10:03 <russell--> mrnuke, hurricos: i captured a dump with pulseview, massaged a little with awk, of plugging in power, and after about 40 seconds, plugging in a PoE load in port 3 and then after a minute or so unplugging in for ~10 seconds and plugging the same load into port 4: https://personaltelco.net/~russell/gs1900-10hp-v2-vendor-fw-trace-03-frames.txt

10:08 <russell--> the methodology is: new view / tabular decoder output view, then save as csv, commas escaped, then pipe the resulting file through: grep -v bit | awk -F, '$4 != last { printf("\n%s %s %s",$2,$4,$6) ; last = $4 } $4 == last { printf(" %s",$6) } END { printf("\n") }'

10:17 <russell--> oops, the first byte is repeated, let me fix

10:19 <russell--> ... then pipe the resulting file through: grep -v bit | awk -F, '$4 != last { printf("\n%s %s",$2,$4) ; last = $4 } $4 == last { printf(" %s",$6) } END { printf("\n") }'

10:21 <russell--> also, pulseview's time stamps are stupid

11:53 <robimarko> Mangix: Any chance you are willing to tackle libmft and mariadb again?

11:53 <robimarko> https://downloads.openwrt.org/snapshots/faillogs/aarch64_cortex-a53/packages/mariadb/compile.txt

11:53 <robimarko> I think its at least partially related to https://github.com/fmtlib/fmt/commit/b90b4bc981ece26c31283ce18f9d50d38d600b60#diff-1441eec7746106c0c58dbb3a207f93cd1cec6d991e4d40b0056b1ad7f5930f58L2292

12:21 <nick[m]> how do you handle backport commits with the github-merge-pr.sh script? It adds another signed-off and link. not sure if that is what we want?

12:22 <nick[m]> and happy new year ;)

14:43 <russell--> looks like cmd 43 is checking port status four ports at a time, all potential 48 ports, each one has a 16-bit status, they start out 0x60 0x10 with nothing connected, about 17 seconds after the first messages, they switch to 0x61 0x10. When i plug in a poe powered device, the corresponding port status changes to 0x42 0x13. when i move the cable from port 3 to port 4, the status follows to the new

14:43 <russell--> slot. the cmd 43's poll all the possible 48 ports approximately every 2 seconds.

15:30 <russell--> plugging in a non-poe device to a port, port status changes to 0x11 0x10.

15:37 <russell--> plugging in a meraki mr24, i get a port status of 0x42 0x14, the 0x42 0x13 came from a tp-link eap225-outdoor

16:35 <hauke> nick[m]: The github-merge-pr.sh script only adds the SoB line if it is not already there

16:36 <hauke> I think more SoB lines are fine

16:36 <hauke> make sure you also have the 'cherry picked from commit' line

16:36 <hauke> nick[m]: happy new year

18:14 <nick[m]> hauke: so the commits here are fine? https://github.com/openwrt/openwrt/pull/17418

18:46 <hauke> nick[m]: looks fine

19:04 <Mangix> robimarko: the gift that keeps on giving

19:05 <Mangix> I think that's the 3rd or 4th fmt related failure with mariadb

19:05 <robimarko> Yeah, I saw the git history

22:54 <SlimeyX> any u-boot peeps here?