Conversation
Signed-off-by: JaylinYu <letrangerjaylin@gmail.com>
Signed-off-by: JaylinYu <letrangerjaylin@gmail.com>
Signed-off-by: JaylinYu <letrangerjaylin@gmail.com>
Signed-off-by: JaylinYu <letrangerjaylin@gmail.com>
Signed-off-by: JaylinYu <letrangerjaylin@gmail.com>
set nano_qos_db as NULL while create new conn_param Signed-off-by: JaylinYu <letrangerjaylin@gmail.com>
reuse old API Signed-off-by: JaylinYu <letrangerjaylin@gmail.com>
Signed-off-by: JaylinYu <letrangerjaylin@gmail.com>
Must seperate Topic & Payload with "/" Signed-off-by: JaylinYu <letrangerjaylin@gmail.com>
Signed-off-by: JaylinYu <letrangerjaylin@gmail.com>
Member
Author
|
@alvin1221 take my work from now on |
Contributor
📝 WalkthroughWalkthroughThis PR replaces BLF configuration with NNG proxy support (pub/sub/bridge nodes), refactors QoS database cleanup from pipe-close to pipe-fini lifecycle phases across multiple transport modules, introduces a SUB0-to-MQTT message adapter function, updates declaration specifiers from Changes
Sequence Diagram(s)sequenceDiagram
participant Client as NNG SUB0 Client
participant Adapter as nng_sub0_msg_adapter()
participant Encoder as nano_encode_publish_msg()
participant MQTT as MQTT Broker/Network
Client->>Adapter: nng_msg (body: "topic/payload")
Adapter->>Adapter: scan for '/' separator
Adapter->>Adapter: allocate and copy topic
Adapter->>Encoder: topic, payload, QoS, MQTT v3.1.1
Encoder->>Encoder: encode PUBLISH packet
Encoder-->>Adapter: nng_msg (MQTT PUBLISH frame)
Adapter-->>Client: transformed message
Client->>MQTT: send PUBLISH
sequenceDiagram
participant Parser as Config Parser
participant CfgStruct as conf_nng_proxy
participant PubNode as conf_nng_pub_node
participant SubNode as conf_nng_sub_node
participant Init as Initializer
Parser->>Parser: parse bridges.nng.pub from JSON
Parser->>PubNode: conf_nng_pnode_parse()
PubNode->>PubNode: populate topic list, validate QoS
PubNode->>CfgStruct: push to pub_nodes vector
Parser->>Parser: parse bridges.nng.sub from JSON
Parser->>SubNode: conf_nng_snode_parse()
SubNode->>SubNode: populate topic list, validate QoS
SubNode->>CfgStruct: push to sub_nodes vector
Init->>PubNode: conf_bridge_pnode_init()
Init->>SubNode: conf_bridge_snode_init()
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 3 | ❌ 2❌ Failed checks (2 warnings)
✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Review rate limit: 0/1 reviews remaining, refill in 60 minutes.Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 7
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@src/sp/protocol/mqtt/mqtt_parser.c`:
- Around line 2154-2179: Check and validate inputs before dereferencing: ensure
origin is non-NULL before calling nng_msg_body/nng_msg_len in the mqtt_parser.c
function, and ensure topic is non-NULL/valid before passing it into
nano_encode_publish_msg for the fallback branch; if origin is NULL return NULL
(or an appropriate error) early, and if topic is NULL use a safe default (or
pass NULL explicitly only if nano_encode_publish_msg supports it) or abort the
fallback publish, updating the early-return path that logs "No valid
topic/payload separator..." to perform these checks and avoid dereferencing
invalid pointers.
In `@src/sp/transport/mqtt/broker_tcp.c`:
- Around line 224-226: The code reads p->npipe->nano_qos_db into nano_qos_db
before checking p->npipe for NULL, which can crash; change the logic to first
ensure p->npipe is not NULL (and p->conf is not NULL) then read
p->npipe->nano_qos_db (or simply move the nano_qos_db assignment below the if
that checks p->npipe and p->conf), and ensure the conditional uses the existing
symbols p->npipe, p->conf, and nano_qos_db so that nano_qos_db is only
dereferenced when p->npipe is valid.
