< nelsonhb>
Guilin High-tech Industrial Development Zone
< nelsonhb>
Nanning Economic & Technological Development Area
< nelsonhb>
Kunming High-tech Industrial Development Zone (biological medicine, new materials, electronic information, photoelectron, agriculture) and Kunming Economic and Technology Development Zone (mechanical equipment production, biological science and food industry, information industry, software).
< nelsonhb>
Chenggong Industrial Park
< nelsonhb>
Anning Industrial Park
< nelsonhb>
Songming Yanglin Industrial Development Zone
< nelsonhb>
Dongchuan Special Industrial Park
< nelsonhb>
Xundian Special Industrial Park
< nelsonhb>
Kunming Haikou Industrial Park.
< sidhujag>
hey guys just wondering is coinbase outputs to witness programs (bech32) considered safe?
< sidhujag>
i thought you normally would pay to legacy addresses in coinbase because witness programs aren't verified in coinbase
< sidhujag>
i was watching blocks found a weird one pop up with 1 tx only.. and saw its outputting to a bech32
< instagibbs>
well, if we're letting a pure utility function in, I'll review...
< provoostenator>
There was some general discussion about that. There's not really a safe alternative for deriveaddresses, because you'd have to implement descriptors. At least until we figure out how to properly library-ify the codebase.
< sdaftuar>
does anyone look at appveyor failures?
< wumpus>
you mean, in general?
< sdaftuar>
yeah
< wumpus>
we do try to have both appveyor and travis pass before merging
< sdaftuar>
ok... i can try to debug then. seems like it fails randomly so wasn't sure how to treat failures on a PR
< wumpus>
is there a problem with that at the moment?
< wumpus>
yea two CI's is double the chance of random failures, unfortunately
< sdaftuar>
i haven't tried to do any debugging yet, i'll see if i can track it down
< wumpus>
ok let me know if you'd like me to have a look
< luke-jr>
it'd be nice if a failure popped up a console on the dev's screen to debug :P
< wumpus>
oh yes
< wumpus>
interactive debug that would be a great feature
< instagibbs>
I've previously tried to get travis to cough up a core dump on various kinds of errors, was unable to
< wumpus>
much more efficient than trying to reproduce the exact environment locally (for appveyor that isn't even an option for me I don't have windows)
< luke-jr>
wumpus: re status of #13339 there's quite a few unresolved issues; I have an older version I use in Knots that ignores the escaping issue entirely, but it seemed like people didn't like that
< bitcoin-git>
bitcoin/master fa1db33 MarcoFalke: travis: Only exit early if compilation took longer than 30 min
< bitcoin-git>
bitcoin/master ddae781 MarcoFalke: Merge #15349: travis: Only exit early if compilation took longer than 30 m...
< wumpus>
(this doesn't mean it can't still make 0.18, but I don't think it makes sense to prioritize it for review as it is)
< bitcoin-git>
[bitcoin] MarcoFalke merged pull request #15349: travis: Only exit early if compilation took longer than 30 min (master...Mf1902-travis30) https://github.com/bitcoin/bitcoin/pull/15349
< wumpus>
luke-jr: I'm surprised and shocked that escaping turns out to be so difficult
< wumpus>
I used to think it was simply replacing ' by '\'' and surrounding by single quotes
< wumpus>
though usually it's preferable to avoid it entirely by not going through the shell and passing the arguments one-to-one directly to execv() etc
< luke-jr>
wumpus: that works fine on *nix, just not Windows
< luke-jr>
wumpus: on Windows, the command line is passed as one big string to the program, and the program decides how to parse it
< sipa>
does it go into an infinite loop? (if so, that's technically UB)
< gmaxwell>
it's not an infinite loop, technically. :)
< kewde[m]>
It caused weird behavior when the value was used afterwards, anyways, it returned a value that was out of the max of the range became 0xFFFFFFFFFFFFFFFF.
< sipa>
well you're using the function wrong
< gmaxwell>
If it were to change at all, it should change to assert. But considering randrange gets used in innerloops and whatnot, I'm dubious of the value of that tradeoff.
< sipa>
returning 0 is equally invalid, as it's outside of the range [0, 0)