Felix Palmen :freebsd: :c64:

Dipl.inform., FreeBSD ports committer, musician, C64 fan/coder, hetero cis white male, not proud, no wing, no ally, leaning liberal/ecologist, pro facts/science

FreeBSD: https://people.freebsd.org/~zirias/
C64 scene: https://csdb.dk/scener/?id=25812
Github: https://github.com/Zirias

Last Notes

2025-06-16 06:11:26 UTC

- reply

plus there's
<optionsname>_SET
and
<optionsname>_UNSET
so you can configure individual ports. The <optionsname> is typically constructed as
<category_portname>
and you can also check it e.g. on freshports. Example:
www_firefox_UNSET= PULSEAUDIO
😏
I personally don't use the TUI-configurable port options at all, to me it's easier to have all the config in one place. It's a matter of taste of course.

2025-06-14 10:28:55 UTC

- reply

try #poudriere to avoid watching shit scroll by. Or more seriously: building directly from the #ports tree will probably always be possible (if just for historical reasons), but opens up lots of opportunity for breakage. poudriere uses #ZFS and #jails to provide #reproducible builds in clean environments.
#FreeBSD

2025-05-04 13:40:52 UTC

- reply

Please don't get me wrong, but I think you should check the link I gave.
In a nutshell, what you say is PART of the reality. There are areas where they excel in their measures and practices, and others where they are way behind (the lack of CI, branching, even basic things like verbose and descriptive commit messages, is e.g. in the "not so nice" area...)

2025-05-04 06:26:17 UTC

- reply

This "#OpenBSD is secure!" claim always annoyed me a lot, mainly because it doesn't tell anything: #Security in IT can only ever be defined in a context of #threat models. Without that, it's meaningless. Somewhat recently, I discovered this:
https://isopenbsdsecu.re/
I should warn it uses some sarcasm and other confrontative language in some parts, unfortunately. But it seems to be a pretty professional analysis and assessment of (mostly) the "mitigations" OpenBSD provides in an attempt to counter "typical" attacks by at least making them harder.
I should also add that I consider this a very interesting and helpful read, and still consider OpenBSD a great project that came up with lots of great stuff (I recently used their #bcrypt code after doing some research on password hashing, for example). And I don't agree with every single criticism on that page either. I just think it's important to build assessments whether something "is secure" on a serious analytical foundation.

2025-03-27 07:11:28 UTC

- reply

Only very few ports are never built by the package builders, typically for legal reasons. ungoogled-chromium is not one of them. On freshports, you'll find a table with available binary packages of every port. For ungoogled-chromium, it's currently indeed missing, and the most likely reason can be seen here:
https://portsfallout.com/fallout?port=www%2Fungoogled-chromium%24
(There's a very recent build failure on amd64 for latest).
Regarding wayland, it's no doubt great to have software reliably ported, but I really dislike the whole concept, part of which is there can't be a "window manager", because you have to implement a compositor instead, responsible for the final compositing of windows, and all rendering of window contents is delegated to each and every client ... both things the X server can do for you. My personal "resistance" was to learn last year how to create an X client "from scratch" (without toolkits) and actually use the rendering facilities X has to offer 😏

2025-03-27 05:56:59 UTC

- reply

There's generally very little need to build from ports yourself when binary packages are available for your target architectures. Some people still like the (IMHO flawed) idea that building everything with "optimizing" CFLAGS (e.g. targeting the specific CPU) would result in better overall performance. That's been historically a widespread idea among Gentoo users. In reality, there's very little code that could significantly profit from that, and for code that does, there's typically a better solution already in place, like runtime CPU detection. And it comes with tons of potential issues, breaking lots of things. So, the only sensible need for building yourself would be when you need non-default port options and similar.
The next issue is, you can easily break stuff by just mixing ports and packages, most obviously because the packages were build from a different state of the ports tree, but also incompatible options can lead to issues.
Therefore, if you really want to build yourself, I recommend setting up poudriere. You'll probably find lots of guides. If you use the "poudriere-devel" package, you get a newer version that can just fetch binary packages *when* they perfectly match what you build, saving you a lot of time for building a whole package repository.

2025-03-26 15:26:59 UTC

- reply

Actually, (almost) every #FreeBSD #port should also be available as a binary #package. When one is missing, this typically indicates a build failure on the latest bulk run. Trying to build it yourself won't help here unless there was a fix committed to the ports tree meanwhile, in which case the package will also "reappear" within few days.

2024-11-12 14:45:48 UTC

- reply

I never went as far as actually using AI-generated code. I thought it's fun to play around and see what I get, so quite a while ago, I showed GPT this interface (from my own code):
https://github.com/Zirias/dos2ansi/blob/master/src/bin/dos2ansi/stream.h
Then I asked it to implement me an example "StreamWriter" that just logs write operations. I immediately got some code that had undefined behavior (crashing). I gave it a gentle hint (this doesn't look quite right because ...), and it told me "thank you, you're right, here's the fixed code" 😅. Well ...
In a nutshell, 2/3 still sounds like a net gain, but be prepared for more debugging of bugs "better hidden" 😏

2024-10-02 16:29:51 UTC

- reply

Maybe I fundamentally misunderstood how the home feed works in #mastodon? I expected it would show me *every* toot from people I follow. Yet here's another one that doesn't appear:
https://snac.bsd.cafe/zirias/p/1727886222.245670
Yes, it's an answer again, without tagging anyone, but I wonder how that could be related, I mean I'm following this account (myself on the snac instance), not the "random" person I might or might not tag when answering. I'm confused 😂

2024-10-02 15:15:08 UTC

- reply

Hm ... could it be related to whether something is an answer or not? I *think* it was answers I didn't see, and I'm pretty sure (?) Mastodon is showing me answers written by other people I follow. 🤔

2024-10-02 10:44:14 UTC

@npub1ef5…gcpc something seems a bit strange with the #snac instance: I followed myself now (both directions), and quite some of my own toots on snac don't arrive in my feed here on #mastodon. But when looking at the snac profile from mastodon, I can read them... 🤷