mirror of
https://github.com/wezm/wezm.net.git
synced 2024-10-22 10:02:32 +00:00
Compare commits
No commits in common. "b834c3401983e96f0b2b9b5160118c7626853915" and "89902e72e6f482e652970d206c711d52a5a73af1" have entirely different histories.
b834c34019
...
89902e72e6
3 changed files with 2 additions and 113 deletions
|
@ -8,8 +8,8 @@ description = "Posts by Wesley Moore"
|
||||||
# Whether to automatically compile all Sass files in the sass directory
|
# Whether to automatically compile all Sass files in the sass directory
|
||||||
compile_sass = true
|
compile_sass = true
|
||||||
|
|
||||||
generate_feeds = true
|
generate_feed = true
|
||||||
feed_filenames = ["rss.xml"]
|
feed_filename = "rss.xml"
|
||||||
|
|
||||||
# Whether to build a search index to be used later on by a JavaScript library
|
# Whether to build a search index to be used later on by a JavaScript library
|
||||||
build_search_index = false
|
build_search_index = false
|
||||||
|
|
|
@ -1,39 +0,0 @@
|
||||||
+++
|
|
||||||
title = "How Much Is a Browser Worth?"
|
|
||||||
date = 2024-07-03T17:22:01+10:00
|
|
||||||
|
|
||||||
# [extra]
|
|
||||||
# updated = 2024-06-04T07:49:36+10:00
|
|
||||||
+++
|
|
||||||
|
|
||||||
Apparently people are excited about funding independent browser efforts this
|
|
||||||
week. I have little interest in funding yet another browser built in C++ in
|
|
||||||
2024 but [Servo] is still alive. Since Mozilla refuse to let us directly
|
|
||||||
fund Firefox I shall set up a recurring donation to Servo.
|
|
||||||
|
|
||||||
The next question is how much is a web browser worth to me? Based on minutes
|
|
||||||
spent using a browser, quite a lot!
|
|
||||||
|
|
||||||
<!-- more -->
|
|
||||||
|
|
||||||
Once upon a time you could pay for Netscape. It seems that [in 1996 Netscape
|
|
||||||
3.0 cost US$49][netscape3]. Adjusting for inflation that's US$97 in 2024. Let's
|
|
||||||
round that up to an even US$100. At the time of writing [Numbat] tells me
|
|
||||||
that's AU$150 or AU$12.50 per month, which seems reasonable.
|
|
||||||
|
|
||||||
Some other services for comparison (I don't actually have an active YouTube
|
|
||||||
Premium subscription):
|
|
||||||
|
|
||||||
- Netflix: AU$18.99/m
|
|
||||||
- YouTube Premium: AU$16.99/m
|
|
||||||
- JetBrains Rust Rover: US$129/y (AU$194, AU$16/m)
|
|
||||||
|
|
||||||
A browser is a least as useful as these, almost certainly more. So, I've set up
|
|
||||||
a US$15/m (AU$22.50) recurring sponsorship on GitHub. I encourage you to [do
|
|
||||||
the same][sponsor-servo].
|
|
||||||
|
|
||||||
[netscape3]: https://web.archive.org/web/19961115062838/http://merchant.netscape.com/netstore/NAVIGATORS/STANDARD/STANDARD_ITEMS/leaf/product1.html
|
|
||||||
[sponsor-servo]: https://github.com/sponsors/servo
|
|
||||||
[Servo]: https://servo.org/
|
|
||||||
[Numbat]: https://numbat.dev/
|
|
||||||
|
|
|
@ -1,72 +0,0 @@
|
||||||
+++
|
|
||||||
title = "Why Chimera Linux"
|
|
||||||
date = 2024-07-03T21:09:53+10:00
|
|
||||||
|
|
||||||
#[extra]
|
|
||||||
#updated = 2024-06-04T07:49:36+10:00
|
|
||||||
+++
|
|
||||||
|
|
||||||
I received a reply to my [Tech Stack 2024](@/posts/2024/tech-stack/index.md)
|
|
||||||
post asking: Why Chimera Linux? I wrote a reply that turned out longer than
|
|
||||||
anticipated and figured I may as well post it here too. I'm not trying to
|
|
||||||
convince you to use Chimera with this post, just note down why it appeals to
|
|
||||||
me. That's really the crux of it: there's dozens of distros out there all with
|
|
||||||
different goals and values and Chimera really speaks to me, for you it might be
|
|
||||||
something else.
