| 30 Jul 2021 |
 jplatte | In reply to @timokoesters:fachschaften.org
I thought Raw was just a wrapper around raw bytes, but I just saw it is already partially parsed / validated * It's a wrapper around a boxed serde_json::value::RawValue which is in turn a wrapper around str | 15:49:17 |
 Timo ⚡️ | To be fair this validation just made me find a bug in conduit, so it's probably a good idea to leave it in 😅 | 15:50:01 |
| jplatte | Yes I think so too 😄 | 15:50:20 |
| Timo ⚡️ | It was one of those "I updated half of the code, but the other half still uses the old version" bugs | 15:50:50 |
| jplatte | What kind of update was this? | 15:51:12 |
| Timo ⚡️ | I changed some slight thing about the database layout | 15:51:42 |
| Timo ⚡️ | This lead to conduit's outgoing sender thread crashing, which lead to conduit not accepting new events anymore as it couldn't talk to the sender thread | 15:52:50 |
|  damccull joined the room. | 16:39:13 |
|  Philip (a-0) set a profile picture. | 18:32:24 |
| 31 Jul 2021 |
|  Ananace | 07:47:51 |
|  @dman:kittenface.studio removed their display name Daniel Carlsson. | 08:44:48 |
 [Jonathan] | @jplatte re: Axum, you wanna make it so that it additionally has FromRequest and IntoRequest support when axum as a feature flag is enabled? o: | 09:40:12 |
| [Jonathan] | * @jplatte re: Axum, you wanna make it so that it additionally has FromRequest and IntoRequest support when axum as a feature flag is enabled? o: | 09:40:27 |
| [Jonathan] | That would be really cool | 09:40:30 |
|  Timo K. changed their display name from Timo to Timo♒. | 11:11:32 |
| Timo K. changed their display name from Timo♒ to Timo 🦫. | 21:25:05 |
| Timo K. changed their display name from Timo 🦫 to Timo 🦄. | 21:27:05 |
| Timo K. changed their display name from Timo 🦄 to ▫️Timo▫️. | 21:31:57 |
| Timo K. changed their display name from ▫️Timo▫️ to Timo K.. | 21:34:26 |
| 1 Aug 2021 |
| jplatte | [Jonathan]: due to authentication, I think implementing additional traits for the types defined by the various api crates doesn't make so much sense, at least for requests. I guess for responses it could work better, though I think for now having a wrapper type (maybe [also] one taking possible errors into account) makes more sense | 10:45:04 |
| [Jonathan] | Having responses be IntoResponse would already help a lot, yeah | 10:45:38 |
| jplatte | * [Jonathan]: due to authentication, I think implementing additional traits for the types defined by the various api crates doesn't make so much sense, at least for requests. I guess for responses it could work better, though I think for now having a wrapper type (maybe [also] one taking possible errors into account) makes more sense | 10:45:59 |
| [Jonathan] | Though I think at least some helper functionality could be offered to make axum experience more smooth | 10:46:30 |
| [Jonathan] | Maybe a small framework of authentication middleware and IntoRequest impls that would require(?) it somehow? | 10:46:51 |
| [Jonathan] | Then on the developer side, the only thing that’s needed is to “plug everything in” (impl logic for the paths and auth middleware) and they’re good to go | 10:47:21 |
| [Jonathan] | * Then on the developer side, the only thing that’s needed is to “plug everything in” (impl logic for the paths and middleware) and they’re good to go | 10:47:41 |
| [Jonathan] | * Then on the developer side, the only thing that’s needed is to “plug everything in” (impl logic for the paths and auth middleware) and they’re good to go | 10:47:46 |
| [Jonathan] | Maybe when trying to develop that, you can give feedback to axum on how to make their API better to accommodate things like this 👀 | 10:53:40 |
| [Jonathan] | * Maybe when trying to develop that, you can give feedback to axum on how to make their API better to accommodate things like this 👀 | 10:53:44 |
| [Jonathan] | I don’t know, I just really like the idea of ruma being instantly usable as a server toolkit in combination with Axum | 10:54:07 |
