https://www.lexaloffle.com/bbs/?pid=183169#p @mayazimmerman@gamedev.lgbt https://gamedev.lgbt/users/mayazimmerman/statuses/116053644566856190

also please boost and/or recommend any cool capsules ❤️

falling into a rabbit hole…
at least my html pages mostly text and easy to convert to #gemini

i try to push shell as far as it reasonably goes. for most tasks, sh, find, grep, sed, and cut is more than enough. only when things become truly complex or painfully slow do i reach for another language.

i used to care a lot about strict portability, but in practice i only run my scripts on macos and #openbsd. that simplifies things: if shellcheck is happy, i'm happy. i'm 99% sure my scripts work on other unix-like systems, but i don't feel the need to check.

what really draws me to shell is that it's always there. it's part of the base system, requires no extra installation, and the runtime has been stable for decades. that stability translates directly into confidence: shell scripts feel future-proof.

i know i can run something like ssg.sh ten years from now and it will still work --- certainly on #openbsd, and hopefully on macos too. there's no dependency churn, no worrying about the "right" version of python or ruby, and no hoping the ecosystem hasn't moved on.

it just runs.

Alt...

built with /bin/sh

i'm thinking how to make smallest possible backups...

html_count: 5
html_size: 1102841
ico_count: 1
ico_size: 1218
id_count: 2
id_size: 10
idx_count: 10416
idx_size: 633369
json_count: 12490
json_size: 50926543
lst_count: 1
lst_size: 33
pid_count: 1
pid_size: 6
png_count: 9
png_size: 557101
tag_count: 2180
tag_size: 21400
txt_count: 7
txt_size: 88
other_count: 669
other_size: 257034

*_size in bytes

but over 30% of nodes have >1 hardlinks

#snac2 is very unixy :)

+260% file count
+16% space

$ du -hd0 /var/snac/data && find /var/snac | wc -l
71.5M	/var/snac/data
   26230

i use shell scripts to configure and deploy a few #openbsd servers:

• generate configuration files locally for each server (sshd, httpd, relayd, acme-client, nsd, and now #snac2)
• deploy with ssh or rsync (set permissions, time)
• clean up if needed
• run health checks that all configurations applied correctly (testing with nc, dig, curl, etc)

$ cat data/queue/1770399482.863468.json
{
    "type": "output",
    "message": {
        "@context": "https://www.w3.org/ns/activitystreams",
        "type": "Follow",
        "id": "https://[redacted]/pub/relay/d/1770399362.561994/Follow",
        "actor": "https://[redacted]/pub/relay",
        "object": "https://fedi-relay.gyptazy.com/actor"
    },
    "retries": 1,
    "ntid": "1770399482.863468",
    "p_status": 500,
    "inbox": "https://fedi-relay.gyptazy.com/inbox",
    "keyid": "https://[redacted]/pub/relay",
...

Feb  6 17:37:47 [redacted] snac[82302]: 17:37:47 queue not yet time for ~/queue/1770399482.863468.json [1770399467]
Feb  6 17:37:50 [redacted] snac[82302]: 17:37:50 queue not yet time for ~/queue/1770399482.863468.json [1770399470]

https://romanzolotarev.com/pub/hi

becomes

https://romanzolotarev.com/pub/hi.rss

how cool is that? 😎

• 12 posts
• 60 following
• over 9000 files (sic!)

$ du -hd0 /var/snac && find /var/snac | wc -l
61.8M	/var/snac
    9532
$

$ curl -s https://romanzolotarev.com/pub/hi.rss | wc -c
    9999

see also
https://comam.es/snac-doc/snac.5.html

# rcctl stop snac 
# cd /var/snac
# tar -cpf /tmp/snac data
# rcctl start snac

# rcctl stop snac 
# cd /var/snac
# tar -xpf /tmp/snac.tgz
# rcctl start snac

any plans to use sqlite or some other database?

also thank you for creating #snac2. i love it!