May 2006 - Inkscape Devel - Inkscape Mailman

some questions about sf.net's RFE and bugtracker (statistical data)
by Khiraly 22 May '06

22 May '06

Hi! How can I obtain some statistical about bugreports and RFEs ? Im thinking of: * how many open bugreports/RFEs are at the moment * how many bugreports/RFEs was closed from a given date (ex: from the latest release) * how many bugreports/RFEs need info, etc Is sf.net is able provide me some of the aboves info? Khiraly

8 18

transparency pdf export
by Peter Dedecker 22 May '06

22 May '06

Hi, I just checked out the snapshot release of Inkscape 0.44, wich is just great, thank you very much! I really love the new pdf export with transparency support and the "crop to image size" button in the page properties. But I also have a little question. Printing such pdf files with transparency isn't simple and transparency is removed. I think there should be two types of pdf export functionality. One of them is the one that's now included, the other one should be a mix between bitmap output and scalable pdf output. I mean: it should be great when you could let Inkscape "render" the image and count the resulting colors when the different semi-transparent layers are added to eachother. The result should be scalable but without transparency. This way, the image printed is the same as how you could see it at the screen. This could also be related to bug 1492670. Acrobat prints files with transparency the right way, Evince doesn't. I think it should be good when it's possible to export to different pdf-versions, with and without transparency. What do you think of this? After some positive feedback from Bryce, I submitted this one as feature request 1492661. Kind regards, Peter Dedecker

2 2

Effects (aka extensions)
by Tavmjong Bah 21 May '06

21 May '06

Hi, I've been working through the extensions in updating my book and have the following comments: 1. The release notes uses "Extensions" while the menu label in the GUI is "Effects". 2. Extract one image: Dialog is confusing. It asks for "Path to save image". I initially though that it would get the file name out of the SVG file (maybe from the id name). The dialog would be more clear if it said "Filename (with path) to save image". (Aaron points out that it would be even better if it used a file chooser dialog.) 3. Envelope: Generates warnings of the type: "Format autodetect failed. The file is being opened as SVG." Otherwise the effect works. 4. Flatten Bezier: "Flatten Path" is the name used in the release notes. 5. Radius Randomize: Minor quibble: The "normal distribution" generated isn't really a normal distribution (as you would find in the error of measuring the position of a point). The x and y coordinates should be uncorrelated and thus one should write: x += abs(random.normalvariate(0.0,0.5*r) y += abs(random.normalvariate(0.0,0.5*r) 6. Latex Formula: pstoedit must be installed with the SVG plug-in. The plug-in is shareware and only available for Windows and Suse Linux 9.3 using g++ 3.4.5. (See: http://www.pstoedit.net/pstoedit.) 7. Fretboard Designer: These effects should really be grouped in a submenu under Render. 8. Effects menus: The new organization of the Effects in submenus is great improvement! I think I would have approached the problem of implementation differently. Rather than embed menu structure inside the effects, it might of been more interesting to extract the menu structure from a directory structure, obtaining the names of the menus and of the effects from the names of the directories and the inx files respectively. For example, the Fretboard effects would be place in a directory named Fretboard_Designer inside a directory named Render. The first Fretboard effect would have an inx file with the name: Single_Length_Equal_Temperament. If would then be easy for a user to add their own effects or to move them around. Of course, one would still have to come up with a method to handle translations. The code could be reused to organize the Templates menu. Tav

