May 2026 - Inkscape Devel - Inkscape Mailman

object-trace: 'stack' parameter ignored in headless/CLI mode — roadmap question
by Mathieu Plessis 28 May '26

28 May '26

Hello Inkscape team, I have been working on a server-side vectorization pipeline using Inkscape 1.4.4 in headless mode on Ubuntu 26.04, and I would like to report a behavior that appears to be a limitation of the CLI interface. When using the object-trace action via --actions in headless mode (with xvfb-run), the 'stack' parameter (position 3 in the argument list) appears to have no effect on the generated SVG output. The command I am using: xvfb-run -a inkscape input.png \ --actions="select-all; object-trace:4,0,1,0,0,1.0,0; export-filename:output.svg; export-do" I tested with both stack=1 and stack=0 and compared the results: The fills are identical in both outputs No clip-path elements are generated in either case The stacking behavior visible in the Desktop UI (Multicolor > Stacked) is not reproduced In the Desktop version, enabling 'Stacked' produces clip-path elements that allow color layers to overlap correctly. This is important for logos and illustrations where adjacent color regions need to stack rather than be cut out. My questions: Is this a known limitation of the current CLI/headless implementation? Is there a workaround to enable stacking behavior from the command line? Is full CLI parity for object-trace parameters on the roadmap for a future release? Thank you for your work on Inkscape. It is a fantastic tool and the CLI capabilities are extremely valuable for server-side automation 🙏🙌 Best regards, <https://lanaworks.com/> Mathieu Plessis — Lanaworks WEB, ECOMMERCE & STRATÉGIE DIGITALE Agence web spécialisée dans la création et le développement de sites WordPress et PrestaShop. Plus de 15 ans d’expertise pour concevoir des plateformes sur mesure adaptées à vos objectifs. Conseil et accompagnement e-commerce, optimisation de votre présence en ligne et stratégies marketing digitales.

1 0

Developer Meeting Notes; May 21st 2026
by Martin Owens 21 May '26

21 May '26

Hi Devels! Event: https://inkscape.org/cals/event/1/ Attending: Martin, Tav, Charlotte, Elendil, Mike Terry, Roland Tarlor, Seán Burke Next: https://inkscape.org/cals/event/64/ We have a backlog of 23 approved but unmerged Merge Requests which need to be dealt with. The windows CI is failing for most of these and blocking the ability to merge effectivly. Martin is tasked with talking with René and Marc to figure out how to solve this problem. Tav has been working on converting opentype to harfbuz, this is especially useful for variable fonts as harfbuz gives better results. We can remove a bunch of code we were using to cope with the lack of functionality. Martin asks about unit testing for the new code, Tav hasn't gotten around to it yet. One limitation is a conflict between font variations and style in file fonts. Eventually the inkscape-font- spec attribute should go away. Martin asked about font variation information being available in the poppler pdf import. Charlotte has been working on her own extension. Has a question about aprovals for merge requests and if the approver should be the one to hit the merge button. The general consensus is that they should for contributions from non-senior contributors, but not for senior developers generally. Though this isn't policy. Roland has given us an update on the first full week of the Bug Accelerator program 2026.1, has been working through the priority list and get down to a short list. Removing issues which have been kicked down the road but are not actually applicable any more. Developers are working on crashes as a priority for this week. Roland is giving usability blockers more priority to help with Inkscape's usability. The end of the month is the targrt for having a solid work list of priorities for the program. Elendil discovered a feature and wants to remove it. It's an obscure feature of the grid which allows incremental moves but was using the Alt key which was inconsistant with other features. The feature is removed as it's not useful and not documented anywhere. The keyboard bindings for some other items are updated too and Elendil wanted to know if the default should be in the xml or the code, Martin says the code unless the default should actually be "disabled". Martin talked about the color work and the new rendering engine. Giving a brief overview of where we are and what the direction of travel is for getting this merged into 1.5. The new static color functionality refactoring is all merged thanks to pbs for the review. The ongoing work to add test cases for the drawing code should prove that the display tree interface for the new rendering engine is the same as the old code and can be swapped out for testing. Roland gives Martin thanks for the PDF work which he's used on multiple projects so far. Martin is proud of the stability of the PDF output as it's been tested by more people and proven itself to be more dependable and faster than the old cairo output. Next meeting is the 29th May and is the earlier meeting. Thanks everyone for coming!

