Cli: add a query for a range of pending snark work #16260
Conversation
src/app/cli/src/init/client.ml
Outdated
| | Some end_idx -> | ||
| let len = | ||
| if len_response <= end_idx then len_response - start_idx | ||
| else max 0 (end_idx - start_idx) | ||
| in | ||
| Array.sub ~pos:start_idx ~len pending_snark_work |
There was a problem hiding this comment.
I prefer having a when clause in the match statement that verifies if start_idx <= end_idx <= len_response. And then having a catchall statement in the end that returns empty list.
Also I think this logic breaks if negative indexes are passed, which is technically possible rn.
There was a problem hiding this comment.
I pushed a a new version that checks if indexes are positives. If not, it now returns [||]
| @@ -1029,6 +1029,38 @@ let to_signed_fee_exn sign magnitude = | |||
| let sgn = match sign with `PLUS -> Sgn.Pos | `MINUS -> Neg in | |||
| Currency.Fee.Signed.create ~sgn ~magnitude | |||
|
|
|||
| let format_pending_snark_work ?(f = Fun.id) graphql_endpoint = | |||
There was a problem hiding this comment.
@mrmr1993 has sent me nits,about passing in functions like this instead of piping, but I don't really mind.
| let start_idx = | ||
| flag "--start-idx" ~aliases:[ "start-idx" ] | ||
| ~doc:"START_IDX first index of the range (included)" | ||
| (required Cli_lib.Arg_type.int16) |
There was a problem hiding this comment.
making this an int allows this to be negative, ik it's a cli command but still seems a bit foot gunny.
There was a problem hiding this comment.
I kept the int, but it is now checked in the function that the indexes are positive.
anne-laure-s
force-pushed
the
al/snark-work-graphql
branch
from
2d977f5
to
d5882ba
Compare
anne-laure-s
force-pushed
the
al/snark-work-cli
branch
from
bd6931b
to
fecdf86
Compare
| let end_idx = | ||
| flag "--end-idx" ~aliases:[ "end-idx" ] | ||
| ~doc:"END_IDX first index of the range (excluded)" | ||
| (optional Cli_lib.Arg_type.int16) |
There was a problem hiding this comment.
nit
i still prefer to enforce this as positive through the
type system instead of runtime, but your logic is good so I will approve. I’m also generally more a fan of returning empty list when the user inputs an extremely high end index instead of treating as a none case, but again the logic is fine.
No description provided.