- Inkscape Devel - Inkscape Mailman

tutorials cleanup: help needed
by bulia byak 10 Jan '07

10 Jan '07

We have tutorial-calligraphy.pt_BR.svg in share/tutorials, but no corresponding xml file in the doc-docbook module in SVN. Please _everyone_ who has ever translated the tutorials, upload the docbook xml sources to that module, or send them to someone with SVN commit rights asking him to upload it for you. Without that, we can't for example fix the tutorial-calligraphy.pt_BR.svg file which is badly broken. Also, we have much more tutorial translations in SVG than we have HTML tutorials here: http://inkscape.org/doc/index.php. Please remember that if you submit a new translation, you need to generate not only SVG but also HTML (use the make-html script) to be put up on the web site and linked from that page. (Again, if you don't have commit rights for the web site, ask someone to do it for you.) Volunteers are welcome and encouraged to help us sync this all up for the release. Thanks! -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

1 0

Announce: Scribus 1.3.3.7 Release
by Plinnell 09 Jan '07

09 Jan '07

The Scribus Team is pleased to announce the release of Scribus 1.3.3.7 This release is purely a further bug fix release to the recently released 1.3.3.6. This release adds more translation updates, some bugs were found and fixed, many improvements to the SVG importer. A new splash screen and other artwork were contributed by Kevin Young. One of two major additions to this release is the porting of Scribus to OS/2 by Paul Smedley. This port is lightly tested and there are some known issues noted in the readme for OS/2, along with notes on installation. No major code changes were required and in testing it has been fairly stable. The other is the availability of a new OpenSuse rpm download repository for all Suse versions 9.3+, including SLED 10. To download add: http://software.opensuse.org/download/home:/mrdocs/ to your yast/apt/smart source list for the corresponding Suse version. We hope have this for Mandriva and Fedora 5+ shortly. As we announced previously, there will be no more development or bug fixes to 1.2.x series. We now recommend all distributions package only Scribus 1.3.3.x The Scribus Team would also like to thank - Anduin.net/Øverby Consulting and Cardia AS for their continued hosting of all of the Scribus websites. The Scribus Team would also like to thank the many end users, translators, testers and contributors who helped us with this release. About Scribus Scribus is a cross-platform open source page layout application with a focus on high quality commercial grade PDF and postscript export. Originally developed on Linux, Scribus now runs natively on MacOSX, OS/2, many forms of Unix and Windows 2000/XP. Underneath the modern and user friendly interface, Scribus supports professional publishing features, such as CMYK color, spot colors, separations, ICC color management and versatile PDF creation. Scribus was the first page layout application on the planet to directly support PDF/X-3 output, a rigorous ISO standard. Scribus did so by almost a year. The Scribus Team is made up of a diverse group of developers with significant experience in pre-press and newspaper publishing. Commercial support for Scribus is available for companies and institutions for implementation, training and migration.

1 0

WIN XP: Preferences dialog crashes
by bulia byak 09 Jan '07

09 Jan '07

On XP, the two latest builds: http://inkscape.modevia.com/win32/Inkscape0701071623.7z http://inkscape.modevia.com/win32/Inkscape0701061515.7z crash when trying to open Preferences dialog. Previous builds are OK. Unfortunately the backtrace I could get on Windows is less than helpful, but on one try I got a message "failed to allocate 1024 bytes"; on other tries it just died silently. Anyone else seeing this? -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

3 2

Python extensions breakage when built against Python 2.5 & later
by Muthiah Annamalai 09 Jan '07

09 Jan '07

This is w.r.t the bug report http://sourceforge.net/tracker/index.php?func=detail&aid=1629444&group_id=9… which complains python extensions break on the new python2.5+ SVN (python's) on the inkscape. It leads from a autogen SWIG script. I dont know if this is the right thing, but I am submitting a diff against the files in /src/extensions/script/inkscape_py_wrap.cpp which fixed the problem for me, on the svn trunk. Cheers Muthu

2 1

PNG export of SVG with Blur
by Nicu Buculei 09 Jan '07

09 Jan '07

I have a not very complex SVG (7 layers, under 20 blurred objects) and it proved impossible to export it as PNG at 1600x1200. CPU usage climbs to 100% (Athlon XP 1800+) and stay here, Inkscape becomes unresponsive. I had to kill it after 30 minutes. Exporting a thumbnail at 128x96 work OK, but the full image not. I tried a workaround: separate the file in 7 different files (one for each layer), export them as PNG and put together in GIMP. It is doable for 6 of the 7 layers (but not nice with some of them). The last layer remaining (attached) is the worst: I can't export it as PNG. It contains only two shapes, each of them with a blur filter. Export to PNG at 1600x1200 freeze Inkscape and I killed it after 30 minutes. Is there a way for me to get the PNG? I guess I can leave it over night to export, but this is ridiculous for such a small and simple file. Used the autopackage version 20070101 (after I got the same bad performance with a slightly older autopackage version). -- nicu

5 10

Tamil translation
by Muthiah Annamalai 08 Jan '07

08 Jan '07

Hello there! I am currently trying to translate parts of Inkscape in Tamil (the ta.po), and I am going to post this later this week. Cheers, Muthu

1 0

Bug Hunt Finished!
by Bryce Harrington 08 Jan '07

08 Jan '07

Today Inkscape achieved the 300 point Bug Hunt goal, thanks to work from Bulia Byak, Johan Engelen, Gustav B, and others. The number of critical bugs has dropped to 20. Due to these efforts, we are now ready to move to the next phase of the release cycle. ID Pts Description Assignee ------------------------------------------------------------------------ 1517414 9 Node join crash: delete segment left subpath bbyak 1630193 9 crash when joining an orphan node bbyak 1591411 9 Paths w/ spaces need to have double quotes johanengelen 1502439 9 Crash :Gtk-CRITICAL **: gtk_style_detach gustav_b 1629994 9 SEGV on CTRL + with nothing selected johanengelen 1620741 3 commit 12417 contains C99 strtof bbyak 1623869 6 clip mask with gradient fill makes objects disappear bbyak 1629384 3 The Post Dates are Wrong cwebb 1537876 6 tutorials in Russian not available nodoby 1628254 9 Localized tutorials do not load nobody 1628295 6 Bottom of the web site page cwebb 1505549 9 drag and delete line crash Inkscape 0.45devel bbyak 1628173 6 Gradient tool not working for flowed text bbyak 1580903 9 crash adjusting slider when spinbutton has focus bbyak 1538282 9 crash when switching to selection tool bbyak 1332888 12 Clones: moving two adjoining objects in chain moves bbyak 1593499 6 Tablet cursor drift when dragging bezier gustav_b 1624880 9 download page ishmal 1588594 9 Link broken - Daniel's Thrice Daily builds bryce 1628085 3 PNG file not created/not recognized by WINXP johanengelen 1598192 6 Download Page Instructions Need Updating bryce 1617112 9 Blur is applied after transform, not before bbyak 1488128 9 Aligning (Snapping) to Grid or Guides does not work nobody 1349761 9 Inkboard: inkscape crashes when trying to connect dwyip 1507710 9 crash after leaving an isolated node handle nobody 1605927 9 Path combine (and other ops) lose transform johanengelen 1610103 6 font-size px missing after font family change johanengelen 1590980 9 CRASH: Markers to Stroke Paint bbyak 1625278 9 Crash when trying to load a SVG file bryce 1501553 3 Can't rotate with transform dialog box bryce 1520625 6 Linked clone as clip path does not update nobody 1516933 9 text stroke crash inkscape nobody 1575685 6 web redirects inkscape.org/faq/ kidproto 1565162 3 Gradient stop color not updated in list nobody 1595030 6 Stroke width after Union changed in a strange way nobody 1620066 6 Lost dialog for exporting and color nobody 1598337 9 Crash when trying to delete invisible layer cyreve 1561416 6 EPS export broken theadib 1597605 6 Fontnames containing spaces are incorrectly written theadib 1572450 3 Dialogs pop up off screen nobody 1604391 3 MouseWheel NodeSelection > Undo = Floating Node nobody 1513218 6 Win32: ask if old menus.xml should be kept when... theadib 1612814 3 Guides ans Gridlines vanish on corresp. alpha... nobody 1576258 3 Color artifact in SVN gaussian blur nobody ------------------------------------------------------------------------ Total 303 Done! Bryce

2 2

Please update Release Notes for Inkscape 0.45
by Bryce Harrington 08 Jan '07

08 Jan '07

One of our most important documents for the release is our Release Notes. Please take a moment to review and ensure that they reflect any noteworthy work you've done for this release: http://wiki.inkscape.org/wiki/index.php/ReleaseNotes045 In particular, we need info added to that page for the following: * [color markers to match stroke - acspike] * [axonometric grid] * [new Help commands] * [sculpt profiles - bbyak] * [new toolbar: squeezable, expansion menu, right-click menus - joncruz] * [union of a single object - acspike] Please edit the wiki page to add info to explain the above changes. Also be sure we have adequately accounted for all "Known Issues" (see the end of the release notes for the list.) Thanks, Bryce

1 0

Translator credits in about dialog
by Luca Bruno 07 Jan '07

07 Jan '07

Hi all, actually the translators list in our about dialog is hardcoded in src/ui/dialog/aboutbox.cpp. It looks quite outdated and without any changes for a while, plus I think it is quite difficult to maintain it as-is and update with latest right data and copyroght at each release. So I propose to adopt the mechanism of "translator-credits" as in use and widespread across the gnome framework, because each contributor knows all local contributors better than us. (see http://www.gnome.org/~malcolm/i18n/tips.html ) I've made a patch for it, but as I'm not a programmer I'm asking for comments/corrections. It's attached here. Any opinions on it? May I commit it now (we aren't yet in string freeze) or should I better wait post-0.45? Cheers, Luca -- .''`. ** Debian GNU/Linux ** | Luca Bruno : :' : The Universal O.S. | luca.br(AT)uno.it `. `'` | GPG Key ID: 3BFB9FB3 `- http://www.debian.org | Proud Debian GNU/Linux User

5 5

RFC: SVG filter transformations
by MenTaLguY 07 Jan '07

07 Jan '07

Well, it looks like Inkscape's behavior is non-compliant at the moment, but only under certain combinations of transformations (generally, rotation + non-uniform scaling). bulia and I have been discussing this, and since reworking filters will take a while, we think we have two options for this release: 1. leave the behavior as it is 2. disable filter rendering when the output would be non-compliant #1 has the advantage that the bug would be less noticeable (fewer user complaints until the next release when it's fixed), and it has the disadvantage that the bug would be less noticeable (people getting bitten when the rendering behavior changes in the next release). #2 has the advantage that it actively discourages people from using filters in situations where Inkscape can't currently render them properly. It has the disadvantage that it might actively discourage them from using filters. Which one should we go with? Other suggestions? -mental

5 15

Bug Hunt for Inkscape 0.45
by Bryce Harrington 07 Jan '07

07 Jan '07

Wow, we're within sight of the end - over 80% of the way there. Bulia closed the first "over a year old" bug. I wonder who'll close the bug that puts us over the 300pt mark? 1537876 6 tutorials in Russian not available nodoby 1628254 9 Localized tutorials do not load nobody 1628295 6 Bottom of the web site page cwebb 1505549 9 drag and delete line crash Inkscape 0.45devel bbyak 1628173 6 Gradient tool not working for flowed text bbyak 1580903 9 crash adjusting slider when spinbutton has focus bbyak 1538282 9 crash when switching to selection tool bbyak 1332888 12 Clones: moving two adjoining objects in chain moves bbyak 1593499 6 Tablet cursor drift when dragging bezier gustav_b 1624880 9 download page ishmal 1588594 9 Link broken - Daniel's Thrice Daily builds bryce 1628085 3 PNG file not created/not recognized by WINXP johanengelen 1598192 6 Download Page Instructions Need Updating bryce 1617112 9 Blur is applied after transform, not before bbyak 1488128 9 Aligning (Snapping) to Grid or Guides does not work nobody 1349761 9 Inkboard: inkscape crashes when trying to connect dwyip 1507710 9 crash after leaving an isolated node handle nobody 1605927 9 Path combine (and other ops) lose transform johanengelen 1610103 6 font-size px missing after font family change johanengelen 1590980 9 CRASH: Markers to Stroke Paint bbyak 1625278 9 Crash when trying to load a SVG file bryce 1501553 3 Can't rotate with transform dialog box bryce 1520625 6 Linked clone as clip path does not update nobody 1516933 9 text stroke crash inkscape nobody 1575685 6 web redirects inkscape.org/faq/ kidproto 1565162 3 Gradient stop color not updated in list nobody 1595030 6 Stroke width after Union changed in a strange way nobody 1620066 6 Lost dialog for exporting and color nobody 1598337 9 Crash when trying to delete invisible layer cyreve 1561416 6 EPS export broken theadib 1597605 6 Fontnames containing spaces are incorrectly written theadib 1572450 3 Dialogs pop up off screen nobody 1604391 3 MouseWheel NodeSelection > Undo = Floating Node nobody 1513218 6 Win32: ask if old menus.xml should be kept when... theadib 1612814 3 Guides ans Gridlines vanish on corresp. alpha... nobody 1576258 3 Color artifact in SVN gaussian blur nobody ------------------------------------------------------------------------ Total 246 Bryce

1 0

Extension Question: Multi-line text box for user input?
by Dustin Vaselaar 07 Jan '07

07 Jan '07

Hello, I'm looking at making a custom extension for Inkscape and would like to know if there is a way to have the user input window for the extension contain a multi-line text box for user input? Basically, I would like a text box that can pass a more complex variable, complete with line-ends. I've looked through the extension documentation, however I can't see a way of creating more that a simple one line user input text box. FYI: I wasn't quite sure if this was the correct mailing list for posting this type of question, if I should be asking a different list please let me know. -Dustin

2 1

Dialog decoration inconsistency
by unknown＠example.com 05 Jan '07

05 Jan '07

There is an inconsistency between decorations of dialogs due to old-C-style dialogs and new C++-style ones. Decorations are the minimize, maximize and close buttons. The old dialogs (eg. Fill&Stroke) have all 3, the new dialogs (eg. Transform) only have the close button. What is the desired decoration style? I vote for only the close button, and perhaps a button to 'collapse' it (minimizing on the spot). - Johan

5 5

docs category in bugs tracker
by Alexandre Prokoudine 04 Jan '07

04 Jan '07

Hi, Bugs tracker now has a new category: "Documentation". Please use it, if you see mistakes in text, tutorials not loading etc. Alexandre

1 0

Bug Stats for 0.45 release
by Bryce Harrington 04 Jan '07

04 Jan '07

Bug fixing activity is really picking up lately. Thanks everyone that's helping on this! We've knocked the total number of critical bugs down to 27. 1593499 6 Tablet cursor drift when dragging bezier gustav_b 1624880 9 download page ishmal 1588594 9 Link broken - Daniel's Thrice Daily builds bryce 1628085 3 PNG file not created/not recognized by WINXP johanengelen 1598192 6 Download Page Instructions Need Updating bryce 1617112 9 Blur is applied after transform, not before bbyak 1488128 9 Aligning (Snapping) to Grid or Guides does not work nobody 1349761 9 Inkboard: inkscape crashes when trying to connect dwyip 1507710 9 crash after leaving an isolated node handle nobody 1605927 9 Path combine (and other ops) lose transform johanengelen 1610103 6 font-size px missing after font family change johanengelen 1590980 9 CRASH: Markers to Stroke Paint bbyak 1625278 9 Crash when trying to load a SVG file bryce 1501553 3 Can't rotate with transform dialog box bryce 1520625 6 Linked clone as clip path does not update nobody 1516933 9 text stroke crash inkscape nobody 1575685 6 web redirects inkscape.org/faq/ kidproto 1565162 3 Gradient stop color not updated in list nobody 1595030 6 Stroke width after Union changed in a strange way nobody 1620066 6 Lost dialog for exporting and color nobody 1598337 9 Crash when trying to delete invisible layer cyreve 1561416 6 EPS export broken theadib 1597605 6 Fontnames containing spaces are incorrectly written theadib 1572450 3 Dialogs pop up off screen nobody 1604391 3 MouseWheel NodeSelection > Undo = Floating Node nobody 1513218 6 Win32: ask if old menus.xml should be kept when... theadib 1612814 3 Guides ans Gridlines vanish on corresp. alpha... nobody 1576258 3 Color artifact in SVN gaussian blur nobody ------------------------------------------------------------------------ Total 180 It's nice to see that most of the critical bugs left are older ones (over 6 months). Also, a few of those may already be fixed, but just need confirmation. Please take a few minutes to go through the level-9 bugs for your platform and see if you can get the same results or not. Bryce

