- Inkscape Devel - Inkscape Mailman

Developers: Refcounting
by MenTaLguY 30 Apr '06

30 Apr '06

I've finally gotten around to writing some developer documentation about memory management. Since nearly all of the memory leaks I have ever had to fix were related to refcounting in some way, I decided to start with an article about refcounting. The latest version of this article is available in SVN as doc/refcounting.txt. Please let me know if anything is unclear or incomplete. Hwæt! --- INTRODUCTION Many objects in Inkscape have lifecycles which are managed by "reference counting". Each such object has a counter associated with it, which is supposed to reflect the number of outstanding references to it. When that counter falls to zero, the object can be freed. This releases the programmer from worrying about having freed an object while somebody else is still using it (or someone else freeing while you're using it!). Simply "ref" the object (increment the counter) to stake your claim, and then "unref" it (decrement the counter) when you don't need it anymore. The ultimate decision to free the object is made safely behind the scenes. You should "ref" an object whenever you plan to hold onto it while transferring control to another part of the code (which might otherwise end up freeing the object out from under you). REFCOUNTING FUNCTIONS Ref and unref functions are provided to manipulate object counter (and make the final decision to free the object for you), but their names will vary depending on the type of object. Examples include g_object_ref()/g_object_unref() (for most GObject-based types), sp_object_ref()/sp_object_unref() (for SPObject-derived classes), and GC::anchor()/GC::release (for garbage-collector managed objects deriving from GC::Anchored -- more on that later). [ note: for code underneath the Inkscape namespace, you need only write GC::anchor(), but in other code you will need to write Inkscape::GC::anchor(), or import the GC namespace to your .cpp file with: namespace GC = Inkscape::GC; Consider this encouragement to start writing new code in the Inkscape namespace. ] REFCOUNTING POLICY Refcounted objects start with a reference count of one when they are created. This means that you do not need to manually ref one that you've just created. However, you will still be responsible for unreffing it when you're done with it. This means that during the lifetime of an object, there should be N refs and N+1 unrefs on it. If these become unbalanced, then you are likely to experience either transient crashing bugs (the object gets freed while someone is still using it) or memory leaks (the object never gets freed). As a rule, an object should be unreffed by the same class or compilation unit that reffed it. Reffing or unreffing an object on someone else's behalf is usually a recipe for confusion (and defeats the point of refcounting, really). If you pass someone a pointer, they should be the ones responsible for reffing it if they need to hold onto it. Similarly, you shouldn't try to make the decision to unref it for them. When you've unreffed the last ref you know about, you should generally assume that the object is now gone. CIRCULAR REFERENCES One disadvantage of reference counting is that a naive application of it breaks down in the presence of objects that reference each other. Common examples are elements in a doubly-linked list with "prev" and "next" pointers, and nodes in a tree, where a parent keeps a list of children, and children keep a pointer to their parent. If both cases, if there is a "ref" in both directions, neither parent nor children can ever get freed. Because of this, circular data structures should be avoided when possible. When they are necessary, try only "reffing" in one direction (e.g. parent -> child) but not the other (e.g. child -> parent). This can sometimes be trickier than it sounds -- circular references don't have to be direct to cause problems. A simple example of an indirect circular reference would be a circular singly-linked list, where the "last" element in the list points back to the "first". In that case, unidirectional reffing isn't sufficient; you'd have no choice but to delegate ref handling to some object which encapsuled the circular list, reffing and unreffing entries as it added and removed them. ANCHORED OBJECTS The garbage collector cannot know about references from some types of objects: * Gtk/gtkmm widgets * SPObjects * other GObject-derived types * Glib data structures * STL data structures that don't use GC::Alloc To accomodate this, I've provided the GC::Anchored class from which garbage-collector managed classes can be derived if they need to be referenced from such places. As noted earlier, ref and unref functions are GC::anchor() and GC::release(), respectively. For most refcounted objects, a nonzero refcount means "this object is in use", and a zero refcount means "this object is no longer in use, you can free it now". For GC::Anchored objects, refcounting is merely an override to the normal operation of the garbage collector, so the rules are relaxed somewhat: a nonzero refcount merely means "the object is in use in places that the garbage collector can't see", and a zero refcount asserts that "the garbage collector can now see every use that matters". While GC::Anchored objects start with an initial refcount of one like any other refcounted type, in most cases it's safe (and convenient) to GC::release the object immediately upon creating it. This is because the garbage collector can see references to the object from parameters or local variables. Trust the collector. One final note: when code is converted from pure refcounting to garbage collection with GC::Anchored, refs and unrefs between GC::Anchored objects should be removed. Refs are no longer necessary, and when circular references are present, reffing will lead to memory leaks. Normally (unlike pure refcounting) the collector has no problem with freeing circular references, but GC::anchor()ing a reference the collector can already see overrides the collector's judgement.

