August 2005 - Inkscape Devel - Inkscape Mailman

Locale settings on windows
by Jon A. Cruz 31 Aug '05

31 Aug '05

There have been a few issues now and then with gettext and friends on Win32. A few related tracker items are [ 1250581 ] WIN32: Inkscape uses locale settings for determining UI lang [ 1246076 ] Mixed languages in UI [ 1246463 ] Language Selection Option / Menu Entry After talking with Peter some last night, it was bubbling around in my brain, and kicked out a work-around. This will need to be tested/refined, but the basic approach could be to hook into things right at program startup (over in the Win32 bits). At that point we could check if LANG and/or the other POSIX environment variables were set. If not, we could then just query the Win32 NLS calls and then build up and set LANG and LC_* properly. Some preferences could be set to affect this, and perhaps override it with explicit user values/choices. The details need to be hammered out, but I think this is a good work- around to allow us to get a better user experience on Win32 without having to get libc changed, etc. I wanted to toss this out in case others were still pondering things today.

1 0

Google Summer of Code projects
by Bryce Harrington 31 Aug '05

31 Aug '05

Hi all, The Google Summer of Code projects are due to wrap up by September 1st. We've been quite thrilled at the progress made from the SOC students, but as they say, all things must come to an end. We've not received direction from Google about the next steps, which is kind of weird... I have expected we'd see some directions by now. Possibly we'll hear something tomorrow? In any case, based on what we've been told in the past, we expect that we're going to need to go through a review phase where the mentors check the student's work and decide whether it is acceptable. I think each mentor will have his own set of criteria to judge from, but there are some expectations that I suspect will be weighed heavily on: * Has the code been checked into Inkscape CVS? * Does the code function well enough that people can at least test it out? * Has the developer participated with other inkscape developers to ensure that bugs are fixed, requested changes made, etc.? * If the developer has "bitten off too much to chew", have they triaged the requirements and implemented an adequate subset? Of course, if the code implements 100% of what was mentioned in the original proposal, that's even better. ;-) But as long as it meets most of the requirements and shows that the participant has been successful at learning Open Source, I would think it would be acceptable. Bryce

3 2

build report /inkscape-20050829-2120
by Bryce Harrington 31 Aug '05

31 Aug '05

= Analysis Report for inkscape = Platform Unpack Config Install WRNS ERRS ------------------------------------------------------------------------ debian x86 0 0 2 5 28 Test Status Score Info ------------------------------------------------------------------------ sut-1 ?? == Errors == attributes.cpp:34: error: syntax error before `<<' token attributes.cpp:37: error: syntax error before `>>' token connector-context.cpp:1195: error: syntax error before `<<' token connector-context.cpp:1199: error: syntax error before `if' connector-context.cpp:1199: error: ISO C++ forbids declaration of `__r' with no connector-context.cpp:1199: error: redefinition of `int __r' connector-context.cpp:1199: error: `gboolean __r' previously declared here connector-context.cpp:1199: error: syntax error before `}' token connector-context.cpp:1212: error: `inkscape_active_desktop' undeclared (first connector-context.cpp:1212: error: (Each undeclared identifier is reported only connector-context.cpp:1230: error: `cc_item_is_shape' undeclared (first use connector-context.cpp:1249: error: syntax error before `>>' token connector-context.cpp:1256: error: `selection' was not declared in this scope connector-context.cpp:1258: error: syntax error before `if' connector-context.cpp:1269: error: redefinition of `GType __t' connector-context.cpp:1199: error: `GType __t' previously declared here connector-context.cpp:1269: error: redefinition of `gboolean __r' connector-context.cpp:1199: error: `int __r' previously declared here connector-context.cpp:1269: error: syntax error before `if' connector-context.cpp:1269: error: ISO C++ forbids declaration of `__r' with no connector-context.cpp:1269: error: redefinition of `int __r' connector-context.cpp:1269: error: `gboolean __r' previously declared here connector-context.cpp:1269: error: syntax error before `}' token connector-context.cpp:97: error: `void sp-item.cpp:324: error: syntax error before `<<' token widgets/toolbox.cpp:2993: error: `cc_selection_set_avoid' undeclared (first use widgets/toolbox.cpp:2993: error: (Each undeclared identifier is reported only widgets/toolbox.cpp:2999: error: `cc_selection_set_avoid' undeclared (first use == Warnings == display/canvas-arena.cpp:91: warning: invalid access to non-static data member display/canvas-arena.cpp:91: warning: (perhaps the `offsetof' macro was used display/sp-canvas.cpp:127: warning: invalid access to non-static data member ` display/sp-canvas.cpp:127: warning: (perhaps the `offsetof' macro was used libnrtype/Layout-TNG-Output.cpp:226: warning: enumeration value ` == Environment Summary ==

