< tradermyx>
Also, if the "wallet/blabla.dat" part is used for a non-wallet-specific call, it will still work and not error out, I hope?
< luke-jr>
tradermyx: it should be regular http there; but in theory https can work for localhost, sure
< tradermyx>
I see.
< tradermyx>
luke-jr: And I assume only the basename() is to be used for the "wallet name", since it can now be a path since v0.17?
< luke-jr>
tradermyx: not sure. please find out and update the wiki? ;)
< tradermyx>
:[
< tradermyx>
Also, not yet any v0.17 API documentation on the site.
< tradermyx>
Why does it feel as if now that I'm finally getting Bitcoin "up and running" and being able to charge people and send them money in an automated fashion, the "party is over", so to speak?
< jonasschnelli>
(maybe anyone else can do the magic with automake)
< booyah>
fyi I noticed reports of corrupted storage (chain etc) seem to went up. But maybe it is just people started to test on very low end (unstable) hardware, most were like usb hdd on rpi dunno
< jonasschnelli>
SSD, SDCard, leveldb is not made for sudden power loss and people just unplug RPI and similar machines.
< jonasschnelli>
There is though a great concept of hardlinking levelDB files to create a utxo set backup in case you corrupt it.
< jonasschnelli>
It needs a shutdown but can be automated (every night or so).
< jonasschnelli>
Since its hardlinks, it won't consume to much space
< promag>
is there a better way of changing depends flags other than changingm, for instance, depends/packages/libevent.mk?
< jcorgan_>
so here's a little oddness from the REST interface when retrieving block headers. Depending on the number of headers at a time one requests, there is a nearly 27x difference in performance to sync the entire chain.
< jcorgan>
GET /rest/headers/1020/xxxx.bin, successively repeated until all of them are retrieved, gets 16.3Mbps over localhost
< jcorgan>
do the same thing but use 1021 at a time, and you get 346Mbps
< jcorgan>
not really going to spend too much time figuring it out, but it was a surprise
< sipa>
jcorgan: that... makes no sense whatsoever
< jcorgan>
it's deterministically repeatable here
< sipa>
are there other numbers where such a discrepancy is observable?
< jcorgan>
i haven't tried to many other ones, but random spot checking it looks like a step function above and below that
< jcorgan>
it tops out at 1900 or so before the REST interface barfs
< jcorgan>
ok, its network related, when i do the same over GbE the step is at 1000/1001
< jcorgan>
take that back, i misread
< jcorgan>
i can't find a discontinuity over GbE
< jcorgan>
seems consistent at about 250Mbps
< jcorgan>
over GbE it's somewhat linear between n=500 @ 136Mbps and n=1900 @ 311Mbps, with a flat region in the middle
< jcorgan>
a real engineer would instrument this and provide you a pretty graph and histograms :)
< sipa>
nah, a real engineer woukd realize that syncing at 15 Mbps is pr9bably sufficient for the application at hand, and stop worrying about it
< jcorgan>
pretty much. when i have other work to do and i need to procrastinate, i'll come back to this one
< bralyclow>
hi everyone, not a development question, but no one can/will answer in #bitcoin, was the New button on the Receiving Addresses screen removed in macOS version of 0.17.0 bitcoin core node gui, it is not there any longer, still on the Send Addresses screen to create a new one there, but not on Receiving Address screen?
< promag>
bralyclow: the button was removed in #12721