In `@src/sp/transport/mqtts/broker_tls.c`:
- Around line 231-233: The code reads p->npipe->nano_qos_db into nano_qos_db
before checking p->npipe for NULL, risking a null-deref; modify the teardown in
broker_tls.c so you only access p->npipe->nano_qos_db after confirming p and
p->npipe are non-NULL (e.g., move the assignment of nano_qos_db inside the if
that checks p->npipe != NULL and p->conf != NULL), and ensure subsequent logic
that uses nano_qos_db still handles a NULL nano_qos_db safely.
In `@src/sp/transport/mqttws/nmq_websocket.c`:
- Around line 1338-1340: The code reads p->npipe->nano_qos_db into nano_qos_db
before checking p->npipe for NULL, which can dereference a NULL pointer during
teardown; fix by guarding p->npipe first (e.g., check p != NULL && p->npipe !=
NULL) before accessing p->npipe->nano_qos_db in the fini/cleanup path, or move
the nano_qos_db assignment to after the existing p->npipe and p->conf null
checks; update the relevant block that references p, p->npipe, p->conf and
nano_qos_db so that nano_qos_db is only read when p->npipe is confirmed
non-NULL.
In `@src/supplemental/nanolib/conf_ver2.c`:
- Around line 1470-1477: The code resets config->nng_proxy.pnodes (and similarly
snodes) to NULL before parsing, leaking previously allocated
conf_nng_pub_node/conf_nng_sub_node objects and their internal strings/topic
lists; add and call destroy helpers (e.g., conf_nng_pub_node_destroy and
conf_nng_sub_node_destroy) to free each node and its internal allocations,
iterate over the existing cvector (config->nng_proxy.pnodes and
config->nng_proxy.snodes) to destroy and clear elements before reassigning or
reinitializing, and then proceed to allocate and push new nodes with
conf_bridge_pnode_init/conf_nng_pnode_parse and cvector_push_back; apply the
same destruction/clear logic to the block that handles snodes (the code around
the other similar parse loop).
- Around line 1192-1225: conf_nng_pnode_parse (and likewise
conf_nng_snode_parse) currently allocates a topics struct and pushes it into
node->pub_list/node->sub_list even when remote_topic or local_topic is missing;
update these functions to guard like conf_bridge_node_parse: after populating s,
check that both s->remote_topic and s->local_topic (or their lengths) are
non-NULL/non-empty before calling cvector_push_back, and if the check fails free
the allocated topics struct and any strings inside it instead of pushing it to
the list so incomplete mappings are skipped.
In `@src/supplemental/nanolib/conf.c`:
- Around line 3094-3115: The init helpers leave some fields uninitialized;
update conf_nng_proxy_init to set proxy->pub_count = 0 and proxy->sub_count = 0
in addition to pub_enable, sub_enable, pnodes, snodes, and update
conf_bridge_pnode_init and conf_bridge_snode_init to explicitly initialize every
field of conf_nng_pub_node and conf_nng_sub_node (set all pointer fields to NULL
and all count/size/integer fields to 0 and booleans to false) so they no longer
rely on the caller providing zeroed storage; modify the bodies of
conf_nng_proxy_init, conf_bridge_pnode_init, and conf_bridge_snode_init
accordingly.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: d95072f3-9c69-4ad5-8285-030255fd6ee9
📒 Files selected for processing (13)
