add /cat endpoint #184
add /cat endpoint #184
Conversation
this passes 9/13 tests in the interface-ipfs-core.
|
Now a new failure: SIGABRT on macos! We have so far had SIGSEGV so we still have a few signals to go until bingo. |
|
Big part of the changes I'll push Listing the directories shouldn't be more than refactoring the common parts of LOC wise the biggest change is probably the error context not to introduce yet more strings into interface-ipfs-core. |
this is probably needed for the directory resolving.
Cache allows us to recover the inner VecDeque which should come handy when resolving over multiple segments.
big ticket item still missing is the two wonky error cases...
ll for low-level
this needs to be done to return error when using cat on directories
dag-pb can have HAMT sharded directories which don't at least directly translate over to IPLD.
|
Great that we have the conformance tests running on CI as I was only running the related |
the remaining issue is how to handle the path resolving for the ugly error mangling case...
also: the walk_path should become ipfs::resolve(IpfsPath).
| pub name: Cow<'static, str>, | ||
| /// Error from the attempted conversion | ||
| pub source: cid::Error, | ||
| /// This is to deny creating these outside of the crate |
There was a problem hiding this comment.
unless we want to expose all the other members outside the crate, making them only pub(crate) would achieve this (just making sure it's not the case)
There was a problem hiding this comment.
My idea was to expose the fields and deny creating these values from outside. Hopefully people would use the fmt::Debug printout ... which will be quite the horror to read, should customize it. And/or perhaps I should make the fmt::Display contain everything in case this ever comes up in a bug report.
Co-authored-by: ljedrz <ljedrz@users.noreply.github.com>
Forgot this from the first batch.. Co-authored-by: ljedrz <ljedrz@users.noreply.github.com>
Initial version passes 9/13 tests in the interface-ipfs-core after ipfs/js-ipfs#3078 which are pretty much all of the tests which doesn't use ipfspath.
Follow-up of #176, part of #172. Test changes rs-ipfs/ipfs-rust-conformance#23.