March 2007 - Inkscape Devel - Inkscape Mailman

NUBIE: extending pstricks output; printing subsystem
by Richard Henwood 20 Mar '07

20 Mar '07

Hi folks, I'm looking into fixing (text doesn't appear in output with latest SVN) and extending it to support embedding text as text. During a preliminary familiarisation hack, I've introduced a function: PrintLatex::text(Inkscape::Extension::Print *mod, char const *text, NR::Point p, SPStyle const *const style) (using ps.cpp as an example). I need the position of the text, and I was hoping that p[NR::X] and p[NR::Y] would give this, but they have the value 0.0000 0.0000 I'm new to cpp and Inkscape. Can you advise where I might be going wrong - and any wiki documents I should be checking in addition to PrintingSubsystem best regards, richard

2 1

Contributing...
by Jeff Schiller 20 Mar '07

20 Mar '07

I've become interested in Inkscape again and I'd like to figure out a good way to contribute. Here's a couple things I thought I'd throw out there: 1) I'm used to Bugzilla, has there been any discussion about moving away from a Sourceforge Bug Tracker into something a little more powerful (in terms of search? 2) In Mozilla, anyone can submit a patch to a bug, then the admins/super-reviewers must accept and merge the patch. Is there a mechanism for this within Inkscape? For instance, I'd like to report a bug against a SVG tutorial and then post a suggested fix for it that an admin can take and merge into the codebase. Similarly, my ambitions would be to get into the source code of Inkscape eventually and attempt another patch (this is more of a long-term goal though). Basically is there a way for the "casual" developer to submit patches in a standardized way? Thanks, Jeff

3 4

Security updates to Inkscape
by Kees Cook 20 Mar '07

20 Mar '07

Just before 0.45.1 was going to release I happened to do an audit of Inkscape's use of printf-style format strings (and the "..." argument). Unfortunately, there were multiple vulnerabilities[1]. Aaron gracefully delayed the final release so I could get CVEs issued, and notify other distributions. The basic problem the use of a *printf call directly on a variable without a format string. For example: printf(variable); vs printf("%s",variable); In the former, printf will parse any format chars in "variable", including the dangerous "%n", which allows for arbitrary memory overwriting. When functions use "..." to take in format strings for later processing by a *printf function, you can end up in the same situation. (Sorry to pick on ishmal's code, but it's a good example.) From the recent security fix commit: - status(event.getData().c_str()); + status("%s", event.getData().c_str()); "status" was defined as: virtual void status(const char *fmt, ...); Now, since the code is *filled* with format string "..." usage, it can be a real pain to find them all. Luckily, gcc has an attribute to flag a function as acting like a printf-family call, and Glib has a macro to create this attribute if gcc is being used as the compiler. So, now "status" is defined as: virtual void status(const char *fmt, ...) G_GNUC_PRINTF(2,3); The two args in G_GNUC_PRINTF point[2] to the format string variable, and the "..." argument. In the case of a class (as above) there is an implicit "this" argument as argument #1, meaning the format string "fmt" is really arg 2, and the "..." happens at position 3, as reflected in the G_GNUC_PRINTF usage. These markings, combined with the gcc compiler flag "-Wformat-security" will issue warnings about missing format strings, and do other strict type-checking. I've added this to the default CFLAGS as well. (And -D_FORTIFY_SOURCE=2, which adds even more protections via glibc.) Overall, many format issues were corrected[3], only a few of which were security-sensitive (37 files changed, 120 insertions(+), 111 deletions(-)). When making new "..."-using function, please review the glib macros for appropriate attributes, and take care to use them safely. I'd especially appreciate it if someone could look at: ./display/sp-canvas.h:SPCanvasItem *sp_canvas_item_new(SPCanvasGroup *parent, GtkType type, const gchar *first_arg_name, ...); This seems to be a NULL-terminated list (and not a format string, obviously), but it should get marked as such[4] if that's true. The dom/js/ tree is stuffed with "..." and I did not audit it, since it is not hooked up yet. However, we need to be very careful with it, and I'd MUCH prefer this code stayed outside of the inkscape code base so we can just link to an external library in the event of security updates to those libraries. Let me know if anyone has any questions; I'd like to keep inkscape as safe as possible. :) Thanks, -Kees [1] Try it yourself: prior to current trunk and 0.45.1, run this: inkscape /tmp/no-such-file-%x.svg And you'll see the "%x" turned into a number off the edge of the stack. The Jabber protocol handler was also vulnerable. [2] http://developer.gnome.org/doc/API/2.0/glib/glib-Miscellaneous-Macros.html#… [3] http://inkscape.svn.sourceforge.net/viewvc/inkscape?view=rev&revision=14554 [4] http://developer.gnome.org/doc/API/2.0/glib/glib-Miscellaneous-Macros.html#… -- Kees Cook @outflux.net

