Got all my site files as html now (with .gz versions for OpenBSD’s httpd), no more php!

@shinyoukai@neko.laidback.moe Not using OpenBSD or httpd? Yeah. It’s been working quite well since ~2017, so, meh, too lazy to switch now. But nothing is set in stone, of course.

@movq@www.uninformativ.de I noticed that your feed’s last modification timestamp was missing in my database. I cannot tell for certain, but I think it did work before. Turns out, your httpd now sends the Last-Modified with UTC instead of GMT. Current example:

Sat, 03 Jan 2026 06:50:20 UTC

I’m not a fan of this timestamp format at all, but according to the HTTP specification, HTTP-date must always use GMT for a timezone, nothing else: https://httpwg.org/specs/rfc9110.html#http.date

I’m so old fashioned that this is the only pod that is behind Apache

And regarding those broken URLs: I once speculated that these bots operate on an old dataset, because I thought that my redirect rules actually were broken once and produced loops. But a) I cannot reproduce this today, and b) I cannot find anything related to that in my Git history, either. But it’s hard to tell, because I switched operating systems and webservers since then …

But the thing is that I’m seeing new URLs constructed in this pattern. So this can’t just be an old crawling dataset.

I am now wondering if those broken URLs are bot bugs as well.

They look like this (zalgo is a new project):

https://www.uninformativ.de/projects/slinp/zalgo/scksums/bevelbar/

When you request that URL, you get redirected to /git/:

$ curl -sI https://www.uninformativ.de/projects/slinp/zalgo/scksums/bevelbar/
HTTP/1.0 301 Moved Permanently
Date: Sat, 22 Nov 2025 06:13:51 GMT
Server: OpenBSD httpd
Connection: close
Content-Type: text/html
Content-Length: 510
Location: /git/

And on /git/, there are links to my repos. So if a broken client requests https://www.uninformativ.de/projects/slinp/zalgo/scksums/bevelbar/, then sees a bunch of links and simply appends them, you’ll end up with an infinite loop.

Is that what’s going on here or are my redirects actually still broken … ?

yggdrasil Network #ygg #httpd #ergo #ipv6

@andros@twtxt.andros.dev Can you reproduce any of this outside of your client? I can’t spot a mistake here:

$ curl -sI 'http://movq.de/v/8684c7d264/.html%2Dindex%2Dthumb%2Dgimp11%2D1.png.jpg'
HTTP/1.1 200 OK
Connection: keep-alive
Content-Length: 2615
Content-Type: image/jpeg
Date: Wed, 19 Mar 2025 19:53:17 GMT
Last-Modified: Wed, 19 Mar 2025 17:34:08 GMT
Server: OpenBSD httpd

$ curl -sI 'https://movq.de/v/8684c7d264/gimp11%2D1.png'
HTTP/1.1 200 OK
Connection: keep-alive
Content-Length: 131798
Content-Type: image/png
Date: Wed, 19 Mar 2025 19:53:19 GMT
Last-Modified: Wed, 19 Mar 2025 17:18:07 GMT
Server: OpenBSD httpd

$ telnet movq.de 80
Trying 185.162.249.140...
Connected to movq.de.
Escape character is '^]'.
HEAD /v/8684c7d264/.html%2Dindex%2Dthumb%2Dgimp11%2D1.png.jpg HTTP/1.1
Host: movq.de
Connection: close

HTTP/1.1 200 OK
Connection: close
Content-Length: 2615
Content-Type: image/jpeg
Date: Wed, 19 Mar 2025 19:53:31 GMT
Last-Modified: Wed, 19 Mar 2025 17:34:08 GMT
Server: OpenBSD httpd

Connection closed by foreign host.
$ 

It should be fixed now. Just needed some unusual quoting in my httpd.conf: https://mail-archive.com/misc@openbsd.org/msg169795.html

@lyse@lyse.isobeef.org Sorry, I don’t think I ever had charset=utf8. I just noticed that a few days ago. OpenBSD’s httpd might not support including a parameter with the mime type, unfortunately. I’m going to look into it.

Quick life hack regarding dead gopherholes: Try to contact the owner. Sometimes there is a httpd listening on the same host which points at a way to contact the owner. So

I don’t like to see #openbsd httpd inserting hardcoded css for autoindexes pages. I’d like my own css stylesheet

Hé oui ^^. Je suis curieux de voir comment ça fonctionne. J’apprécie de voir une configuration très semblable à httpd. L’idée de reléguer les cgi à slowcgi me semble intelligente. Il y a des logs plus courants à parser. Et surtout, ça évite le cumul de relayd+inetd. <@https://im-in.space/web/@solene@bsd.network/112020802069845381>

Je teste gmid pour remplacer vger+inetd+relayd. Reste à configurer le cgi, il faut que je prépare un slowcgi en parallèle de celui pour httpd :/. https://gmid.omarpolo.com #openbsd #gemini

Oh look, someone already wrote patches for httpd headers: https://github.com/mpfr/httpd-plus <@https://im-in.space/@prx/111941320383362603>

There’s something really annoying me on #openbsd httpd: to set custom headers, one need to use relayd. Yes, it works, but it is so complex for this task. And it breaks logs for analysis with goaccess or webalizer: the forwarded log format put the original IP at the end, but there is still 127.0.0.1 at the beginning :/. httpd should be able to set headers. ‘hsts’ instruction already does