include/nng/protocol/mqtt/mqtt_parser.hinclude/nng/supplemental/nanolib/conf.hinclude/nng/supplemental/nanolib/nanolib.hsrc/sp/protocol/mqtt/mqtt_parser.csrc/sp/protocol/mqtt/nmq_mqtt.csrc/sp/transport/mqtt/broker_tcp.csrc/sp/transport/mqtts/broker_tls.csrc/sp/transport/mqttws/nmq_websocket.csrc/supplemental/nanolib/CMakeLists.txtsrc/supplemental/nanolib/blf/CMakeLists.txtsrc/supplemental/nanolib/blf/blf.ccsrc/supplemental/nanolib/conf.csrc/supplemental/nanolib/conf_ver2.c
💤 Files with no reviewable changes (3)
- src/supplemental/nanolib/CMakeLists.txt
- src/supplemental/nanolib/blf/CMakeLists.txt
- src/supplemental/nanolib/blf/blf.cc
Comment on lines
+2154
to
+2179
| const uint8_t *body = nng_msg_body(origin); | ||
| size_t body_len = nng_msg_len(origin); | ||
|
|
||
| if (body == NULL || body_len == 0) { | ||
| log_error("Empty origin message"); | ||
| return NULL; | ||
| } | ||
|
|
||
| // Define '/' to separate topic and payload. | ||
| // NNG SUB msg format: "topic/path/payload" -> Topic: "topic", Payload: "path/payload" | ||
| const char *ptr = (const char *)body; | ||
| char *sep = NULL; | ||
|
|
||
| for (size_t i = 0; i < body_len; i++) { | ||
| if (ptr[i] == '/') { | ||
| sep = &ptr[i]; | ||
| break; | ||
| } | ||
| } | ||
|
|
||
| // If no separator found, or starts with '/', fallback to using the whole message as payload. | ||
| if (sep == NULL || sep == ptr) { | ||
| log_warn("No valid topic/payload separator found in NNG sub0 msg."); | ||
| mqtt_msg = nano_encode_publish_msg(MQTT_PROTOCOL_VERSION_v311, | ||
| 0, false, false, body, body_len, NULL, topic, NULL); | ||
| return mqtt_msg; |
Contributor
There was a problem hiding this comment.
Validate adapter inputs before dereference/fallback publish.
origin is dereferenced before any null check, and fallback publish can use topic without validating it. Both can trigger crashes on bad inputs.
Suggested fix
nng_msg *
nng_sub0_msg_adapter(nng_msg *origin, char *topic)
{
nng_msg *mqtt_msg = NULL;
+ if (origin == NULL) {
+ log_error("origin message is NULL");
+ return NULL;
+ }
const uint8_t *body = nng_msg_body(origin);
size_t body_len = nng_msg_len(origin);
@@
if (sep == NULL || sep == ptr) {
+ if (topic == NULL || topic[0] == '\0') {
+ log_error("Default topic is required when separator is missing");
+ return NULL;
+ }
log_warn("No valid topic/payload separator found in NNG sub0 msg.");
mqtt_msg = nano_encode_publish_msg(MQTT_PROTOCOL_VERSION_v311,
0, false, false, body, body_len, NULL, topic, NULL);
return mqtt_msg;
}🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/sp/protocol/mqtt/mqtt_parser.c` around lines 2154 - 2179, Check and
validate inputs before dereferencing: ensure origin is non-NULL before calling
nng_msg_body/nng_msg_len in the mqtt_parser.c function, and ensure topic is
non-NULL/valid before passing it into nano_encode_publish_msg for the fallback
branch; if origin is NULL return NULL (or an appropriate error) early, and if
topic is NULL use a safe default (or pass NULL explicitly only if
nano_encode_publish_msg supports it) or abort the fallback publish, updating the
early-return path that logs "No valid topic/payload separator..." to perform
these checks and avoid dereferencing invalid pointers.
Comment on lines
+224
to
+226
| void *nano_qos_db = p->npipe->nano_qos_db; | ||
| if (p->npipe != NULL && p->conf != NULL && | ||
| !p->conf->sqlite.enable && nano_qos_db != NULL) { |
Contributor
There was a problem hiding this comment.
Guard p->npipe before accessing nano_qos_db.
Line 224 dereferences p->npipe before the null check. In teardown/error paths this can crash.