1 0

WIN32 locale setting infiltrates Inkscape UI
by momo 04 Jan '07

04 Jan '07

Hi, Not sure if this is closely related to bug [ 1250581 ] "WIN32: Inkscape uses locale settings for determining UI lang" so I post it to the list. On an ENGLISH WinXP system with Russian locale set I get russian buttons in Inkscape dialogs (all the rest of Inkscape's UI is in English as it should). I post a screenshot of the bug attached to this mail. Is this an Inkscape bug, or a GTK one? Thanks! Molumen

1 0

Bug confirmation needed
by Bryce Harrington 04 Jan '07

04 Jan '07

Hi all, I've added two new categories 'NEEDS CONFIRMATION' and 'NEEDS CONFIRMATION ON SVN HEAD' to our bug tracker, and have begun marking a few bugs that need this. If you have a recent inkscape installation and are on either Win32 or OSX, you can help greatly by seeing if you can recreate these bugs. Please go to this page: http://sourceforge.net/tracker/?group_id=93438&atid=604306 Search on Category = NEEDS CONFIRMATION ON SVN HEAD, and set the Group to the platform you're on. Also look in Groups 'Any' or 'All Platforms'. Try to recreate the problem, and then please write in the comment section if you could or could not reproduce the problem. Some of these are very, very old bugs, so just having people confirm that the bugs are no longer present would be extremely helpful in enabling us to focus on bugs that really need our attention. Thanks! Bryce

2 1

Tutorials
by Daniel Pope 04 Jan '07

04 Jan '07

I notice in general that the tutorials in the SVN package do not correspond at all well to the current Inkscape feature set. We don't really have a tutorial for the more complicated features (cloning, clipping, masking, blurring) and tutorial-tips.svg explicitly mentions that we don't support blur yet. Attached is an SVG highlighting some of the effects that can be achieved using Inkscape 0.45. Perhaps something along these lines would be a useful thing to bundle? Dan

6 9

Bug Hunt for 0.45
by Bryce Harrington 03 Jan '07

03 Jan '07

1488128 9 Aligning (Snapping) to Grid or Guides does not work nobody 1349761 9 Inkboard: inkscape crashes when trying to connect dwyip 1507710 9 crash after leaving an isolated node handle nobody 1605927 9 Path combine (and other ops) lose transform johanengelen 1610103 6 font-size px missing after font family change johanengelen 1590980 9 CRASH: Markers to Stroke Paint bbyak 1625278 9 Crash when trying to load a SVG file bryce 1501553 3 Can't rotate with transform dialog box bryce 1520625 6 Linked clone as clip path does not update nobody 1516933 9 text stroke crash inkscape nobody 1575685 6 web redirects inkscape.org/faq/ kidproto 1565162 3 Gradient stop color not updated in list nobody 1595030 6 Stroke width after Union changed in a strange way nobody 1620066 6 Lost dialog for exporting and color nobody 1598337 9 Crash when trying to delete invisible layer cyreve 1561416 6 EPS export broken theadib 1597605 6 Fontnames containing spaces are incorrectly written theadib 1572450 3 Dialogs pop up off screen nobody 1604391 3 MouseWheel NodeSelection > Undo = Floating Node nobody 1513218 6 Win32: ask if old menus.xml should be kept when... theadib 1612814 3 Guides ans Gridlines vanish on corresp. alpha... nobody 1576258 3 Color artifact in SVN gaussian blur nobody ------------------------------------------------------------------------ Total 138 (Goal: 300) We're almost halfway there! I notice that due in part to our bug activity, Inkscape has gone up to rank #20 on SourceForge's most active list. Still, there are 30 level-9 critical bugs still open. I don't see any that are absolute release killers, but the fewer of those left in inkscape when we release, the better. Bryce

1 0

Re: [Inkscape-devel] Inkscape - hyperlinking xlink
by Jon Phillips 03 Jan '07

03 Jan '07

On Tue, 2006-12-19 at 17:31 +1100, Phil Rhoades wrote: > Jon, > > > On Mon, 2006-12-18 at 17:14 -0800, Jon Phillips wrote: > > On Tue, 2006-12-19 at 03:14 +1100, Phil Rhoades wrote: > > > Jon, > > > > > > > > > > Date: 2006-12-15 12:02 > > > > Sender: kidproto > > > > Logged In: YES > > > > user_id=914868 > > > > Originator: YES > > > > > > > Phil, please contact me and/or on the inkscape-devel mailing list > > > > about this. I'm jon@...235... (jon phillips). > > > > > > > Great! > > > > > > > > > Have you got any idea what this might take in time/cost? > > > > > > Regards, > > > > > > Phil. > > > > > > Could you give me a bit of background? I can help put you in touch with > > inkscape developers who could put the work with the $$$... > > > Your comment: "it would be killer to be able to click on an area and it > zooms/pans you to another part of the document...or to another > document." pretty much sums up what I want. The actual application I > have in mind is something resembling a "mind map" or "concept map" for > organising my various projects, which are all related to Life Extension > of some sort (ie Biomedical/Scientific Research and Commercial). I have > looked around at available software but I particularly want to use > Inkscape and Ruby on Rails so it looks like I need to do it myself . . I > could generate the svg from the DB but it would not be as nice, flexible > and easy as drawing the relationships in Inkscape and linking objects to > the data pages. Yes, I agree. This hyperlinking is what you want. It is useful for a lot of things...would you like to work on this? > > > linking is already in the codebase, but there are some specifics which > > need to be enabled for what you want... > > > I tried manually putting in the xlinks in the svg but it causes > namespace problems - in any case, I would immediately lose anything that > was added manually if I wanted to use Inkscape to re-edit the svg. It > seems like such a useful addition to Inkscape that I thought it might be > worth paying for . . depending on the cost. I'm cc'ing the inkscape-devel list to see if others have played with this. Yes, someone in Inkscape would be interested in coding and/or working for pay ;) > The first (easier) step would be just to link to an external page which > would do what I want immediately but your extended idea of zooming and > panning within the svg would be an attractive second stage. > > Is this what you need? Yes, sounds great! Jon > Regards, > > Phil. -- 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...

2 1

Inkscape 0.45 About Screen Contest
by Joshua A. Andler 02 Jan '07

02 Jan '07

The contest begins today and the deadline is December 31st. As with the past few contests we've had, it is being done on DeviantART. You can find out more about it at: http://inkscapers.deviantart.com/journal/11071733/ Good Luck! -Josh

6 10

Inkscape and transient windows (on top dialogs)
by unknown＠example.com 02 Jan '07

02 Jan '07

Short introduction for the maillinglist: There is a discussion ongoing about transient dialogs (dialogs on top) in win32 builds of Inkscape. They do not work for current builds, but can be made to work with (afaik) only one drawback: when Inkscape's mainwindow is minimized while having a dialog on top, the mainwindow cannot simply be restored by clicking the taskbar button as this will give focus to the dialog that was opened. The dialog is not minimized when Inkscape is minimized which causes the problem. The mainwindow is restored by rightclicking the taskbar button and clicking on "Restore". The ongoing discussion is about how to solve this in a way that respects current standards. This mail is about a temporary solution that would greatly greatly greatly (I cannot stress enough) enhance Inkscape's useability. Bulia Byak wrote: > Johan wrote: > > The rightclick-'restore' procedure is not really annoying > > if known, but is very confusing for users and therefore I > > think a complete fix is needed before I can commit the > > #ifndef WIN32 removal. > > Yes, definitely. At the moment, the "dialogs on top" option does not work on win32 anyway, so we might as well remove it from win32 builds. However, I have a better idea: (Note this is for win32 builds only!!!) We could concatenate the options "Dialogs are hidden in taskbar" and "Dialogs on top" into one option "Dialogs on top (experimental)" (checkbox). If unchecked, Inkscape will use the old options "dialogs NOT hidden in taskbar" and "Dialogs on top: none", if checked Inkscape will use "dialogs hidden in taskbar" and "dialogs on top:normal". The "(experimental)" is added so that users know that it is likely to work dissatisfactory, and an explanatory text should be added on how to restore the main window when Inkscape was minimized (i.e. the 'rightclick-restore' procedure). Regards, Johan

1 0

First win32 build of the year
by Bob Jamison 02 Jan '07

02 Jan '07

Nothing important. It just looks good to have "070101". http://inkscape.modevia.com/win32/Inkscape0701011510.7z bob

4 3

Bug Hunt for 0.45
by Bryce Harrington 02 Jan '07

02 Jan '07

About a third of the way there now: 1590980 9 CRASH: Markers to Stroke Paint bbyak 1625278 9 Crash when trying to load a SVG file bryce 1501553 3 Can't rotate with transform dialog box bryce 1520625 6 Linked clone as clip path does not update nobody 1516933 9 text stroke crash inkscape nobody 1575685 6 web redirects inkscape.org/faq/ kidproto 1565162 3 Gradient stop color not updated in list nobody 1595030 6 Stroke width after Union changed in a strange way nobody 1620066 6 Lost dialog for exporting and color nobody 1598337 9 Crash when trying to delete invisible layer cyreve 1561416 6 EPS export broken theadib 1597605 6 Fontnames containing spaces are incorrectly written theadib 1572450 3 Dialogs pop up off screen nobody 1604391 3 MouseWheel NodeSelection > Undo = Floating Node nobody 1513218 6 Win32: ask if old menus.xml should be kept when... theadib 1612814 3 Guides ans Gridlines vanish on corresp. alpha... nobody 1576258 3 Color artifact in SVN gaussian blur nobody ------------------------------------------------------------------------ Total 96 Bryce

1 0

Bug Hunt for 0.45
by Bryce Harrington 01 Jan '07

01 Jan '07

More bugs fixed! About a quarter of the way there. Please remember to set the assignee to yourself (or whomever fixed it) when you close bugs, so we know who to give credit to. 1625278 9 Crash when trying to load a SVG file bryce 1501553 3 Can't rotate with transform dialog box bryce 1520625 6 Linked clone as clip path does not update nobody 1516933 9 text stroke crash inkscape nobody 1575685 6 web redirects inkscape.org/faq/ kidproto 1565162 3 Gradient stop color not updated in list nobody 1595030 6 Stroke width after Union changed in a strange way nobody 1620066 6 Lost dialog for exporting and color nobody 1598337 9 Crash when trying to delete invisible layer cyreve 1561416 6 EPS export broken theadib 1597605 6 Fontnames containing spaces are incorrectly written theadib 1572450 3 Dialogs pop up off screen nobody 1604391 3 MouseWheel NodeSelection > Undo = Floating Node nobody 1513218 6 Win32: ask if old menus.xml should be kept when... theadib 1612814 3 Guides ans Gridlines vanish on corresp. alpha... nobody 1576258 3 Color artifact in SVN gaussian blur nobody ------------------------------------------------------------------------ Total 87 It would be great if we could get 0.45 released before the end of January, so if you have some time, please help investigate open bugs! Bryce

2 1

Re: [Inkscape-devel] Pattern along path crashes
by Michèle Garoche 31 Dec '06

31 Dec '06

Le 30 déc. 2006 à 09:54, Bryce Harrington a écrit : > On Sat, Dec 30, 2006 at 09:37:29AM +0100, Mich?le Garoche wrote: >> Sorry if this is not the right mailing list to ask, just thought that >> there is more chance to touch Mac OS X users here. >> >> I've tried the new pattern along path extension from svn on Mac OS >> X 10.4. >> >> It crashes reproducibly. >> >> Did anyone else try it on Mac successfully? Or is the crash >> confirmed? >> >> Note: I've just copied the extension directory onto the old one from >> 0.44 installation via fink, it worked for all other extensions, >> but it >> could make a difference for this one. > > Please doublecheck that there is a bug in our bug tracker about this. > Attach any example files where the issue occurs, along with steps to > reproduce it. I did not find any bug in the tracker, so I've created one: <http://sourceforge.net/tracker/index.php? func=detail&aid=1625230&group_id=93438&atid=604306> The examples are taken from the extension author's web site (url in the tracker). But it is the same for any group of images. Cheers, Michèle <http://micmacfr.homeunix.org>

1 0

Pattern along path crashes
by Michèle Garoche 30 Dec '06

30 Dec '06

Sorry if this is not the right mailing list to ask, just thought that there is more chance to touch Mac OS X users here. I've tried the new pattern along path extension from svn on Mac OS X 10.4. It crashes reproducibly. Did anyone else try it on Mac successfully? Or is the crash confirmed? Note: I've just copied the extension directory onto the old one from 0.44 installation via fink, it worked for all other extensions, but it could make a difference for this one.

2 1

saving in windows friendly format
by tj 30 Dec '06

30 Dec '06

How can I save an inkscape svg file into a format that Windows can use? I tried saving it as an .eps, but Windows Paint Shop Pro barfed on it. Is there a linux native way to create a jpeg, bitmap, etc? tj

2 1

our filters may be incompatible with SVG
by bulia byak 29 Dec '06

29 Dec '06

I would like to discuss a potentially very serious compatibility issue with our filters implementation. The question is, If an object has a transform, do the filters apply to it _before_ or _after_ transfomation? To illustrate this, do the following: - draw a circle - blur it - squeeze it vertically (note the blur is also squeezed, i.e. x-blur is greater than y-blur) - rotate the object At this stage in Inkscape, the blur remains horizontal, so it is no longer parallel to the object itself, so the appearance of the object changes very visibly after the rotation. If however you now save your file and view it in Batik, you will see that it rotates the blur too, so the object looks exactly the same as before rotation. Note that Inkscape differs from Batik only after you add rotation. For non-uniform scaling, Inkscape emulates Batik's behavior by scaling the x and y deviations of the blur filter when applied to a scaled object. However, such emulation is not possible for rotation or shear, hence the discrepancy. Now, I could not find a definitive indication in SVG 1.1 standard as to which of the two behaviors is correct. (I may have missed it, of course.) Some arguments can be found in favor of both behaviors: Batik: - It is more visually consistent because rotation/shear does not alter the appearance of a filtered object in a confusing way - So far, we always considered Batik an authoritative renderer (though it may be in error too, of course) Inkscape: - The idea of transforming an already filtered object (i.e., in essence, a bitmap!) goes absolutely against our architecture, in which rasterization and filtering are always the last step after all things that can be done in vector (such as transforms) are done. If we try to fix this to match Batik, we will need to implement a separate bitmap transformer (!) which will have to be run for filtered objects _only_ (otherwise we'll end up with pixelated renditions of any scaled-up object). This will be a total mess from the programmer's viewpoint if you ask me. - Our behavior preserves the same appearance of a path regardless of whether its transform is embedded or given in a transform= attribute (but this is currently only true for rotations, not for non-uniform scales). This is a principle which we're kinda trying to uphold, although it's not always possible. I _think_ that even if our behavior is an error, it's not a reason enough to stop 0.45, because this only affects squeezed _and_ rotated objects. However it's still very serious because it's a visible rendering discrepancy. So I would like to discuss this issue and the approach we should eventually pursue. -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

4 4

please test the proposed perspective extension
by bulia byak 27 Dec '06

27 Dec '06

We have a new Perspective extension in our patch tracker: http://sourceforge.net/tracker/index.php?func=detail&aid=1600298&group_id=9… It's very useful potentially, but for some reason, I can't make it work properly, although it reportedly works for the author and another user. I would appreciate if more people could try it out and report their results. Note that to use it, you will need to install Numeric Python: http://numpy.scipy.org (in addition to regular Python). To install the extension, place the inx and py file from perspective_v0.3.tgz into your extensions directory. Please share your experience - thanks! -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

1 0

Bug Hunt for 0.45
by Bryce Harrington 27 Dec '06

27 Dec '06

Unfortunately the numbers haven't changed since the last time I posted them, but here's where we are: 1576258 3 Color artifact in SVN gaussian blur nobody 1612814 3 Guides ans Gridlines vanish on corresp. alpha... nobody 1513218 6 Win32: ask if old menus.xml should be kept when... theadib 1604391 3 MouseWheel NodeSelection > Undo = Floating Node nobody 1572450 3 Dialogs pop up off screen nobody ------------------------------------------------------------------------ Total 18 I'm a little concerned because we have a *lot* of high criticality bugs that lead to crashes. How can we get more of these closed? Bryce

4 6

Problems with configure file
by Pierre-Luc Auclair 27 Dec '06

27 Dec '06

Hi, I'm having problems with the configure file. I tried using --with-xft and --without-xft but it doesn't change a thing. Here's the problem: ############################################################# ############################################################# ## ## Error ## ########### ./configure: line 9675: syntax error near unexpected token `XFT,' ./configure: line 9675: ` PKG_CHECK_MODULES(XFT, xft, xft_ok=yes, xft_ok=no)' ############################################################# ############################################################# ## ## configure file ## ############ # Check whether --with-xft or --without-xft was given. if test "${with_xft+set}" = set; then withval="$with_xft" with_xft=$withval else with_xft=auto fi; if test "x$with_xft" != "xno" ; then PKG_CHECK_MODULES(XFT, xft, xft_ok=yes, xft_ok=no) if test "x$xft_ok" != "xyes"; then PKG_CHECK_MODULES(XFT, fontconfig, xft_ok=yes, xft_ok=no) if test "x$xft_ok" != "xyes"; then if test "${ac_cv_header_X11_Xft_Xft_h+set}" = set; then echo "$as_me:$LINENO: checking for X11/Xft/Xft.h" >&5 echo $ECHO_N "checking for X11/Xft/Xft.h... $ECHO_C" >&6 if test "${ac_cv_header_X11_Xft_Xft_h+set}" = set; then echo $ECHO_N "(cached) $ECHO_C" >&6 fi echo "$as_me:$LINENO: result: $ac_cv_header_X11_Xft_Xft_h" >&5 echo "${ECHO_T}$ac_cv_header_X11_Xft_Xft_h" >&6 else # Is the header compilable? ############################################################# ############################################################# Thanks, Pierre-Luc

