2022 annual report – GIMP

Pursuing the newfound custom began a year
ago, right here is my
report for previous yr 2022.

In 2022, we had:

• 1 secure releases (GIMP
  2.10.32)
• 3 growth releases (GIMP
  2.99.10,
  2.99.12 and
  2.99.14).
• 1461 commits on the unstable growth department (2.99.x, future
  3.0) and 276 commits on the secure growth department (2.10.x) of
  the primary repository.
• 87 contributors on the primary repository, together with (some individuals belong
  to a number of classes):
  • 35 builders
  • 47 translators
  • 26 contributors to sources (icons, themes, in-code documentation) or
    construct enhancements
• 7 core builders contributed 10 or extra commits in GIMP’s principal
  repository:
  • Jehan: 649 commits
  • Jacob Boerema: 64 commits
  • Nikc: 50 commits
  • Daniel Novomeský: 25 commits
  • lloyd konneker: 25 commits
  • Lukas Oberhuber: 18 commits
  • Niels De Graef: 15 commits
• 115 commits to babl by 10 contributors, with 3
  builders contributing 10 or extra commits:
  • Øyvind Kolås: 86 commits
  • Axel Viala: 10 commits
  • Jehan: 10 commits
• 138 commits to GEGL by 32 contributors, with 5
  builders contributing 5 or extra commits:
  • Øyvind Kolås: 47 commits
  • Behnam Momeni: 9 commits
  • Michael Drake: 7 commits
  • Thomas Manni: 7 commits
  • Jehan: 5 commits
• 1042 commits to ctx by 2 contributors (principally
  Øyvind Kolås).
• 492 commits in gimp-help (our handbook) by 29 contributors, with 11
  individuals contributing 10 or extra commits (this record mixes documenters, construct
  upkeep and translators):
  • Jacob Boerema: 229 commits
  • Anders Jonsson: 47 commits
  • Rodrigo Lledó: 38 commits
  • Jehan: 28 commits
  • Jordi Mas: 25 commits
  • Tim Sabsch: 19 commits
  • Nathan Follens: 17 commits
  • Marco Ciampa: 16 commits
  • Yuri Chornoivan: 15 commits
  • Andre Klapper: 13 commits
  • Hugo Carvalho: 11 commits
• 178 commits in gimp-macos-build (our macOS construct) by 3 contributors
  (principally Lukas Oberhuber).
• 33 commits within the secure department of our Flathub/Flatpak package deal and 23
  commits on the beta department by 6 contributors, together with 4 core contributors:
  Jehan, Ondřej Míchal, Hubert Figuière and Daniel Novomeský.
• 227 commits to GIMP’s web site (gimp.org, i.e. proper right here) by 10
  contributors (principally Jehan).
• 158 commits to our new developers website
  by 4 contributors:
  • Jehan: 104 commits
  • Pat David: 38 commits
  • Robin Swift: 15 commits
  • Lukas Oberhuber: 1 commit
• 178 experiences mounted and 206 merge requests built-in in our 2022
  releases. A whole lot extra experiences dealt with, triaged, answered to, labored on…
• Many patches contributed by GIMP contributors in numerous different tasks we use
  (not less than GLib, GTK, Cairo, meson, Mirrorbits…) and an uncountable variety of
  points reported by our contributors to different tasks.
• And extra!

In comparison with final yr:

• The whole quantity of labor is sort of comparable, and whereas that tendency had already
  began a yr in the past, the work has clearly been shifting much more in the direction of the
  growth department (future 3.0), which nows accounts for 84% of commits
  (towards 74% final yr), whereas the secure department is absolutely moving into
  maintenance-only mode.
• Much less work on GEGL occurred however extra work on babl. The current work on computerized
  LUT creation and SIMD optimizations explains it.
• ctx stays closely developed.
• Whereas Øyvind and myself nonetheless stay the two heavy-lifters, we get extra individuals
  round clearly pulling their weight. It’s thrilling to see extra contributors keep.
• Jacob is working extra on the documentation which is absolutely rising in high quality.

babl and GEGL

On the facet of our graphics engine, the automatic LUT
creation
for coloration conversion in babl is clearly a giant step ahead, launched in GIMP
2.99.10 (then within the secure model 2.10.32).

