< luke-jr>
sipa: any obvious way to perhaps make the dnsseed.dat files compress and/or xdelta better? with 1.2 GB every day, it's starting to add up (~200 GB since 2012, compressed)
< luke-jr>
maybe there's some kind of sorting that would help
< sipa>
luke-jr: the file is 276 MiB for me
< luke-jr>
sipa: 1.2 GB for me XD
< luke-jr>
and the xdelta for today to yesterday is 295 MB
< midnightmagic>
is there that much churn?
< luke-jr>
dunno, I guess if I started over, mine would look closer to sipa's, but it seems more useful to have the diversity of whatever old data is in there
< ProfMac>
Not totally unrelated to dnsseed.dat, I wrote my first awk script and have been looking at debug.log to see who is connecting to me. I especially looked at IPv6 addresses, and I notice a lot of alias addresses for the same machine. I assume my data is almost identical to the peers.dat data.
< aj>
jtimon: sorry, i keep forgetting to send you test case code for getblockstats
< jtimon>
aj: feel free to open a pr with my commits plus your tests, I really think it's the best wat to test it in concept but never wrote or seen the code
< jtimon>
and of course if you do ping me for review and testing
< aj>
BlueMatt: iceweasel rebranding was trademark issues (maybe "can't use firefox trademark if you ship old shit" though, i forget; thought it was just that debian did patches though), but they did away with that and introduced "extended support releases", so firefox-esr is in debian stable these days, and iceweasel is long gone
< jimpo>
Do people think that the best-effort getrawtransaction lookup from CCoinsCacheView if the txindex is disabled is a useful thing?
< jimpo>
As opposed to requiring either the txindex for the RPC to work or passing in the block_hash argument?
< jtimon>
aj: looks very good at a glance, I'm going to bed but I will check it out and test it
< aj>
jtimon: yeah, feel free to squash and tidy it up to your preferred style, but i think the approach should work fine
< jtimon>
yeah, thanks
< aj>
jtimon: np, sorry for taking so long to get back to it :)
< jtimon>
np, I didn't have time to code it myself either, but this looks cool
< bitcoin-git>
[bitcoin] promag opened pull request #12836: Make WalletInitInterface and DummyWalletInit private (master...2018-03-walletinit) https://github.com/bitcoin/bitcoin/pull/12836
< promag>
aj: did you delete the GH comment?
< aj>
promag: yeah, the compile i was running with that change failed
< promag>
that PR is mostly moving code
< promag>
but I think we could make g_wallet_init_interface const
< aj>
promag: yeah, it looks very good
< aj>
promag: since it doesn't have any member variables, i guess that's right?
< aj>
promag: would need to mark all the functions const too which is arguably a bit weird
< promag>
I don't mean const in that way
< promag>
I mean, don't allow g_wallet_init_interface to take a new value
< promag>
let's see what others say
< aj>
promag: i think marking it const-pointer fails because it can get optimised away in the module its declared in, while other modules stil try referencing it? (that's what the compile failure looked like anyway)
< aj>
promag: it all got confusing anyway, which is why i deleted my comment :)
< bitcoin-git>
[bitcoin] joemphilips opened pull request #12837: fix type mistmatch in `listreceivedbyaddress` (master...fix_listreceivedbyaddress_type) https://github.com/bitcoin/bitcoin/pull/12837
< BlueMatt>
aj: yea, it was "cant use firefox trademark if you dont do X, Y, Z" where one of those, iirc, was "if you ship out-of-date shit and then try to backport changes"
< bitcoin-git>
bitcoin/master 8394300 Roman Zeyde: [Tests] fix a typo in TestNode.assert_start_raises_init_error()...
< bitcoin-git>
bitcoin/master de7e586 MarcoFalke: Merge #12810: [Tests] Fix a typo at assert_start_raises_init_error() and use specific exception for initialization error...
< 18WAAOAGJ>
[bitcoin] MarcoFalke closed pull request #12810: [Tests] Fix a typo at assert_start_raises_init_error() and use specific exception for initialization error (master...fix-blocksdir-test) https://github.com/bitcoin/bitcoin/pull/12810
< gribble>
https://github.com/bitcoin/bitcoin/issues/12810 | [Tests] Fix a typo at assert_start_raises_init_error() and use specific exception for initialization error by romanz · Pull Request #12810 · bitcoin/bitcoin · GitHub
< sipa>
is it my comment on the utxo lookup in getrawtransaction
< sipa>
?
< BlueMatt>
yes, your comment was "I dont like utxo lookup in getrawtransaction", the pr is about the utxo db itself
< BlueMatt>
it doesnt touch the interface to it
< sipa>
jimpo made some comment somewhere about whether it was useful to have the utxo based lookup
< sipa>
and i commented on that
< sipa>
but from my phone i couldn't directly respond to the review comment, so it probably became a top level comment on the pr
< BlueMatt>
ah, I see the discussion thread, anyway, his point about the pr being large is well-made, then you followed up with "well, you could add a new rpc command"
< BlueMatt>
anyway, whatever, i was just confused
< sipa>
i wasn't suggesting anything change in the PR itself; just giving my opinion on the usefulness of that feature :)
< bitcoin-git>
[bitcoin] jnewbery opened pull request #12843: [tests] Test starting bitcoind with -h and -version (master...test_help) https://github.com/bitcoin/bitcoin/pull/12843
< U039bmz>
.-. .-.