1 1

Update on www.inkscape.org performance issues
by Bryce Harrington 24 Dec '06

24 Dec '06

I've been continuing to investigate the problems we've been having with the website. The root issue is that we get so many hits to our website that it is overwhelming apache. Apache has been hitting its MAX_CLIENTS limit. Unfortunately, simply increasing MAX_CLIENTS isn't doable because we only have 1G of memory available. We've also been maxing out our ip_conntrack_count limit periodically (which is really bad, because it disables all the other sites sharing the box with us at OSUOSL). I've hooked up squid as a transparent caching proxy, which hopefully should help a good deal, as it'll cache our php pages and avoid reexecuting those scripts. Should make the site work faster too. Whether it'll actually solve the problem remains to be seen though, so please let me know if you notice the site down again. Bryce

2 2

Blur Radius Scales
by Daniel Pope 23 Dec '06

23 Dec '06

Hi, last weekend I took a moment to download an SVN build and posted a few observations to the tracker. One particular wishlist item which I suggested was tweaking the blur scale from linear to quadratic, which in my experimentation seemed to give me a much more even control over how much blur to apply, because I was finding the current scale, and certainly the slider, useless for the subtle effects. I've found as I've used Inkscape more over the course of the week that generally the upper 70% of the scale is useless for my purposes. I've uploaded a comparison of the current and my proposed blur scales[1] to the tracker and would like to request comments. Merry Xmas/whatever else you are celebrating, Dan [1] http://sf.net/tracker/download.php?group_id=93438&atid=604309&file_id=20833…

1 0

Re: [Inkscape-devel] Would it be possible to add support for opacity greater than 100%
by Kevin Baragona 22 Dec '06

22 Dec '06

It would be a straight multiplier to the alpha of the pixels, just like how (i think) it is now except you could also make something 'more' opaque than normal. This would be useful for say blurred objects that became too transparent. I am pretty sure this would invalidate the css according to w3c but isn't that why there is inkscape-svg and plain-svg? When saving to plain-svg a file with something like 250% opacity, it could be layered like 2 100% objects and a 50% object which I believe would then be ok. Thanks, Kevin

3 2

Layers Opacity %?
by John Bintz 22 Dec '06

22 Dec '06

I noticed that, while all of the other 0..1 ranges have been changed to 0..100%, the Layers Dialog opacity has not. I have the code to make this change almost ready, but I wanted to ask two things before I committed anything: 1. Since it's close to a release, should I hold off on this change? 2. The Fill & Stroke dialog for Opacity and Blur places the percent signs outside of the SpinButtons, in the labels. Should the Layer Opacity be the same (the label now becoming "Opacity %:") or should it be like the Zoom control (% in the SpinButton itself)? Thanks, John

3 2

[Fwd: SCALE to host Birds of a Feather Sessions!]
by jtaber 21 Dec '06

21 Dec '06

I don't think many of us were able to submit a paper/full presentation but how about an Inkscape BOF? It would be cool to have a few of us give 15 min demos of how we're using some Inkscape features in our line of work - I'll volunteer to show off how I'm using it with GIS (nodes, layers, potrace). Anyone else interested ? John -------- Original Message -------- Subject: SCALE to host Birds of a Feather Sessions! Date: Wed, 20 Dec 2006 11:23:00 -0800 From: Gareth J. Greenaway <g@...1153...> Organization: Southern California Linux Expo To: Gareth J. Greenaway <g@...1153...> “BoF”s are “Birds of a Feather” sessions. They're intended for those who would like to get together with others with interest in a similar subject. Linux Expo of SoCal would like to host your BoF at the Fifth Annual So Cal Linux Expo February 10th & 11th, 2007. The BoF sessions will be 45 minutes to 1 hour long. If you'd like to organize a BoF at SCALE, the Expo will publicize it and host it for you. Please send the following info for consideration to g@...1153...: - Title of the BoF - Brief description of BoF topic - Group hosting the BoF - Contact person for BoF (discussion leader/moderator,etc) Thank you and we look forward to seeing you at SCALE 5x Gareth -- Gareth J. Greenaway | g@...1153... Voice - 877-831-2569 x130 Southern California Linux Expo http://www.socallinuxexpo.org

3 3

Re: [Inkscape-devel] [Fwd: SCALE to host Birds of a Feather Sessions!]
by jtaber 21 Dec '06

21 Dec '06

Gareth J. Greenaway wrote: > I think this would make a good BoF, showing Inkscape and SVG in real life > situation so people can see an example of open source in the real world. > > Would someone mind emailing me the details that I had requested for in the > original email and I'll add this idea to the list of potential BoFs for > consideration? Thanks. > > Gareth > If someone else wants to take the lead here, fine. Otherwise, I'll set this up with Gareth. Josh, if you're coming, one of your tablet drawing demos would be great. Maybe someone could show off the gradient, filters stuff. btw, Happy Holidays to all

1 0

Re: [Inkscape-devel] [Fwd: SCALE to host Birds of a Feather Sessions!]
by Ted Gould 21 Dec '06

21 Dec '06

On Thu, 2006-12-21 at 09:06 -0800, Gareth J. Greenaway wrote: > I think this would make a good BoF, showing Inkscape and SVG in real life > situation so people can see an example of open source in the real world. > > Would someone mind emailing me the details that I had requested for in the > original email and I'll add this idea to the list of potential BoFs for > consideration? Thanks. John, Do you want to write this up? I'm putting in a vote for you being in charge ;) --Ted

2 1

release until January 12th ? was [Fwd: Inkscape auf c't-Software-DVD]
by Adib Taraben 20 Dec '06

20 Dec '06

Hello, the german computer mag "ct" wants to bundle inkscape again on one of their add CD. So the question is: Is there a chance to get the release out until January 12th ? I think the answer is NO but want to make shure and wait for your answer. Thx, Adib. -------- Original-Nachricht -------- Betreff: Inkscape auf c't-Software-DVD Datum: Tue, 19 Dec 2006 16:55:52 +0100 Von: Peter König <pek@...883...> An: Colin Marquardt <colin@...384...>, Adib Taraben <taraben.a@...50...> Hallo, ich weiß, die Lizenz erlaubt es auch ohne Nachfrage, und wir haben auch schon mal die generelle Freigabe bekommen, aber ich wollte doch gerne Bescheid geben: 8-) Die Heft-DVD der c't-Ausgabe 4/2007 soll wieder einmal eine aktuelle Version von Inkscape enthalten. "Einsendeschluss" der Software ist der 12.01.2007. Ist bis dahin ein neuer Build zu erwarten? (Leider erreiche ich gerade inkscape.org nicht, ist die Seite down?) Danke und mit vielen Grüßen, Peter König c't - Magazin für Computertechnik http://www.heise.de/ct/ Heise Zeitschriften Verlag GmbH & Co.KG phone: 0511-5352-763 Helstorfer Str. 7 fax: 0511-5352-417 D-30625 Hannover, Germany eMail: pek@...883...

4 3

Missing cairomm when cross-compiling
by Daniel Díaz 19 Dec '06

19 Dec '06

Hello, all! I think I downloaded the latest bundle for cross-compiling to win32, but I'm getting these errors. Since it's been a while since I last made a build, I'm sure I'm missing something but can't find exactly what. Here's an excerpt of the errors when compiling: -----8<----------8<----------8<----------8<----------8<----- make[1]: Entering directory `/home/inkscape/source/inkscape.win32/src' i686-pc-mingw32-g++ -Wall -g -O3 -mms-bitfields -DVERSION=\"0.45+devel\" -DXP_WIN -D_INTL_REDIRECT_INLINE -DHAVE_CONFIG_H -DWITH_INKBOARD -DHAVE_SSL -I. -I.. -I../.. -I../cxxtest -DGLIBMM_DLL -I/target/include/glibmm-2.4 -I/target/lib/glibmm-2.4/include -I/target/include/gtkmm-2.4 -I/target/lib/gtkmm-2.4/include -I/target/include/gdkmm-2.4 -I/target/lib/gdkmm-2.4/include -I/target/include/pangomm-1.4 -I/target/include/atkmm-1.6 -I/target/include/cairo -I/target/include/sigc++-2.0 -I/target/lib/sigc++-2.0/include -I/target/include/gtk-2.0 -I/target/lib/gtk-2.0/include -I/target/include/atk-1.0 -I/target/include/pango-1.0 -I/target/include/glib-2.0 -I/target/lib/glib-2.0/include -I/target/include -Wno-comment -I/target/perl/lib/CORE -I/target/python/include -I/target/include/libxml2 -I/target/include/freetype2 -I. -Idom -Idom/svg -Idom/js -Idom/js/fdlibm -Idom/io -Idom/odf -Idom/util -Ilibcroco -Idisplay -Iwidgets -Iextension -Iextension/internal -Iextension/internal/libwpg -Iextension/implementation -Iextension/script -Iremoveoverlap -Igraphlayout -Ilibnrtype -Isvg -Ilibcola -Ixml -Idebug -Iapplication -Ihelper -Idialogs -Ilibvpsc -Ilibvpsc/pairingheap -Itraits -Ipedro -Ilivarot -Ilibavoid -Itrace -Itrace/potrace -Iui -Iui/widget -Iui/dialog -Iui/view -Ijabber_whiteboard -Ijabber_whiteboard/dialog -Ialgorithms -Ilibnr -Iutil -Iinkjar -Iio -c -o application/editor.o application/editor.cpp In file included from /target/include/glibmm-2.4/glibmm/ustring.h:32, from ./document.h:29, from application/editor.cpp:31: /target/lib/glibmm-2.4/include/glibmmconfig.h:21:1: warning: "GLIBMM_DLL" redefined <command line>:7:1: warning: this is the location of the previous definition In file included from /target/include/gtkmm-2.4/gtkmm/widget.h:28, from /target/include/gtkmm-2.4/gtkmm/container.h:29, from /target/include/gtkmm-2.4/gtkmm/box.h:29, from ./ui/view/edit-widget.h:20, from application/editor.cpp:39: /target/include/pangomm-1.4/pangomm/context.h:42:29: cairomm/context.h: No such file or directory /target/include/pangomm-1.4/pangomm/context.h:43:33: cairomm/fontoptions.h: No such file or directory In file included from /target/include/gtkmm-2.4/gtkmm/widget.h:28, from /target/include/gtkmm-2.4/gtkmm/container.h:29, from /target/include/gtkmm-2.4/gtkmm/box.h:29, from ./ui/view/edit-widget.h:20, from application/editor.cpp:39: /target/include/pangomm-1.4/pangomm/context.h:333: error: `Cairo' has not been declared /target/include/pangomm-1.4/pangomm/context.h:333: error: expected `,' or `...' before '<' token /target/include/pangomm-1.4/pangomm/context.h:333: error: ISO C++ forbids declaration of `RefPtr' with no type /target/include/pangomm-1.4/pangomm/context.h:336: error: `Cairo' has not been declared /target/include/pangomm-1.4/pangomm/context.h:336: error: expected `,' or `...' before '&' token /target/include/pangomm-1.4/pangomm/context.h:336: error: ISO C++ forbids declaration of `FontOptions' with no type /target/include/pangomm-1.4/pangomm/context.h:339: error: `Cairo' has not been declared /target/include/pangomm-1.4/pangomm/context.h:339: error: ISO C++ forbids declaration of `FontOptions' with no type /target/include/pangomm-1.4/pangomm/context.h:339: error: expected `;' before "get_font_options" ----->8---------->8---------->8---------->8---------->8----- Where can I get this glib-/cairomm stuff? Thanks and greetings! Daniel Díaz yosoy@...31...

1 0

Feature request
by Vladimir Savic 18 Dec '06

18 Dec '06

Because of some, to me unknown reason, sourceforge wont let me log in. So, I'm going to post this to the list directly and ask someone to fill regular report in my name. If the idea is worth enough, of course. What I really need is one more information in secondary toolbar of selection tool. It's rotation angle. I know it's available through Transform dialog, but it's incapable of noticing any changes happening on using second click and rotating object(on-canvas editing). I see Transform dialog as a perfect tool for a technical drawings but not for casual illustrative jobs. Right now, if I accidentally rotate an object (line for example) it's practically impossible to bring it back into original position. Of course, I'm aware of Undo, but that's not really a solution. Thanks for your attention, Vlada

2 2

Require GTK 2.8+ for Inkscape 0.45
by Aaron Spike 18 Dec '06

18 Dec '06

If you think back to our last release a decision was made to hold back on inflating our GTK requirements until the 0.45 release. 0.45 time has come and we will now officially be requiring GTK 2.8+. Is this correct? If I am not mistaken GTK 2.8 requires Cairo. This is good for us. I've been researching creating runtime checks for cairo's availability. If we agree on GTK 2.8, I won't have to worry about that. We may, however, want to consider runtime checks for the PDF backend. I've been looking into relaytool as a method for doing this on linux. http://autopackage.org/apbuild-relaytool.php I would appreciate assistance investigating this. Aaron Spike

10 27

Pango and Inkscape.app on mac os x
by jiho 17 Dec '06

17 Dec '06

Hello, This question is more oriented toward Michael Wybrow or any other developper who helped with the packaging of Inkscape inside an .app bundle on Mac OS X. I am trying to use DarwinPorts rather than Fink to provide Inkscape with its dependencies and everything seems to work except for Pango. What I did to get osx-app.sh working with DarwinPorts was simply to: - change SW=/sw to SW=/opt/local (DarwinPorts install root) - change the version numbers of GTK and Pango (2.4.0 and 1.4.0) to those used in DarwinPorts (2.10.0 and 1.5.0) I checked the app bundle which seems to be OK: all dependecies were correctly pulled out DarwinPorts tree and copied into the bundle. The pango.modules, pangorc and pangox.aliases files are similar to those found in a working bundle produced with Fink, as are the files in .inkscape-etc and all only reflect the version change in Pango. To sum up, everything looks to be similar to what works with Fink. However, when starting Inkscape from inside the app bundle and I get this message: (inkscape-bin:18144): Pango-WARNING **: No builtin or dynamically loaded modules were found. Pango will not work correctly. This probably means there was an error in the creation of: '/opt/local/etc/pango/pango.modules' You should create this file by running pango-querymodules. [... many other Pango-WARNINGs about bad fonts...] and in Inkscape all UI fonts are replaced by rectangles. Before running Inkscape binary a shell script sets PANGO_RC_FILE to "$HOME/.inkscape-etc/pangorc" (and .inkscape-etc/pangorc points to the libs inside the bundle). My understanding is that it makes Pango libraries/modules independent from where they were compiled, and it works with Fink-provided Pango librairies. The reference to /opt/ local/ here makes me think that there is still a link between Pango libs inside the bundle and the place where they were compiled. Was anything else than setting PANGO_RC_FILE necessary to get Fink Pango libs to work from inside Inskcape.app? Any additional knowledge on the bundling process would probably help because I only know what gathered from the code and some Google searches. Thanks in advance. NB: the file '/opt/local/etc/pango/pango.modules' exists and is only root writable, hence the error. But this would be true for its Fink equivalent also and Inkscape.app should not try to write anything in this location anyway. JiHO --- http://jo.irisson.free.fr/

1 0

Book update
by Tavmjong Bah 16 Dec '06

16 Dec '06

Hello, I have uploaded a new version of my book "A Guide to Inkscape" to my web site. Although there is only a small amount of new material, there is one change of note: the html files now use the names of the chapters and/or sections. This will prevent breakage of links to the pages (e.g. from the mailing lists) when new material is added and the chapter or section numbers change. Recent links in the mailing lists should not be broken as I've included copies of the html files at the old addresses. There is also the beginnings of a Dutch translation by Tom Nijns. So far only the introduction is translated. I've set up a Lulu account where PDF versions of the book can be downloaded (both in A4 and Letter size). The downloads will be available for free for the next few days. After that I plan on charging for the PDF version of the book (probably $7.50), the web version will, of course, remain free. Please let me know of any problems with the downloads. Book's website: http://tavmjong.free.fr/INKSCAPE/ Lulu website: http://stores.lulu.com/tavmjong Tav

1 0

Hardware Technical Questions
by unknown＠example.com 16 Dec '06

16 Dec '06

Hi, all. I've been a regular Inkscape user since shortly after the fork from sodipodi. I love the program, transitioning from Xara on Windows to Inkscape on Linux has been good, if not always easy. So I'm always excited to get the newest releases of Inkscape with all kinds of great usable features. (Blur is a lifesaver for me - I've been using the nightly builds for a lot of work recently because of that!) I'm interested in helping the project, but I'm limited in what I can do. I'm not a coder, and I'd love to learn but don't have the time at the moment. I'm actually a graphic designer, which is why I use Inkscape heavily. Anyway, one of the ways that I would like to help, at least peripherally, is to write an article about building a Linux graphic workstation that's optimized for 2-d graphic apps, primarily Inkscape and the Gimp. I have been trying to understand the basics of how hardware rendering of graphic data really works in order to get a handle on this myself, but I haven't found anything readily available to tell me this info. That's why I want to write it. I'm wondering if any of you (singular or multiple people) would be interested in helping me understand some basics so that I can write a very useful article. Please contact me if you are interested in helping. Thanks.