3 6

[Fwd: Re: [Inkscape-devel] Added CREATE support for gradient, pattern, swatches directories]
by Jon Phillips 30 Apr '06

30 Apr '06

-------- Forwarded Message -------- > From: Jon Phillips <jon@...235...> > To: Joshua Blocher <verbalshadow@...400...> > Cc: inkscape-devel(a)lists.sourceforge.net, create@...626... > <create@...626...> > Subject: Re: [Inkscape-devel] Added CREATE support for gradient, > pattern, swatches directories > Date: Sat, 29 Apr 2006 15:13:39 -0700 > > On Thu, 2006-04-27 at 22:43 -0600, Joshua Blocher wrote: > > I add support for the above directories for Linux and Win32, no Mac > > support yet as the directories are not final yet( according to the > > wiki), but as soon as it is i will add it. > > > > This needs some testing to make sure it works as expected. They are > > setup as fallback directories to the normal Inkscape directories. > > > > Any Problems please let me know. > > > > verbalshadow > > Joshua L. Blocher > > > > Heya Josh, could you add to the CREATE wiki the level of support you > added to Inkscape: > http://create.freedesktop.org/wiki/index.php/Shared_resources_spec_support_… > > Also, please add any other level of support here: > http://create.freedesktop.org/wiki/index.php/Shared_resources > > I also updated the latest spec with the suggestion for where to look on > MAC OS X: > http://create.freedesktop.org/wiki/index.php/Shared_resources#Proposed_chan… > > Jon > -- Jon Phillips San Francisco, CA USA PH 510.499.0894 jon@...235... http://www.rejon.org MSN, AIM, Yahoo Chat: kidproto Jabber Chat: rejon@...896... IRC: rejon@...897... Inkscape (http://inkscape.org) Open Clip Art Library (www.openclipart.org) Creative Commons (www.creativecommons.org) San Francisco Art Institute (www.sfai.edu)

1 0

PATCH: Export dialog improvements
by Jonathon Jongsma 30 Apr '06

30 Apr '06

Hi all, I've attached a patch to improve the look of the Export dialog window (mostly following the guidelines of the GNOME HIG and somewhat inspired by this article: http://www.gnomejournal.org/article/44/three-simple-tips-for-interface-desi…). While I was modifying the layout of the window, I couldn't help but move some of the widgets to gtkmm since I find it so much easier to write and to understand than GTK+. I know Inkscape uses some gtkmm, but I know you also use a lot of plain GTK+. I don't know if the goal / policy is to move more toward gtkmm, so I'm not sure whether you'll appreciate my gtkmm-ification or not :) Of course, this was by no means a complete gtkmm-ification, it's very incremental. In any case, if this sort of patch is welcomed, I may work on a couple other dialogs that could use a bit of improvement as well. A couple of notes about the patch: Most of the GTK+ widgets that I replaced with gtkmm widgets were being allocated with gtk_*_new() but I didn't notice them being freed anywhere. So my code follows the same pattern. I'm not particularly familiar with the code-base of Inkscape, but from what I was able to tell, these kinds of dialogs are basically singletons that are managed by some sort of dialog factory. So while we aren't really keeping track of these widgets that are getting allocated, there's not much danger since they'll only be instantiated once and live for the life of the application (??). Am I understanding this more-or-less correctly? Or am I way off? I should also add that this patch changes some translatable strings by adding some Pango markup to the 'section headers'. Is this the accepted way to do this or is there a better way (i.e. just marking the words between the markup as translatable)? BTW, I'm not subscribed to Inkscape-devel, so please cc me on replies. Jonner

2 2

[elliott.bignell@...1255...: Re: [Clipart] Use of ClipArt in HeadCase]
by Bryce Harrington 29 Apr '06