1 0

Re: [Inkscape-devel] DXF export
by Matt Squires 31 Aug '05

31 Aug '05

DXF export is now possible... At my website (http://ucsu.colorado.edu/~squiresm/), that hopefully will not be needed much more, I have posted .inx and .sh files that will export to dxf using pstoedit. I found that dxf is already supported, there are a fair number of features that I haven't explored. The conversion seems to be ok. I also have more functionality in the dxf import. Layers are supported and I have turned linetypes back on. The linetype is a little buggy because it is defined in two places and I have tried to make a good decision about which definition takes precedence. I also fixed problems with bulges in plines and lwplines. The files are there for color conversion, but I haven't included the color yet in the SVG. I am open to people sending me their favorite dxf and then I test it against AutoCAD (there is a license at school) or if you test a file and you don't see what you want then let me know. Eric: Along those lines... you are right the projection of a tilted circle on a plane below should look elliptical. I will fix that sometime later. Matt > Generally, our policy is to provide at least one patch, and if good, > then we give you cvs...it is our primary policy. I will get a patch tomorrow when I have better facilities

1 0

please test: svgslice as an extension
by Lee Braiden 31 Aug '05

31 Aug '05

Hi all, Now that 0.42 is done, I thought I'd send the extension version of svgslice for others to test and consider for inclusion. It's pretty basic: just saves marked sections of the image into separate PNG "slices", and generates a basic XHTML file that lists the slices. Most notably, no equivalent layout is done in the XHTML file, so that it looks like the original SVG drawing. Problem is, there are many ways to do that in CSS, and the most obvious way (absolute positioning) is probably not much use to people. I suspect it would lead to bad web pages, or just get in the way of proper hand-crafted layout, and people who want to use a WYSIWYG CSS editor to move them around easily can still do so. Note also, that the slices will be saved in Inkscape's current directory. If you load inkscape by double-clicking on an image in a project folder, and save the XHTML file in the same folder, then this is fine -- your startup directory will be that folder, and images will be right alongside the XHTML file. But if you start inkscape from your home directory on the command line, or from just running it via a menu entry, and then load and save a file in another location, the PNG slices will end up in your home directory (at least on unix; windows might save then in C: or "C:\Program Files\Inkscape", or something like that. At the moment, there's nothing I can do about that, since inkscape doesn't CD to the output directory before calling the extension, and doesn't seem to provide the extension with any idea of where it is actually saving stuff. Personally, I'd like to see that issue fixed for 0.43. But either way, it'll be good to hear if this works for others, if there are any bugs, etc. So please give it a test, and let me know :) Installation ============ Save both attached files in your inkscape extensions directory. If that directory is somewhere other than /usr/share/inkscape/extensions, then you'll need to edit the paths in svgslice.inx to reflect that. Obviously, this will have a different path on Windows. Maybe someone could edit the svgslice.inx file, and post it here for other windows users? Usage ===== * Draw something. * Create a layer called "slices". * Draw rectangles in this layer representing the areas you want to slice (note: it helps if you make the rectangles semi-transparent) * Optionally, name any/all slices by using the built-in XML editor to edit the "id" attributes of the rectangles. * Hide the slices layer. * Choose "Save as..." from the File menu. * Select "XHTML file with PNG slices" as the file type to use. * Enter a filename (e.g., "test.xhtml"), and save. Result ====== * You should have the test.xhtml file on your disk, alongside the PNG slices for each area of the drawing that you made a rectangle for. -- Lee Braiden http://www.DigitalUnleashed.com

