nanocoap_sock: always use coap_opt_put_uri_pathquery()

[benpicco]

Contribution description

This is pretty subtle: coap_opt_put_uri_pathquery() will detect the presense of a query parameter, coap_opt_put_uri_path() will not.

That means that when you request path/to/resource?option=1 with coap_opt_put_uri_path() you get the path resource?option=1 whereas with coap_opt_put_uri_pathquery() you get the path resource with query parameter option=1.

We really want the latter.

Testing procedure

Issues/PRs references

[chrysn]

LGTM in the sense that it is probably what people need and looks correct from an implementation PoV.

The docs will need updating in all the functions like nanocoap_sock_get that previously speak of a "remote path" to now say "remote path and query".

And it is an API change, moreover, one that is not trivially raising compiler warnings. Given that it's likely reflecting any user's intention, it should suffice to add the right labels to make sure it gets a prominent note in the change log.

[riot-ci]

Murdock results

✔️ PASSED

eb76b8e nanocoap: allow lastonum=NULL in coap_opt_put_uri_pathquery()

Success	Failures	Total	Runtime
8101	0	8101	10m:26s

Artifacts

• Documentation preview

[benpicco]

To me it feels more like a bug fix than an API change - when I added the functions I was not aware of the difference between coap_opt_put_uri_pathquery() and coap_opt_put_uri_path() (or that there are in fact two different functions) 😅

[benpicco]

Is there anything I can do to move this forward?

[benpicco]

Thank you!