1 0

Hi everyone!!
by Celina Jorge 20 Mar '07

20 Mar '07

I arrive just now in this mailing list. At first, I have to say that my mother language isn't english, but portuguese, so I ask you to apologize any grammar/spell error. Here, in Brazil, I'm very interested in Inkscape plugins development, although I'm still trying write the first and studying things like SVG and XML. You also can find me in Inkscape-Brasil mailing list. It's a pleasure to be here. [ ]'s -- +..........................................................+ . Gmail:celina.jorge@...400... . . JID: celj@...1696... . . Msn: celina_jorge@...864... . . . . http://celde.wordpress.com . . http://www.cin.ufpe.br/~cinlug/planet/ . +...........................................................+

2 2

Re: [Inkscape-devel] Summer of Code 2007 - Possible application?
by Gail Banaszkiewicz 20 Mar '07

20 Mar '07

Hey Max! You seem to share the same enthusiasm as I do. You have beat me to the bug fixing though... I was hoping to do that this week along with putting together some proposals ;) > On second thought, the shared interest of two people could also be a > great advantage because maybe there is a way to split the project into > smaller parts which can be tackled more or less independently. I think that if this is possible, it's a really great idea. Our abilities seem to match up well, because as you say, you have a (probably) deeper understanding of the math, and I (may) have more programming experience given that my major is computer science rather than minor. Yet I imagine we are both sufficiently proficient in both aspects. I guess we can wait and see what the good mentoring folks say, but if we can come up with complementary proposals where we each have separate and sufficient work, I imagine our chances are good! I should note, though, that there are many suggested projects I am interested in, which somewhat deters me from wanting to come up with my own ideas ;) Instead, I may just try to put together a killer proposal for the suggested project ideas. We'll see what happens. My point here is mostly that if you turn out to be more suited for the 3D tool, there are no hard feelings. We should both put forward multiple proposals if possible and it would be nice if we could both get in. Looking forward to seeing how this will play out, Gail