|
|
||||||
|
|
||||||
<!-- more -->
|
|
||||||
|
|
||||||
---
|
|
||||||
|
|
||||||
I like [Chimera Linux] because it's the closest distro I've found to what I would build
|
|
||||||
if I was building my own (something I've tinkered with a few times over the
|
|
||||||
years).
|
|
||||||
|
|
||||||
I like that it is a comparatively small and easy to understand system without
|
|
||||||
giving up quite as much as you do with Alpine Linux, which to be clear I like
|
|
||||||
as well and use on my server. The [userland from FreeBSD][userland] is capable and easy
|
|
||||||
to understand code wise, [Dinit] provides an init system with process monitoring,
|
|
||||||
dependency tracking, and a service file format that doesn't require writing
|
|
||||||
shell scripts like in FreeBSD and Alpine. [musl] is designed to be secure and
|
|
||||||
uses quite straightforward implementations of libc functions while sticking
|
|
||||||
closely to the POSIX standard. See the recent OpenSSH vulnerability where
|
|
||||||
[remote code execution was not possible on musl based systems](https://fosstodon.org/@musl/112711796005712271).
|
|
||||||
|
|
||||||
I really like the [cbuild] system for building packages. It uses a real
|
|
||||||
programming language (Python) to define packages and share library code, which
|
|
||||||
makes them easier to write and understand over Make and shell based packaging
|
|
||||||
systems. Packages are built in an isolated sandbox, preventing them depending
|
|
||||||
on the host system accidentally—this is definitely an advantage over building
|
|
||||||
packages on Arch. Most run-time dependencies are automatically determined so
|
|
||||||
you don't have to list all those out in the package template.
|
|
||||||
|
|
||||||
[apk] is fast (although not as fast as Pacman when doing updates). It has a
|
|
||||||
clever way of tracking packages where [the world file][world] specifies all the packages
|
|
||||||
that should be present and it uses a solver to determine what needs to be
|
|
||||||
installed/removed. The neat bit is that when you `apk del` a package it can remove all
|
|
||||||
packages that are no longer specifically requested, whereas in Arch it's easy
|
|
||||||
to end up with orphaned packages that are dead weight and require [manual
|
|
||||||
maintenance to clean up](https://wiki.archlinux.org/title/Pacman/Tips_and_tricks#Removing_unused_packages_(orphans)).
|
|
||||||
|
|
||||||
Chimera is also a rolling distro (like Arch) so things stay up to date. There's
|
|
||||||
a low barrier to submitting new packages and updates, you don't have to be a
|
|
||||||
special "committer", you can just open a pull request. Packages in the repo are
|
|
||||||
built automatically with a build bot server for all supported architectures,
|
|
||||||
whereas I believe Arch is still working towards automated packaging.
|
|
||||||
Additionally first class support for multiple CPU architectures allows me to
|
|
||||||
run the same system on different devices I use such as Raspberry Pis, RISC-V
|
|
||||||
single board computers, and hopefully eventually my new ARM based Snapdragon X
|
|
||||||
Elite laptop.
|
|
||||||
|
|
||||||
Having written that all out I guess Chimera feels like a distro that is
|
|
||||||
full-featured but also simple enough that you can poke around and understand
|
|
||||||
all the parts. It's also easy to get involved with the project.
|
|
||||||
|
|
||||||
[cbuild]: https://github.com/chimera-linux/cports/blob/master/Usage.md
|
|
||||||
[cports]: https://github.com/chimera-linux/cports
|
|
||||||
[musl]: https://musl.libc.org/
|
|
||||||
[Dinit]: https://davmac.org/projects/dinit/
|
|
||||||
[apk]: https://gitlab.alpinelinux.org/alpine/apk-tools
|
|
||||||
[userland]: https://github.com/chimera-linux/chimerautils
|
|
||||||
[Chimera Linux]: https://chimera-linux.org/
|
|
||||||
[world]: https://chimera-linux.org/docs/apk/world
|
|
Loading…
Reference in a new issue