5 4

Time to unpause 0.45 release?
by Bryce Harrington 16 Dec '06

16 Dec '06

Hi all, A few weeks back we decided to put the release on pause to give time for additional work on the PDF extension. Could someone outline the status of this work, and if it is still necessary for this to be blocking the release? If not, would anyone have an issue if we resumed the release process at this point? I find myself with some unexpected extra freetime this month that I could devote to assisting with the release. Bryce

4 11

Difference between inkview and inkscape -s?
by Tim Dwyer 15 Dec '06

15 Dec '06

Hi All, In the past I have found inkview to be a very handy tool for quickly looking at svg, but lately it seems horribly broken (see bug 1592467). Then I discovered that "inkscape -s" seems to be more or less the same thing... but it doesn't crash. What's the difference between inkview and "inkscape -s"? Do we really need the maintenance and bloat of two separate binaries? Sorry if I've missed any on-going discussion on this. Thanks heaps Tim

3 5

[Fwd: Re: [Inkscape-user] Inkscape 0.45 About Screen Contest]
by Jon Phillips 15 Dec '06

15 Dec '06

-------- Forwarded Message -------- > From: Jon Phillips <jon@...235...> > To: Inkscape User Community <inkscape-user(a)lists.sourceforge.net> > Cc: Inkscape Devel List <inkscape-devel(a)lists.sourceforge.net> > Subject: Re: [Inkscape-user] [Inkscape-devel] Inkscape 0.45 About > Screen Contest > Date: Fri, 15 Dec 2006 11:24:14 -0800 > > On Fri, 2006-12-15 at 12:14 -0700, Joshua A. Andler wrote: > > Jon Phillips wrote: > > > On Fri, 2006-12-15 at 11:47 -0700, Joshua A. Andler wrote: > > >> The contest begins today and the deadline is December 31st. As with the > > >> past few contests we've had, it is being done on DeviantART. You can > > >> find out more about it at: > > >> > > >> http://inkscapers.deviantart.com/journal/11071733/ > > >> > > >> Good Luck! > > >> > > >> -Josh > > > > > > What, I thought you were going to email me the announcement first before > > > launching this to coordinate? > > > > Hmmm... I had sent an email yesterday morning and since I had not heard > > back in 24 hours I went ahead and posted it to give people enough time > > before the end of the month. It doesn't look like it got returned > > either. :-/ > > > > Are there changes you'd like? > > > > -Josh > > I never received it...nor was I pinged...quoting the last email I > received from you: > > "If your end is ready to go, I'll shoot you a direct email with the > announcement prior to posting it so we can make sure everything looks > good. It never hurts to have another set of eyes look it over." > > I guess its done now and out...total bummer :( > > Jon > > > > > > > ------------------------------------------------------------------------- > > Take Surveys. Earn Cash. Influence the Future of IT > > Join SourceForge.net's Techsay panel and you'll get the chance to share your > > opinions on IT & business topics through brief surveys - and earn cash > > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > > _______________________________________________ > > Inkscape-user mailing list > > Inkscape-user(a)lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/inkscape-user -- 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...

1 0

[Fwd: Re: [Inkscape-user] Inkscape 0.45 About Screen Contest]
by Jon Phillips 15 Dec '06

15 Dec '06

-------- Forwarded Message -------- > From: Jon Phillips <jon@...235...> > To: Aaron Spike <aaron@...749...> > Cc: Inkscape User Community <inkscape-user(a)lists.sourceforge.net>, > Inkscape Devel List <inkscape-devel(a)lists.sourceforge.net> > Subject: Re: [Inkscape-user] [Inkscape-devel] Inkscape 0.45 About > Screen Contest > Date: Fri, 15 Dec 2006 11:32:30 -0800 > > On Fri, 2006-12-15 at 13:29 -0600, Aaron Spike wrote: > > Joshua A. Andler wrote: > > > Are there changes you'd like? > > > > Accept entries only if they are accompanied by the original SVG! We've > > lost way too much time in the past waiting on a usable format. (Wish I > > could remember if I had any other complaints. I should write this stuff > > down.) > > > > Aaron Spike > > ccHost totally can handle all these things and has it all built-in...I > think though since its already announced publicly, its too hard to > change... > > 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...

1 0

If a feature falls in a wood and no one hears it ... ?
by Alan Horkan 15 Dec '06

15 Dec '06

How many people knew Inkscape had spell checking? (Rhetorical.) https://sourceforge.net/tracker/index.php?func=detail&aid=1605259&group_id=… How can such a valuable feature be omitted and can it be enabled by default in time for the next release? Normally I avoid crossposting but I feel I am doing Inkscape users a service by informing or reminding them about this feature. Are there any other fantastic features like this hidden away and awaiting rediscovery? Sincerely Alan Horkan http://advogato.org/person/AlanHorkan/ http://alanhorkan.livejournal.com/

1 0

Bughunting
by Engelen, J.B.C. (Johan) 14 Dec '06

14 Dec '06

Bryce wrote: > 1576258 3 Color artifact in SVN gaussian blur > nobody > 1612814 3 Guides ans Gridlines vanish on corresp. alpha... > nobody > 1513218 6 Win32: ask if old menus.xml should be kept > when... theadib > -------------------------------------------------------------- > ---------- > Total 12 I've fixed another bug that (as far as I looked) is not in the bugtracker. Does that count? ;) When ungrouping, the transformation of the group was applied to the individual items, but not correctly written back to XML. Noticeable when undo+redo'ing. It's fixed now To see what I mean try: LaTeX render 'ab' ungroup ungroup undo redo Incidentally it meant fixing a "FIXME" in the code, of which there are a lot :S Happy hunting! Johan

2 1

Re: [Inkscape-devel] Bughunting
by John Bintz 14 Dec '06

14 Dec '06

Engelen, J.B.C. (Johan) said: > I've fixed another bug that (as far as I looked) is not in the bugtracker. > Does that count? ;) Along the same lines, I've gotten one in the bugtracker that is ready to be closed, but I don't have permissions to do so: [ 1604391 ] MouseWheel NodeSelection > Undo = Floating Node http://sourceforge.net/tracker/index.php?func=detail&aid=1604391&group_id=9… Thanks, John

2 1

autosave on crash
by Henning Lorenz 14 Dec '06

14 Dec '06

During an Inkscape crash a popup-window was displayed announcing that backups of the open files have been created automatically. This is a very nice feature, however I can't find the files on the system (I tried with locate after an update of its database). Is there a known problem with the function or the filename? Cheers, Hennig

2 1

closing bug [ 1513218 ] Win32: ask if old menus.xml should be kept when installing
by Adib Taraben 14 Dec '06

14 Dec '06

Just inform you all that I commited a patch that solves [ 1513218 ] Win32: ask if old menus.xml should be kept when installing priority 7 bug. I found the solution while installing latest vlc ;-) Adib.

2 1

Re: [Inkscape-devel] Inkscape 0.45 - About Screen and moving forward
by Joshua A. Andler 13 Dec '06

13 Dec '06

Jon Phillips wrote: > On Wed, 2006-12-13 at 12:00 -0700, Joshua A. Andler wrote: >> I know you said if I don't hear from you, which I didn't, so I just >> wanted to verify that I should I move forward with it. >> >> -Josh > > No, lets launch it through OCAl with the deviantart community. > > What will it entail... So you got CCHost up and running? On the DA end, basically all I need to do is announce it via both inkscape accounts and that would be all. If your end is ready to go, I'll shoot you a direct email with the announcement prior to posting it so we can make sure everything looks good. It never hurts to have another set of eyes look it over. -Josh

2 1

Re: [Inkscape-devel] Inkscape 0.45 - About Screen and moving forward
by Joshua A. Andler 13 Dec '06

13 Dec '06

I know you said if I don't hear from you, which I didn't, so I just wanted to verify that I should I move forward with it. -Josh Jon Phillips wrote: > Heya, give me till the end of the day to get ccHost up and running so we > can do through OCAL. If you don't hear from me, then we can move it onto > deviant art...cool? > > Jon > > > On Tue, 2006-12-12 at 15:04 -0700, Joshua A. Andler wrote: >> Hi All, >> >> I'm going to be setting up the About Screen contest on deviantART again. >> >> When should the deadline be? >> >> Any special requests for required new techniques/features other than >> using Blur in it? >> >> One more thing... before I can start this contest I need to make sure we >> have reasonably up-to-date packages available for the big 3 platforms. >> >> Win32 is 2 days old which is definitely okay. >> Linux Autopackage is 10 days old which isn't too terribly bad. >> OSX is a problem though as it's 2 months old as of today. >> >> Anyone know who to contact for the OSX stuff? >> >> The sooner I get responses, the sooner we can get the ball rolling on >> this. :) >> >> -Josh >> >> >> ------------------------------------------------------------------------- >> Take Surveys. Earn Cash. Influence the Future of IT >> Join SourceForge.net's Techsay panel and you'll get the chance to share your >> opinions on IT & business topics through brief surveys - and earn cash >> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV >> _______________________________________________ >> Inkscape-devel mailing list >> Inkscape-devel(a)lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/inkscape-devel

1 0

will there be new test versions for 0.45/win32?
by momo 13 Dec '06

13 Dec '06

Hello, Since the bug hunting has begun for the 0.45 version, I have a question about the download locations listed in http://www.inkscape.org/download/ - The versions at http://inkscape.modevia.com/win32-snap/?M=D are desperately old (25 oct 2005) - http://cortijodelrio.net/~inkscape/win32/ is dead for more than a month already - http://inkscape.modevia.com/win32/?M=D the most fresh version is from the 10 dec 2005 Will there be everyday compiled fresh win32 test versions to play with in the close future? Thanks! Molumen

4 3

svg fonts
by Alexandre Prokoudine 13 Dec '06

13 Dec '06

Hi, Just to make sure... http://wiki.inkscape.org/wiki/index.php/SVG_Tiny_Compliance states that massifr started working on support for SVG fonts a while ago. Is there any update on this? Alexandre

1 0

Inkscape 0.45 - About Screen and moving forward
by Joshua A. Andler 13 Dec '06

13 Dec '06

Hi All, I'm going to be setting up the About Screen contest on deviantART again. When should the deadline be? Any special requests for required new techniques/features other than using Blur in it? One more thing... before I can start this contest I need to make sure we have reasonably up-to-date packages available for the big 3 platforms. Win32 is 2 days old which is definitely okay. Linux Autopackage is 10 days old which isn't too terribly bad. OSX is a problem though as it's 2 months old as of today. Anyone know who to contact for the OSX stuff? The sooner I get responses, the sooner we can get the ball rolling on this. :) -Josh

3 2

Re: [Inkscape-devel] Inkscape 0.45 - About Screen and moving forward
by Joshua A. Andler 12 Dec '06

12 Dec '06

Sure man, no problem... :) Jon Phillips wrote: > Heya, give me till the end of the day to get ccHost up and running so we > can do through OCAL. If you don't hear from me, then we can move it onto > deviant art...cool? > > Jon > > > On Tue, 2006-12-12 at 15:04 -0700, Joshua A. Andler wrote: >> Hi All, >> >> I'm going to be setting up the About Screen contest on deviantART again. >> >> When should the deadline be? >> >> Any special requests for required new techniques/features other than >> using Blur in it? >> >> One more thing... before I can start this contest I need to make sure we >> have reasonably up-to-date packages available for the big 3 platforms. >> >> Win32 is 2 days old which is definitely okay. >> Linux Autopackage is 10 days old which isn't too terribly bad. >> OSX is a problem though as it's 2 months old as of today. >> >> Anyone know who to contact for the OSX stuff? >> >> The sooner I get responses, the sooner we can get the ball rolling on >> this. :) >> >> -Josh >> >> >> ------------------------------------------------------------------------- >> Take Surveys. Earn Cash. Influence the Future of IT >> Join SourceForge.net's Techsay panel and you'll get the chance to share your >> opinions on IT & business topics through brief surveys - and earn cash >> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV >> _______________________________________________ >> Inkscape-devel mailing list >> Inkscape-devel(a)lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/inkscape-devel

1 0

Re: [Inkscape-devel] [Inkscape-user] Inkscape site slow?
by Bryce Harrington 12 Dec '06

12 Dec '06

On Sun, Dec 10, 2006 at 03:40:23PM +0100, Jasper van de Gronde wrote: > Is it just me or do other people also experience rediculous load times > for the Inkscape site? (homepage and Wiki main page, I didn't dare go > any further) With rediculous I mean something like a minute, or even > several minutes. I noticed there were problems yesterday and restarted apache. Looking at it now, I see an error: [Sun Dec 10 05:59:12 2006] [error] server reached MaxClients setting, consider raising the MaxClients setting I went ahead and raised the MaxClients settings in the config by a factor of about 4 and restarted. Seems to be ok now. Please keep an eye on the site and report here if it again becomes a problem; getting things properly tuned could take a few iterations. Bryce

