Added support for building/running alternate architectures using Qemu/binfmt today.
Replacing a shell script with a C program Show more
It's kind of cute how constructing calls with the GVariant and GDBus APIs in C is kind of more convenient than figuring out how to correctly call D-Bus methods using any of dbus-send, gdbus, and busctl.
On the other hand, command line option handling using GOptionContext to have “program [options] <command> [command-options]” in C is clumsy and painful.
C can be a neat language but string manipulation and undefined behaviour kill the joy of programming.
I thought Emacs opening image files like an image viewer was a bit of overkill... but it's just PERFECT for quickly going down Cairo's test results.
It's probably because I can leave windows in place and replace the images shown in them, much like a window manager that actually doesn't get in your way.
Don't @ me about tiling window managers please 🙉
Okay, so one test gets the Fuck It treatment and I force-regen its reference file; another one gets the XFAIL treatment since I don't know why it gets color fringing, and don't really want to find out right now.
There's no way this Cairo test could ever have worked. It fails if checked out at the commit when it was introduced. The dependencies haven't changed the relevant code.
I'm regenerating the test image the hard-ass way and calling Fuck It.
@catonano I'm debugging extended-blend-mask right now. Glad that you discovered test/output/* :) Need to review test/README to see if it actually mentions that.
@calvin I'm suspecting that. About to try something.
(Annoyingly, it's in the PDF blend modes... which librsvg doesn't care about.)
@catonano Ah! I've already regenerated that test file. Try this branch? https://gitlab.com/federicomenaquintero/cairo/tree/normalize-image-test-suite
It's where I'm seeing if test files just need to be regenerated, or if there are actual bugs to be fixed.
I can't bisect a problem in Cairo's test suite because I can't find an old revision that passes the test.
It *could* have been due to changes in pixman, but that code in pixman hasn't changed in the amount of time since the oldest revision in Cairo that I'm looking at.
Of course with plain pkg-config dependencies, there's no way to tie a git revision to the actual version of the dependency that it was tested with.
I.e. if there was Cargo.lock for C, this would be a lot easier.
programmer horror Show more
Tired: git bisect
Wired: git vivisect
Making more progress. You can push your app as a Flatpak to a device such as this tablet. Still needs more automation, but getting closer to something useful.
you know that saying that goes like
"be there or be square"?
i just realized why it's like that
it's cuz if you're not there, you're not around.
you're not a ROUND
A SQUARE IS NOT ROUND
@catonano thanks; it looks fine!
to conceive of mech recursion, imagine remy from ratatouille piloting linguini piloting a part of voltron
RT @mgattozzi: I've just updated https://t.co/IvA5ln5DKK! It now contains some of the first stories of @rustlang's history: "Why do we have… source: https://twitter.com/rustlang/status/971542000339177472
GNOME co-founder. Amateur woodworker, cook, cyclist, cartographer, gardener.
Follow friends and discover new ones. Publish anything you want: links, pictures, text, video. This server is run by the main developers of the Mastodon project. Everyone is welcome as long as you follow our code of conduct!