10 15

tablet bugs
by bulia byak 21 May '06

21 May '06

I think I fixed now the various bugs reported by Mental and others about the use of tablet pen in various tools: http://sourceforge.net/tracker/index.php?func=detail&aid=1455685&group_id=9… (and probably other reports too). Turns out "it's not a bug, it's a feature". This affected everything that dragged knots plus the rect tool; this also probably affected guideline dragging but nobody noticed that. The reason is that only knot.cpp, rect-context.cpp and the ruler/guideline code grabbed canvas with a mask that included GDK_POINTER_MOTION_HINT_MASK. And this mask, per documentation, is exactly for this: it passes only the first motion event and suppresses all others, until you call gdk_window_get_pointer. Only ruler code cared to call; there's also a call in canvas motion event handler but it seems to have no effect. Now, I have no time nor desire to figure out why this worked before (the hint mask was there since Sodipodi times), or why it broke with newer GTK versions, or why it broke only for tablet pens and not mice. Someone who likes GTK more than me might enjoy the exercise :) For now I simply removed the hint masks wherever I could find them, and everything seems to work fine for me (though for me it always worked fine, because I have a rather old gtk) and for a person on IRC who had troubles with pen before. Of course this change is somewhat risky, so please test all tools and all kinds of dragging thoroughly. Let me know at once if you find something broken. -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

2 1

Inkboard in 0.44
by Ted Gould 21 May '06

21 May '06

Okay, So, I know that everyone is excited about the growing maturity of the Inkboard code, but I'm going to suggest that it not be included by default in the 0.44 release. Why? I think that with David reworking the protocol, and the potential for a SoC project to help with that, it's likely that there will be some protocol changes between 0.44 and 0.45. I don't think we really want to support the 0.44 protocol forever (if nothing else, it won't be as well documented), so I think we should leave it for people who are actually on the mailing list and update their Inkscape binaries regularly. Thoughts? --Ted

3 2

SWF Export - Supporting static library
by Brian 21 May '06

21 May '06

I was able to take the .swf export patch on the Sourceforge page and update it so it compiles with the current Inkscape/ming codebase. Now comes the fun of linking libming in as a static library. Out of the box, ming ships with a make target that builds a static version of the lib. I added the following entry into src/Makefile.am noinst_LIBRARIES = \ extension/internal/ming/libming.a \ But this wants to build libming. What do I need to do to simply link the libming lib in statically? Or would it be best to set up a Makefile_insert that builds a custom version of the Ming library? Brian

3 2

license and attribution
by Aaron Spike 21 May '06

21 May '06

I'm a chronic code copier, I know. But I just realized that I have made a few mistakes moving code around in our codebase. I moved PointToCurvilignPosition from libnrtext/Layout-TNG-OutIter.cpp to livarot/Path.cpp as was indicated in the orignial comment, "this function really belongs to Path. I'll probably move it there eventually, hence the Path-esque coding style." In doing this I forgot to copy the original copyright in the code. But much worse, I just realised that the files that I moved this code between are under different licenses. I moved a GPLed function into a PD file. This cannot happen. The possibility for other accidents in our codebase seems huge. I don't think PD and GPL can coexist well in a codebase that desires so much refactoring. Can we and should we place the PD files under the GPL? As for attribution, what is proper and customary? I suppose I need to chase down the orignial author through the commit logs any time I want to move a function or copy code. Very often people alter code without adding a copyright notice. I haven't added a notice to any of the files I have edited. Am I endangering people of doing the same copying without attribution to me? When should I attribute myself? Right now the act of adding my name to a source file seems rather proud to me. And right now I can't imagine ill will tward someone copying my code without attribution, but as I put more and more sweat and tears into the code that may become possible. Aaron Spike

