WIP: Protocol 0.8 #223
WIP: Protocol 0.8 #223
Conversation
| int8_t _z_bytes_init(_z_bytes_t *bs, size_t capacity); | ||
| _z_bytes_t _z_bytes_make(size_t capacity); | ||
| _z_bytes_t _z_bytes_wrap(const uint8_t *bs, size_t len); | ||
| _z_bytes_t _z_bytes_steal(_z_bytes_t *b); |
There was a problem hiding this comment.
Wasn't '_z_bytes_move` already implementing the same behavior?
There was a problem hiding this comment.
it was indeed, didn't spot it because the behaviour differs to other moves. I think _z_bytes_move will be renamed to the steal version, to avoid being confused for the move convention
include/zenoh-pico/net/query.h
Outdated
| @@ -23,7 +25,7 @@ | |||
| typedef struct { | |||
| _z_value_t _value; | |||
| _z_keyexpr_t _key; | |||
| _z_zint_t _qid; | |||
| uint32_t _rid; | |||
There was a problem hiding this comment.
z_declare_resource returns an uint16_t-size id.
Any reason for a wider range here?
There was a problem hiding this comment.
Here, rid stands for request id. Changing the name to avoid confusions
include/zenoh-pico/net/query.h
Outdated
| @@ -33,7 +35,7 @@ typedef struct { | |||
| * Return type when declaring a queryable. | |||
| */ | |||
| typedef struct { | |||
| _z_zint_t _id; | |||
| uint32_t _id; | |||
There was a problem hiding this comment.
Similar comment here, but I did not yet check the qid return size.
There was a problem hiding this comment.
This here is an entity id, they're u32
include/zenoh-pico/net/subscribe.h
Outdated
| #include "zenoh-pico/net/session.h" | ||
| #include "zenoh-pico/protocol/core.h" | ||
|
|
||
| /** | ||
| * Return type when declaring a subscriber. | ||
| */ | ||
| typedef struct { | ||
| _z_zint_t _id; | ||
| uint32_t _id; |
There was a problem hiding this comment.
Same comment here as for the queryable
src/link/link.c
Outdated
| @@ -148,8 +148,7 @@ size_t _z_link_recv_exact_zbuf(const _z_link_t *link, _z_zbuf_t *zbf, size_t len | |||
|
|
|||
| int8_t _z_link_send_wbuf(const _z_link_t *link, const _z_wbuf_t *wbf) { | |||
| int8_t ret = _Z_RES_OK; | |||
|
|
|||
| for (size_t i = 0; (i < _z_wbuf_len_iosli(wbf)) || (ret == -1); i++) { | |||
| for (size_t i = 0; (i < _z_wbuf_len_iosli(wbf)); i++) { | |||
There was a problem hiding this comment.
if an iosli fails to send and a subsequent one succeeds, this function will return _Z_RES_OK and it shall not be the case.
There was a problem hiding this comment.
so here, the behaviour should depend on whether or not the link is streamed:
- on streamed links, keep trying to send what hasn't been sent.
- on datagram links, give up after the first failure, as only fragmented payloads have multi-slice wbufs, which means missing one is a full on transmit failure for the whole thing
| return ret; | ||
| } | ||
| _z_id_t _z_id_empty() { | ||
| return (_z_id_t){.id = { |
There was a problem hiding this comment.
Please confirm, but I think: return (_z_id_t){.id = {0}}; would initialize the entire array with 0s.
There was a problem hiding this comment.
I know that works for structures, I have a small doubt on arrays... needs double-checking just in case
There was a problem hiding this comment.
This was a quick and light review on my side.
I suggest @Mallets to do a more detailed one, including double checking the message formats and constant values with the Rust implementation.
No description provided.