3 3

Inkscape and barcodes - Inkbar (fwd)
by Alan Horkan 19 Mar '07

19 Mar '07

I was going to send this to inkscape users to help inform them about the Inkbar Barcode extension and ask if anyone know who was the author of Inkbar. Instead I thought maybe it would be better to direct it here so that developers might share their thoughts on making extensions easier to develop since Inkbar does suffer some of the common problems? ---------- Forwarded message ---------- Date: Mon, 19 Mar 2007 22:50:05 +0000 (GMT) From: Alan Horkan <horkana@...44...> Subject: Inkscape and barcodes - Inkbar I was thinking a barcode program could make some of the work I'm doing at the moment a lot easier, especially one I could also run on windows (Kbarcode is a good barcode AND label program but it is not availalbe for windows). On a whim I searched google for Inkscape + Barcodes and turned up a nice page in the Wiki listing some extensions* which are not included by default (doesn't make sense to include all extensions* by default). http://inkscape.org/wiki/index.php/ExtensionsRepository In particular it looks like "InkBar" might do what I need http://www.inkbar.lineaire.net/ My french is a bit rusty and I cannot see any contact or author information either on that page or included with the extension but I would apprecaite some help and perhaps my concerns with this script* are applicatable to others. I've had similar concerns before when writing scripts for the GNU Image Manipulation Program (GIMP) and some Javascript bookmarklets so these suggestions are based on a fair measure of experience, and some of things I wished I'd had to help me write my scripts. First there is the problem of translation. I'm not sure how we can help translate third party extensions/plugins/scripts/whatever*. I would guess that perhaps certain strings would keep coming up over and over again such as Angles, Detail, Gradient, Path, or something like that so maybe Inkscape could provide some kind of stock labels which would get translated once then any script writer could reuse them (something like how Gnome stock lables are supposed to work)? The other problem - and I've seen this a million times with GIMP scripts - is input checking. The inkbar script fails if the required 12 or 13 characters are not provided, which isn't very forgiving. (In the case of inkbar the rest could probably be padded with spaces.) In general scripts will require input checking, and hopefully this isn't something each script developer would need to writer their own collection of checks and tests for (because in that case testing is unlikely to happen at all). In the case of Python scripts (and any other scripts) there is something of a delay between the time the user sets the values and then hits OK (lets call that stage 1) to execute the script and then another delay before the script runs (stage 2). In practice this delay is usually quite short but delay and errors could be avoided if there was some immediate error checking as the user types to make sure a letter is not typed into a box where only numbers are allowed even before the OK button is hit (stage 1). After the OK button has been hit other checks may also be necessary. (Sorry very tired, I'll stop and try and explain better later, but I think I should probably have phrased things for developers in terms of preprocessing and postprocessing like on Web based forms as an example of how much easier extensions might be.) Anyway, I could do with some help contacting the author of the Inkbar barcode extensions and I hope to get a lot of use out of it (and help make it easier to use, provide an English translation, and anything else i can think of). I do believe it could gain inkscape a lot of attention in small businesses and more. (Useful as Inkbar could be for me I'm not saying I'd want to include it by default, not yet anyway. Inkscape for artists.) Off to see if I can modify the script to group the output to keep things tidy and stop them from falling apart. Sincerely Alan Horkan http://advogato.org/person/AlanHorkan/ http://www.linkedin.com/in/alanhorkan http://alanhorkan.livejournal.com/ * cannot use the jargon consistently, brain hurts, long day P.S. Looking at the command line options I see they are also in French, hmm.

