April 2026 - Inkscape Devel - Inkscape Mailman

Developer Team Meeting, April 19th 2026
by Martin Owens 19 Apr '26

19 Apr '26

Hey everyone, Event: https://inkscape.org/cals/event/1/ Attending: Martin, Vaibhav, Ishaan, Philip, René, Tav Next: https://inkscape.org/cals/event/64/ CI builds with 26.04 from Ubuntu 24.04 and there are some breakages which will need attention. This was highlighted by a PPA failure with test failures. There were some ASAN issues, these might have effected our CI, but don't appear to be. The test failures appear to be related to fonts. Martin talks about how capypdf works with fonts and how the system in the CI might be failing in it. We talk about the test failures for antialiasing and improving robustness. Vaibhav has been working on backports. Wants to talk about a merge request for a crash ready for review. Ishaan wanted to talk about the issue in Gtk4 clipboard and the missing functionality. We talk about the C++ bug fixing program and how much the contractor will have to fix bugs in upstream projects like Cairo, Gtk4 etc. There's Gtk4 demos which we can test to see what can be fixed in Inkscape for 1.5. Martin has continued working on the new rendering engine with a full separation of SPstyle using templating and a consistent API use they are now developing a way to allow the current rendering engine to print out the code for a running svg file so we can copy paste those outputs into test files to make sure we are getting the same output in the new engine. Philip and Martin talk about the new engine's place, and how the discipline could help with using a GPU rendering engine in the future. Pillip wanted to ask René about the CI improvements for pinning. The linux CI is not pinned and asks about the setup in use for the repository for ci building. There are details which this note taker does not understand, but it looks like a bunch of work are possible but would take time with GitLab. Philip asks Tav about his color font merge request. There's an open merge request which can be fixed; any cairo version before 2027 would deadlock instead which we can patch in windows and macOS but not on Linux. Martin requests checks to avoid the new fix from causing the deadlock for users who might paste emoji into their document. Philip finishes with the lib2geom test case issues; there's been a build up of test issues for various platforms. There's some slight differences between archetectures and Philip believes some of the features can't be fixed at all. Martin advises to contact previous lib2geom contributors at the same time as investigating the issues and seeing what is possible to fix and what tests will have to be disabled. Tav is planning on a refactoring branch to cleanup the namespace use of Inkscape. Martin suggests that using namespace Inkscape is valid for a lot of code since it's part of Inkscape. Ishaan reports that the 1.4.4 release in on schedule and thanks ltlnx (Willy) for their work on backporting. René would like to understand how the python interaction works for macOS for 1.4.4 but this might not be possible in the time. Thank you everyone for coming to this week's meeting.

1 0

Inkscape Developer Meeting Notes, April 11th 2026
by Martin Owens 11 Apr '26

11 Apr '26

Hey everyone, A smaller meeting from last week, but some heavily topics. Event: https://inkscape.org/cals/event/64/ Attending: Martin, karliss, Mikekov, Tav, Elendil Next: https://inkscape.org/cals/event/1/ Tav has been fixing bugs with the svg font, a crash. May have found a problem in the regex where strings are defreferenced. We talk about cairo bugs and the need to pay close attention to the versions. We will need fixes from code that isn't even released upstream yet, but we should still specify we need that future version. With a critical problem hanging around for developers, testers and future 1.5 users. Elendil talks about the poppler drama with versions constantly needing IFDEFs and Martin talks about the problems with the upstream poppler design making font interfaces private. We have an indepth discussion about the problem with depedencies and upstreams and how receptive they are to accommodating our needs. Karliss has been doing some pixel alignment and some other issues. Martin has been working on unpicking the Gradient paint servers and trying to allow the rendering engine to work without needing to include th whole of the SPObject tree and SPStyle. These includes prevent proper unit testing. Mikekov has been working on his object properties dialog. We had a longer discussion about the AI policy. Mikekov will probably stop contributing to the project if and when we stop allowing AI contributions. He believes that the project would benifit from having merge request reviewed by AI and that the capabilities of these tools are quite a bit more capable than the project believes. Martin has his own issues with AI and a frank discussion is had about some of the many dimentions involved in the discussion. We get into merge requests and the history of trying to have enough people to review code being submittd to the project. If a proposal is before the project about using a new tool, then we'd be looking for something ethical, pen source and capable; no matter what it's tooling was. Thanks everyone for coming to this week's meeting!

1 0

Developer Meeting Notes; April 3rd 2026
by Martin Owens 03 Apr '26

03 Apr '26

Hey everyone, Big thanks to everyone who was able to join our dev meeting which happened at the same time as the PLC meeting. Event: https://inkscape.org/cals/event/64/ Attending: René, Karliss, Charlotte, mgmax, Philip, Tav, Martin, Vaibhav Next Meeting: https://inkscape.org/cals/event/1/ René; update on hiring process for bugfix accelerator, applications are coming in and official deadline is soon. Also updating dependencies for Win/macOS for upcoming release. Pointless reporting of security issues by AI agents who don't understand the codebase and can't answer questions about the problem. MC provides access to Windows ARM, René will test that will look into offloading Windows CI with budget of 150 Euros to pay for cloud provider. Tav asks if Inkscape will run on new "cheap" mac, Rene says yes it should, the architecture is the same. Less powerful chip and 8 GB RAM, but Apple claims to use memory more efficiently. Vaibhav has no problem on 8 GB low-end macbook. Charlotte, has been doing some windows testing. Failed successfully. Philip tried to do a fix for this specific static threading windows bug but it's quite a difficult problem. A couple of people have tried to fix it. He thanks Charlotte for the testing and asks follow up questions. Has a limited machine to do development on. Karliss, pixel alignment related issues. Links us to an issue related to security issues which allows svgs with unsafe problems in them. Philip has looked into this issue too; it's partly an upstream issue in Gio and similar pathways. Martin asks if it's possible to deal with the issues on a case by case basis. Tav; getting back into some old merge requests. Some specific libnrtype changes which now have a merge conflict thanks to some recent unit testing work. mgmax has an old merge request which they were asked to present at the meeting. It's a windows script which sets up everything for you. (https://gitlab.com/inkscape/inkscape/-/merge_requests/7557) it also allows a virtual machine for testing. Martin has already looked into it, and wants to get more visibility on the changes. René wants to know if this is based on the CI toolchain which is based on making a fixed image with better stability. This new script uses the developer instructions which is a rolling release. Max and Rene will get into contact to get the MR merged and to get the scripts and documentation changed to use a fixed MSYS image instead of rolling release. Vaibhav has continued their performance optimisations, in existing issues. An LPE specific issue using simplify (https://gitlab.com/inkscape/inkscape/-/merge_requests/7833) the reason is because of update signals that happen on every selection change. Martin has been working on text on path and text in shape to catch up to the big improvements in Krita text handling. Tav and Martin talk about the SVG spec support, and Krita using SVG for their vector layers giving the spec two implemetations which is important. The new text functionality is basically ready, but will need more testing. We talk about the steam issues and move on to talking about our relationship with the SFC. See you all next time! Best Regards, Martin Owens

1 0