ci: Delete QNS PR comment on successful run

[larseggert]

Because we only add one when there is a failure, and when things then later get fixed we don't update it again (because we skip the update on success). So the outdated, failed state keeps being shown.

Because if we only update on failure, we keep showing the last-failed state after success.

[mxinden]

Unfortunate to have it show up on each successful pull request then, i.e. more noise. That said, maybe better than the above mentioned confusion.

[larseggert]

Do you see a way to only update an existing comment, rather than always add or updated?

[mxinden]

On success, we could use mode delete and ignore a failure for the case where it doesn't exist.

https://github.com/thollander/actions-comment-pull-request?tab=readme-ov-file#action-inputs

[larseggert]

On success, we could use mode delete

Oh nice!

[mxinden]

Neat! Thank you!

[github-actions]

Benchmark results

Performance differences relative to 7ff76c7.

• drain a timer quickly time: [358.04 ns 366.46 ns 374.46 ns]
  change: [-12.410% -6.4320% -2.3016%] (p = 0.01 < 0.05)
  💚 Performance has improved.

• coalesce_acked_from_zero 1+1 entries
  time: [193.97 ns 194.43 ns 194.92 ns]
  change: [-0.6249% -0.1017% +0.4323%] (p = 0.72 > 0.05)
  No change in performance detected.

• coalesce_acked_from_zero 3+1 entries
  time: [238.04 ns 238.66 ns 239.29 ns]
  change: [-0.4518% -0.0535% +0.3563%] (p = 0.80 > 0.05)
  No change in performance detected.

• coalesce_acked_from_zero 10+1 entries
  time: [236.81 ns 237.48 ns 238.31 ns]
  change: [-0.5826% -0.0475% +0.5704%] (p = 0.87 > 0.05)
  No change in performance detected.

• coalesce_acked_from_zero 1000+1 entries
  time: [220.11 ns 220.36 ns 220.66 ns]
  change: [-0.2855% +0.4598% +1.2258%] (p = 0.24 > 0.05)
  No change in performance detected.

• RxStreamOrderer::inbound_frame()
  time: [118.61 ms 118.77 ms 119.01 ms]
  change: [+0.1207% +0.3145% +0.5440%] (p = 0.00 < 0.05)
  Change within noise threshold.

• transfer/Run multiple transfers with varying seeds
  time: [120.63 ms 120.92 ms 121.22 ms]
  thrpt: [32.998 MiB/s 33.079 MiB/s 33.160 MiB/s]
  change:
  time: [+0.2568% +0.5781% +0.9562%] (p = 0.00 < 0.05)
  thrpt: [-0.9471% -0.5748% -0.2561%]
  Change within noise threshold.

• transfer/Run multiple transfers with the same seed
  time: [121.43 ms 121.62 ms 121.82 ms]
  thrpt: [32.835 MiB/s 32.889 MiB/s 32.940 MiB/s]
  change:
  time: [+0.4401% +0.6719% +0.8881%] (p = 0.00 < 0.05)
  thrpt: [-0.8803% -0.6674% -0.4382%]
  Change within noise threshold.

• 1-conn/1-100mb-resp (aka. Download)/client
  time: [1.0999 s 1.1082 s 1.1176 s]
  thrpt: [89.477 MiB/s 90.238 MiB/s 90.916 MiB/s]
  change:
  time: [-2.1695% -0.8177% +0.5102%] (p = 0.28 > 0.05)
  thrpt: [-0.5077% +0.8244% +2.2176%]
  No change in performance detected.

• 1-conn/10_000-parallel-1b-resp (aka. RPS)/client
  time: [428.34 ms 430.52 ms 432.72 ms]
  thrpt: [23.109 Kelem/s 23.228 Kelem/s 23.346 Kelem/s]
  change:
  time: [-0.9869% -0.2271% +0.4441%] (p = 0.54 > 0.05)
  thrpt: [-0.4422% +0.2276% +0.9967%]
  No change in performance detected.

• 1-conn/1-1b-resp (aka. HPS)/client
  time: [51.027 ms 51.375 ms 51.749 ms]
  thrpt: [19.324 elem/s 19.465 elem/s 19.598 elem/s]
  change:
  time: [-0.7265% +0.3420% +1.3927%] (p = 0.53 > 0.05)
  thrpt: [-1.3736% -0.3408% +0.7318%]
  No change in performance detected.

Client/server transfer results

Transfer of 134217728 bytes over loopback.

Client	Server	CC	Pacing	Mean [ms]	Min [ms]	Max [ms]	Relative
msquic	msquic			637.2 ± 202.9	412.6	1127.2	1.00
neqo	msquic	reno	on	896.2 ± 189.5	758.6	1380.4	1.00
neqo	msquic	reno		844.3 ± 150.7	765.1	1263.8	1.00
neqo	msquic	cubic	on	775.4 ± 21.7	753.4	823.8	1.00
neqo	msquic	cubic		796.2 ± 39.5	752.3	870.4	1.00
msquic	neqo	reno	on	4415.2 ± 270.0	4038.9	4949.0	1.00
msquic	neqo	reno		4442.5 ± 259.8	4118.3	4994.2	1.00
msquic	neqo	cubic	on	4503.8 ± 227.6	4185.7	4868.7	1.00
msquic	neqo	cubic		4483.7 ± 205.3	4169.1	4874.3	1.00
neqo	neqo	reno	on	3827.2 ± 319.9	3251.1	4374.7	1.00
neqo	neqo	reno		3349.4 ± 162.3	3086.4	3644.9	1.00
neqo	neqo	cubic	on	4387.3 ± 234.9	4119.4	4987.6	1.00
neqo	neqo	cubic		4431.5 ± 163.9	4120.8	4733.1	1.00

⬇️ Download logs