29 Apr '06

----- Forwarded message from Elliott Bignell <elliott.bignell@...1255...> ----- Date: Sat, 29 Apr 2006 10:54:25 +0200 From: Elliott Bignell <elliott.bignell@...1255...> To: clipart@...626... Subject: Re: [Clipart] Use of ClipArt in HeadCase Couple of mails that should have gone to the group and I sent to Bryce instead. (Duh.) : 1) FYI this is an example of evolved shapes from the HeadCase gallery: http://www.loanedgenius.com/evolution.gif /////////////////////////////// 2) Yes, I couldn't just drop it in fully-formed right now. Firstly, conversion to SVG is a new area for me and I suspect the learning curve is fairly steep. The components in question render to a Windows DC. Secondly, they are COM components and will need some adaptation if they are to run cross-platform. I've set the mutation rates to levels that are really macromutations, which would offend an evolutionary purist, but that's a trivial thing to alter. Also, I deliberately used COM components with a common interface which can be loaded from a list in the registry, in order for third parties to be able to create their own components. I think that feature would be well worth retaining. The evolutionary aspect works on the same principle as Richard Dawkins' "biomorphs" - the component itself generates a set of 25 "descendants" which are displayed in a gallery exactly like all the other ClipArt. When a descendant is selected, the component is notified of the choice of a numbered descendant and adopts its properties then renders itself AND a new set of descendants. Not rocket science, really. What remains to be done is to convert the primitives such as flower and eye to a composite pattern that can be aggregated into whole faces and plants, or whatever. Then it will be really powerful. I'll start working out what is needed in a public release for people to be able to work with the components. Best wishes... Elliott ----- Original Message ----- > > Presently, Inkscape is getting ready to do a release, and it'll be a > month or two before it's back to a state where brand new features can be > added, but I'm guessing you'd want a bit of time to rework things > anyway? > > Bryce > /////////////////////////////// I'll seek the lists out and participate. Now that I know someone appreciates the work, I feel a bit more motivated! The evolutionary side is pretty standard stuff - Dawkins described the pattern in The Selfish Gene, which just had its 30th anniversary. For a more in-depth discussion of evolutionary software, the standard work appears to be "Genetic Programming", by Banzhaf, Nordin, Keller and Francone. The flower shapes I came up with myself and coded by hand. There are a set of numeric parameters that are tweaked up and down within limits in the "Mutate" method, and these specify the number of whorls, the way the spacing of whorls changes and a set of numbers that define how the Bezier curves that form the petals are formed. Basically it draws two symmetrical Bezier curves to form a petal, rotates a little around the axis and moves inward either immediately or after a whorl has been completed and then does it again after applying the appropriate transformations.I was hoping to get quite complex results, but the orchid and ammonite shapes were a pleasant surprise. There's another aspect that would be worth retaining, and that is that the shapes can take Decorators. Everything in HeadCase is basically assembled out of decorators, and these include fills and outlines, so one can apply colour schemes to the pictures produced by the evolutionary process. Also, HeadCase inherits properties down the tree structure and is intended eventualy to support parameterised XML templates, whch has interesting implications for using SVG. Okay, I'll get thinking on how to publish this stuff. Cheers... Elliott ----- Original Message ----- > > Wow, these actually look surprisingly good. You should send this url to > the openclipart and inkscape mailing lists, I think folks would be > impressed by them. > > Btw, how did you come up with the algorithm to do these? Looks > extremely cool. > > Bryce > _______________________________________________ clipart mailing list clipart@...626... http://lists.freedesktop.org/mailman/listinfo/clipart ----- End forwarded message -----

1 0

New translators mailing list for Inkscape
by Bryce Harrington 29 Apr '06

29 Apr '06

Heya all, We've set up a brand new mailing list for Inkscape translators. You can subscribe to it here: https://lists.sourceforge.net/lists/listinfo/inkscape-translator The purpose of this list is to discuss localization and translation issues of Inkscape and its documentation. Here's a sampling of specific topics the translators have proposed discussing on the list: - where are we ? (what are the translation status ? Which languages need some more commitment...) - where do we (translators) usually face some issues ? (specific vocabulary, synchronization with development...) - get some best practices from the ML and then, write them down on the wiki - share experiences on specific blocking points - take a look at the various tasks, on different levels : (core software strings, tutorials, text docs, web pages...) Bryce