Suggested fix
- void *nano_qos_db = p->npipe->nano_qos_db;
- if (p->npipe != NULL && p->conf != NULL &&
+ void *nano_qos_db = NULL;
+ if (p->npipe != NULL) {
+ nano_qos_db = p->npipe->nano_qos_db;
+ }
+ if (p->npipe != NULL && p->conf != NULL &&
!p->conf->sqlite.enable && nano_qos_db != NULL) {📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| void *nano_qos_db = p->npipe->nano_qos_db; | |
| if (p->npipe != NULL && p->conf != NULL && | |
| !p->conf->sqlite.enable && nano_qos_db != NULL) { | |
| void *nano_qos_db = NULL; | |
| if (p->npipe != NULL) { | |
| nano_qos_db = p->npipe->nano_qos_db; | |
| } | |
| if (p->npipe != NULL && p->conf != NULL && | |
| !p->conf->sqlite.enable && nano_qos_db != NULL) { |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/sp/transport/mqtt/broker_tcp.c` around lines 224 - 226, The code reads
p->npipe->nano_qos_db into nano_qos_db before checking p->npipe for NULL, which
can crash; change the logic to first ensure p->npipe is not NULL (and p->conf is
not NULL) then read p->npipe->nano_qos_db (or simply move the nano_qos_db
assignment below the if that checks p->npipe and p->conf), and ensure the
conditional uses the existing symbols p->npipe, p->conf, and nano_qos_db so that
nano_qos_db is only dereferenced when p->npipe is valid.
Comment on lines
+231
to
+233
| void *nano_qos_db = p->npipe->nano_qos_db; | ||
| if (p->npipe != NULL && p->conf != NULL && | ||
| !p->conf->sqlite.enable && nano_qos_db != NULL) { |
Contributor
There was a problem hiding this comment.
Fix null-deref risk in QoS DB teardown.
Line 231 reads p->npipe->nano_qos_db before confirming p->npipe is non-null.
Suggested fix
- void *nano_qos_db = p->npipe->nano_qos_db;
- if (p->npipe != NULL && p->conf != NULL &&
+ void *nano_qos_db = NULL;
+ if (p->npipe != NULL) {
+ nano_qos_db = p->npipe->nano_qos_db;
+ }
+ if (p->npipe != NULL && p->conf != NULL &&
!p->conf->sqlite.enable && nano_qos_db != NULL) {📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| void *nano_qos_db = p->npipe->nano_qos_db; | |
| if (p->npipe != NULL && p->conf != NULL && | |
| !p->conf->sqlite.enable && nano_qos_db != NULL) { | |
| void *nano_qos_db = NULL; | |
| if (p->npipe != NULL) { | |
| nano_qos_db = p->npipe->nano_qos_db; | |
| } | |
| if (p->npipe != NULL && p->conf != NULL && | |
| !p->conf->sqlite.enable && nano_qos_db != NULL) { |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/sp/transport/mqtts/broker_tls.c` around lines 231 - 233, The code reads
p->npipe->nano_qos_db into nano_qos_db before checking p->npipe for NULL,
risking a null-deref; modify the teardown in broker_tls.c so you only access
p->npipe->nano_qos_db after confirming p and p->npipe are non-NULL (e.g., move
the assignment of nano_qos_db inside the if that checks p->npipe != NULL and
p->conf != NULL), and ensure subsequent logic that uses nano_qos_db still
handles a NULL nano_qos_db safely.
Comment on lines
+1338
to
+1340
| void *nano_qos_db = p->npipe->nano_qos_db; | ||
| if (p->npipe != NULL && p->conf != NULL && | ||
| !p->conf->sqlite.enable && nano_qos_db != NULL) { |
Contributor
There was a problem hiding this comment.
Guard p->npipe before reading nano_qos_db in fini.
Line 1338 dereferences p->npipe before the null check, which can crash in partial-init teardown paths.
Suggested fix
- void *nano_qos_db = p->npipe->nano_qos_db;
- if (p->npipe != NULL && p->conf != NULL &&
+ void *nano_qos_db = NULL;
+ if (p->npipe != NULL) {
+ nano_qos_db = p->npipe->nano_qos_db;
+ }
+ if (p->npipe != NULL && p->conf != NULL &&
!p->conf->sqlite.enable && nano_qos_db != NULL) {🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/sp/transport/mqttws/nmq_websocket.c` around lines 1338 - 1340, The code
reads p->npipe->nano_qos_db into nano_qos_db before checking p->npipe for NULL,
which can dereference a NULL pointer during teardown; fix by guarding p->npipe
first (e.g., check p != NULL && p->npipe != NULL) before accessing
p->npipe->nano_qos_db in the fini/cleanup path, or move the nano_qos_db
assignment to after the existing p->npipe and p->conf null checks; update the
relevant block that references p, p->npipe, p->conf and nano_qos_db so that
nano_qos_db is only read when p->npipe is confirmed non-NULL.
