to UC for importing old AI files. Do you think it's the right time to do it now? If yes, do we need a separate verb for this extension so that both PS-based and PDF-based AI importers would work? Alexandre

3 2

Copy-and-paste of defs/use
by Michael Droettboom 26 Mar '08

26 Mar '08

First -- I'd like to take the opportunity to thank you for Inkscape... Truly an important application to have as open source. I am one of the matplotlib developers, and somewhat recently extended its SVG output backend to save the glyph outlines directly in the SVG file. (To remove the need to install obscure math fonts everywhere the SVG file will be viewed.) It uses <defs>/<symbol> to define the character and then <use> to render them. Yes, SVG fonts would probably be better for this, but the lack of support for them in Mozilla and other tools makes them a no-go for now. One of our users recently came across a problem where copying from one SVG file into another mangles the text. See his blog here: http://assorted-experience.blogspot.com/2008/03/inkscape-matplotlibs-svg-on… From poking around the resulting XML, what seems to be happening is that the <symbol> elements in the pasted data are reassigned new ids, but the <use> elements are not updated accordingly, and thus the <use> elements end up referring to symbols in the existing (pasted onto) file. Matplotlib, at present, assigns ids to glyph outlines sequentially as they are encountered, so there is no attempt to assign "universally" unique names, and id clashes across different files are guaranteed. There's a simple workaround I'm considering at our end, which is to use hashes on the path content itself as ids. This would also ensure that glyphs that are common between files will have the same id, for whatever that's worth. It does seem to at least make things behave as expected in Inkscape. However, I wanted to bring this to your attention because a) maybe this is an Inkscape bug, and it should be correcting the defs/use links on a copy/paste operation. b) maybe my understanding/expectations aren't correct and I'm just horribly misusing the SVG spec... I've attached two example SVG files. If you open both, and then copy the contents of one into the other, you can see that the title text will be incorrect. Cheers, Mike -- Michael Droettboom Science Software Branch Operations and Engineering Division Space Telescope Science Institute Operated by AURA for NASA

4 4

Font Selector widget - Font size unit issue
by tayfun_biyikoglu 26 Mar '08

26 Mar '08

Hi, I need some help on font selector widget ? My exact question is how do we set the font name and size values on the toolbar right above the canvas after selecting some values from font selector(Text and Font Dialog). I am java developer and investigating the codes to add unit functionality and i am trying to get familier with c++. I really appreciate brief explanation of codes and also any technical approach for this issue is welcome. I am planning to add a new tab to document properties maybe named Font Defaults. I want to add two things there one is default font name and the other is default font size unit. If for example pt is selected i want to treat the font sizes as pt. I am sure this is a very simple issue for the ones who knows the codes and c++ very well. But as i go deeper i get more confused. Thanks, Sincerely, Tayfun -- View this message in context: http://www.nabble.com/Font-Selector-widget---Font-size-unit-issue-tp1631764… Sent from the Inkscape - Dev mailing list archive at Nabble.com.

1 0

Gsoc 2008 for 2geom: towards making up a proposal
by Marco 26 Mar '08

26 Mar '08

As already stated in previous mails the goal of my application is to implement 2geom or lpe features. After having browsed a lot of 2geom source code I started implementing a proof of concept version of SVGEllipticalArc, the result of my effort is attached at this mail (see svg_elliptical_arc.cpp). In order to get it compiled you have to put the file under the toy folder and add the file name into the list of source file in CMakeList.txt. The file provides a EllipticalArc class and a toy class for test it. (Please feedback!) I have always to write down a complete proposal but here are some ideas (any critique or suggestion is welcome) : (1) give my help in the integration task (2) make elliptical arc a first citizen in 2geom library I'd like also to implement basic shape support ( regular poligon, circle, ellipse, ..) and routines to support classical geometric construction methods (circonference by 3 point, circonference by 2 point and tangent to a given line, ...). Anyway I saw Maximilian Albert's mail and I'm not sure if this last point may overlap with his proposal or as I hope it could provide a basis for his project. Moreover I'm trying to complete my proposal with some other algorithm to be implemented so if you have something in mind I'd like to hear it. Kind Regards, Marco Cecchetti PS: are there mentors avaible for my project ? -- Using Opera's revolutionary e-mail client: http://www.opera.com/mail/

2 1

Re: [Inkscape-devel] Cmake build stuff
by Aaron Spike 26 Mar '08

26 Mar '08

Kai-Uwe Behrmann wrote: > # CMAKE generated file: DO NOT EDIT! > # Generated by "Unix Makefiles" Generator, CMake Version 2.4 > > # Default target executed when no arguments are given to make. > default_target: all > > ... > > all: cmake_check_build_system > $(CMAKE_COMMAND) -E cmake_progress_start /data3/CD/programme/fltk/djv_74b11/CMakeFiles $(CMAKE_ALL_PROGRESS) > $(MAKE) -f CMakeFiles/Makefile2 all > $(CMAKE_COMMAND) -E cmake_progress_start /data3/CD/programme/fltk/djv_74b11/CMakeFiles 0 > > ... > > Do you have an other version or output? The above is just an example. Here's a Makefile I have lying around from June of 2007: # Include the progress variables for this target. include progress.make # CMAKE generated file: DO NOT EDIT! # Generated by "MinGW Makefiles" Generator, CMake Version 2.4 #============================================================================= # Set environment variables for the build. SHELL = cmd.exe # The CMake executable. CMAKE_COMMAND = c:\cmake\bin\cmake.exe # The command to remove a file. RM = c:\cmake\bin\cmake.exe -E remove -f # The program to use to edit the cache. CMAKE_EDIT_COMMAND = c:\cmake\bin\CMakeSetup.exe # The top-level source directory on which CMake was run. CMAKE_SOURCE_DIR = C:\src\lib2geom # The top-level build directory on which CMake was run. CMAKE_BINARY_DIR = C:\src\2geom_build # Default target executed when no arguments are given to make. default_target: all ... # The main all target all: cmake_check_build_system $(CMAKE_COMMAND) -E cmake_progress_start C:\src\2geom_build\CMakeFiles $(CMAKE_ALL_PROGRESS) $(MAKE) -f CMakeFiles/Makefile2 all $(CMAKE_COMMAND) -E cmake_progress_start C:\src\2geom_build\CMakeFiles 0 ... Aaron Spike

