Object store iface #165
Object store iface #165
Conversation
lann
force-pushed
the
object-store-iface
branch
3 times, most recently
from
262d8a6
to
5a7a618
Compare
lann
force-pushed
the
object-store-iface
branch
2 times, most recently
from
75a4af1
to
175f3dc
Compare
lann
force-pushed
the
object-store-iface
branch
from
175f3dc
to
e248c28
Compare
|
This needs tests but I'd like feedback too. |
lann
force-pushed
the
object-store-iface
branch
from
e248c28
to
cf4976b
Compare
| } | ||
|
|
||
| /// Prepare an object with the given key to be written. See `object-writer`. | ||
| get-object: function(key: string) -> expected<object-reader, error> |
There was a problem hiding this comment.
I thought we were going to have a store resource rather than putting the object accessors directly on the storage interface? I know we are skipping multiple store implementations on day one but it would be good to have the design aligned to the eventual plan so as to minimise change and avoid baking in assumptions.
| use crate::wit::spin_object_store; | ||
| use cap_std::{ambient_authority, fs::OpenOptions}; | ||
|
|
||
| pub type FileObjectStoreData = ( |
There was a problem hiding this comment.
I don't quite get the different roles of FOS and FOSData - are clearer names available or is this an artefact of the bindgen stuff? (I vaguely recall being puzzled by something similar in HTTP so maybe the latter?)
There was a problem hiding this comment.
Mostly an artifact of wit-bindgen, though we could organize them better.
crates/object-store/src/file.rs
Outdated
| let size = file.read(buf).map_err(|err| err.to_string())?; | ||
| u64::try_from(size).map_err(|err| err.to_string()) | ||
| } | ||
| None => Err("file closed".to_string()), |
There was a problem hiding this comment.
Context or anything else to help with diagnostics, or does that belong at a higher level?
There was a problem hiding this comment.
See above for this specific error, but for errors in general they are being passed to the guest so have to pass through the wasm barrier. I went for string errors for simplicity.
lann
force-pushed
the
object-store-iface
branch
from
cf4976b
to
0810702
Compare
lann
force-pushed
the
object-store-iface
branch
3 times, most recently
from
d1498a6
to
5edb438
Compare
- Includes a simple file-based implementation.
re: #48