5 4

object snapping (bug?), related RFEs
by Khiraly 20 May '06

20 May '06

Hi! Is object snapping supposed to work? Im talking about the recent svn version, and File->Document properties->[Snap] The options: [ ] Snap bounding boxes to objects [ ] Snap nodes to objects [ ] Snap to object paths [ ] Snap to object nodes Snap sensitivity: [-|-------------] [ ] Always Snap It took some time the figure out how it is working. 1. So If I want to snap one object node to an another object node I must check the following options: [ ] Snap bounding boxes to objects [X] Snap nodes to objects [ ] Snap to object paths [X] Snap to object nodes Snap sensitivity: [-|-------------] [ ] Always Snap 2. If I want to snap one object node to an another object path I must check the following options: [ ] Snap bounding boxes to objects [X] Snap nodes to objects [X] Snap to object paths [ ] Snap to object nodes Snap sensitivity: [-|-------------] [ ] Always Snap 3. If I check the above two cases it work a little buggy way. Particulary on a curved path. So I pull the mousepointer and sometimeworks sometime just jumping a little distance. Around the node is really hard to place the node. 4. If I check the following to options, it does snapping, but Im not able to figure out its snapping to what. [X] Snap bounding boxes to objects [ ] Snap nodes to objects [ ] Snap to object paths [X] Snap to object nodes Snap sensitivity: [-|-------------] [ ] Always Snap It does snapping, because the Snap sensitivity have its behaviour. 5. If I want to snap the bounding box around an object paths, I must check the following options: [X] Snap bounding boxes to objects [ ] Snap nodes to objects [X] Snap to object paths [ ] Snap to object nodes Snap sensitivity: [-|-------------] [ ] Always Snap ------------------ The problems: The bigger problem is the not clear UI design, its not marked anywhere what I must check If I want some behavoiour. And it does not marked, that one checkbox *depends* an another checkbox. 'Always snap' button does not seems to work. What is supposed to do? Its not clear that the 'Snap sensitivity' GtkHScale element affects ALL the other checkboxes. I was thinking yesterday, that the snappings are not implemented, because I was unable to figure out what checkbox must I checked. So I think it must seriously reconsider this ui design, because this EXCELLENT feature is really a MUSTHAVE. I have found and figured out accidently, and I have already written my feature request for this feature, which is already implemented! (just figured out a Snapping title in the recent changelog) ---------------- Question/rfe about the snapping issues: There is an option (like the new show/hide bezier handles option) to show the nodes of all the objects? It would be easier to snapping. And more speedy to create nice diagrams, flowcharts, etc. ---------------- This feature CLOSE the following RFE (I think): [1014477]Snap (is already closed, but now is fully implemented): https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1326224] Snap to curve and snap to node: https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [907478] Snap to (nodes of) existing objects: https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1252641] Snap objects to nodes of other objects (DUP of #907478): https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1108238] snap bounding boxes to each other: https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1306873] In node tool, nodes/handles must always snap w/ grid on https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [903432] Snap selection to grid command: https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1424653] Snap behavior instability (This is not implemented, but it can easily achieve with the recent snapping possibilities): https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1418315] Grid: option to snap to upper left corner only (DUP of #1424653) https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… ----------------- Related RFEs, which is NOT IMPLEMENTED, but depends on the above implementation: [1104630] Special fixing points to allow objects to snap together (Its like in Microsoft Visio, or DIA, where one object can snap some autogenerated fixing points. It can workaround by adding additional nodes, but there arent any more elegant way?): https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1421855] Snap to Grid (There should be an option to active/deactivate the Snap to Grid without going to Document Preferences): https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1328200] snap icons in toolbar (DUP of #1421855) https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [863737] Snap indicator+ snap/guides dialog/palette (DUP of #1421855) https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1213890] Guides to Snap to Grid (almost everything can be snapping except GUIDES to GRID/BOUNDING BOX/NODE): https://sourceforge.net/tracker/?func=detail&atid=604309&aid=1213890&group_… [1328174] snap guides to nodes (DUP of #1213890, we can merge the two rfe) https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1213890] Guides to Snap to Grid (DUP of #1213890): https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [928027] advanced snap features (missing features, but it can easily workaround by adding some nodes to the object): Snap to Endpoint, Snap to Midpoint, Snap to Intersection): https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1390687] snap to path imprecise in some cases: https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… ------------- RFEs what is related, but I cant decide is it resolved or not. [1215241] snap to guides with screen pixels (I cant understand what is the difference between screen and document pixels) https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1119020] snap only to visible gridlines, snap distance in screen px (DUP of #1215241): https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… [1390676] On-screen-pixel distance as snap to... unit (DUP of #1215241) https://sourceforge.net/tracker/?group_id=93438&atid=604309&func=detail&aid… -------------- Proposal: Can we consider to close ALL the feature request about snapping? And create one new feature request, where we summarize what snapping feautres was implemented, and what needs to be implemented. So we can close the feature request what is implemented, and mark as DUPLICATE of the above RFE all the other feature requests. However with this RFE, we can track all the snapping issues and roadmap. Sorry for my english, and thanks if you have read all my email;) Best regards, Khiraly