1 0

Simple tugging curve editing etc
by Bill Baxter 26 Mar '08

26 Mar '08

The tweak/sculpt tool is nice, but sometimes it seems to introduce a little too much distortion. It's also quite slow, and setting the appropriate width can be a bit annoying. How about implementing something like this? http://www-ui.is.s.u-tokyo.ac.jp/~takeo/research/rigid/curve/index.html The two big ideas in there are as far as I know: 1) automatically setting region of influence based on maximum amount dragged. 2) angle preserving deformation -- for instance if you try Inkscape's sculpt tool on a crown shape like in the lower right corner of the applet there it's pretty impossible not to turn the straight lines into curves. The smoothing thing in that demo isn't bad either, I don't know of any way to "brush on" smoothing in Inkscape right now. Maybe the tweak/sculpt tool can do it somehow though? A couple of other line drawing enhancements that I think would be nice are * a mode where scribbling back and forth over a line erases it. * a mode where drawing over top of a portion of an existing curve would automatically splice in the new part. I've seen these latter two techniques used in some of Igarashi's papers, and they look handy. The first is used in Teddy. I'm not sure where I've seen the second. I think this could maybe even be a nice little set of mini-projects for a GSoC. --bb

4 6

Re: [Inkscape-devel] Cmake build stuff
by Aaron Spike 26 Mar '08

26 Mar '08

Kai-Uwe Behrmann wrote: > And the problem I see with current cmake "Makefiles" is not so much about > setting up a simple or big project, rather than to debug it. > In cmake I have no idea where veriables come from. In Make I do, at least > mostly. Learning debuging new feels a bit like a burden as it will need > quite extra time. On the other side make is pretty common. That's weird. I seem to have opposite experience. Perhaps we need a more concrete example. It has been a while now since I did any work with CMake, but I remember thinking that it was great because the generated makefiles had comments that point to the line of the source CMake file that generated them. In my most recent trip though trying to debug autotools makefiles I was missing this help. Aaron Spike

1 0

SOC 2008 - project - Guidlines feature
by Vangelis Katsikaros 26 Mar '08

26 Mar '08

Hello I am interested in the Guidelines features proposed in the SOC 2008 ideas (http://wiki.inkscape.org/wiki/index.php/SpecGuidesImprovement) and I would like to discuss some aspects before submitting my proposal (if this is ok). I guess this feature fits in the category of projects that doesn't mess much with the base code, since it uses already existing features. Moreover, I tried to find info about the guidelines: * From a quick search I did I saw that thre is no "guide line" element in SVG 1.1 . * I saw that in svg files the guides are defined as sodipodi:guide element. So, guidelines are Inkscape specific. * Guidelines are defined in src/sp-guide.cpp . SPGuide handles the svg file info. (Does this mean reading and writing from the file and modifying the svg abstraction?) * The rendering of the guidelines is handled in src/display/guideline.cpp . SPGuideLine holds the guideline info The the sp_guideline_* functions are responsible to deliver to the renderer the correct into? (not sure about this last one). * SnapManager manages the snapping. However, I haven't figured out if snapping and guidelines are directly connected. (However. this isn't relevant with the UI feature) This is an early search and I am just getting in touch with Inkscape internal. So, if I get something wrong or if you have any hints and directions they are welcome. (I've seen the dev docs you have on the wiki) As far as the feature specification is concerned, I would like to ask about the Design -> On-canvas editing section. The first 3 actions are already implemented. The last 2 were not implemented for a reason or you simply didn't have the time? Regards Vangelis Katsikaros __________________________________________________ ×ñçóéìïðïéåßôå Yahoo!; ÂáñåèÞêáôå ôá åíï÷ëçôéêÜ ìçíýìáôá (spam); Ôï Yahoo! Mail äéáèÝôåé ôçí êáëýôåñç äõíáôÞ ðñïóôáóßá êáôÜ ôùí åíï÷ëçôéêþí ìçíõìÜôùí http://mail.yahoo.gr

4 10

Transformations using the selector tool
by Diederik van Lierop 26 Mar '08

26 Mar '08

Hi all, Please be keen on any regressions in the way the selector tool handles transformations. If you see something weird then drop me a mail. This is all because I've given seltrans.cpp an overhaul to improve the handling of transformations, their origins and the snapping. For details, see this bug and revision https://bugs.launchpad.net/inkscape/+bug/174046 http://inkscape.svn.sourceforge.net/viewvc/inkscape?view=rev&revision=18015 Johan, I will now finally start working on the snapping indicator in the selector tool! Regards, Diederik

2 1

python extensions - lxml version 2
by Terry Brown 26 Mar '08

26 Mar '08

lxml version 2, from http://codespeak.net/lxml/lxml2.html#incompatible-changes "Some API functions now require passing options as keyword arguments, as opposed to positional arguments." e.g. the Edge 3d extension is broken because xpath is now ...xpath(path, namespaces=MyNameSpaces) rather than ...xpath(path, MyNameSpaces) so does that mean I should fix the extension to work out which version of lxml it's using? Or assume 2.x? Cheers -Terry

2 1