1 0

Help about Animate Document in Inkscape
by Federico 19 Mar '07

19 Mar '07

Hello Bryce and M-List member, you remembar that I'm an Italian student and I'm writing my Thesis of bachelor in engineering on the diagrams animated in inkscape. I explain you my problem: I would have the necessity to modify the attributes of field STYLE of an object to regular intervals of time. More in the detail, I would have necessity to write a “extension” of Inkscape that reads the data of an object (rect, path, star, etc.) from a DBMS and he applies them to the present object on the screen. As I said to you in one mail of some day makes (the bringing back under): --------------------------------- On Tue, Mar 13, 2007 at 12:08:15PM +0100, Federico wrote: > Hello, remembered of me! > I'm a student graduating in Computer science Engineering and I'm leading a > thesis on the Vectorial Diagrams Animated in Inkscape. > > I had successful to insert of inkscape a "parallel" Thread to the main > program, that it reads of the regarding data the characteristics of an > object and then assigns them to the object by means of the endured function > > sp_repr_set_attr(nodes[i],style.c_str(),attrib.c_str()); > > nodes[i] -> it is a carrier of gunlayers "Node" > (std::vector<Inkscape::XML::Node *> nodes;) > style & attrib -> are strings. > > But after this call, the program crasha and it is closed without some > message. By means of the debuggher they are successful to understand that > there is an error of Segmentation Fault: > > Program received signal SIGSEGV, Segmentation fault. > ---Type <return> to continue, or q <return> to quit--- > [Switching to thread 3236.0x6a8] > GC_clear_stack_inner (arg=0x47429a0 "", limit=2284336) at misc.c:293 > 293 misc.c: No such file or directory. > in misc.c > Current language: auto; currently c > > You can help me to resolve this problem being given to me some council to > follow? When you are in the debugger, issue the command 'bt'; this should give you a backtrace to show the functions that were being called to reach this crash. You will probably also need to specify which thread you're debugging; since Inkscape is single threaded we don't have directions on this, however they can be found on the internet - Ubuntu has good directions for debugging multithreaded applications. A technique that I find useful to help identify bugs like this is to set a break point in the code somewhere prior to the segmentation fault, and then step through the lines until the segmentation fault, and look at the values of some variables that might be causing the problem. Generally segmentation faults are caused by mis-used pointers. For instance, trying to use a pointer with a NULL or undefined value. Regarding adding multi-threaded capabilities to Inkscape, this is something which we have avoided. In part, it makes debugging more complex, and it opens a large class of bugs that are challenging to deal with. We've also heard that some of the libraries we use may not behave as well when used in a multithreaded app. If it is very important for you to have the multithreaded capabilities, please explain why; otherwise you may wish to go single threaded to save yourself a lot of difficult work. Either way though, it's exciting to hear of your work on animated diagrams in Inkscape! I'd love to hear more about it. Bryce --------------------------------- I have used a Thread parallel that goes to write field STYLE of the rappresentazione of an object. But in this way a Segmentation Fault is generated. How I can avoid this problem? For me it is not binding that my extension is multithread, only interests to me that to regular intervali of time they can be applied of the changes to of the present objects on the canvas and that they have a particular ID or one particular LABEL. You can give me of the straight ones in order to resolve the problem? Thanks Federico Naviga e telefona senza limiti con Tiscali Scopri le promozioni Tiscali adsl: navighi e telefoni senza canone Telecom http://abbonati.tiscali.it/adsl/