3 2

refcount bugs: one down...
by MenTaLguY 29 Apr '06

29 Apr '06

I've just fixed one refcount bug in UI::View::View, so we don't appear to leave dangling refs on SPDocuments anymore. However, someone's still leaving a dangling ref on SPRoot, so we're still leaking quite a bit of memory after a document's closed. I'll stab at it further tomorrow. -mental

1 0

Re: [Inkscape-user] Question regarding draw mode
by bulia byak 29 Apr '06

29 Apr '06

On 4/28/06, Jon A. Cruz <jon@...18...> wrote: > Except there's no visual feedback to tell what's going on... We discussed the way to give such feedback long ago, and I had an idea, but was too lazy to file a RFE, so now I just implemented it so it's not forgotten :) Now each object-creating tool has a standard style widget at the right end of its controls bar, reflecting the style that the next created object will have. This gives you an idea of your "brush" before you start drawing, and it updates when you click on the palette with nothing selected. Both last-set and tool-fixed styles are correctly shown, and the widget is updated when you switch between the two. Now, this thing also needs a text label before it, but I think this could wait until our toolbars are not fixed-width and thus adding stuff does not blow up the window. -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

1 0

Re: [Clipart] Use of ClipArt in HeadCase
by Bryce Harrington 29 Apr '06

29 Apr '06

On Fri, Apr 28, 2006 at 06:10:54PM +0200, Elliott Bignell wrote: > Hello, Nicu, > > Thanks for the clarification. My feeling is that even if nothing is > expected, something is due. I closed the company due to limited sales and am > now only developing HeadCase as a hobby and marginal source of income, but > it is still technically a commercial app. I may in fact go Open Source to > breathe some life into the project, but I invested a great deal of > perspiration in it and effectively admitting that it will never make me a > living is a painful step! Still, I feel that it is unique enough to be worth > saving and Open Source could be the way. Just for your information, it is a > Windows-only app written 98% in C++ using ATL COM with a thin VB wrapper, is > largely undocumented and in a few areas in need of serious refactoring, and > was created with the best of object-oriented intentions but has in a few > places sprawled! It will need tidying up. > > What I could perhaps do is open the evolutionary algorithm and associated > picture-creation objects for re-use, perhaps in Inkscape, and perhaps with > help adapt it to write SVG files for the archive? That'd be a wonderful way to make a contribution. There's a couple different ways it could be integrated into Inkscape: One is to add the code to inkscape itself, which would make it fairly easy to create GUI dialogs and so forth for the user to set options. The other is to implement it as an extension program, in which case it distinct from Inkscape itself, but can be used with Inkscape. In either case, I think this would be quite handy in Inkscape, and I'd encourage you to investigate it further. Presently, Inkscape is getting ready to do a release, and it'll be a month or two before it's back to a state where brand new features can be added, but I'm guessing you'd want a bit of time to rework things anyway? Bryce

1 0

Would a translator mailing list be an appropriate tool ?
by Matiphas 28 Apr '06

28 Apr '06

Hi Inkscape developpers sorry for polluting the dev-ML, especially in the beginning of a release cycle, but i'd like to ask you a question : would a translator mailing list be an appropriate tool ? I guess it may help a little bit, but I'm not sure it would be a decisive tool. My main issue with translations is the fact that the software, strings and docs undergo the most important changes only a few days before the release and this does not give a lot of time to translators. Updating strings is quite ok, as strings evolve parallely to feature development. But docs (such as tutorials and keys) can become a very heavy workload to be dealt in a very few days. I have been very sorry not being able to bring an update of the keys for last release, and i would be very disappointed not being able to offer this for 0.44 release (especially because this release looks like the very best one up to now). I'm very happy to see that - developpers bring lots of nice features and a stable/portable software - translation is taken care of in the release cycle - documentation is taken care of during the same release cycle I'd be even happier the doc translation was also integrated in the release cycle so that users can get a huge package (software + translation of the software + doc + translated doc + extensions + doc for extensions ...). Regards, matiphas

9 16

Added CREATE support for gradient, pattern, swatches directories
by Joshua Blocher 28 Apr '06

28 Apr '06

I add support for the above directories for Linux and Win32, no Mac support yet as the directories are not final yet( according to the wiki), but as soon as it is i will add it. This needs some testing to make sure it works as expected. They are setup as fallback directories to the normal Inkscape directories. Any Problems please let me know. verbalshadow Joshua L. Blocher

2 1