On the similar time, all babl, GEGL and ctx bought good SIMD
optimization
which allowed good efficiency boosts.

Øyvind Kolås is absolutely doing a tremendous job, as normal.

It is usually fascinating to notice how the idea of “GEGL plug-ins” took off in 2022.
It actually simply refers to third-party GEGL operations which you merely set up
in a folder and GIMP will see them at subsequent restart, together with all the flamboyant UI,
equivalent to on-canvas preview with cut up view (and after we’ll have non-destructive
layer results, these operations may even be usable!).
Amongst individuals spearheading such neighborhood growth, we must always cite
LinuxBeaver and Liam
Quin.
For anybody , I counsel to learn the 3-part
tutorial written by Liam (“Using GEGL
Plug-Ins”,
“GEGL
Graph”
and “Writing C
Plug-Ins”).

Checking objects ✅ in GIMP roadmap

We proudly checked-off a number of objects within the GIMP 3.0.0
roadmap.

Amoung 2022 achievements, we certainly…

• ✔ ported away from intltool to gettext solely (technical debt cleanup);
• ✔ completed the meson construct: the autotools construct nonetheless exists however is now
  thought-about secondary;
• ✔ completed the final items for multi-layer choice (a transfer started early
  2020).
  together with rewriting fully the interplay within the previously horrible align
  and distribute tool.

These are 3 big items in our roadmap which we fortunately marked as accomplished
(aside from possible bugs).

On the getting nearer facet:

• We almost completed the “Much less floating choice” transfer (some use circumstances stay,
  which we’d like to consider extra).
• The Wayland help remains to be kinda wonky at instances (even disregarding all of the
  points we can’t do something about — equivalent to coloration administration not applied
  but in Wayland —, now we have bizarre windowing points), nevertheless it improved in 2022.
• The API work is absolutely shifting ahead; Lloyd Konneker helped rather a lot on this.
• The GTK+3 port is sort of completed, as we’re dealing with today the final
  annoying warnings (although it’s extra a January 2023 factor!).
• House invasion: good elements of it have been finished for the reason that CMYK push made us take a look at
  particular items of code extra in particulars. Although rather a lot nonetheless must be finished
  and coloration science is at instances a really head-scratching a part of the work.

Now anybody following our development news is aware of
that much more occurred.
This report isn’t going to repeat what we already wrote about in numerous information objects.

One specific contributor to encourage this yr is Nikc who got here to us with a
few patches at first then proposed a Google Summer of Code
project, and determined
to remain round. Due to them, rather a lot occurred for CMYK help in GIMP and our
“House Invasion” mission additionally moved ahead additional.
They’re now a really prolific core contributor. This may solely imply great things for
the future!

Packaging

Clearly our macOS help has by no means been higher: good steady
integration, computerized DMG package deal creation, and now we even bought an Apple
Silicon package!
The standard of upkeep and updates for this package deal is excellent.
Lukas Oberhuber is the one to thank for this. But the bus issue for our macOS
package deal stays extraordinarily low so we at all times welcome extra contributors.

On Home windows facet, GIMP is now formally distributed on the Windows
Store,
after getting contacted by a developer relations crew at Microsoft. That is
nice as too many non-trusted packages was distributed there and now they
appear to have principally disappeared with the official one eclipsing them with its very
good score.

On Flathub (GNU/Linux), the burden is getting lightened as we now bought
automation in dependency model examine, because of Ondřej Míchal. The flatpak
package deal crew can be getting greater, with 4 recurring contributors.