2 1

Summer of Code 2007 - Possible application?
by Maximilian Albert 19 Mar '07

19 Mar '07

Hi everyone at the Inkscape team, my name is Maximilian Albert and I am currently a student of mathematics with minor subject computer science at the university of Heidelberg, Germany. This email serves the purpose of introducing myself to the Inkscape community because I am considering applying for Google Summer of Code 2007. For your convenience I have split this email into several sections. I apologize in advance for its length, but I figured it might be helpful to be a bit more verbose and give you a chance to get an impression of me (at least, as far as this is possible via such a rather impersonal medium as email). 1) Current stage of my education. Right now I am in the final stage of my Diplom thesis in mathematics (the Diplom is roughly equivalent to a combined bachelor and master degree). I expect to complete my thesis in the next few weeks. Subsequently, there are four oral examinations the last of which will probably take place in late autumn or even further towards the end of the year. Passing these will finish my studies. 2) How I got interested in the project. I learned about Google summer of Code only about one week ago when I accidentally stumbled across the corresponding webpages. I was fascinated by the concept right from the outset, but with regard to the current state of my studies quickly jettisoned the idea of applying. I figured it would be illusionary to try and tackle an additional project of this extent, given that I will probably be spending most of the time in summer learning for my exams. However, the idea kept lurking in the back of my mind, and since I found myself more and more excited about the project's concept, I thought it over and came to the conclusion that there actually is a very good - and, most importantly, realistic - chance to successfully work on it. By now I am very much willing to give it a try. The point is that I can probably choose the dates of my exams in such a way as to keep the period during which SoC is supposed to take place completely free from any other duties. At this point you will probably be thinking that I am much too over-enthusiastic and therefore running the risk of completely misjudging my capacities. Please be assured that I definitely won't submit an application unless I can 100% guarantee that I will be able to commit my full time and power to the project. I am well aware that this requires very careful planning and organization. But I proved more than once in the past that precisely this kind of "double burden" makes me tap my full potential (this is also why I am mentioning it at all instead of keeping you in the dark about it). I am sure that in case I should really come to the conclusion that an application is realistic, and should I be chosen, I would be of great benefit for the project. 3) Why Inkscape? Having decided to wholeheartedly think about getting involved, I browsed the list of mentoring organizations and was delighted to find Inkscape among them. Way back in the last millenium ;) when I was still working under MS Windows I had made my first experiences with Adobe Illustrator and got even more fascinated about computer graphics and visualization than I already was. After my switch to GNU/Linux many years ago I discovered Inkscape and since then have used it working on a handful of projects, most of which were fun ones only of interest to myself. But some examples are actually used as the logos on my choir's website. See the main page at http://www.kammerchor-alumni.de as well as the several subcategories (unfortunately the site is in German, but clicking on the icons shouldn't require any advanced language knowledge ;)). Recently I have been following the improvements and breathtaking new developments announced on the Inkscape website. More than once I thought about making a contribution (in fact, I already made a tiny one in the form of reporting bug #1598337, but I guess that doesn't really count). Lacking a concrete starting point, however, and above all the necessary time during the preparation of my thesis, I had to postpone the idea. Therefore I was excited to learn about the opportunity of getting involved in the framework of GSoC. 4) My experience in (graphics) programming and open source development To tell the truth, in Heidelberg the demands for computer science as a minor subject with regard to concrete programming tasks are not the highest. Therefore most of the programming I have done so far were smaller projects for my own amusement. However, when working on a practical at the visualization group of the Interdisciplinary Center of Scientific Research [1] which was required for my Vordiplom (a kind of pre-degree), I was lucky in that I could bring in an idea of my own which developed into a very fruitful and quite extensive project. It was concerned with the development of a program I dubbed HyperView, which serves as a tool for 3-dimensional visualization of 4-dimensional spheres, using the Hopf fibration. The projects' documentation can be found at http://pille.iwr.uni-heidelberg.de/~kugeln/ (Luckily, I prepared it in English at the time.) Be warned, however, that the code has not been maintained since the completion of the project in 2003, and I am not even certain that it will currently compile. In recent times my extensive hobby of making music (singing in particular) and in this connection preparing score arrangements and even composing short pieces has led me to get involved with Lilypond,[2] a free(TM) musical typesetter and automated engraving software. My first small contributions were the creation of a new notehead (requested by a composer on the mailing list), a small bugfix and one or two bug reports. Right now I am working on a couple of small to medium-sized projects ranging from the creation of further noteheads and other symbols to the implementation of a new style for microtonal accidentals (used, for example, in contemporary pieces as well as Eastern European and Asian music). All of these are close to being finished, and I hope them to be included in one of the next few development releases (as soon as I will have submitted them). As far as my general programming knowledge is concerned, I am proficient in C, C++, HTML, (shell) scripting, and even such an excentric thing as metafont. ;) Furthermore, I am knowledgeable to different extents in a couple of other languages/tools, including Scheme/Lisp, Python, Perl, sed, and awk. 5) Ideas for a SoC project The most severe problem I am facing right now is that in the short period since mid-week (when I discovered SoC) I haven't yet been able to develop really great ideas about possible projects I could tackle. My primary intention so far has been to get acquainted with the source code (see also point 6 below) and to find out for myself if it would make sense to submit an application (which the way I see it now I would almost certainly answer in the affirmative). Of course, I browsed your website where first and foremost the description of the planned 3D box tool captured my fascination. I believe that all of the ideas related to it that are mentioned on the webpage are certainly both tractable (although the whole lot is doubtlessly too much for a single person during a SoC project) and promising to be fun projects. The underlying mathematics are definitely comparatively trivial from a mathematician's point of view, but judging from my experience with HyperView (see point 4 above) - albeit only on a gut level, without having thought about details yet - I suppose that a lot of small problems can be avoided by choosing the right design approach. I think that my knowledge of the underlying algebra would definitely prove very useful in this regard. Unfortunately (for me, at least ;)), I saw that Gail Banszkiewicz has already exchanged her ideas with you about this project (I may have missed other peoples' contributions on the subject since I only skimmed over the mailing list archives). I certainly do not want to come into conflict with her, especially since she has contacted you much earlier and seems to have more precise ideas. So I will have a deeper look at the other projects described on your website and will also try to come up with and crystalize out my own ideas. On second thought, the shared interest of two people could also be a great advantage because maybe there is a way to split the project into smaller parts which can be tackled more or less independently. If it were possible to agree in some way on a certain design or common interface prior to the start of the project, the two of us (or even more students if interested and accepted) could work independently on their specified task(s). In the end the various features could be combined to yield a much more powerful tool than would have been possible if worked on by a single person alone. Does any one of the mentors who is more into it than I currently am see a possibility for this? Of course, this is only a vague idea and would certainly require Gail's and other people's approval. And I would like to reemphasize that I definitely don't want to thwart anyone. Just thinking aloud ... 6) Completed tasks and open issues related to an application Following the recommendations on your website, I have tried to get acquainted with Inkscape's source code and internals. This also served the purpose of deciding for myself if it would make sense to apply. IMHO, the results have been encouraging. ;) Yesterday I checked out the svn version and compiled it on my system, which worked more or less smoothly. After agreeing on very short notice to participate in a singing project last night, ;) I have been concerned today with fixing bug #1671665 (randomly chosen from the list). This afternoon I was already about to write that my progress looks promising. Waiting until tonight turned out to be a good decision, however, because in the meantime I was able to actually fix the bug. The fix is submitted as patch #1683338. Please feel free to have a look at it and make comments if you wish. This completes half of the required bugfix task, I believe. It is unlikely that I can work on a second one tomorrow but I will do it in the next couple of days and keep you posted on my progress. As I said, it will also take me a few more days before I can finally decide if working on a project this summer really makes sense. Definitely the last thing I want to do is waste both my and your time with a healf-hearted application that is doomed to failure due to misjudgement and bad planning. I will get in touch with you again as soon as there are any news, such as ideas for possible projects. In the meantime, if there are any questions or proposaly from your part, please do not hesitate to contact me. I am very much looking forward to hearing your comments, thoughts and oponions. Until then thank you very much for struggling through this extremely long email! Best regards, Max