7 26

Gradient ids change between 0.43 and SVN in my files
by Jean-François Lemaire 20 May '06

20 May '06

Hello, I've got a serious problem regarding gradients with last days' Inkscape SVN. Something changed that screws up the gradients ids of files I created with previous SVN builds. To make things clearer: If I create today an object with gradient fill with, say, Inkscape 0.43 stable and open it in Inkscape SVN, the gradient is correctly displayed. If on the other hand I open the files I created and worked on earlier with Inkscape 0.43 in Inkscape SVN, *all* the gradients are mis-attributed and do not display. They are present in the gradients drop down menu, but their ids no longer correspond to what is specified in <defs>. If I vacuum the defs, the gradients are deleted since they are not used by any object. Here's a link to a file that contains 24 objects, almost all of them with gradients: http://users.skynet.be/jflemaire/files/gradients_issue.svg (Incidentally, this file displays correctly in Konqueror.) I hope someone with more knowledge of SVG/Inkscape will be able to help, because I dread the moment I'll have to reassign all gradients to each objects manually! Cheers, JFL -- Jean-François Lemaire

5 5

Inkscape UI
by Kevin Cannon 20 May '06

20 May '06

Hi all, I've been following Inkscape's development for some time, and it's always looked like it had quite a lot of promise. I was looking for an open source vector tool to recommend to someone today and thought I'd download it to see how it's coming along. I'm happy to say that it's improved a lot since I last used it. I've very impressed. The one thing I'm concerned about is the UI. I'm a web designer by trade and while Inkscape is quite full featured, the UI is a bit of a leap to what I'm used to, which makes transition very hard. I read your wiki, and I understand wanting an Inkscape specific UI is important and I agree that just aiming to be an Illustrator clone wouldn't be a worthy goal. However, I think you could win a lot of hearts and minds by conforming to general UI elements that are used in graphics packages. During Firefox's early development, they made the conscious decision to match some of IE's shortcuts (alt+d, backspace for example) and some of it's conventions. The aim was to reduce the barrier to entry, and to make it feel more familiar, and it was one of the factors in it's success. I would strongly suggest you make designers feel more at home when they give Inkscape a try. Things like pressing spacebar to pan around a page is a de facto standard amongst image editing tools. I'd say it would be trivial to make Inkscape work in the same manner. Similarly, your path nodes/handles tool are the same as the 'white arrow' tool in other apps. Changing that so it matches would again dramatically increase the accessibility of the software, and I believe seriously increase your usage figures. These are some small examples of what I'm talking about. As a designer, I'm designing interfaces every day, and creating a low barrier to entry is crucial for my clients. I innovate in certain areas, but know that some things need to stick to conventions or else people will never use the sites I build. Inkscape is a great tool, but it's broken some of those conventions, and that will frustrate and annoy users when they try to apply their current knowledge to this application, and that is a waste. You needn't lose potential users when Inkscape is so very good, just because they're confused. I would strongly suggest that you make efforts to remove those barriers while staying true to your goals. I'm not a coder so I cannot offer my services in that regard, but if you want to change some things to make the UI more friendly to new users, then I will happily offer my services in identifying issues. I sincerely hope you'll take me up on my offer, Kind Regards, - Kevin Cannon

11 27