3 4

Faster gaussian blur
by Jasper van de Gronde 11 Dec '06

11 Dec '06

I have modified Inkscape's blur filter to operate in linear time using a method that gives a very good approximation of a gaussian blur using a small IIR filter (falling back on a heavily modified version of the old code for very small kernel sizes, as the IIR filter doesn't handle those well). This means that the time needed to compute a blur doesn't depend on the size of the kernel (except for the fact that larger borders are needed for larger kernels). I'm posting it here for feedback, if people think this is a good idea I'll clean up some last things and submit it to the patch tracker. Also, I wouldn't mind some testing, I only tested with a simple image with a blurred rectangle and gallardo.svgz (the latter has quite a nice collection of blurs). When rendering gallardo.svgz using the new code cuts the total time spent in FilterGaussian::render roughly in half (on my computer it is 30-40%, so 50% should be quite attainable), compared to the SVN implementation using average(!) quality, while being virtually indistinguishable from best(!) quality rendering. I have seen some differences on certain test drawings in my test application, but for a rendering of gallardo.svgz the maximum difference was 1 and didn't occur often (and could be due to slightly different rounding, numerical errors, etc.). In my patch I completely eliminated the old resampling stuff, but I could also imagine keeping it to let people speed up gaussian blur even more. Some things I know of that could be done to make it faster/smaller/etc.: - Currently it uses either the IIR filter or the FIR filter for both the X and Y directions, which could in theory result in bad performance on blurs where the deviation is high in only one direction. There is no technical reason preventing this decision to be made separately for the X and Y directions though (I just haven't come around to doing that). - Code for the X and Y directions could be merged, as they are so similar that only the bounds and some strides differ (might have a negative impact on performance though, would have to be timed). - A fixed-point format could be used for the intermediate values in the IIR filter (I already experimented with this in a test application, it works quite well and is slightly faster, probably more so on older hardware). - Currently temporary storage is used to store the intermediate values for an entire row/column for the IIR filter. This isn't strictly necessary (but probably provides slightly better quality and without it the values would need to be clipped when storing them, at least in my tests). BTW, I also made a small change to how the normal gaussian kernel is calculated which results in slightly higher quality blurs but can make direct comparison with the old code harder (the step size wasn't very accurate for small kernel sizes, resulting in a blur that could be quite visibly different from the correct one). So when evaluating the performance of the IIR filter it should be compared to the output of the FIR filter in the new code. Index: src/display/nr-filter-gaussian.cpp =================================================================== --- src/display/nr-filter-gaussian.cpp (revision 13436) +++ src/display/nr-filter-gaussian.cpp (working copy) @@ -6,13 +6,16 @@ * Authors: * Niko Kiirala <niko@...1267...> * bulia byak + * Jasper van de Gronde <th.v.d.gronde@...528...> * * Copyright (C) 2006 authors * * Released under GNU GPL, read the file 'COPYING' for more information */ +#include <algorithm> #include <cmath> +#include <complex> #include <glib.h> using std::isnormal; @@ -24,6 +27,37 @@ #include "libnr/nr-matrix.h" #include "prefs-utils.h" +// IIR filtering method based on: +// L.J. van Vliet, I.T. Young, and P.W. Verbeek, Recursive Gaussian Derivative Filters, +// in: A.K. Jain, S. Venkatesh, B.C. Lovell (eds.), +// ICPR'98, Proc. 14th Int. Conference on Pattern Recognition (Brisbane, Aug. 16-20), +// IEEE Computer Society Press, Los Alamitos, 1998, 509-514. +// +// Using the backwards-pass initialization procedure from: +// Boundary Conditions for Young - van Vliet Recursive Filtering +// Bill Triggs, Michael Sdika +// IEEE Transactions on Signal Processing, Volume 54, Number 5 - may 2006 + +// Number of IIR filter coefficients used. Currently only 3 is supported. +// "Recursive Gaussian Derivative Filters" says this is enough though (and +// some testing indeed shows that the quality doesn't improve much if larger +// filters are used). +const size_t N = 3; + +template<typename InIt, typename OutIt, typename Size> +void copy_n(InIt beg_in, Size N, OutIt beg_out) { + std::copy(beg_in, beg_in+N, beg_out); +} + +typedef double Value; // Type used for filter coefficients (can be 10.21 fixed point) +template<typename T> inline T sqr(T const v) { return v*v; } + +template<typename Tt, typename Ts> +Tt round_cast(Ts const& v) { + static Ts const rndoffset(.5); + return static_cast<Tt>(v+rndoffset); +} + namespace NR { FilterGaussian::FilterGaussian() @@ -45,211 +79,39 @@ { int length_x = _effect_area_scr_x(trans); int length_y = _effect_area_scr_y(trans); - return _max(length_x, length_y) * 2 + 1; + return std::max(length_x, length_y) + 1; } void FilterGaussian::_make_kernel(double *kernel, double deviation, double expansion) { - double length = deviation * 3.0; - int scr_len = (int)std::floor(length * expansion); - if(scr_len < 1) scr_len = 1; - double d_sq = deviation * deviation * 2; - double step = length / scr_len; + int const scr_len = (int)std::ceil(deviation * 3.0 * expansion); + double const d_sq = sqr(deviation * expansion) * 2; + // Compute kernel and sum of coefficients + // Note that actually only half the kernel is computed, as it is symmetric double sum = 0; - for ( int i = 0; i < scr_len * 2 + 1 ; i++ ) { - double i_sq = (step * i - length) * (step * i - length); - sum += (std::exp(-i_sq / d_sq) / std::sqrt(M_PI * d_sq)); + for ( int i = 0; i <= scr_len ; i++ ) { + kernel[i] = std::exp(-sqr(i) / d_sq); + sum += kernel[i]; } + sum = 2*sum - kernel[0]; // the sum of the complete kernel is twice as large (minus the center element to avoid counting it twice) - for ( int i = 0; i < scr_len * 2 + 1 ; i++ ) { - double i_sq = (step * i - length) * (step * i - length); - kernel[i] = (std::exp(-i_sq / d_sq) / std::sqrt(M_PI * d_sq)) / sum; + // Normalize kernel + for ( int i = 0; i <= scr_len ; i++ ) { + kernel[i] /= sum; } } int FilterGaussian::_effect_area_scr_x(Matrix const &trans) { - int ret = (int)std::floor(_deviation_x * 3.0 * trans.expansionX()); - if(ret < 1) ret = 1; - return ret; + return (int)std::ceil(_deviation_x * 3.0 * trans.expansionX()); } int FilterGaussian::_effect_area_scr_y(Matrix const &trans) { - int ret = (int)std::floor(_deviation_y * 3.0 * trans.expansionY()); - if(ret < 1) ret = 1; - return ret; + return (int)std::ceil(_deviation_y * 3.0 * trans.expansionY()); } -int FilterGaussian::_effect_subsample_step(int scr_len_x, int quality) -{ - switch (quality) { - case BLUR_QUALITY_WORST: - if (scr_len_x < 8) { - return 1; - } else if (scr_len_x < 32) { - return 4; - } else if (scr_len_x < 64) { - return 8; - } else if (scr_len_x < 128) { - return 32; - } else if (scr_len_x < 256) { - return 128; - } else if (scr_len_x < 512) { - return 512; - } else if (scr_len_x < 1024) { - return 4096; - } else { - return 65536; - } - break; - case BLUR_QUALITY_WORSE: - if (scr_len_x < 16) { - return 1; - } else if (scr_len_x < 64) { - return 4; - } else if (scr_len_x < 120) { - return 8; - } else if (scr_len_x < 200) { - return 32; - } else if (scr_len_x < 400) { - return 64; - } else if (scr_len_x < 800) { - return 256; - } else if (scr_len_x < 1200) { - return 1024; - } else { - return 65536; - } - break; - case BLUR_QUALITY_BETTER: - if (scr_len_x < 32) { - return 1; - } else if (scr_len_x < 160) { - return 4; - } else if (scr_len_x < 320) { - return 8; - } else if (scr_len_x < 640) { - return 32; - } else if (scr_len_x < 1280) { - return 64; - } else if (scr_len_x < 2560) { - return 256; - } else { - return 1024; - } - break; - case BLUR_QUALITY_BEST: - return 1; // no subsampling at all - break; - case BLUR_QUALITY_NORMAL: - default: - if (scr_len_x < 16) { - return 1; - } else if (scr_len_x < 80) { - return 4; - } else if (scr_len_x < 160) { - return 8; - } else if (scr_len_x < 320) { - return 32; - } else if (scr_len_x < 640) { - return 64; - } else if (scr_len_x < 1280) { - return 256; - } else if (scr_len_x < 2560) { - return 1024; - } else { - return 65536; - } - break; - } -} - -int FilterGaussian::_effect_subsample_step_log2(int scr_len_x, int quality) -{ - switch (quality) { - case BLUR_QUALITY_WORST: - if (scr_len_x < 8) { - return 0; - } else if (scr_len_x < 32) { - return 2; - } else if (scr_len_x < 64) { - return 3; - } else if (scr_len_x < 128) { - return 5; - } else if (scr_len_x < 256) { - return 7; - } else if (scr_len_x < 512) { - return 9; - } else if (scr_len_x < 1024) { - return 12; - } else { - return 16; - } - break; - case BLUR_QUALITY_WORSE: - if (scr_len_x < 16) { - return 0; - } else if (scr_len_x < 64) { - return 2; - } else if (scr_len_x < 120) { - return 3; - } else if (scr_len_x < 200) { - return 5; - } else if (scr_len_x < 400) { - return 6; - } else if (scr_len_x < 800) { - return 8; - } else if (scr_len_x < 1200) { - return 10; - } else { - return 16; - } - break; - case BLUR_QUALITY_BETTER: - if (scr_len_x < 32) { - return 0; - } else if (scr_len_x < 160) { - return 2; - } else if (scr_len_x < 320) { - return 3; - } else if (scr_len_x < 640) { - return 5; - } else if (scr_len_x < 1280) { - return 6; - } else if (scr_len_x < 2560) { - return 8; - } else { - return 10; - } - break; - case BLUR_QUALITY_BEST: - return 0; // no subsampling at all - break; - case BLUR_QUALITY_NORMAL: - default: - if (scr_len_x < 16) { - return 0; - } else if (scr_len_x < 80) { - return 2; - } else if (scr_len_x < 160) { - return 3; - } else if (scr_len_x < 320) { - return 5; - } else if (scr_len_x < 640) { - return 6; - } else if (scr_len_x < 1280) { - return 8; - } else if (scr_len_x < 2560) { - return 10; - } else { - return 16; - } - break; - } -} - /** * Sanity check function for indexing pixblocks. * Catches reading and writing outside the pixblock area. @@ -264,6 +126,155 @@ } } +static void calcFilter(double const sigma, double b[N]) { + assert(N==3); + std::complex<double> const d1_org(1.40098, 1.00236); + double const d3_org = 1.85132; + double qbeg = 1; // Don't go lower than sigma==2 (we'd probably want a normal convolution in that case anyway) + double qend = 2*sigma; + double const sigmasqr = sqr(sigma); + double s; + do { // Binary search for right q (a linear interpolation scheme is suggested, but this should work fine as well) + double const q = (qbeg+qend)/2; + // Compute scaled filter coefficients + std::complex<double> const d1 = pow(d1_org, 1.0/q); + double const d3 = pow(d3_org, 1.0/q); + double const absd1sqr = std::norm(d1); + double const re2d1 = 2*d1.real(); + double const bscale = 1.0/(absd1sqr*d3); + b[2] = -bscale; + b[1] = bscale*(d3+re2d1); + b[0] = -bscale*(absd1sqr+d3*re2d1); + // Compute actual sigma^2 + double const ssqr = 2*(2*(d1/sqr(d1-1.)).real()+d3/sqr(d3-1.)); + if ( ssqr < sigmasqr ) { + qbeg = q; + } else { + qend = q; + } + s = sqrt(ssqr); + } while(qend-qbeg>(sigma/(1<<30))); + printf("real sigma: %g, q: %g\n", s, qend); +} + +static void calcTriggsSdikaM(double const b[N], double M[N*N]) { + assert(N==3); + double a1=b[0], a2=b[1], a3=b[2]; + double const Mscale = 1.0/((1+a1-a2+a3)*(1-a1-a2-a3)*(1+a2+(a1-a3)*a3)); + M[0] = 1-a2-a1*a3-sqr(a3); + M[1] = (a1+a3)*(a2+a1*a3); + M[2] = a3*(a1+a2*a3); + M[3] = a1+a2*a3; + M[4] = (1-a2)*(a2+a1*a3); + M[5] = a3*(1-a2-a1*a3-sqr(a3)); + M[6] = a1*(a1+a3)+a2*(1-a2); + M[7] = a1*(a2-sqr(a3))+a3*(1+a2*(a2-1)-sqr(a3)); + M[8] = a3*(a1+a2*a3); + for(size_t i=0; i<9; i++) M[i] *= Mscale; +} + +template<unsigned int SIZE> +static void calcTriggsSdikaInitialization(double const M[N*N], Value const uold[N][SIZE], Value const uplus[SIZE], Value const vplus[SIZE], Value const alpha, Value vold[N][SIZE]) { + for(size_t c=0; c<SIZE; c++) { + double uminp[N]; + for(size_t i=0; i<N; i++) uminp[i] = uold[i][c] - uplus[c]; + for(size_t i=0; i<N; i++) { + double voldf = 0; + for(size_t j=0; j<N; j++) { + voldf += uminp[j]*M[i*N+j]; + } + // Properly takes care of the scaling coefficient alpha and vplus (which is already appropriately scaled) + // This was arrived at by starting from a version of the blur filter that ignored the scaling coefficient + // (and scaled the final output by alpha^2) and then gradually reintroducing the scaling coefficient. + voldf *= alpha; + vold[i][c] = voldf; + vold[i][c] += vplus[c]; + } + } +} + +// TODO: Make filterX and filterY the same function by doing something with strides +template<typename PT, unsigned int PC> +void filterX(NRPixBlock *const in, NRPixBlock *const out, Value const b[N], double const M[N*N], Value *const tmpdata) { + size_t const w=in->area.x1-in->area.x0, h=in->area.y1-in->area.y0; + // Blur image (x) + for ( size_t y = 0 ; y < h ; y++ ) { + // corresponding line in the source and output buffer + PT const *const srcimg = reinterpret_cast<PT const *>(NR_PIXBLOCK_PX(in) + y*in->rs); + PT *const dstimg = reinterpret_cast<PT *>(NR_PIXBLOCK_PX(out) + y*out->rs); + // Border constants + Value imin[PC]; copy_n(srcimg + (0)*PC, PC, imin); + Value iplus[PC]; copy_n(srcimg + (w-1)*PC, PC, iplus); + // Forward pass + Value u[N+1][PC]; + for(size_t i=0; i<N; i++) copy_n(imin, PC, u[i]); + for ( size_t x = 0 ; x < w ; x++ ) { + for(size_t i=N; i>0; i--) copy_n(u[i-1], PC, u[i]); + copy_n(srcimg+x*PC, PC, u[0]); + for(size_t c=0; c<PC; c++) { + u[0][c] *= b[0]; + for(size_t i=1; i<N+1; i++) u[0][c] += u[i][c]*b[i]; + } + copy_n(u[0], PC, tmpdata+x*PC); + } + // Backward pass + Value v[N+1][PC]; + calcTriggsSdikaInitialization(M, u, iplus, iplus, b[0], v); + for(size_t c=0; c<PC; c++) dstimg[(w-1)*PC+c] = round_cast<PT>(v[0][c]); + size_t x=w-1; + while(x-->0) { + for(size_t i=N; i>0; i--) copy_n(v[i-1], PC, v[i]); + copy_n(tmpdata+x*PC, PC, v[0]); + for(size_t c=0; c<PC; c++) { + v[0][c] *= b[0]; + for(size_t i=1; i<N+1; i++) v[0][c] += v[i][c]*b[i]; + } + for(size_t c=0; c<PC; c++) dstimg[x*PC+c] = round_cast<PT>(v[0][c]); + } + } +} + +template<typename PT, unsigned int PC> +void filterY(NRPixBlock *const in, NRPixBlock *const out, Value const b[N], double const M[N*N], Value *const tmpdata) { + size_t const w=in->area.x1-in->area.x0, h=in->area.y1-in->area.y0; + size_t const in_rs=in->rs/sizeof(PT), out_rs=out->rs/sizeof(PT); + // Blur image (y) + for ( size_t x = 0 ; x < w ; x++ ) { + // corresponding line in the source and output buffer + PT const *const srcimg = reinterpret_cast<PT const *>(NR_PIXBLOCK_PX(in) + x*PC); + PT *const dstimg = reinterpret_cast<PT *>(NR_PIXBLOCK_PX(out) + x*PC); + // Border constants + Value imin[PC]; copy_n(srcimg + (0)*in_rs, PC, imin); + Value iplus[PC]; copy_n(srcimg + (h-1)*in_rs, PC, iplus); + // Forward pass + Value u[N+1][PC]; + for(size_t i=0; i<N; i++) copy_n(imin, PC, u[i]); + for ( size_t y = 0 ; y < h ; y++ ) { + for(size_t i=N; i>0; i--) copy_n(u[i-1], PC, u[i]); + copy_n(srcimg+y*in_rs, PC, u[0]); + for(size_t c=0; c<PC; c++) { + u[0][c] *= b[0]; + for(size_t i=1; i<N+1; i++) u[0][c] += u[i][c]*b[i]; + } + copy_n(u[0], PC, tmpdata+y*PC); + } + // Backward pass + Value v[N+1][PC]; + calcTriggsSdikaInitialization(M, u, iplus, iplus, b[0], v); + for(size_t c=0; c<PC; c++) dstimg[(h-1)*out_rs+c] = round_cast<PT>(v[0][c]); + size_t y=h-1; + while(y-->0) { + for(size_t i=N; i>0; i--) copy_n(v[i-1], PC, v[i]); + copy_n(tmpdata+y*PC, PC, v[0]); + for(size_t c=0; c<PC; c++) { + v[0][c] *= b[0]; + for(size_t i=1; i<N+1; i++) v[0][c] += v[i][c]*b[i]; + } + for(size_t c=0; c<PC; c++) dstimg[y*out_rs+c] = round_cast<PT>(v[0][c]); + } + } +} + int FilterGaussian::render(FilterSlot &slot, Matrix const &trans) { /* in holds the input pixblock */ @@ -288,262 +299,313 @@ return 0; } - /* Blur radius in screen units (pixels) */ - int scr_len_x = _effect_area_scr_x(trans); - int scr_len_y = _effect_area_scr_y(trans); + // TODO: Find the best threshold (should not be lower than 1) + // TODO: Split into separate x and y parts. + //if (_deviation_x * trans.expansionX() >= 2 && _deviation_y * trans.expansionY() >= 2 && std::abs(_deviation_x * trans.expansionX()-2.37749)>0.00001) { + // This condition was chosen based on the old code that was used to decide on how much subsampling to use. + // (Using the condition below the old code would never have used subsampling.) + if (3*std::min(_deviation_x * trans.expansionX(), _deviation_y * trans.expansionY()) >= 8) { + // new buffer for the final output, same resolution as the in buffer + NRPixBlock *out = new NRPixBlock; + nr_pixblock_setup_fast(out, in->mode, in->area.x0, in->area.y0, + in->area.x1, in->area.y1, true); + if (out->data.px == NULL) { + // alas, we've accomplished a lot, but ran out of memory - so abort + return 0; + } + size_t numChannels; + switch(in->mode) { + case NR_PIXBLOCK_MODE_A8: ///< Grayscale + numChannels = 1; + break; + case NR_PIXBLOCK_MODE_R8G8B8: ///< 8 bit RGB + numChannels = 3; + break; + case NR_PIXBLOCK_MODE_R8G8B8A8N: ///< Normal 8 bit RGBA + numChannels = 4; + break; + case NR_PIXBLOCK_MODE_R8G8B8A8P: ///< Premultiplied 8 bit RGBA + numChannels = 4; + break; + default: + assert(false); + } + Value *const tmpdata = new Value[std::max(in->area.x1-in->area.x0,in->area.y1-in->area.y0)*numChannels]; + if (tmpdata == NULL) { + nr_pixblock_release(out); + delete out; + return 0; + } - // subsampling step; it depends on the radius, but somewhat nonlinearly, to make high zooms - // workable; is adjustable by quality in -2..2; 0 is the default; 2 is the best quality with no - // subsampling - int quality = prefs_get_int_attribute ("options.blurquality", "value", 0); - int stepx = _effect_subsample_step(scr_len_x, quality); - int stepx_l2 = _effect_subsample_step_log2(scr_len_x, quality); - int stepy = _effect_subsample_step(scr_len_y, quality); - int stepy_l2 = _effect_subsample_step_log2(scr_len_y, quality); - int stepx2 = stepx >> 1; - int stepy2 = stepy >> 1; + // Filter variables + Value b[N+1]; // scaling coefficient + filter coefficients (can be 10.21 fixed point) + double bf[N]; // computed filter coefficients + double M[N*N]; // matrix used for initialization procedure (has to be double) - /* buffer for x-axis blur */ - NRPixBlock *bufx = new NRPixBlock; - /* buffer for y-axis blur */ - NRPixBlock *bufy = new NRPixBlock; + // Compute filter (x) + calcFilter(_deviation_x * trans.expansionX(), bf); + for(size_t i=0; i<N; i++) bf[i] = -bf[i]; + b[0] = 1; // b[0] == alpha (scaling coefficient) + for(size_t i=0; i<N; i++) { + b[i+1] = bf[i]; + b[0] -= b[i+1]; + } - // boundaries of the subsampled (smaller, unless step==1) buffers - int xd0 = (in->area.x0 >> stepx_l2); - int xd1 = (in->area.x1 >> stepx_l2) + 1; - int yd0 = (in->area.y0 >> stepy_l2); - int yd1 = (in->area.y1 >> stepy_l2) + 1; + // Compute initialization matrix (x) + calcTriggsSdikaM(bf, M); - // set up subsampled buffers - nr_pixblock_setup_fast(bufx, in->mode, xd0, yd0, xd1, yd1, true); - nr_pixblock_setup_fast(bufy, in->mode, xd0, yd0, xd1, yd1, true); - if (bufx->data.px == NULL || bufy->data.px == NULL) { // no memory - return 0; - } + // Filter (x) + switch(in->mode) { + case NR_PIXBLOCK_MODE_A8: ///< Grayscale + filterX<unsigned char,1>(in, out, b, M, tmpdata); + break; + case NR_PIXBLOCK_MODE_R8G8B8: ///< 8 bit RGB + filterX<unsigned char,3>(in, out, b, M, tmpdata); + break; + case NR_PIXBLOCK_MODE_R8G8B8A8N: ///< Normal 8 bit RGBA + filterX<unsigned char,4>(in, out, b, M, tmpdata); + break; + case NR_PIXBLOCK_MODE_R8G8B8A8P: ///< Premultiplied 8 bit RGBA + filterX<unsigned char,4>(in, out, b, M, tmpdata); + break; + default: + assert(false); + }; - //mid->visible_area = in->visible_area; - //out->visible_area = in->visible_area; + if ( std::abs(_deviation_x*trans.expansionX() - _deviation_y*trans.expansionY()) > 0.000001 ) { + // Compute filter (y) + calcFilter(_deviation_y * trans.expansionY(), bf); + for(size_t i=0; i<N; i++) bf[i] = -bf[i]; + b[0] = 1; // b[0] == alpha (scaling coefficient) + for(size_t i=0; i<N; i++) { + b[i+1] = bf[i]; + b[0] -= b[i+1]; + } - /* Array for filter kernel, big enough to fit kernels for both X and Y - * direction kernel, one at time */ - double kernel[_kernel_size(trans)]; - - /* 1. Blur in direction of X-axis, from in to bufx (they have different resolution)*/ - _make_kernel(kernel, _deviation_x, trans.expansionX()); + // Compute initialization matrix (y) + calcTriggsSdikaM(bf, M); + } - for ( int y = bufx->area.y0 ; y < bufx->area.y1; y++ ) { + // Filter (y) + switch(in->mode) { + case NR_PIXBLOCK_MODE_A8: ///< Grayscale + filterY<unsigned char,1>(out, out, b, M, tmpdata); + break; + case NR_PIXBLOCK_MODE_R8G8B8: ///< 8 bit RGB + filterY<unsigned char,3>(out, out, b, M, tmpdata); + break; + case NR_PIXBLOCK_MODE_R8G8B8A8N: ///< Normal 8 bit RGBA + filterY<unsigned char,4>(out, out, b, M, tmpdata); + break; + case NR_PIXBLOCK_MODE_R8G8B8A8P: ///< Premultiplied 8 bit RGBA + filterY<unsigned char,4>(out, out, b, M, tmpdata); + break; + default: + assert(false); + }; - // corresponding line in the source buffer - int in_line; - if ((y << stepy_l2) >= in->area.y1) { - in_line = (in->area.y1 - in->area.y0 - 1) * in->rs; - } else { - in_line = ((y << stepy_l2) - (in->area.y0)) * in->rs; - if (in_line < 0) - in_line = 0; + delete[] tmpdata; + + out->empty = FALSE; + slot.set(_output, out); + } else { + /* Blur radius in screen units (pixels) */ + int scr_len_x = _effect_area_scr_x(trans); + int scr_len_y = _effect_area_scr_y(trans); + + int const PC = NR_PIXBLOCK_BPP(in); + typedef unsigned char PT; + + /* buffer for blur */ + NRPixBlock *out = new NRPixBlock; + nr_pixblock_setup_fast(out, in->mode, in->area.x0, in->area.y0, + in->area.x1, in->area.y1, true); + if (out->data.px == NULL) { // no memory + return 0; } - // current line in bufx - int bufx_line = (y - yd0) * bufx->rs; + /* Array for filter kernel, big enough to fit kernels for both X and Y + * direction kernel, one at time */ + double kernel[_kernel_size(trans)]; - int skipbuf[4] = {INT_MIN, INT_MIN, INT_MIN, INT_MIN}; + // Past pixels seen (to enable in-place operation) + PT history[_kernel_size(trans)][PC]; + + int const width=in->area.x1-in->area.x0, height=in->area.y1-in->area.y0; - for ( int x = bufx->area.x0 ; x < bufx->area.x1 ; x++ ) { + /* 1. Blur in direction of X-axis, from in to bufx (they have different resolution)*/ + _make_kernel(kernel, _deviation_x, trans.expansionX()); - // for all bytes of the pixel - for ( int byte = 0 ; byte < NR_PIXBLOCK_BPP(in) ; byte++) { + for ( int y = 0 ; y < height; y++ ) { - if(skipbuf[byte] > x) continue; + // corresponding line in the source buffer + int const in_line = y * in->rs; - double sum = 0; - int last_in = -1; - int different_count = 0; + // current line in bufx + int const bufx_line = y * out->rs; - // go over our point's neighborhood on x axis in the in buffer, with stepx increment - for ( int i = -scr_len_x ; i <= scr_len_x ; i += stepx ) { + int skipbuf[4] = {INT_MIN, INT_MIN, INT_MIN, INT_MIN}; - // the pixel we're looking at - int x_in = (((x << stepx_l2) + i + stepx2) >> stepx_l2) << stepx_l2; + // history initialization + PT imin[PC]; copy_n(NR_PIXBLOCK_PX(in) + in_line, PC, imin); + for(int i=0; i<scr_len_x; i++) copy_n(imin, PC, history[i]); - // distance from it to the current x,y - int dist = x_in - (x << stepx_l2); - if (dist < -scr_len_x) - dist = -scr_len_x; - if (dist > scr_len_x) - dist = scr_len_x; + for ( int x = 0 ; x < width ; x++ ) { - if (x_in >= in->area.x1) { - x_in = (in->area.x1 - in->area.x0 - 1); - } else { - x_in = (x_in - in->area.x0); - if (x_in < 0) - x_in = 0; - } + int const disp = PC * x; - // value at the pixel - _check_index(in, in_line + NR_PIXBLOCK_BPP(in) * x_in + byte, __LINE__); - unsigned char in_byte = NR_PIXBLOCK_PX(in)[in_line + NR_PIXBLOCK_BPP(in) * x_in + byte]; + // update history + for(int i=scr_len_x; i>0; i--) copy_n(history[i-1], PC, history[i]); + copy_n(NR_PIXBLOCK_PX(in) + in_line + disp, PC, history[0]); - // is it the same as last one we saw? - if(in_byte != last_in) different_count++; - last_in = in_byte; + // for all bytes of the pixel + for ( int byte = 0 ; byte < PC ; byte++) { - // sum pixels weighted by the kernel; multiply by stepx because we're skipping stepx pixels - sum += stepx * in_byte * kernel[scr_len_x + dist]; - } + if(skipbuf[byte] > x) continue; - // store the result in bufx - _check_index(bufx, bufx_line + NR_PIXBLOCK_BPP(bufx) * (x - xd0) + byte, __LINE__); - NR_PIXBLOCK_PX(bufx)[bufx_line + NR_PIXBLOCK_BPP(bufx) * (x - xd0) + byte] = (unsigned char)sum; + double sum = 0; + int last_in = -1; + int different_count = 0; - // optimization: if there was no variation within this point's neighborhood, - // skip ahead while we keep seeing the same last_in byte: - // blurring flat color would not change it anyway - if (different_count <= 1) { - int pos = x + 1; - while(((pos << stepx_l2) + scr_len_x) < in->area.x1 && - NR_PIXBLOCK_PX(in)[in_line + NR_PIXBLOCK_BPP(in) * ((pos << stepx_l2) + scr_len_x - in->area.x0) + byte] == last_in) - { - _check_index(in, in_line + NR_PIXBLOCK_BPP(in) * ((pos << stepx_l2) + scr_len_x - in->area.x0) + byte, __LINE__); - _check_index(bufx, bufx_line + NR_PIXBLOCK_BPP(bufx) * (pos - xd0) + byte, __LINE__); - NR_PIXBLOCK_PX(bufx)[bufx_line + NR_PIXBLOCK_BPP(bufx) * (pos - xd0) + byte] = last_in; - pos++; + // go over our point's neighbours in the history + for ( int i = 0 ; i <= scr_len_x ; i++ ) { + // value at the pixel + unsigned char in_byte = history[i][byte]; + + // is it the same as last one we saw? + if(in_byte != last_in) different_count++; + last_in = in_byte; + + // sum pixels weighted by the kernel + sum += in_byte * kernel[i]; } - skipbuf[byte] = pos; - } - } - } - } + // go over our point's neighborhood on x axis in the in buffer + for ( int i = 1 ; i <= scr_len_x ; i++ ) { - /* 2. Blur in direction of Y-axis, from bufx to bufy (they have the same resolution) */ - _make_kernel(kernel, _deviation_y, trans.expansionY()); + // the pixel we're looking at + int x_in = x + i; + if (x_in >= width) { + x_in = width - 1; + } else if (x_in < 0) { + x_in = 0; + } - for ( int x = bufy->area.x0 ; x < bufy->area.x1; x++ ) { + // value at the pixel + _check_index(in, in_line + PC * x_in + byte, __LINE__); + unsigned char in_byte = NR_PIXBLOCK_PX(in)[in_line + PC * x_in + byte]; - int bufy_disp = NR_PIXBLOCK_BPP(bufy) * (x - xd0); - int bufx_disp = NR_PIXBLOCK_BPP(bufx) * (x - xd0); + // is it the same as last one we saw? + if(in_byte != last_in) different_count++; + last_in = in_byte; - int skipbuf[4] = {INT_MIN, INT_MIN, INT_MIN, INT_MIN}; + // sum pixels weighted by the kernel + sum += in_byte * kernel[i]; + } - for ( int y = bufy->area.y0; y < bufy->area.y1; y++ ) { + // store the result in bufx + _check_index(out, bufx_line + disp + byte, __LINE__); + NR_PIXBLOCK_PX(out)[bufx_line + disp + byte] = (unsigned char)sum; - int bufy_line = (y - yd0) * bufy->rs; + // optimization: if there was no variation within this point's neighborhood, + // skip ahead while we keep seeing the same last_in byte: + // blurring flat color would not change it anyway + if (different_count <= 1) { + int pos = x + 1; + while(pos + scr_len_x < width && + NR_PIXBLOCK_PX(in)[in_line + PC * (pos + scr_len_x) + byte] == last_in) + { + _check_index(in, in_line + PC * (pos + scr_len_x) + byte, __LINE__); + _check_index(out, bufx_line + PC * pos + byte, __LINE__); + NR_PIXBLOCK_PX(out)[bufx_line + PC * pos + byte] = last_in; + pos++; + } + skipbuf[byte] = pos; + } + } + } + } - for ( int byte = 0 ; byte < NR_PIXBLOCK_BPP(bufx) ; byte++) { - if (skipbuf[byte] > y) continue; + /* 2. Blur in direction of Y-axis, from bufx to bufy (they have the same resolution) */ + _make_kernel(kernel, _deviation_y, trans.expansionY()); - double sum = 0; - int last_in = -1; - int different_count = 0; + for ( int x = 0 ; x < width; x++ ) { - for ( int i = -scr_len_y ; i <= scr_len_y ; i += stepy ) { + int const disp = PC * x; - int y_in = ((((y << stepy_l2) + i + stepy2) >> stepy_l2) - yd0); + int skipbuf[4] = {INT_MIN, INT_MIN, INT_MIN, INT_MIN}; - int dist = ((y_in + yd0) << stepy_l2) - (y << stepy_l2); - if (dist < -scr_len_y) - dist = -scr_len_y; - if (dist > scr_len_y) - dist = scr_len_y; + PT imin[PC]; copy_n(NR_PIXBLOCK_PX(out) + disp, PC, imin); + for(int i=0; i<scr_len_y; i++) copy_n(imin, PC, history[i]); - if (y_in >= (yd1 - yd0)) y_in = (yd1 - yd0) - 1; - if (y_in < 0) y_in = 0; + for ( int y = 0; y < height; y++ ) { - _check_index(bufx, y_in * bufx->rs + NR_PIXBLOCK_BPP(bufx) * (x - xd0) + byte, __LINE__); - unsigned char in_byte = NR_PIXBLOCK_PX(bufx)[y_in * bufx->rs + NR_PIXBLOCK_BPP(bufx) * (x - xd0) + byte]; - if(in_byte != last_in) different_count++; - last_in = in_byte; - sum += stepy * in_byte * kernel[scr_len_y + dist]; - } + int bufy_line = y * out->rs; - _check_index(bufy, bufy_line + bufy_disp + byte, __LINE__); - NR_PIXBLOCK_PX(bufy)[bufy_line + bufy_disp + byte] = (unsigned char)sum; + for(int i=scr_len_y; i>0; i--) copy_n(history[i-1], PC, history[i]); + copy_n(NR_PIXBLOCK_PX(out) + bufy_line + disp, PC, history[0]); - if (different_count <= 1) { - int pos = y + 1; - while((pos + (scr_len_y >> stepy_l2) + 1) < yd1 && - NR_PIXBLOCK_PX(bufx)[(pos + (scr_len_y >> stepy_l2) + 1 - yd0) * bufx->rs + bufx_disp + byte] == last_in) - { - _check_index(bufx, (pos + (scr_len_y >> stepy_l2) + 1 - yd0) * bufx->rs + bufx_disp + byte, __LINE__); - _check_index(bufy, (pos - yd0) * bufy->rs + bufy_disp + byte, __LINE__); - NR_PIXBLOCK_PX(bufy)[(pos - yd0) * bufy->rs + bufy_disp + byte] = last_in; - pos++; - } - skipbuf[byte] = pos; - } + for ( int byte = 0 ; byte < PC ; byte++) { - } - } - } + if (skipbuf[byte] > y) continue; - // we don't need bufx anymore - nr_pixblock_release(bufx); - delete bufx; + double sum = 0; + int last_in = -1; + int different_count = 0; - // interpolation will need to divide by stepx * stepy - int divisor = stepx_l2 + stepy_l2; + for ( int i = 0 ; i <= scr_len_y ; i++ ) { + // value at the pixel + unsigned char in_byte = history[i][byte]; - // new buffer for the final output, same resolution as the in buffer - NRPixBlock *out = new NRPixBlock; - nr_pixblock_setup_fast(out, in->mode, in->area.x0, in->area.y0, - in->area.x1, in->area.y1, true); - if (out->data.px == NULL) { - // alas, we've accomplished a lot, but ran out of memory - so abort - return 0; - } + // is it the same as last one we saw? + if(in_byte != last_in) different_count++; + last_in = in_byte; - for ( int y = yd0 ; y < yd1 - 1; y++ ) { - for ( int x = xd0 ; x < xd1 - 1; x++ ) { - for ( int byte = 0 ; byte < NR_PIXBLOCK_BPP(bufy) ; byte++) { + // sum pixels weighted by the kernel + sum += in_byte * kernel[i]; + } - // get 4 values at the corners of the pixel from bufy - _check_index(bufy, ((y - yd0) * bufy->rs) + NR_PIXBLOCK_BPP(bufy) + (x - xd0) + byte, __LINE__); - unsigned char a00 = NR_PIXBLOCK_PX(bufy)[((y - yd0) * bufy->rs) + NR_PIXBLOCK_BPP(bufy) * (x - xd0) + byte]; - if (stepx == 1 && stepy == 1) { // if there was no subsampling, just use a00 - _check_index(out, ((y - yd0) * out->rs) + NR_PIXBLOCK_BPP(out) * (x - xd0) + byte, __LINE__); - NR_PIXBLOCK_PX(out)[((y - yd0) * out->rs) + NR_PIXBLOCK_BPP(out) * (x - xd0) + byte] = a00; - continue; - } - _check_index(bufy, ((y - yd0) * bufy->rs) + NR_PIXBLOCK_BPP(bufy) * (x + 1 - xd0) + byte, __LINE__); - unsigned char a10 = NR_PIXBLOCK_PX(bufy)[((y - yd0) * bufy->rs) + NR_PIXBLOCK_BPP(bufy) * (x + 1 - xd0) + byte]; - _check_index(bufy, ((y + 1 - yd0) * bufy->rs) + NR_PIXBLOCK_BPP(bufy) * (x - xd0) + byte, __LINE__); - unsigned char a01 = NR_PIXBLOCK_PX(bufy)[((y + 1 - yd0) * bufy->rs) + NR_PIXBLOCK_BPP(bufy) * (x - xd0) + byte]; - _check_index(bufy, ((y + 1 - yd0) * bufy->rs) + NR_PIXBLOCK_BPP(bufy) * (x + 1 - xd0) + byte, __LINE__); - unsigned char a11 = NR_PIXBLOCK_PX(bufy)[((y + 1 - yd0) * bufy->rs) + NR_PIXBLOCK_BPP(bufy) * (x + 1 - xd0) + byte]; + for ( int i = 1 ; i <= scr_len_y ; i++ ) { - // iterate over the rectangle to be interpolated - for ( int yi = 0 ; yi < stepy; yi++ ) { - int iy = stepy - yi; - int y_out = (y << stepy_l2) + yi; - if ((y_out < out->area.y0) || (y_out >= out->area.y1)) - continue; - int out_line = (y_out - out->area.y0) * out->rs; + int y_in = y + i; + if (y_in >= height) { + y_in = height - 1; + } else if (y_in < 0) { + y_in = 0; + } - for ( int xi = 0 ; xi < stepx; xi++ ) { - int ix = stepx - xi; - int x_out = (x << stepx_l2) + xi; - if ((x_out < out->area.x0) || (x_out >= out->area.x1)) - continue; + _check_index(out, y_in * out->rs + disp + byte, __LINE__); + unsigned char in_byte = NR_PIXBLOCK_PX(out)[y_in * out->rs + disp + byte]; + if(in_byte != last_in) different_count++; + last_in = in_byte; + sum += in_byte * kernel[i]; + } - // simple linear interpolation - int a = (a00*ix*iy + a10*xi*iy + a01*ix*yi + a11*xi*yi) >> divisor; + _check_index(out, bufy_line + disp + byte, __LINE__); + NR_PIXBLOCK_PX(out)[bufy_line + disp + byte] = (unsigned char)sum; - _check_index(out, out_line + NR_PIXBLOCK_BPP(out) * (x_out - out->area.x0) + byte, __LINE__); - NR_PIXBLOCK_PX(out)[out_line + NR_PIXBLOCK_BPP(out) * (x_out - out->area.x0) + byte] = (unsigned char) a; + if (different_count == 0) { + int pos = y + 1; + while((pos + scr_len_y) < height && + NR_PIXBLOCK_PX(out)[(pos + scr_len_y) * out->rs + disp + byte] == last_in) + { + _check_index(out, (pos + scr_len_y) * out->rs + disp + byte, __LINE__); + _check_index(out, pos * out->rs + disp + byte, __LINE__); + NR_PIXBLOCK_PX(out)[pos * out->rs + disp + byte] = last_in; + pos++; + } + skipbuf[byte] = pos; } + } } } + + out->empty = FALSE; + slot.set(_output, out); } - nr_pixblock_release(bufy); - delete bufy; - - out->empty = FALSE; - slot.set(_output, out); - return 0; } @@ -551,7 +613,7 @@ { int area_x = _effect_area_scr_x(trans); int area_y = _effect_area_scr_y(trans); - return _max(area_x, area_y); + return std::max(area_x, area_y); } void FilterGaussian::set_deviation(double deviation) Index: src/display/nr-filter-gaussian.h =================================================================== --- src/display/nr-filter-gaussian.h (revision 13436) +++ src/display/nr-filter-gaussian.h (working copy) @@ -7,6 +7,7 @@ * Authors: * Niko Kiirala <niko@...1267...> * bulia byak + * Jasper van de Gronde <th.v.d.gronde@...528...> * * Copyright (C) 2006 authors * @@ -62,17 +63,6 @@ void _make_kernel(double *kernel, double deviation, double expansion); int _effect_area_scr_x(Matrix const &trans); int _effect_area_scr_y(Matrix const &trans); - int _effect_subsample_step(int scr_len_x, int quality); - int _effect_subsample_step_log2(int scr_len_x, int quality); - - inline int _min(int const a, int const b) - { - return ((a < b) ? a : b); - } - inline int _max(int const a, int const b) - { - return ((a > b) ? a : b); - } };

