--- Log opened Tue Nov 02 00:00:21 2021 11:00 < andytoshi> heads up i am going to merge https://github.com/rust-bitcoin/rust-secp256k1/pull/327 today 11:00 < andytoshi> (well, i am reviewing it now anyway) 18:31 < ariard> BlueMatt: meh i've done 200k iterations of chanmon_consistency_target, didn't get a failure on 1054, if you found one on your side i would be happy to fix it! 18:31 < BlueMatt> ariard: yea lol you need good seeds. to find crashes in there you need like a day or two on my 20 core workstation with month-iteration seeds to start :p 18:32 < BlueMatt> ariard: i thought I sent you one, though? 18:32 < BlueMatt> let me look 18:32 < ariard> BlueMatt: oh man i don't have 20 core available rn :) 18:32 < ariard> yeah the seed was fixed with the latest update 18:32 < BlueMatt> ariard: can always ask chaincode :p 18:33 < ariard> BlueMatt: i'll do my own fuzzing cluster, though not now now 18:34 < ariard> yeah let's not delay 0.0.103 for #1054, especially if we have users pending on new sccoring thing 18:34 < BlueMatt> ariard: yea, 103 just needs to be shipped so folks can test pre-prod, we can probably get another release in before folks go to prod with $$$ 18:34 < BlueMatt> but, that means landing 1054 next week when I'm back from atlanta, shouldn't delay 1054 either :p 18:35 < BlueMatt> ok, let me see, I have a new seed I see I checked in, let me see if it fails on latest pr + git 18:35 < BlueMatt> its possible i forgot to send it, been a completely insane few weeks :( 18:36 < ariard> BlueMatt: i've time now to fix 1054 and 1056 during you're in atlanta, just finishing coinpool those days 18:37 < BlueMatt> yes, it seems I have crashes that work against the latest pr 18:37 < BlueMatt> i'd thought the previous version of the pr had fixed them but maybe they regressed? 18:38 < ariard> yeah there was a lot of back-and-forth because find a simpler solution to test the non-enough-feerate for outbound update fee 18:38 < BlueMatt> oh, it would help if i merge correctly 18:38 < ariard> by modifying the test itself instead of test-only path 18:38 < BlueMatt> yea 18:39 < BlueMatt> yes, ok, i still have a single test that fails 18:39 < BlueMatt> 040000ffff343081646400093d180064093d0000000000ffffffffffffff640000010c0021181800000011008080804d00006464646400000000000001000000ffff01000000204d4d4d000018014d004d4d4d4d4d4d00008000000000000000000038323238303464646464646464646464ff8080808080808080cf85ffffff 18:40 < BlueMatt> I haven't minimized it, i can probably make it much smaller if you prefer 18:40 < ariard> let me try that 18:42 < BlueMatt> I can also add your ssh key to this vm that is only used to fuzz rust-lightning 18:42 < BlueMatt> if you want to test there 18:42 < BlueMatt> or make you a vm, but i guess that may be easier at chaincode 18:43 < ariard> I've new glorious hardware available with a bunch more of CPUs, just need time to install the Right Setup (tm) 18:44 < ariard> but yeah that could be a solution the fuzzing vm 18:45 < ariard> BlueMatt: good i've got a failure, now I've something to fix :) 18:45 < BlueMatt> lol, yea... 18:45 < BlueMatt> cool, thanks 18:45 < BlueMatt> sorry it appears i forgot to send that one to you 18:46 < ariard> neeeh, this PR has been the mess with us realizing you can't right the LN state machine on update_fee, lol 18:47 < BlueMatt> yea, well at least I think we're confident athwe know there are a few actually-impossible cases. means we can just make the fuzzer accept otherwise-crash conditions and move on :p 18:51 < ariard> BlueMatt: yeah wanna ACK on 0.0.103 cut-off before i go dinner? 18:51 < BlueMatt> sure 18:52 < BlueMatt> if I get to rewrite the changelog for release tonight I'll be very happy :p 18:52 < BlueMatt> (it currently says tomorrow) 19:07 < ariard> BlueMatt: https://github.com/rust-bitcoin/rust-lightning/pull/1153#pullrequestreview-796049297, maybe just add 1104? 19:09 < BlueMatt> ariard: hmm, right, i forgot the bug is basically user-visible without retires, i was thinking of it as only a retries fix 19:11 < ariard> BlueMatt: yeah it's really that kind of minor changes, dunno where we should put the cursor to be worthy for a change log mention 19:11 < BlueMatt> right, yea, its worth mentioning 19:11 < BlueMatt> generally I've been broad with whats "worth mentioning" in release notes 19:11 < BlueMatt> we do releases often enough its not a big deal for people to read twenty entries. 19:12 < ariard> sure, if we start to do bigger one, we could stretch 19:12 < BlueMatt> yea 19:12 < BlueMatt> anyway, fixed your nits and pushed a fixup 19:12 < ariard> i'll go for tonight, re-acking tmrw yearly 19:13 < BlueMatt> can you re-ack and I'll merge? 19:13 < ariard> oky 19:13 < ariard> it's alrrady late here :p 19:13 < BlueMatt> i know lol but if we cut tonight i can get bindings builds out tomorrow easier 19:14 < ariard> and it's done! 19:14 < ariard> bonne nuit 19:20 < BlueMatt> night! --- Log closed Wed Nov 03 00:00:22 2021