1 0

Developer Meeting; May 13th 2026
by Martin Owens 13 May '26

13 May '26

Hey everyone! First week with our new hires attending with us. Event: https://inkscape.org/cals/event/64/ Attending: Martin, Seán Burke, Ishaan Arora, Mc, Mike Terry, René, Roland Taylor, Tav, Charlotte Next meeting: May 21st, later time, https://inkscape.org/cals/event/1/ We kick off today's meeting with introductions of the new contract hires and all the volunteer developers. Martin explains how the developer meeting works, cadence, format and questions and what the developer team is. We now kick off into the meeting proper: René has been doing some infrastructure work for windows and macOS. An open merge request to look at. Has been packaging a different Gtk4 application to improve support for Gtk4 on macOS as a side quest so we don't have just Gimp and Inkscape using it. Martin has been continuing their work on the rendering engine. Has a new color API merge request which needs review. It allows the calling of internal functions for color conversion by the rendering engine much faster than the std::vector pathway as well as various other fixes. lcms2 is still slow though in comparison. Marc asked about windows arm support in the build system. René pledges to work on it during the long weekend. Marc says that the system might be a good option for x86 as well as arm. Ishaan asked about PDF export and if we will support cairo going forwards. Martin says that the plan is to support CapyPDF going forwards and not fix issues with the cairo export. Marc asked about 1.4.5 and if that was going to happen. ltlnx wasn't here to answer so Marc will proceed as if there will be no more 1.4.x releases wrt packaging maintenance. Marc has fixed some communication issues on the website as downstreams were packaging the arm build by default which was not great. Marc asked Martin about the website and creating a new platform for x86-64bit which is more specific that the current platform `64bit`. Martin says yes and redirects can be created, though only for one at a time. Charlotte wanted to look into py2geom, it's not ready yet. Decided to use the addon system in C++ instead of a python based system. Because of a deadline will put the python functionality to one side. Perceptual grouping, imported from pdf which can group based on them being together visually. Tav has been going down a rabbit hole with memory leaks in the color font; while searching found other memory problems in freetype. Rewriting using Harfbuz directly but then found an issue with the optical size values in variable fonts. Removing the fudge we have in the sizing for hinting actually makes us more consistant with firefox. Martin gives some instruction for the bug accelerator program and what it will look like and how people will be meeting. Ishaan talks about an issue with text alignment, where Inkscpe is now more compatible, but the user experience is now broken as we appear to be including extra spaces at the ends of lines which then result in ugly spacing. This certainly is a regression, but it can't be fixed by making Inkscape less compatible with other programs. So the xml we produce for text must be changed to remov the ugly spacing instead. We talked about the recent AI announcement by GitLab and what the risks are. Developers reported that they see the AI UX even though it's switchef off for our project. Members inccouraged to report the issue to GitLab. Thanks everyone for coming to this week's busy meeting! Best Regards, Martin Owens

1 0

Inkscape 1.4.4 is out!
by Maren Hachmann 06 May '26

06 May '26

Inkscape 1.4.4 ============== Inkscape 1.4.4 is now available for all three supported operating systems (Linux, macOS, Windows)! You can download it from: https://inkscape.org/release/inkscape-1.4.4 BUGFIX and MAINTENANCE RELEASE ------------------------------ Inkscape 1.4.4 is be a bugfix and maintenance release, which brings you 20 crash fixes, more than 25 bug fixes and performance improvements, a couple small new features, and comes with almost 20 updated translations. The most important fixes and changes in this release are: * 20 crash fixes, among them for three nasty bugs where Inkscape wouldn't even start * almost 20 other bug fixes * 6 changes that improve performance * a new palette * a new button for rotating stars and polygons into their 'neutral' or 'upright' position * 27 updated interface translations * 15 updated documentation translations * installation files for Windows on Arm See the release notes for more information (also about known issues): https://inkscape.org/doc/release_notes/1.4.4/Inkscape_1.4.4.html Like its predecessor, it is also a bridge release in the sense that it can be used to convert the planned Inkscape 1.5 multipage file format to the pre-1.5 multipage format. Versions lower than Inkscape 1.4.3 will not be able to interpret pages created in Inkscape versions 1.5 and upwards. Opening a document in Inkscape 1.4.3 and saving it will convert it to the current ('old') page format (MR #7608). Background: While the 'old' format of pages in Inkscape is a custom addition that only works in Inkscape, the new format will make use of the svg:view element, which is standardized and can work in other SVG viewers, too. Find more information about this in MR #7525. Report Bugs ----------- Please report all bugs you find in our Inbox bug tracker on GitLab! https://inkscape.org/report ----------------------------- Join an Inkscape project team ----------------------------- Inkscape is always happy to welcome new contributors in various areas: code, docs, translation, UX, bug triaging, or even outreach! The main ways to get involved are listed on https://inkscape.org/contribute/ and you can get in touch with us on https://chat.inkscape.org/ Thanks to everyone involved in this release! -- The Inkscape team

1 0

Developer Meeting; May 5th 2026
by Martin Owens 05 May '26

05 May '26

Hey Everyone, No attendees for this week's meeting. Next meeting is the earlier meeting: https://inkscape.org/cals/event/64/ hope to see you all there, it's the earlier one. Best regards, Martin Owens

1 0