Infrastructure

We additionally bought some infrastructure adjustments, equivalent to our mirroring system, now primarily based
on Mirrorbits. That is one thing I’m planning to speak once more about quickly, so I
gained’t go into particulars.

On neighborhood facet, our mailing lists have been discontinued, collectively with all
of GNOME mailing
lists
whose infrastructure we’re on. We now
recommend 2 boards for the neighborhood:

Web sites

Our documentation
website
is getting a variety of love, because of Jacob Boerema, with computerized updates,
statistics displaying… and naturally, the contents is getting severe scrutiny to
enhance documentation high quality. In comparison with 2021, there was almost double
the variety of commits in 2022, which is revealing of the massive step up.

In the meantime we revived the developers
website
which was in a dire state for over 10 years.

We nonetheless have a pending mission to port the primary web site to the Hugo framework
as effectively. Sadly this might not occur in 2022.

GIMP 3.0.0?

I shouldn’t give dates, so don’t take it as a promise. Possibly it’s only a
silly dream by a silly man:
I’m at the moment planning GIMP 3.0.0 launch in 2023, or not less than our first
launch candidates.

Right here. I mentioned it. If it doesn’t occur, keep in mind that it was not a promise. ????

There may be nonetheless rather a lot to be finished, so I hope I’m not making a idiot of myself. However
sooner or later, not with the ability to launch simply will get irritating. In fact, we
are nonetheless inside acceptable growth durations (GIMP 2.8 to 2.10 took 6
years; we’re nonetheless within the fifth yr since 2.10) however I actually need to get it over with.

Now to get this deadline to work, I’ve determined to delay some components out of
our 3.0 roadmap. In specific:

• Extensions administration: mission expensive to me as I began it and developed what
  is already applied, but to get a secure on-line infrastructure to deal with
  extension search and obtain, we’ll want time.
• Paint Choose software: very superior software, however its contributor, Thomas Manni, is
  not pleased with the efficiency (it requires immediate canvas suggestions to be
  usable) and is at the moment investigating different algorithms.

In the identical time, I’ve been pushing apart some good new code contributed to us
after I notice reviewing it and making backwards and forwards corrections will take us
weeks. For example, a few of you could have seen the good “vector
layers” demo by Nikc (primarily based on
work by Hendrik Growth and Jacob Boerema) on social networks. This gained’t make it
to GIMP 3.0.

This can be a rule which I apply to my very own code. Some individuals would possibly certainly keep in mind
my very own link layer experiments for
occasion, which I ended engaged on 2 years in the past, already for a similar purpose.

These will nonetheless occur, I’m solely shifting these targets away into additional
releases, which I’m explaining within the subsequent part.

Rethinking our roadmaps

This leads me to an organizational work I’ve been doing these days on our roadmaps
and on planification of releases. As much as today, you should have learn rather a lot about
our bi-version planification: GIMP 3.0 for GTK+3 port then 3.2 for
superior non-destructive enhancing.

Whereas this second goal remains to be undoubtedly a giant plan in our roadmap, I don’t
assume that making it once more an enormous growth cycle with dozens of options and
taking a number of years is the wisest factor. This outdated growth mannequin made sense
again within the day, however much less these days in my opinion.

My objective for GIMP is to launch extra typically, with quicker growth cycles, perhaps
much less options directly, but good options at every launch. That is one thing I
had been pushing for, ever since 2014, after I was nonetheless a newcomer (I first
evoked that we must always have the ability to publish new options even in micro variations in
a meeting during LGM
2014).
This finally led to our release policy
change,
ranging from GIMP 2.10.0. And that is what I need to proceed pushing additional.

So my level is that targetting for a “GIMP 3.2” someplace within the distant future
doesn’t make sense anymore. The non-destructive enhancing options, equivalent to
non-destructive layer results, will occur, however will or not it’s GIMP 3.2.0? Or some
3.0.x model as a substitute? We’ll see. It’s all simply numbers anyway. We could doubtless
break this down in smaller releases within the finish.