4 5

Trace dialog tweaks
by Bob Jamison 10 Dec '06

10 Dec '06

Hi all, I really like the redesign of the Trace dialog, and I loved it when I first saw it. But later, I see two things that need to be adjusted before the release. 1. The SIOX button needs to be on the same view as the preview, since it alters how the selection is made. Normally, we want only an image and nothing else. With the SIOX checkbox pressed, we want the image, and also any curves which are above it in Z-order. 2. Much more important: The "Credits" frame is missing. We need to put it back. I (we) promised Peter Selinger that as long as we used Potrace, he would have obvious attribution for his stuff. He is a treasured benefactor of Inkscape, and we need to make sure that users know whence the technology comes. I am an active opponent of software patents, but if anyone really deserved a patent, it would be Mr. Selinger. Potrace is cutting technology, yet instead of trying to profit from it, he altruistically allows people like us to use it. Behaviour such as this needs to be encouraged. For example, the difference between Potrace 1.6 and 1.7 is a set of optimizations that he did explicitly for us. Cool guy! And all we need is a (3cm x 2cm) box with the name and URL on it. bob

2 3

Gaussian blur fix
by Jasper van de Gronde 10 Dec '06

10 Dec '06

I found bug 1576258 in the tracker and made a patch for Gaussian blur that fixes it (attached it to the bug). Since no one was assigned to the bug I thought I'd send a message to the mailing list so someone can review/commit the fix. BTW, the IIR gaussian blur patch I posted earlier also solves this problem, as it doesn't use subsampling anymore, but this patch is much less invasive.

1 0

Re: [Inkscape-devel] [Inkscape-user] bbox calculation
by bulia byak 09 Dec '06

09 Dec '06

On 12/9/06, Jasper van de Gronde <th.v.d.gronde@...528...> wrote: > "The bounding box is computed *exclusive* of any values for clipping, > masking, *filter effects*, opacity and *stroke-width*." > > See: > http://www.w3.org/TR/SVG11/coords.html#ObjectBoundingBox > > I filed this as bug 1612108 (with a test case which shows the problem). Thanks, that's a good point - it concerns SVG compliance and therefore must be fixed regardless of what we decide for other visual and UI aspects of bboxes. Unfortunately this means that at the moment, we display gradients, patterns, clippaths and masks slightly wrong when they have objectBoundingBox units and apply to an object with stroke, and their display will change when we fix this. But the good news is that Inkscape does not create objectBoundingBox itself (except filter areas I think), so such files can only come from elsewhere. If a gradient has objectBoundingBox units, it is converted to absolute units as soon as you edit the gradient or the object that uses it. So, this fix is likely to have a relatively low impact. -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

1 0

Re: [Inkscape-devel] Time to unpause 0.45 release?
by Bryce Harrington 09 Dec '06

09 Dec '06

On Wed, Dec 06, 2006 at 01:25:35PM -0800, Jon Phillips wrote: > On Wed, 2006-12-06 at 12:57 -0800, Bryce Harrington wrote: > > Hi all, > > > > A few weeks back we decided to put the release on pause to give time for > > additional work on the PDF extension. Could someone outline the status > > of this work, and if it is still necessary for this to be blocking the > > release? > > > > If not, would anyone have an issue if we resumed the release process at > > this point? I find myself with some unexpected extra freetime this > > month that I could devote to assisting with the release. > > > > Bryce > > Funny...I had the near same email ready to send...lets haul on the > release. Activity seems to have lowered...I think now rather than wait > any more... Okay, well one of the really big items is getting the about screen contest underway. What steps are necessary to get this going? Any chance we can get this ready to start accepting submissions by this weekend? Bryce

4 5

Re: [Inkscape-devel] [Inkscape-user] bbox calculation
by bulia byak 09 Dec '06

09 Dec '06

On 12/8/06, MenTaLguY <mental@...3...> wrote: > I think some things (e.g. selection outlines) should probably always use the visual bbox. But the line between "technical" and "artistic" can be very fine. Even if we ignore technical users, there are artistic users (including myself) who sometimes require geometric bounds -- tiling is one of the most common cases, but there are others as well. > > We need to have a better solution in those situations, because there, the lack of geometric bounds isn't just annoying, it's debilitating. I just wish I knew what that solution was. :/ I can change it back, but I don't want to get back all the problems we had with no-stroke bboxes, and I don't really want to introduce a user option which will be hard to explain to the user and a maintenance burden. On the other hand, we now have an even worse option in Options > Tools > Selector > Default scale origin, which is highly obscure and rather buggy. It seems to me that if we add a "bounding box includes stroke" option we'll be able to get rid of that "default scale origin" which is a good thing in itself. It would be still better to find a compromise which would work for everyone and therefore not require an option. Suggestions are welcome. Please also consider that we now have blur, and in at least some cases, it makes perfect sense to include the blur margin in the bbox too. -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

3 3

wiimote+inkscape (YouTube)
by MenTaLguY 07 Dec '06

07 Dec '06

Inkscape featured (briefly) as a demo application for this guy's project to develop Linux drivers for the Wiimote. http://www.youtube.com/watch?v=Vp1_pZ-AteA&eurl= -mental

4 5

Runtime checks for SSL
by Aaron Spike 06 Dec '06

06 Dec '06

One of the problems I've been running up against with the autopackage is ensuring that the users have SSL installed. SSL is required for securing jabber login and whiteboard traffic. The autopackages I am currently generating have SSL linked to inkscape statically. This is a huge security nightmare IMHO and I do not plan to do this for the release packages because I do not plan on undertaking the responsibility of ensuring that we update the packages for any and every SSL vulnerability. I would like to user relaytool to make a runtime check for the presence of libssl and conditionally disable features. I would appreciate assistance with this. http://autopackage.org/apbuild-relaytool.php Aaron Spike

1 0

Boolean operations on lines, bug or feature?
by Jasper van de Gronde 04 Dec '06

04 Dec '06

When I use the difference or intersection operation on an open path (and a rectangle) the path suddenly gets new edges and is closed, see attachment (which also shows my intention). Is this intentional or should I file a bug?

2 2

Kompiling Inkscape
by Loi 03 Dec '06

03 Dec '06

Hi, Today I made a svn checkout of inkscape and tried to compile it. But it doesn't want: ./autogen.sh ============================================================================= When you report a trouble about building SVN version of inkscape, Please report following information about distro and tools version, too. --1. distribution------------------------------------------------------------ Debian GNU/Linux release 4.0 --2. tools------------------------------------------------------------------- /usr/bin/m4: /usr/bin/autoconf: autoconf (GNU Autoconf) 2.60a /usr/bin/autoheader: autoheader (GNU Autoconf) 2.60a /usr/bin/automake: automake (GNU automake) 1.10 automake-1.7: not found automake-1.8: not found automake-1.9: not found /usr/bin/aclocal: aclocal (GNU automake) 1.10 aclocal-1.7: not found aclocal-1.8: not found aclocal-1.9: not found /usr/bin/intltoolize: intltoolize (GNU intltool) 0.35.0 /usr/bin/gettextize: /usr/bin/gettextize (GNU gettext-tools) 0.15 --3. environment variables--------------------------------------------------- PATH=/usr/local/bin:/usr/bin:/bin:/usr/bin/X11:/usr/games LANG=de_CH LANGUAGE=de_CH:de_DE:de:en_GB:en ============================================================================= I am testing that you have the required versions of autoconf, automake, glib-gettextize and intltoolize. This test is not foolproof and if anything goes wrong, there may be guidance in the file HACKING.txt checking for autoconf >= 2.52 ... yes (version 2.60) checking for automake >= 1.7 ... yes (version 1.10) checking for glib-gettextize >= 2.0.0 ... yes (version 2.12.4) checking for intltool >= 0.17 ... yes (version 0.35.0) Running aclocal ... configure.ac:44: error: AC_PROG_CC cannot be called after AM_PROG_CC_C_O /usr/share/aclocal-1.10/ccstdc.m4:25: AM_PROG_CC_STDC is expanded from... configure.ac:44: the top level autom4te: /usr/bin/m4 failed with exit status: 1 aclocal: autom4te failed with exit status: 1 Please fix the error conditions and try again. I use: autoconf (GNU Autoconf) 2.60a automake (GNU automake) 1.10 All builddeps given in debian/control are installed.

2 1

Buildtool : 'console' option
by Johan Engelen 01 Dec '06

01 Dec '06

Thanks alot for making the buildtool! It's pretty nice. However, now I manually changed build.xml:255 into "-lgc -mconsole -lws2_32 -lintl -lm -lgdi32 -lcomdlg32" to create a console version so I can see errors and other messages. Could build.xml be changed such that adding 'console' as commandline option would build a console version of Inkscape? ("buildtool dist console" for example). Or do I have to make a copy-paste of targets link and dist and rename them linkconsole and distconsole such that "buildtool distconsole" would do what I want? Thanks, Johan

1 0

Re: [Inkscape-devel] Problem with the Latex Formula module
by Stéphane Gimenez 30 Nov '06

30 Nov '06

Sorry, my first message was meant to be addresed to the list. > >Could you try with the latest svn version of the extension script and > >give us feedback? > > > >http://inkscape.svn.sourceforge.net/viewvc/inkscape/inkscape/trunk/share/ex… > > > It is better, I have now a window message :) : > > dvips: ! DVI file can't be opened. […] Yes, it's better! Now I guess you got this message because the formula input you gave to the extension was either not valid LateX, or it became invalid because of a shell expansion which occured when the script was called by the extensions manager (this is a bug). Can you try with a simple "$ \cos(x) $" LaTeX input? To know how your imput will be expanded you can use the following shell command: $ echo -E "$$\cos(x)$$" 8012\cos(x)8012 For example '$$' is expanded to the pid of the processus. As a workarround, try not to use '$' or '$$' but '\(' and '\[' instead to enter math mode. Stéphane. PS: I'll also try to improve error reports when the given input is not valid LaTeX.