1 1

Re: [Inkscape-devel] Error during building the Inkscape 0.45.
by Areozol 19 Mar '07

19 Mar '07

On 2007-03-16, at 09:20, jiho wrote: >> I think that with gcc 3.3 on Mac you had to use: >> #define isNaN(_a) (__isnan(_a)) >> #define isFinite(_a) (__isfinite(_a)) >> This is judging by the state of isnan.h before I started committing >> Tiger >> changes to it. >> >> Could you please see if it compiles with: >> #define isFinite(_a) (__isfinite(_a)) >> >> If it does I will add the following case to src/isnan.h for SVN and >> for >> the release branch: >> #if defined(__APPLE__) && __GNUC__ == 3 >> # define isFinite(_a) (__isfinite(_a)) // MacOSX definition gcc 3.x > > Hi areozol, > so michael looked at your issue with his usual efficiency. here is the > modified file (attached). you need to put it in src/ in the inkscape > directory (replacing the version that is already there) and then try > to recompile (./osx-build b). > In addition my build script is now in Inkscape code base so will be in > the next version (in 0.45.1). If your compilation issue is solved, we > are close to having Inkscape packages for 10.3.9! > > <isnan.h> > Keep me informed. > I have substituted the appropriate header files with this attachements, and I still get this error: object-edit.cpp: In function `void sp_spiral_outer_set(SPItem*, const NR::Point&, const NR::Point&, unsigned int)': object-edit.cpp:921: error: `isfinite' undeclared in namespace `std' make[2]: *** [object-edit.o] Error 1 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2 I have started the compilation just from scratch, from freshly unpacked archives, just for sure.

2 1

Google SoC
by Ayoze Roberto Bernardo 18 Mar '07

18 Mar '07

Hi everybody, I'm new in this list :) I'm writing in order to know some things about the SoC. I thought about sending some of the inkscape proposed ideas, so I need a mentor, or any dev to talk about them, to improve the ideas. Anybody? :) Cheers, Ayoze.

2 1