Comment on lines
+1192
to
+1225
| void | ||
| conf_nng_pnode_parse( | ||
| conf_nng_pub_node *node, cJSON *obj) | ||
| { | ||
| node->name = nng_strdup(obj->string); | ||
| hocon_read_str(node, clientid, obj); | ||
| hocon_read_str(node, pub_url, obj); | ||
| cJSON *forward = NULL; | ||
| cJSON *forwards = hocon_get_obj("forwards", obj); | ||
| cJSON_ArrayForEach(forward, forwards) | ||
| { | ||
| topics *s = NNI_ALLOC_STRUCT(s); | ||
| s->retain = NO_RETAIN; | ||
| s->qos = NO_QOS; | ||
| hocon_read_str(s, remote_topic, forward); | ||
| hocon_read_str(s, local_topic, forward); | ||
| hocon_read_num(s, qos, forward); | ||
| cJSON *jso_key2 = cJSON_GetObjectItem(forward, "qos"); | ||
| if (cJSON_IsNumber(jso_key2) && | ||
| (jso_key2->valuedouble == 0 || jso_key2->valuedouble == 1 || | ||
| jso_key2->valuedouble == 2)) { | ||
| s->qos = jso_key2->valuedouble; | ||
| } else { | ||
| if (jso_key2 != NULL) | ||
| log_warn("invalid qos level detected in pub list"); | ||
| } | ||
| if (s->remote_topic) | ||
| s->remote_topic_len = strlen(s->remote_topic); | ||
| if (s->local_topic) | ||
| s->local_topic_len = strlen(s->local_topic); | ||
| cvector_push_back(node->pub_list, s); | ||
| } | ||
| node->forwards_count = cvector_size(node->pub_list); | ||
| } |
Contributor
There was a problem hiding this comment.
Skip incomplete NNG topic mappings instead of storing them.
Unlike conf_bridge_node_parse(), these helpers push entries even when either topic is missing. That leaves null-bearing topics objects in pub_list / sub_list, which will break later bridge handling.
♻️ Suggested fix
cJSON_ArrayForEach(forward, forwards)
{
topics *s = NNI_ALLOC_STRUCT(s);
s->retain = NO_RETAIN;
s->qos = NO_QOS;
hocon_read_str(s, remote_topic, forward);
hocon_read_str(s, local_topic, forward);
+ if (!s->remote_topic || !s->local_topic) {
+ log_warn("remote_topic/local_topic not found");
+ nng_strfree(s->remote_topic);
+ nng_strfree(s->local_topic);
+ NNI_FREE_STRUCT(s);
+ continue;
+ }
hocon_read_num(s, qos, forward);
...
}Apply the same guard in conf_nng_snode_parse().
Also applies to: 1227-1260
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/supplemental/nanolib/conf_ver2.c` around lines 1192 - 1225,
conf_nng_pnode_parse (and likewise conf_nng_snode_parse) currently allocates a
topics struct and pushes it into node->pub_list/node->sub_list even when
remote_topic or local_topic is missing; update these functions to guard like
conf_bridge_node_parse: after populating s, check that both s->remote_topic and
s->local_topic (or their lengths) are non-NULL/non-empty before calling
cvector_push_back, and if the check fails free the allocated topics struct and
any strings inside it instead of pushing it to the list so incomplete mappings
are skipped.
Comment on lines
+1470
to
+1477
| config->nng_proxy.pnodes = NULL; | ||
| cJSON_ArrayForEach(node_item, node_array) | ||
| { | ||
| conf_nng_pub_node *node = NNI_ALLOC_STRUCT(node); | ||
| conf_bridge_pnode_init(node); | ||
| conf_nng_pnode_parse(node, node_item); | ||
| cvector_push_back(config->nng_proxy.pnodes, node); | ||
| config->nng_proxy.pub_enable = true; |
Contributor
There was a problem hiding this comment.
