Tagging a new Dia release?

Hello,

Last time Dia got an official tag, it was in 2011 (0.97.2). For some reason, an untagged 0.97.3 Dia release has been pushed to the Gnome download center ( Index of /sources/dia/0.97/ ) in 2014. Since then, a lot of useful patches/fixes have been merged into the main tree but none of the new contributions received the honor to appear in a new release yet.

The problem is that most mainstream distributions do package the latest release and not the latest HEAD, so they are stuck with a 12 years old -slightly buggy- version. Only rolling distributions such as ArchLinux or Gentoo embed the latest (stable/fixed) version of Dia.

As an example, Dia has issues generating SVGs with PNGs inside, this has been fixed upstream for quite some time. One solution would be to ask each and every package maintainer of every distribution to cherry-pick a number of patches, but it would be much cleaner for the whole ecosystem if Dia official version could be bumped.

Thanks a lot for considering this request,

Miquèl

Hi and welcome to the forum. Problem with Dia isn’t the lack of releases, but more the fact that it’s dead. No one maintains it.

Edit: actually I take it back, it seems @zbrown have recent commit history in Dia. What’s your take here?

The problem is Dia isn’t really in a state suitable for stable release, I certainly share the frustrations outlined and Debian et al shipping random tags certainly isn’t optimal, but it’s tricky to know what else to do there.

FWIW I do have, to my knowledge at least :sweat_smile:, a somewhat good relationship with the Debian and Gentoo maintainers, which has been largely mutually beneficial.

Hello, thanks a lot for your answers!

May I challenge a bit “Dia isn’t really in a state suitable for stable release”? At the moment, people using Dia (and there are still users) are unfortunately defaulting to a so called stable release that is much less stable and much more buggy than the one available on the tip of the repository. There have been efforts to stabilize it further and patches have been reviewed and merged; why not allowing these extra minor changes to be released and packaged? I must say that I do not understand the rationale at the moment. Dia is not under active development, I understand, and maintaining tools in the open source community is a load, but how is status quo better than making a a more recent release on top of what already exists (assuming it is not a too complex process and you still have the tools for that, of course)? The release note does not need to sell new fancy features, it’s just a stabilization step that is needed.

For the record, we share training slide decks and their sources, we use Dia to draw figures. People can download the sources and recompile the decks themselves, which regularly raises issues that have been solved a decade ago.

Hello,

I support Miquel’s point (I also work on the training slide decks he is mentioning). To add a bit more details, I stumbled upon the PNG embedding issue mentioned above in a Fedora 44 container, which is packaging this not-really-released 0.97.3 with a few out-of-tree patches. I managed to get through the issue by re-building DIA in this very same container, on the same SHA1 as the one currently used by the `dia-git` package from Arch (pdf-import: poppler broke API again (3ce95975) · Commits · GNOME / Dia · GitLab), without any OoT patch on top of it.

Not sure what point about slide decks you are referring to?

If you want to use newer Dia on Fedora might getting the flatpak from gnome-nightly be sufficient?

I replied 5h ago, my message was visible, Alexis wrote his answer, then my message got flagged SPAM by “the community” (?). So some context is now missing. Can a moderator (@ebassi ?) please have a loot and unmark it?

A few messages in this topic got flagged automatically for review because they were written by different users but shared the same IP address. They should all be visible, now.

I agree with the others, why not just do a simple release update to include the latest fixes and updates because if the program would be more useful, why not? I have been using DIA for years and I find it easy to use and I have even started making some of my own objects to draw with. For what it does it is a very fun program that can do a lot more.

Hello @zbrown , can you now access the post that got shadowed for a few hours? Does it clarify a bit the need for that stable release?

Whoops, yes sorry heatwave, infra issues, IMAP nonsense, all got a little silly there. So sorry about all that.

This is certainly a fair point, my comment came from the perspective that it’s still mid-gtk3 port (the current target is loosely gtk3.4, or about 2012, ideally we’d get to 3.24, if not gtk4…) and there’s various things I really do keep meaning to get ironed out, which puts me in a position where I’m not really comfortably saying ‘this is a shippable product I can support for $time’. But on the other hand I do see what you are getting at there, that even this incomplete state is more useful in 2026 than the last release.

Yes contra to what @tragivictoria said Dia’s not dead, it is ‘maintained’ and I’m committed to avoiding total bit-rot, but I do also have to be upfront in that it’s pretty much just me, and I’ve simply not been able to put in as much time as I’d like. I really do regret this, and can only apologise.

Oh nice, that’s neat. Great to hear your using Dia and finding it useful :purple_heart:.

That’s a really nice use-case.

The problem, from my angle, is that the main branch, where all those fixes exist, is also where we’ve gotten someway into porting to gtk3 (the last release was gtk2 based) amongst other things. A ‘simple bugfix release’ isn’t really viable.

I am wondering though, especially as with how Debian has been packaging Dia, if perhaps a sorta ‘snapshot’ release that Fedora and others could package would be a way forward? That way I’m not ‘making promises’ in the way that a full release, or even say a beta, might imply, but still gives distros something to synchronise to?


What do we all think, would a ‘Dia Snaphot 2026’ style tag suit everyone?

Apologies again for this. Should had checked it better

@mraynal @Test_Droidx86 @Tropicao any thoughts?