Posts by bugaevc@floss.social
(DIR) Post #AoPGygPynBYlo4KlLE by bugaevc@floss.social
2024-11-25T16:17:09Z
1 likes, 0 repeats
New week, new round of wrestling against GTK layout
(DIR) Post #AqsIVQyErxdJ0Qe45Q by bugaevc@floss.social
2025-02-07T08:35:27Z
1 likes, 0 repeats
"zero-knowledge proof" is what happens when I manage to prove a statement in Lean by repeatedly applying standard tactics without having any actual understanding of the problem domain
(DIR) Post #Asa4OA2CTWRIJrt62q by bugaevc@floss.social
2025-03-28T15:18:40Z
1 likes, 0 repeats
$ pldd 9345593455:/usr/bin/sshdlinux-vdso.so.1/lib64/libcrypto.so.3/lib64/libz.so.1/lib64/libc.so.6/lib64/ld-linux-x86-64.so.2look ma, no libsystemd, no libxz (and yet Type=notify).
(DIR) Post #Asgdruess0KafAfCVM by bugaevc@floss.social
2025-04-02T16:56:00Z
3 likes, 1 repeats
Replaced OpenRC with systemd, now my phone no longer boots👌this-is-fine.service
(DIR) Post #AtAq6GSUVAq9UHnaYi by bugaevc@floss.social
2025-04-04T07:25:29Z
0 likes, 0 repeats
What systems developers want:🤷 algebraic data types🤷 trait-based generics🤷 Hindley–Milner type system❌ npm-style package ecosystem❌ language-specific build system✅ shared libraries, dynamic linking✅ stable, well-defined ABI✅ precise control over symbols (exports, versioning, visibility, relocations)✅ seamless FFI with C and C++
(DIR) Post #AuMZdZCktcGyWjlEIK by bugaevc@floss.social
2025-05-22T20:19:13Z
1 likes, 0 repeats
@fale what about accessibility?
(DIR) Post #Av5D9UFzOlVpWRI7EW by bugaevc@floss.social
2025-06-11T19:02:05Z
0 likes, 1 repeats
Seriously, any post/announcement talking about removing X.Org should not argue with made up strawman, but instead talk about accessibility. If you think removing X11 is still worth it despite the accessibility regressions, freaking own it.
(DIR) Post #Av68plXDFFmUereEHQ by bugaevc@floss.social
2025-06-13T12:31:00Z
0 likes, 0 repeats
@tbernard we should look into cutting down the size of our apps too. 5-10 MB is large for when the platform (GTK, GLib etc) is the one doing the heavy lifting.
(DIR) Post #Av68pndfQE8HBSgPcO by bugaevc@floss.social
2025-06-13T19:14:36Z
0 likes, 0 repeats
@tbernard I don't think anyone did. And with some tongue in cheek, the current practices around the Rust programming language unfortunately result in overblown binary sizes; and Rust is increasingly popular in GNOME app dev. And again I don't think anyone but me actually cares.Could we have like a GNOME-wide initiative for this?
(DIR) Post #AvFvEjLt2dGFLnxYCu by bugaevc@floss.social
2025-06-18T12:42:15Z
0 likes, 0 repeats
@domi what I always found weird is that Java doesn't use ELF for its object file format (like Guile does for instance), given both has ties to Sun.
(DIR) Post #AwQZUkUAykmI7fQdbE by bugaevc@floss.social
2025-07-23T14:03:34Z
1 likes, 0 repeats
(DIR) Post #AwUMJNarEJ2nANdrf6 by bugaevc@floss.social
2025-07-25T10:07:05Z
0 likes, 0 repeats
peel was born out of my frustration trying to write idiomatic :gtk: in gtkmm.I felt that, while gtkmm was indeed letting me use GTK in a way that felt very C++'y, it was severely limiting what I could do, and was making my widgets feel extremely third-class, compared to the C-written ones in GTK proper — I went into some more details about this in the talk.
(DIR) Post #AwUMJPMkdiwsfD3cx6 by bugaevc@floss.social
2025-07-25T10:07:42Z
0 likes, 0 repeats
And yet C++ so clearly had useful language features like closures and smart pointers and class syntax that could be used to significantly clean up the C GObject code.So I had these various ideas about how it could be possible to do things better in various aspects compared to gtkmm, to make use of C++ language features while not introducing any limitations or runtime overhead.
(DIR) Post #AwUMJQG3K7BVQivn9M by bugaevc@floss.social
2025-07-25T10:09:04Z
0 likes, 0 repeats
Perhaps we could do custom classes like this. Perhaps something like this could be done for signals. Perhaps we could have a RefPtr that actually managed GObject reference count.But these ideas, they were largely theoretical (I'd spend some time prototyping them in Godbolt at best), and while they were covering various aspects, they were not connected to one another, they were just bits and pieces, and a larger picture was missing.
(DIR) Post #AwUMJQrH5ixVIALWAS by bugaevc@floss.social
2025-07-25T10:10:31Z
1 likes, 0 repeats
It took a lot of design work to combine them into one cohesive vision, and to actually implement things, and make them work good enough in practice, to the point where peel could actually supplant gtkmm.But this now exists and works, and it turned out better than I could dream of.
(DIR) Post #AweRqd9F3Ro8JfokAy by bugaevc@floss.social
2024-05-25T14:48:57Z
0 likes, 0 repeats
Remember to hydrate, fellow humans 💧
(DIR) Post #AweRqelZ2S3XKokrY0 by bugaevc@floss.social
2025-07-30T07:02:51Z
0 likes, 0 repeats
Fellow humans, remember to hydrate!
(DIR) Post #AwpWI14gk0pothYpxg by bugaevc@floss.social
2025-08-04T13:48:10Z
0 likes, 0 repeats
Do you use Git CLI, despite there being a large variety of UI apps for Git? (So do I!)Why is that? What are they all missing? What would make a Git UI app work for you?Is it "not exposing important Git concepts"? (which ones?) "Dumbing things down"? (how?)
(DIR) Post #AwpZblWkYUfoKenMg4 by bugaevc@floss.social
2025-08-04T15:50:44Z
0 likes, 0 repeats
@lanodan okay yeah, remote dev is a good point. So potentially this would be better if more UIs for Git supported remote repos over SSH or something?
(DIR) Post #AwpaKPNZkQ22fln7fk by bugaevc@floss.social
2025-08-04T16:01:16Z
1 likes, 0 repeats
@lanodan yes, but actually no. Git does a lot of file access (naturally), both in the working tree and inside .git, and doing it sequentially is going to be dog slow over sshfs. You either need to run a backend (e.g. git CLI) on the remote, so it can run its logic there and only send the useful output (e.g. the list of modified files) over the network, or have some latency-aware Git implementation that would access all the files concurrently.And yeah good point about the toolchain too.