Don't drop the previous NNG proxy config on reparse.
Resetting config->nng_proxy.pnodes / snodes to NULL here loses ownership of the previously parsed nodes without freeing their strings, topic lists, or any attached runtime state. On hot reload this leaks the old proxy config and orphans its resources.
Add destroy helpers for NNG pub/sub nodes and clear the existing side before resetting the vector.
Also applies to: 1491-1498
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/supplemental/nanolib/conf_ver2.c` around lines 1470 - 1477, The code
resets config->nng_proxy.pnodes (and similarly snodes) to NULL before parsing,
leaking previously allocated conf_nng_pub_node/conf_nng_sub_node objects and
their internal strings/topic lists; add and call destroy helpers (e.g.,
conf_nng_pub_node_destroy and conf_nng_sub_node_destroy) to free each node and
its internal allocations, iterate over the existing cvector
(config->nng_proxy.pnodes and config->nng_proxy.snodes) to destroy and clear
elements before reassigning or reinitializing, and then proceed to allocate and
push new nodes with conf_bridge_pnode_init/conf_nng_pnode_parse and
cvector_push_back; apply the same destruction/clear logic to the block that
handles snodes (the code around the other similar parse loop).
Comment on lines
+3094
to
+3115
| static void | ||
| conf_nng_proxy_init(conf_nng_bridge *proxy) | ||
| { | ||
| proxy->pub_enable = false; | ||
| proxy->sub_enable = false; | ||
| proxy->pnodes = NULL; | ||
| proxy->snodes = NULL; | ||
| } | ||
|
|
||
| void | ||
| conf_bridge_pnode_init(conf_nng_pub_node *node) | ||
| { | ||
| node->pub_url = NULL; | ||
| node->pub_list = NULL; | ||
| } | ||
|
|
||
| void | ||
| conf_bridge_snode_init(conf_nng_sub_node *node) | ||
| { | ||
| node->sub_url = NULL; | ||
| node->sub_list = NULL; | ||
| } |
Contributor
There was a problem hiding this comment.
Fully initialize the new NNG config structs.
These helpers currently rely on the caller passing zeroed storage. conf_nng_proxy_init() leaves pub_count/sub_count indeterminate, and conf_bridge_pnode_init() / conf_bridge_snode_init() only clear a subset of the public struct fields. That makes stack-allocated callers easy to break.
♻️ Suggested fix
static void
conf_nng_proxy_init(conf_nng_bridge *proxy)
{
- proxy->pub_enable = false;
- proxy->sub_enable = false;
- proxy->pnodes = NULL;
- proxy->snodes = NULL;
+ *proxy = (conf_nng_bridge) { 0 };
}
void
conf_bridge_pnode_init(conf_nng_pub_node *node)
{
- node->pub_url = NULL;
- node->pub_list = NULL;
+ *node = (conf_nng_pub_node) { 0 };
}
void
conf_bridge_snode_init(conf_nng_sub_node *node)
{
- node->sub_url = NULL;
- node->sub_list = NULL;
+ *node = (conf_nng_sub_node) { 0 };
}🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/supplemental/nanolib/conf.c` around lines 3094 - 3115, The init helpers
leave some fields uninitialized; update conf_nng_proxy_init to set
proxy->pub_count = 0 and proxy->sub_count = 0 in addition to pub_enable,
sub_enable, pnodes, snodes, and update conf_bridge_pnode_init and
conf_bridge_snode_init to explicitly initialize every field of conf_nng_pub_node
and conf_nng_sub_node (set all pointer fields to NULL and all count/size/integer
fields to 0 and booleans to false) so they no longer rely on the caller
providing zeroed storage; modify the bodies of conf_nng_proxy_init,
conf_bridge_pnode_init, and conf_bridge_snode_init accordingly.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary by CodeRabbit
New Features
Bug Fixes
Removals