With this in thoughts, I reviewed our after-3.0 roadmaps into smaller items,
per logical classes of tasks we wish and which will certainly occur.

• Hyperlink and vector layers are actually into a brand new “non-destructive layer
  types”
  class. The code is so effectively superior that it will be a waste and whereas
  these gained’t make it to GIMP 3.0.0, it can undoubtedly turn into one of many prime
  targets instantly after launch. Possibly in GIMP 3.0.2?
  By the way in which, this additionally opens the door to the long-awaited form options:
  with vector layers, we might have non-destructive form drawing.
  I imply, on-canvas shapes must be a vector options to make it proper!
• Non-destructive layer
  effects
  (previously the primary goal for 3.2) is clearly a mission on its personal.
• Macro
  help can be one thing we’ve needed for a very long time and with GIMP 3.0, we
  have began to put the foundations for this characteristic. This could hopefully
  quickly turn into a actuality.
• Animation
  support,
  which as most of you already know is one thing I’ve labored on for years, must
  be in GIMP sometime. So it’s additionally its personal class. It’ll additionally deliver
  multi-page help (not simply layers as pages).
• Our extension platform
  remains to be very a lot deliberate!
• The Space Invasion
  mission will proceed: for 3.0, we deal with correctness of coloration fashions we
  already help; after 3.0, we’d look into going additional with new coloration
  fashions backends, equivalent to core CMYK or L*a*b* help, but in addition spot coloration
  channels and whatnot…
• We now have now a bunch of unfinished tools
  in our playground space and it will be good if we took the time to complete
  them. In fact, we even have concepts for good new instruments. And at last there are
  instruments which we actually need to enhance, equivalent to our Textual content software which deserves
  extra love.
• Lastly now we have began to boost the idea of
  “canvas“, with the “Present
  all” characteristic since GIMP
  2.10.14.
  We at all times needed to go additional, and in addition to transform the idea of layer
  dimension (e.g. with auto-growing layers, and even infinite layer talents).
• …

And that is how I fully rewrote our roadmap web page. Hopefully some individuals
will take pleasure in studying the brand new web page and can discover it thrilling. Notice that contents
didn’t change that a lot, besides that it has been reorganized to place extra
emphasis on the larger strokes for GIMP evolution after GIMP 3.0 launch, making
it extra apparent (hopefully) which route present contributors are pushing
GIMP to go.

That is the place we’re at. I’m anticipating 2023 to be an eventful yr. 2022 has
been fairly superior too, but in addition tiring to the purpose that there have been weeks when
I couldn’t work on something, particularly quickly after coding bursts for releases. I
additionally targeted a bit extra on getting more healthy work habits, equivalent to working with a
height-adjustable desk (for sitting and standing work periods) and doing
common walks.

That is additionally why I work on procedures to get quicker releases, higher
infrastructure and higher documentation for onboarding new contributors. I’m
aiming for a extra organized path whereas retaining the marginally ????️ chaotic ❤️‍???? core
which actually makes working in our crew so gratifying. ☺️

As I used to be saying in final yr’s report, GIMP isn’t solely a Free Software program, it
can be a Neighborhood Software program: random human beings doing one thing good
collectively and sharing it with everybody. Why? As a result of we are able to, as a result of we wish. And
that’s why I like our small neighborhood, with simply the correct amount of chaos and
madness, sparkled with simply the correct amount of group!

Lastly don’t overlook you may ???? donate to the project and personally fund several
GIMP developers, as a strategy to give again
and speed up the event of GIMP. As you already know, myself as
maintainer of GIMP (by “ZeMarmot” mission) and Øyvind as maintainer of GEGL
are crowdfunding the work this report is about.
Any help is appreciated to assist us achieve such endeavour.

I want you all a cheerful, humorous ???? and wholesome yr 2023 and/or yr of the rabbit ????!