resolve: use unixfs ResolveOnce #5484
Conversation
License: MIT Signed-off-by: Łukasz Magiera <magik6k@gmail.com>
ghost
assigned 
|
Can we switch to to the CoreAPI? It should select the correct resolver based on the namespace. |
|
Good point, I'll see how that works |
License: MIT Signed-off-by: Łukasz Magiera <magik6k@gmail.com>
magik6k
force-pushed
the
fix/resolve-hamt
branch
from
448dfc4
to
92a302e
Compare
License: MIT Signed-off-by: Łukasz Magiera <magik6k@gmail.com>
magik6k
force-pushed
the
fix/resolve-hamt
branch
from
92a302e
to
f915730
Compare
|
Done |
| // DhtRecordCount is an option for Name.Resolve which specifies how many records | ||
| // we want to validate before selecting the best one (newest). Note that setting | ||
| // this value too low will have security implications | ||
| func (nameOpts) DhtRecordCount(rc int) NameResolveOption { |
There was a problem hiding this comment.
Should we have DHT specific options like this or a more general purpose way to pass routing options (that is, go-libp2p-routing/options.Option). That is:
func RoutingOption(r ropts.Option) NameResolveOption ...Or is this leaking too much into this API?
My worry here is adding a ton of special-purpose options for each router we may implement.
| Local: false, | ||
| Cache: true, | ||
|
|
||
| DhtRecordCount: 16, |
There was a problem hiding this comment.
Not sure if we should hard-code this. Can we use the value from the DHT (unless we go with the "resolve options" solution in which case this point is moot).
|
Note: I'm fine keeping this as it is, I'd just like to consider a more generalized solution to adding more and more API options. |
License: MIT Signed-off-by: Łukasz Magiera <magik6k@gmail.com>
|
@magik6k This is wonderful. These issues have been plaguing us for ages. |
(This is assuming we want to make
ipfs resolveoperate on unixfs paths instead of ipld)Fixes #5270 (comment)