There are a few toots which I feel need to be re-boosted from time to time, because just boosting them once is not enough.
-
Lars Wirzeniusreplied to Gergely Nagy 🐁 last edited by
@algernon Docs should probably say an instance of what the program connects to.
-
Gergely Nagy 🐁replied to Lars Wirzenius last edited by
@liw ACK.
The docs are quite sparse at the moment, while I work out some kinks, so there will be shortcomings and missing information here and there. Highlighting those - as you just did - is useful, though, thank you!
-
Lars Wirzeniusreplied to Gergely Nagy 🐁 last edited by
@algernon Mostly I want to know if I can even try this.
-
Gergely Nagy 🐁replied to Lars Wirzenius last edited by
@liw nod
I'd suggest waiting a day or so, until I write proper docs, and iron out a few things.
If you really want to try it, I'd suggest trying it with a test account (Mastodon & GoToSocial accounts both work) that doesn't have followers, to not spam anyone.
You can then throw together a config file like this:
[instance]
url = "https://toot.liw.fi"
token = "<access-token-of-the-test-account>"
[timing]
delay = "5s"
[[toot]]
url = "https://toot.liw.fi/@liw/113038516930914728"
interval = "1m"Then, from the test account, boost
https://toot.liw.fi/@liw/113038516930914728
, and then runRUST_LOG=belated-retooter=trace cargo run -- -c config.toml
, and watch it do... something.If all goes well, it should unboost/reboost the configured toot every minute.
-
Gergely Nagy 🐁replied to Gergely Nagy 🐁 last edited by
@liw (I think I'll add a
--dry-run
mode, where it will pretend to reboost, but won't actually do it.) -
Gergely Nagy 🐁replied to Gergely Nagy 🐁 last edited by
Meh. Turns out, I will need a database, because crawling statuses exhausts rate limits quickly, and is also incredibly slow.
I could use sqlite, but... I might prefer using a pure-rust crate instead, like
redb
.I'm gonna sleep on it.
-
Gergely Nagy 🐁replied to Gergely Nagy 🐁 last edited by
@liw FYI, I updated
belated-retooter
, the documentation should be clearer and more complete now. Also fixed a bunch of issues.Unfortunately, it requires a local (sqlite) db now, because scraping the statuses ended up being prohibitively expensive (slow, lots of traffic, and exhausts rate limits very fast).