Add an 'id' field to commando requests

[niftynei]

Tested for the golang one, the commando one probably needs some more test runs.

Fixes the error where commando requires an id field at the top level for calls.

[jb55]

this won't push a string, this pushes a raw u64

[niftynei]

Is that ok if it’s encased in quotes? I should have read the underlying routine…

…

On Tue, Jul 4, 2023 at 20:00 William Casarin ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In commando.c <#22 (comment)>: > @@ -27,6 +27,8 @@ int EXPORT commando_make_rpc_msg(const char *method, const char *params, cursor_push_str(&msgbuf, params) && cursor_push_str(&msgbuf, ",\"rune\":\"") && cursor_push_str(&msgbuf, rune) && + cursor_push_str(&msgbuf, "\",\"id\":\"") && + cursor_push_u64(&msgbuf, (u64)req_id) && this won't push a string, this pushes a raw u64 — Reply to this email directly, view it on GitHub <#22 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIMAKPIKEDV2H7AEEEZJL3XOSVCZANCNFSM6AAAAAAZ6G5J3I> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[jb55]

On Tue, Jul 04, 2023 at 05:20:20PM -0700, neil saitug wrote: Is that ok if it’s encased in quotes? I should have read the underlying routine…

you would need to sprintf it like you did on the golang side

[niftynei]

Updated: now the c code uses char * for ids (which matches underlying update)

[jb55]

On Tue, Jul 04, 2023 at 07:09:40PM -0400, niftynei wrote: turn the type from int to string -> print directly to CLN Closes: #22 --- commando.c | 7 +++++-- commando.h | 2 +- dist/js/lnsocket.js | 6 +++--- dist/js/lnsocket.wasm | Bin 147919 -> 142399 bytes 4 files changed, 9 insertions(+), 6 deletions(-) diff --git a/commando.c b/commando.c index 9112498..6fbfd9f 100644 --- a/commando.c +++ b/commando.c @@ -4,9 +4,10 @@ #include "endian.h" #include "commando.h" #include "export.h" +#include <stdio.h> int EXPORT commando_make_rpc_msg(const char *method, const char *params, - const char *rune, unsigned int req_id, unsigned char *buf, int buflen) + const char *rune, const char *req_id, unsigned char *buf, int buflen) { struct cursor msgbuf; int ok; @@ -27,7 +28,9 @@ int EXPORT commando_make_rpc_msg(const char *method, const char *params, cursor_push_str(&msgbuf, params) && cursor_push_str(&msgbuf, ",\"rune\":\"") && cursor_push_str(&msgbuf, rune) && - cursor_push_str(&msgbuf, "\",\"id\":\"d0\",\"jsonrpc\":\"2.0\"}"); + cursor_push_str(&msgbuf, "\",\"id\":\"") && + cursor_push_str(&msgbuf, req_id) &&

Is this still a number? If so we can just use a temporary buffer to sprintf a number like we do on the golang side? Otherwise this breaks the API.

[niftynei]

Is this still a number?

I think any string is valid now? Yeah it definitely breaks the API for users downstream of the C library.

[jb55]

On Thu, Nov 30, 2023 at 11:32:51AM -0800, neil saitug wrote: > Is this still a number? I think any string is valid now? Yeah it definitely breaks the API for users downstream of the C library.

ok, good to know. thanks!

[jb55]

On Tue, Jul 04, 2023 at 06:30:50PM -0400, niftynei wrote: diff --git a/go/lnsocket.go b/go/lnsocket.go index 7947866..7083e09 100644 --- a/go/lnsocket.go +++ b/go/lnsocket.go @@ -57,7 +57,9 @@ func (msg *CommandoMsg) Encode(buf *bytes.Buffer, pver uint32) error { buf.WriteString(msg.Params) buf.WriteString(",\"rune\":\"") buf.WriteString(msg.Rune) - buf.WriteString("\",\"id\":\"d0\",\"jsonrpc\":\"2.0\"}") + buf.WriteString("\",\"id\":\"") + buf.WriteString(fmt.Sprintf("%d", msg.RequestId))

If RequestId is a string, I think we need to update RequestId to a string like we do in the C version.