Add 1s to each timestamp and drop negatives.

[kixelated]

Fixes audio priming.

[coderabbitai]

Walkthrough

The change modifies PTS (presentation timestamp) calculation for audio and video frame publishing. Instead of using raw packet timestamps, the code now applies an offset-based approach by adding timebase_den/timebase_num to packet->pts and converting the result to microseconds. A validation guard was added to drop frames with negative adjusted PTS values. Comments explain the approximately 1-second offset adjustment for audio priming and audio/video synchronization. Error handling for publish failures remains unchanged, with one new early-return path for invalid timestamps.

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title 'Add 1s to each timestamp and drop negatives' accurately describes the main technical changes in the pull request: adding a 1-second offset to timestamps and filtering out negative PTS values.
Description check	✅ Passed	The description 'Fixes audio priming' is directly related to the changeset, which implements timestamp offset logic specifically to address audio priming issues as documented in the PR objectives.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

---

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.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Fix all issues with AI agents

In `@src/moq-output.cpp`:
- Around line 159-169: The pts calculation uses packet->timebase_den /
packet->timebase_num which can divide by zero if packet->timebase_num == 0; add
a defensive guard before computing pts (e.g., check packet->timebase_num != 0
and packet->timebase_den != 0) and handle the error path by logging a warning
(using LOG_WARNING) and returning early so util_mul_div64 and
moq_publish_media_frame are not called with invalid values; update the block
around the pts calculation that references packet->timebase_num,
packet->timebase_den, pts, util_mul_div64(...) and moq_publish_media_frame(...)
to perform this check.

🧹 Nitpick comments (1)

src/moq-output.cpp (1)

188-198: Consider extracting duplicate PTS adjustment logic.

The PTS calculation and validation logic (lines 188-196) is identical to AudioData (lines 159-167), differing only in the log message. Extracting this to a helper function would reduce duplication and ensure consistent behavior if the offset calculation changes.

♻️ Suggested helper extraction

// Add as a private helper method or static function
std::optional<int64_t> MoQOutput::AdjustPts(struct encoder_packet *packet, const char *type)
{
    if (packet->timebase_num == 0) {
        LOG_WARNING("Invalid %s timebase (num=0), dropping frame", type);
        return std::nullopt;
    }
    int64_t pts = packet->pts + packet->timebase_den / packet->timebase_num;
    if (pts < 0) {
        LOG_WARNING("Dropping %s frame with negative PTS: %lld", type, (long long)packet->pts);
        return std::nullopt;
    }
    return util_mul_div64(pts, 1000000ULL * packet->timebase_num, packet->timebase_den);
}

Then in AudioData/VideoData:

auto pts_us = AdjustPts(packet, "audio");
if (!pts_us) return;
auto result = moq_publish_media_frame(audio, packet->data, packet->size, *pts_us);

[coderabbitai]

⚠️ Potential issue | 🟡 Minor

Guard against potential division by zero.

The calculation packet->timebase_den / packet->timebase_num on line 161 will cause undefined behavior if timebase_num is ever 0. While valid encoder packets should always have a non-zero timebase numerator, adding a defensive check would improve robustness.

🛡️ Suggested defensive guard

 	// Add ~1 second offset to handle negative PTS from audio priming frames.
 	// TODO: This is slightly wrong when den is not evenly divisible by num, but close enough.
+	if (packet->timebase_num == 0) {
+		LOG_WARNING("Invalid audio timebase (num=0), dropping frame");
+		return;
+	}
 	int64_t pts = packet->pts + packet->timebase_den / packet->timebase_num;

🤖 Prompt for AI Agents

In `@src/moq-output.cpp` around lines 159 - 169, The pts calculation uses
packet->timebase_den / packet->timebase_num which can divide by zero if
packet->timebase_num == 0; add a defensive guard before computing pts (e.g.,
check packet->timebase_num != 0 and packet->timebase_den != 0) and handle the
error path by logging a warning (using LOG_WARNING) and returning early so
util_mul_div64 and moq_publish_media_frame are not called with invalid values;
update the block around the pts calculation that references
packet->timebase_num, packet->timebase_den, pts, util_mul_div64(...) and
moq_publish_media_frame(...) to perform this check.