< U039bmz>
.-. .-.
< U039bmz>
.-. .-.
< U039bmz>
.-. .-.
< U039bmz>
/ \ / \
< U039bmz>
/ \ / \
< U039bmz>
/ \ / \
< U039bmz>
/ \ / \
< U039bmz>
| _ \ / _ |
< U039bmz>
| _ \ / _ |
< U039bmz>
| _ \ / _ |
< U039bmz>
| _ \ / _ |
< U039bmz>
; | \ \ / / | ;
< U039bmz>
; | \ \ / / | ;
< U039bmz>
; | \ \ / / | ;
< U039bmz>
; | \ \ / / | ;
< U039bmz>
\ \ \ \_.._/ / / /
< U039bmz>
\ \ \ \_.._/ / / /
< U039bmz>
\ \ \ \_.._/ / / /
< U039bmz>
\ \ \ \_.._/ / / /
< U039bmz>
'. '.;' ';,' .'
< U039bmz>
'. '.;' ';,' .'
< U039bmz>
'. '.;' ';,' .'
< U039bmz>
'. '.;' ';,' .'
< U039bmz>
'./ _ _ \.'
< U039bmz>
'./ _ _ \.'
< U039bmz>
'./ _ _ \.'
< U039bmz>
'./ _ _ \.'
< U039bmz>
.' a __ a '.
< U039bmz>
.' a __ a '.
< U039bmz>
.' a __ a '.
< U039bmz>
.' a __ a '.
< U039bmz>
'--./ _, \/ ,_ \.--'
< U039bmz>
'--./ _, \/ ,_ \.--'
< U039bmz>
'--./ _, \/ ,_ \.--'
< U039bmz>
'--./ _, \/ ,_ \.--'
< U039bmz>
----| \ /\ / |----
< U039bmz>
----| \ /\ / |----
< U039bmz>
----| \ /\ / |----
< U039bmz>
----| \ /\ / |----
< U039bmz>
.--'\ '-' '-' /'--.
< U039bmz>
.--'\ '-' '-' /'--.
< U039bmz>
.--'\ '-' '-' /'--.
< U039bmz>
.--'\ '-' '-' /'--.
< U039bmz>
_>.__ -- _.- `;
< U039bmz>
_>.__ -- _.- `;
< U039bmz>
_>.__ -- _.- `;
< U039bmz>
_>.__ -- _.- `;
< U039bmz>
.' _ __/ _/
< U039bmz>
.' _ __/ _/
< U039bmz>
.' _ __/ _/
< U039bmz>
.' _ __/ _/
< U039bmz>
/ '.,:".-\ /:,
< U039bmz>
/ '.,:".-\ /:,
< U039bmz>
/ '.,:".-\ /:,
< U039bmz>
/ '.,:".-\ /:,
< U039bmz>
| \.' `""`'.\\
< U039bmz>
| \.' `""`'.\\
< U039bmz>
| \.' `""`'.\\
< U039bmz>
| \.' `""`'.\\
< U039bmz>
'-,.__/ _ .-. ;|_
< U039bmz>
'-,.__/ _ .-. ;|_
< U039bmz>
'-,.__/ _ .-. ;|_
< U039bmz>
'-,.__/ _ .-. ;|_
< U039bmz>
/` `|| _/ `\/_ \_|| `\
< U039bmz>
/` `|| _/ `\/_ \_|| `\
< U039bmz>
/` `|| _/ `\/_ \_|| `\
< U039bmz>
/` `|| _/ `\/_ \_|| `\
< U039bmz>
| ||/ \-./` \ / || |
< U039bmz>
| ||/ \-./` \ / || |
< U039bmz>
| ||/ \-./` \ / || |
< U039bmz>
| ||/ \-./` \ / || |
< U039bmz>
\ ||__/__|___|__|| /
< U039bmz>
\ ||__/__|___|__|| /
< U039bmz>
\ ||__/__|___|__|| /
< U039bmz>
\ ||__/__|___|__|| /
< U039bmz>
\_ |_Happy Easter_| /
< U039bmz>
\_ |_Happy Easter_| /
< U039bmz>
\_ |_Happy Easter_| /
< U039bmz>
\_ |_Happy Easter_| /
< U039bmz>
jgs .' \ = _= _ = _= /`\
< U039bmz>
jgs .' \ = _= _ = _= /`\
< U039bmz>
jgs .' \ = _= _ = _= /`\
< U039bmz>
jgs .' \ = _= _ = _= /`\
< U039bmz>
/ `-;----=--;--' \
< U039bmz>
/ `-;----=--;--' \
< U039bmz>
/ `-;----=--;--' \
< U039bmz>
/ `-;----=--;--' \
< U039bmz>
\ _.-' '. /
< U039bmz>
\ _.-' '. /
< U039bmz>
\ _.-' '. /
< U039bmz>
\ _.-' '. /
< jimpo>
BlueMatt: The fact that the block index may not be synced to fs before a BlockConnected/BlockDisconnected callback is quite annoying
< jimpo>
Makes the interface quite difficult to work with/reason about
< BlueMatt>
jimpo: true, but otherwise block connection would be way slower :/
< BlueMatt>
jimpo: there is a separate callback for the wallet that indicates things have been flushed
< BlueMatt>
so you kinda have to go the direction of the wallet and use a locator
< jimpo>
Yeah, the locator is probably the way to go