Daniel G. Taylor

Very interesting! I wonder if things could "just work" if we generate a real cert that browsers would accept and make local.rest.sh (or something similar) resolve to localhost. Any other...

Example configuration (`~/.restish/apis.json`) which triggers this bug: ```json { "do": { "base": "https://api.digitalocean.com/v2", "profiles": { "default": { "headers": { "authorization": "Bearer ***REDACTED***" } } }, "spec_files": [ "https://raw.githubusercontent.com/digitalocean/openapi/main/specification/DigitalOcean-public.v2.yaml" ] }...

Interesting. The `client_id` is right there in the URL as a query param, but it seems to want it as part of the body. Maybe we need a way to...

@elsewhat please have a look at https://github.com/danielgtaylor/restish/commit/e1d9b7d7226cfbb5982a7fec781911f59ec3a0d5 which has more robust URI and POST payload generation code added. I believe the latest release should fix both problems, I am asked...

@dreverri thanks for the PR! I did not know that `127.0.0.1` is recommended over `localhost`, so thank you for that! As for the random port, this seems to not work...

For example, see https://community.auth0.com/t/random-local-ports-on-redirect-uri/28623

@dreverri you could try adding a `--rsh-port` parameter for this purpose!

Yes, this is one of the shortcomings of the shorthand syntax as it exists today. I'm in the process of rewriting the parser and making some changes to the language...

@ServerlessArchitect the name should be getting set from the operation ID, via this code: https://github.com/danielgtaylor/restish/blob/main/openapi/openapi.go#L214 Do you have a short example which exhibits the incorrect behavior that I could debug?...

@ServerlessArchitect > Just to confirm, are you saying the command names are still based on my operationIds, just normalised? Yes, it is based on the `operationId` if present. It is...