3 6

Secondary toolbar size?
by Andreas Nilsson 30 Nov '06

30 Nov '06

Dear Inkscape developers. Myself and some other folks are currently working on some icons for Inkscape [1], but have ran into some issues. Apparently my build (13378 ...the latest one I got to work) wants to use 16x16 images for the icons in both the topmost (although the ones it gets from the desktop theme is 24x24) and the secondary toolbar. The tools toolbar however still use 24x24 images (and that is the correct behavior I suspect). However, other peoples toolbars use the 24x24 size (and so did mine until recently), this leads to the problem that I don't know what size to aim at. If I use 24x24 images for the toolbar, they get blurry on my system, and if I use 16x16 it gets blurry on others. What is the correct size used by these toolbars, and is there some way I can have icons in several sizes? The several sizes issue is even more troublesome, as the icons from the tools toolbar gets all blurry in the undo history dialog among other places [2]. 1.http://ramnet.se/~nisse/diverse/temp/inkscape-icons.tar.gz 2. http://ramnet.se/~nisse/diverse/temp/inkscape-iconsizes.png - Andreas

4 10

New C++ find dialog
by Johan Engelen 29 Nov '06

29 Nov '06

Hello all, (this is about the current SVN-trunk version) I've been busy implementing the find dialog as a c++ dialog (ui\dialog\find.cpp). I hope it inspires others to spend some off-time on implementing the other old (non gtkmm) dialogs! It's pretty lame work though. Maybe good work for new people to get started? But very very unexciting I guess. But refactoring is a noble cause! (Mail me if you need help) If you have time to test the find dialog for bugs, please go to src/verbs.cpp:1618 and remove the comment //. I know the UI is not what it should be. When I tried adding the HSeparator as in the old find dialog, Inkscape would crash so I gave up on the issue at the moment. Thanks, Johan

1 0

markers in arena shapes?
by Bryce Harrington 28 Nov '06

28 Nov '06

Mental or Bulia, I've got a question... In looking at where marker stuff is hooked in I notice a lot of stuff in nr-arena-shape.cpp, which appears to be special casing marker elements as shape children. I'm attempting to add some documentation, but it's a little confusing to me. Could either of you explain what the design intent is for this? In particular, I'm wondering why markers have this special case; it feels like an odd fit in nr-arena-shape since everything else is so generalized, and I'm wondering if there could be a way to generalize the markers so no special case is needed for them? Bryce

4 7

g2png
by bulia byak 28 Nov '06

28 Nov '06

Hi Cedric, you have committed the "g2png" extension. It has many problems. 1. It wasn't added to Makefile.am and so ignored during installation. 2. We don't use capital letters in filenames so G2Pngs.py had to be renamed to g2pngs.py. 3. The python file had some French Unicode symbols on which my Python interpreter choked. Please translate the comments to English. 4. It was added to a "Recursive actions" submenu which makes little sense from the user viewpoint. I changed that to "Export". 5. It has a wrong copyright statement. You are not Aaron Spike, are you? :) 6. The URI should also be changed - don't use org.ekips, use org.inkscape or your own domain. 7. Most importantly, this extension just goes and exports ALL groups in the document, recursively. This does not seem to make sense to me. In most documents, it results in an awful lot of pngs, most of which are not needed. To be useful, instead, it should just export _all selected objects_ (not just groups!) without any recursion. Correspondingly, rename it to "Export selected objects" or something like that. 8. It ignores the filename and resolution hints stored for objects, always exporting them at the default resolution with the filename from the id. This is not acceptable. See the man page for how to use export hints. I have fixed problems 1 to 4 in SVN, but the rest are for you to fix. I'd suggest that you should have submitted this as a patch to our patch tracker, or at least have it discussed on the list, before committing to SVN. -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

3 2

PDF Output Atlas
by Aaron Spike 26 Nov '06

26 Nov '06

I think it is generally agreed that PDF output is the last outstanding blocker on the release process. Anyone who can spare a few cycles to help perfect PDF output will be much appreciated. It would be nice if the wise could supply some commentary on the state of PDF output to taper off the learning curve for newcomers to the field (like myself). As I sit here staring blankly at the files in src/extension/internal, I could really use an atlas that explains which files are important to pdf output, which particular effort they are part of, how they relate to each other, and where my effort should be focused. http://inkscape.svn.sourceforge.net/viewvc/inkscape/inkscape/trunk/src/exte… Thanks, Aaron Spike

1 0

Re: [Inkscape-devel] NEW: RadioButtons for extensions (.INX)
by Johan Engelen 26 Nov '06

26 Nov '06

> Date: Sat, 25 Nov 2006 15:41:16 +0000 (GMT) > From: Alan Horkan <horkana@...44...> > > > Date: Fri, 24 Nov 2006 22:23:31 +0100 > > From: Johan Engelen <goejendaagh@...1571...> > > To: inkscape-devel(a)lists.sourceforge.net > > Subject: [Inkscape-devel] NEW: RadioButtons for extensions (.INX) > > > > I've added radiobuttons to the parameter types for > extensions :) They are > > called: 'optiongroup'; English is not my native tongue > which makes it often > > difficult to come up with a label..., please rename if desired. > > A developer example is available as a placeholder until a > real extension > > uses it. > > Have you seperated the option group concept from the > imlementation as a radio list? I've thought about that (since the last discussion about enums), and it is the reason that the syntax is exactly equal to enums. I think it is safe to answer 'Yes'. (I will rename <item> to <option>) > For example you might want to take the same options and > reformat them as > drop down list rather than a radio list depending on how much > space you > have available or how important that option really is. Yes that is possible (for example: if the number of choices is more than 3, display dropdown, otherwise radiobuttons. However I must say that at this moment I prefer to give full control to the extension developer.

2 1

Closing bugs/requests/etc
by Johan Engelen 26 Nov '06

26 Nov '06

I think I am not able to close bugreports and the like? Or if so, please tell me how to! Thanks, Johan

2 1

Inkscape / win32 with gtk2.10
by Bob Jamison 25 Nov '06

25 Nov '06

Hi, all. Starting with Bulia's need for me to make an Inkscape build with Tor's patch for transient dialogs on win32, there was a need for me to change the build to use gtk2.10. Since I was going to change the build, I thought it would be good to modernize it, thus the Buildtool. Whew. Well, I got the Buildtool working yesterday, and I just now got a gtk2.10 build working. Tor's patch tomorrow? This copy was made with gtk2.10: http://inkscape.modevia.com/win32/Inkscape0611161701.7z This is not just a dll drop-in replacement, but a complete build with new libraries. It seems to work ok here, but people might want to give it a try to see if it works for them. bob

5 10

Re: [Inkscape-devel] Compile error with FreeType >= 2.2.0 (caused by ps export t
by bulia byak 25 Nov '06

25 Nov '06

On 11/25/06, Cyril Dangerville <cyrildangerville@...19...> wrote: > Hi! > I uploaded a new version that does not rely on freetype internal headers. I > patched my svn copy (revision 13361) and compiled successfully with Freetype > 2.2.1. But I need some help on one issue. Please read my last comment on the > tracker for details: > http://sourceforge.net/tracker/index.php?func=detail&aid=1219308&group_id=9… I am far from an expert on Freetype, but from the name of the function, it does indeed seem that it's X11-only. If so, I have no idea how to get font type in a portable way. Can you perhaps get the font file name and figure out from its extension? Just an idea. Any help from anyone familiar with this will be much appreciated! -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

1 0

NEW: grow/shrink selection in node tool
by bulia byak 25 Nov '06

25 Nov '06

* You can grow or shrink node selection by hovering the mouse pointer over a node and using mousewheel (up = grow, down = shrink) or the keys PageUp (grow) and PageDown (shrink). Growing adds the closest unselected node to the selection; shrinking deselects the farthest selected node. There are two modes that differ by how the closest/farthest nodes are chosen: o Spatial selection (Mousewheel, PageUp/PageDown): distances to nodes are measured directly, regardless of which subpath a node belongs to. o Linear selection (Ctrl+Mousewheel, Ctrl+PageUp/Ctrl+PageDown): node distances are measured along the path, and only the nodes belonging to the same subpath as the hovered node are considered (i.e. other subpaths are never selected). This technique is convenient for quickly selecting an area in a complex path starting from a center - for example, for node sculpting. Let me know if you think linear mode will be more often used than spatial - I can swap the shortcuts so that the more common mode can be used without Ctrl. -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

2 1

stupid question
by tj 25 Nov '06

25 Nov '06

I am a new user to inkscape and I have a stupid question. How do you close windows like the document and inkscape preferences windows? I see no close, cancel, etc buttons? tj

5 5

NEW: RadioButtons for extensions (.INX)
by Johan Engelen 25 Nov '06

25 Nov '06

I've added radiobuttons to the parameter types for extensions :) They are called: 'optiongroup'; English is not my native tongue which makes it often difficult to come up with a label..., please rename if desired. A developer example is available as a placeholder until a real extension uses it. Regards, Johan

2 1

Re: [Inkscape-devel] Two function plotters?
by bulia byak 24 Nov '06

24 Nov '06

On 11/24/06, Johan Engelen <johane@...1482...> wrote: > The "function plotter" is meant to be a more general plotter than "wavy". > "wavy" let's you specify the x-range in multiples of pi, which is very > convenient for plotting periodic functions. "function plotter" does not let > you do this which can be annoying. I'd like to keep them separate, but we > can also merge them by adding for example a checkbox to "function plotter" > to multiply the x-range by 2pi. Yes, I think such a checkbox would be a good idea. Can you please do it, and then we can retire the Wavy. -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

2 3

Re: [Inkscape-devel] Two function plotters?
by Johan Engelen 24 Nov '06

24 Nov '06

> Date: Thu, 23 Nov 2006 17:46:32 -0400 > From: "bulia byak" <buliabyak@...400...> > > I notice that we have two extensions to do function plotting: > old "wavy" and new "function plotter". Johan, since you added the > latter, please clean this up by removing one of them (make sure no > functionality is lost!). (Since I graduated two days ago, I'm back working on Inkscape again!) The "function plotter" is meant to be a more general plotter than "wavy". "wavy" let's you specify the x-range in multiples of pi, which is very convenient for plotting periodic functions. "function plotter" does not let you do this which can be annoying. I'd like to keep them separate, but we can also merge them by adding for example a checkbox to "function plotter" to multiply the x-range by 2pi. Another option is to use the read-out possibility of the notebook parameter: "wavy" could be added as a tab and the extension can determine which tab was selected when processing... (definately not clear to users and this has not my preference)

1 0

Re: [Inkscape-devel] g2png
by bulia byak 24 Nov '06

24 Nov '06

On 11/24/06, cedric GEMY <radar.map35@...8...> wrote: > my purpose was exactly to export groups, only. See the release Notes if > you want o know why. I don't see how your code fits that explanation, in particular the icons example. There may be much more groups in an icons file (examine our own icons.svg for example), not only the top-level groups corresponding to icons. Icons may contain some inner groups, which will get exported too. On the other hand, some simple icons may be represented by a single shape, not a group. And finally, I may want to export not ALL icons in the file but only some of them (more than one but not all). Your code can't do this either. > i don't understand what your proposal have > different from the export to bitmap dialog (option "selection" checked) It is different in that it exports each selected object SEPARATELY, creating as many PNG files as you have objects selected. -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

1 0

Compile error with FreeType >= 2.2.0 (caused by ps export type 1 patch)
by Gustav Broberg 23 Nov '06

23 Nov '06

I'm having problems compiling since revision 13379 which applied the (rather old) patch 1219308 [1] for embedding Type 1 fonts into exported PostScript files. The reason for the problems is that the patch relies on internal FreeType headers which are no longer exposed as of version 2.2.0, see [2] for details. Unless the patch can be rewritten any time soon, I think it needs to be rolled back -- FreeType 2.2.* is pretty common, Ubuntu 6.10 uses 2.2.1 for instance. A bug is filed here... http://sourceforge.net/tracker/index.php?func=detail&aid=1600900&group_id=9… Thanks! 1.http://sourceforge.net/tracker/index.php?func=detail&aid=1219308&group_id… 2.http://freetype.sourceforge.net/freetype2/freetype-2.2.0.html -- Gustav Broberg

2 1

win32 build tool
by Adib Taraben 23 Nov '06

23 Nov '06

Hello Bob Jamison, I think your build tool is much faster than make ... there is a problem with the l10n: Build should create for all po/*.mo files a directory in inkscape/local/ with the name of the file (without extension) and copy the file to that location. Is that possible ? Thx, Adib.

2 1

Two function plotters?
by bulia byak 23 Nov '06

23 Nov '06

I notice that we have two extensions to do function plotting: old "wavy" and new "function plotter". Johan, since you added the latter, please clean this up by removing one of them (make sure no functionality is lost!). -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

1 0

Help Wanted: Testing Autopackages
by Aaron Spike 23 Nov '06

23 Nov '06

The quality of the Autopackage builds could be greatly increased with more consistent testing. There is a need for volunteers running various distros (Slackware, Suse, Mandrake, Fedora, Debian) and various versions of these distros to work with the autopackager to ensure compatibility throughout the release cycle. This work involves regular downloading and installing of inkscape autopackages (about 10MB), submitting detailed bug reports to the autopackager and testing fixes as they are prepared. Snapshot builds for testing live at: http://inkscape.modevia.com/ap/?C=M;O=D Thanks, Aaron Spike

2 3

Problem with the Latex Formula module
by Jean-Luc Menut 23 Nov '06

23 Nov '06

Hello, I have a problem when I try to use the latex formula module: nothing appears, I have this message in my terminal : ** (.proxy.inkscape:7980): CRITICAL **: Last <defs> removed I'm using the latest inkscape autopackage (05 nov), pstoedit 3.44 and ubuntu (edgy) Any idea ? thanks, Jean-Luc

2 4

[Fwd: [Scale-announce] SCALE Call For Papers closes Monday!]
by Ted Gould 23 Nov '06

23 Nov '06

So, I think we're going to have a booth at SCALE this year. Josh has volunteered to take a leadership role in that -- so I'm going to play SCALE cheerleader. Which means, SUBMIT ABSTRACTS! :) It's February, and unlikely that the weather is better where you are than it is going to be in LA. Southwest flies to LAX so it means you can probably get a cheap flight. You can walk to the hotel from LAX[1]. The conference is on the weekend. I think it would be very cool to have a set of Inkscape presentations. Forward this to your GIMP/Blender/Scribus friends, perhaps we can get a Linux Graphics track :) SCALE is a cool conference, small and laid back. Still a volunteer run thing, so the registration fee isn't crazy. But, every year the speakers have gotten better, and the overall quality has gone up. I don't expect this year to be different. --Ted [1] - Not that I would recommend this. The traffic around LAX is awful, so unless you're invincible to cars, take the hotel shuttle. -------- Forwarded Message -------- From: scale-announce@...1566... Reply-To: nobody@...1566... To: SCALE-announce <scale-announce@...1566...> Subject: [Scale-announce] SCALE Call For Papers closes Monday! Date: Thu, 16 Nov 2006 14:18:31 -0800 If you're considering making a presentation at the 5th annual So Cal Linux Expo, be advised that the window to submit your CFP closes Monday, November 20th. Get your proposal in now! In the same vein, SCALE is sponsoring a one-day Open Source Health Care Summit on Friday, February 9th, just prior to the Expo. If you use Open Source software in the health care arena, here's your chance to talk about it. The CFP for the Summit closes December 29th. For more information, or for the details on the CFP, see http://socallinuxexpo.com/healthcare07/index.php. _______________________________________________ Scale-announce mailing list Scale-announce@...1566... http://mail.socallinuxexpo.com/mailman/listinfo/scale-announce

3 4

Test please ignore
by Ted Gould 22 Nov '06

22 Nov '06

I've been having trouble sending to Sourceforge. Hopefully this message will get through. --Ted

1 0

Re: [Inkscape-devel] Secondary toolbar size?
by Jon A. Cruz 22 Nov '06

22 Nov '06

On Nov 22, 2006, at 6:08 AM, David Christian Berg wrote: >> Mmmmmh, I can't find the Print Large theme you talking about. I have >> only have: David_icons.svg, icons.svg and legacy_icons.svg >> Where is "the size resquested" you are talking about? In the .svg >> file? > > Talking about the GTK Theme here, not the icon theme used by > Inkscape. I > didn't check but I guess you'll find the Large Print theme at > art.gnome.org Yes, the GTK+ theme. It's generally shipped with most RedHat derived distros at the least.

1 0

Fit selection to page
by Nika Vidal 20 Nov '06

20 Nov '06

Hi all, I've been reading yout list for some time but this is my first message to this list. I think it could be useful a "fit selection to page" button, specially when you get/made an artwork and you want to fit it into some specific format (a3, a4...), for example working for a printing work. I saw something like this to anybody was working with Macromedia Freehand. This is only an idea, I don know if it's already implemented at devel version. Excuse my poor English. Regards, Nika.

2 1

"All roads lead to bitmap" icon blog entry
by unknown＠example.com 20 Nov '06

20 Nov '06

Hi, I just came across a blog entry discussing vector vs. bitmap for icons. http://islayer.com/blog/?p=87 It also mentions a limitation with Illustrator, but that's a side one. Anyway, just wanted to point this out as something we might look at. I know I'll follow up on the icon aspects, as it falls in line with things I've discussed with various icon authors in addition to the librsvg team, ksvg peoples, etc.

3 2

NEW: Trace Bitmap dialog redesign and new parameters
by bulia byak 19 Nov '06

19 Nov '06

* This dialog now provides access to three more tracing parameters: o Suppress speckles: If set, spots or speckles larger than the given size are suppressed in the trace. o Smooth corners: This parameter controls how much smoothing is applied to corners in the traced path. o Optimize paths: If set, trace paths are optimized by joining adjacent Bezier segments with the given tolerance. * The dialog is redesigned to use three tabs: Mode, Parameters, and Preview. The controls are reorganized to be easier to find. Most labels and tooltips are rewritten for clarity. * The trace preview image is made twice larger. -- bulia byak Inkscape. Draw Freely. http://www.inkscape.org

8 10