4 13

Follow up with Nuno
by Dan Ostrowski 31 Aug '05

31 Aug '05

Follow up with Nuno Pinheiro: > Im using 0.42, but i did not figured the correct way to use the > swatches, i can see them, but i cant "paste" a color in a gradient, > and also i don't know how to make a swatch, (but i can obviously > learn), and the feature im requesting is also useful for a lot of > other things, (such as creating a perfect gradient from one solid > object to an other solid object, setind qa color for an outline and > more.) regards, dan o

4 4

releasing Clip Art Browser 0.4!!
by Greg Steffensen 30 Aug '05

30 Aug '05

Hey all, I'm releasing version 0.4 of the Clip Art Browser. This, of course, is the release of the browser thats meant to be the final product for the purposes of Google's Summer of Code. Grab it from http://download.berlios.de/ocalhelper/clipartbrowser-0.4.tar.gz Hopefully, installation is even simpler now. The indexing process for local clip art, which could be a bit confusing, has been integrated into the GUI, so that now, you really only have to extract and run. The README has been updated, but if you have trouble getting it to work, let me know. Ironically, the one feature that no longer works is, uh, working as an Inkscape extension. That should be very simple to do... it just need to figure out how to have the executable that Inkscape calls launch the browser as a separate process, but this has been surprisingly tricky in python. If people would like to put this in 0.43, I can get this fixed tonight if needed, but I assumed that it probbaly wouldn't be. I also figured that the effects menu isn't really the appropriate place for something like this, and we'd have to semi-hardcode it into some other menu later anyway (I think Jon mentioned something about this earlier). This isn't a final product, of course. I look forward to continuing to improve it, and hopefully getting more into the code of Inkscape proper over the coming months. Thanks to all the developers who provided feedback, and Bryce and Jon in particular. Also, java sucks. Greg

2 1

inkscape:connector-avoid vs inkscape:avoid
by Peter Moulder 30 Aug '05

30 Aug '05

Some places in inkscape use inkscape:connector-avoid (attributes-test.cpp, and the spelling of SP_ATTR_CONNECTOR_AVOID), and some places use inkscape:avoid (sp_object_read_attr and sp_object_setAttribute). I think I prefer inkscape:connector-avoid. Or is it intended for the same attribute to be used for removing node-node overlaps? (Whether as a once-off operation, e.g. using Tim Dwyer's implementation, or the dynamic linear approximation stuff that njh & I did for UIST last year.) If so, then I'd go for inkscape:avoid, and change the enum to SP_ATTR_INKSCAPE_AVOID. pjrm.

2 1

mjwybrow, bug in CVS with AMD64, 'Error: clock_t is not 4 bytes'
by Edwin de Jong 29 Aug '05

29 Aug '05

Hi mjwybrow, A small bug exists in the code of the file: libavoid/timer.cpp After running the code on an AMD64 system: Error: clock_t is not 4 bytes Aborted This snippet has a bug: if (sizeof(clock_t) != 4) { fprintf(stderr, "Error: clock_t is not 4 bytes\n"); abort(); } Its always a good idea to never compare a size_of with a constant (unless in things like configure scripts or with very lowlevel program code). I havnt read the rest of the file yet (so I dont know why you insist on having it 4 bytes), but it is obvious that such structures have different sizes on different architectures. If necessary, hop me an email and we can work this problem out via IRC (if you have not an 64-bit system available). Greets, Edwin de Jong

2 1

Inkscape 0.43
by Bryce Harrington 29 Aug '05

29 Aug '05

Hi all, Last month we talked a bit about accelerating our releases a bit. Since it's been a tad over a month since the 0.42 release, and feels like now would be a good point to begin the 0.43 release process. Based on prior releases I expect it may take several weeks to a month to go through this release process. I'd like to propose that we target Sept 1st as our feature freeze point. Does anyone feel strongly that we should wait longer than this? Bryce

12 25