fix: properly handle return codes in pack_timestamp - #672
Conversation
Sorry, something went wrong.
| } else { | ||
| /* seconds is signed or >34bits */ | ||
| unsigned char buf[12]; | ||
| _msgpack_store32(&buf[0], nanoseconds); | ||
| _msgpack_store64(&buf[4], seconds); | ||
| msgpack_pack_ext(x, -1, 12); | ||
| msgpack_pack_raw_body(x, buf, 12); | ||
| /* seconds is signed or >34bits */ | ||
| unsigned char buf[12]; | ||
| _msgpack_store32(&buf[0], nanoseconds); | ||
| _msgpack_store64(&buf[4], seconds); | ||
| ret = msgpack_pack_ext(x, -1, 12); | ||
| if (ret != 0) | ||
| return ret; | ||
|
|
||
| return msgpack_pack_raw_body(x, buf, 12); | ||
| } | ||
| return 0; | ||
| } |
There was a problem hiding this comment.
If this was my code, I would probably prefer doing something like
if (condition) {
// whatever
return something;
}
// the contents of the elseto make it clear to the reader that we never not return a value from the function. This is even more relevant now that I've removed the return 0 at the end. But it's not my code so... let me know.
Sorry, something went wrong.
There was a problem hiding this comment.
Pull request overview
This PR ensures msgpack_pack_timestamp() correctly propagates failures from the underlying packing helpers, so callers can observe and handle allocation/write errors instead of silently succeeding.
Changes:
- Capture and check the return value from msgpack_pack_ext() in all timestamp encoding branches.
- Propagate the return value from msgpack_pack_raw_body() (instead of always returning 0).
- Normalize indentation in the timestamp96 branch to match the surrounding style.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Sorry, something went wrong.
What is this PR?
This PR makes it so that return codes of helpers called in msgpack_pack_ext are properly checked and surfaced if they're indicative of a failure.
Note the "big" diff at the end of the function is an indentation change on top of the actual change -- the code used to be indented with three spaces instead of four. I changed that (should be clear looking at the diff without whitespace changes).