--- Log opened Wed Apr 01 00:00:40 2020
06:40 < adiabat> kcalvinalvin: hey is PR #106 mergable?  I'd like to merge before changing everything around in cmd/
06:42 < kcalvinalvin> adiabat yes
06:42 < adiabat> ok cool will test it out & merge
06:42 < kcalvinalvin> oh wait wait
06:42 < adiabat> (in like an hour or two)
06:42 < kcalvinalvin> maybe it un-does the fix
06:42 < kcalvinalvin> let me rebase
06:43 < adiabat> ok sure no rush
06:57 < kcalvinalvin> ok I made a new PR
06:57 < kcalvinalvin> #107 is the one to merge
09:10 < adiabat> ok will check that one
19:54 < adiabat> Hey - started using the rev.go code and there are some problems
19:55 < adiabat> I'm looking at it and will push to a WIP PR if you want to take a look as you are probably more familiar with this weird format
19:55 < adiabat> also genproofs seems to be height off by 1 
20:21 < kcalvinalvin> adiabat right the genproofs height+1
20:22 < kcalvinalvin> About that... remember when the whole cmd was with text
20:22 < kcalvinalvin> The code skipped genesis but it was fine because it was just parsing a txt file
20:23 < kcalvinalvin> But I just carried over the logic and so the offsetdata *doesn't* store the offset for the genesis block
20:23 < kcalvinalvin> so if you ask for block 0, it'll give you block 1
20:24 < kcalvinalvin> So if you fetch from the files you need to do -1 of actual height
20:25 < kcalvinalvin> Yeah I did put this in comments somewhere but I guess I was the only one that knew about it
20:25 < kcalvinalvin> I'll take a look at rev.go
20:27 < kcalvinalvin> If you look at places in genproofs, you'll see "height +1" a lot. That's because the fetched block's height is -1 of the actual height
20:32 < adiabat> ok the off by 1 thing is whatever
20:33 < adiabat> can fix I guess, but the real issue is rev is... I have no idea what it's doing
20:33 < adiabat> like there's rev data, but it seems to have little or nothing to do with the txins in the block
20:34 < adiabat> like on block 962
20:35 < adiabat> whether +1 or -1, nothing around 962 has any transactions at all
20:36 < adiabat> but the rev data for 962 has 35 rev txs
20:36 < adiabat> and they're utxos from, like, block 118015, what the heck
20:36 < adiabat> this is coming from rev00000.dat
20:37 < adiabat> so why is a utxo from block 118015 in the first rev file argh
20:41 < kcalvinalvin> hmmmm...
20:41 < kcalvinalvin> I'll give it a looksie
20:42 < kcalvinalvin> Maybe I'm deserializing it wrong
20:42 < adiabat> maybe it's not *that* crazy in that blk00000.dat goes up to that height
20:42 < adiabat> so there could be rev data from that in there
20:42 < kcalvinalvin> The Blocks should be right because the magicbytes woudl fail
20:42 < adiabat> but maybe the rev stuff is in a totally different order
20:43 < kcalvinalvin> *magicbytes checking*
20:43 < adiabat> blocks are right, yeah, that's been working for a while
20:43 < adiabat> this is the first time we've actually trying *using* rev data thought right?
20:43 < kcalvinalvin> yes
20:43 < adiabat> and the whole time we're like .. hopefully it's in the same order as the blocks
20:43 < adiabat> because it'd be crazy for it not to be.... right....???
20:44 < adiabat> but...
20:44 < kcalvinalvin> Well sipa did say that the only way it could be wrong is if you've had an orphaned block or something
20:44 < kcalvinalvin> or a reorg
20:44 < adiabat> maybe, but like
20:45 < adiabat> its not a little off, it's way off
20:45 < adiabat> ?ERROR? block 962 0 inputs but 35 revs
20:45 < adiabat> and the revs are all these utxos from the end of blk00000.dat
20:45 < adiabat> like almost the exact end
20:46 < adiabat> also I should not use mainnet data because it's super depressing, every utxo is worth like millions of dollars
20:47 < adiabat> anyway I will be up a little longer but will go to bed somewhat soon, and I'm guessing I will not figure it out by then heh :)
20:47 < kcalvinalvin> lol
20:47 < adiabat> so if you want to look at it, cool; I'll try figuring it out tomorrow as well
20:48 < adiabat> like I'm pretty sure the rev data is what we think it is
20:48 < adiabat> because sometimes it's right. or close
20:48 < kcalvinalvin> what do you mean close
20:48 < adiabat> but the rev block you get like, txs, and then txins, but why?  why not just a slice of txins... why organize them into txs...
20:48 < kcalvinalvin> like an extra tx is in there?
20:49 < adiabat> also the rev data doesn't have outpoints, which, OK fine it'd be redundant
20:50 < adiabat> but that means I have to figure out what the utxos are without the outpoints which is tricky
20:50 < adiabat> also sometimes the pubkey is blank which 
20:50 < adiabat> well... also sometimes the pubkey is blank in the actual output so what's that about...
20:50 < adiabat> or... no wait it's not, the block explorer just says it's blank but it's not
20:51 < adiabat> anyway
20:51 < kcalvinalvin> > but the rev block you get like, txs, and then txins, but why? oh that's how Core used to do it. I could change it
20:51 < kcalvinalvin> Now they just use the Coin class
20:52 < adiabat> do the rev txs mean they correspond to which tx is spending them
20:52 < adiabat> err
20:52 < adiabat> what I mean is, if a bunch of txins are grouped into a single rev tx, that's not because those utxos were created at the same time
20:52 < adiabat> it's because they all get spent at the same time, right?
20:52 < adiabat> yeah because they have different heights
20:54 < kcalvinalvin> Yeah it's because they're spent
20:56 < adiabat> ok the rev stuff is... somewhat coherent in that like, I get a rev block
20:57 < adiabat> and it works, but it's the rev data for block 119965
20:57 < adiabat> hmmm
20:57 < adiabat> and the lastIndexOffsetHeight is 119964
20:58 < adiabat> so... something weird is going on, like rev is counting down backwards or something, that can't be a coincidence
20:59 < adiabat> hmm yeah something switched or backwards; the last block we're going to process is the first rev block we get.
21:00 < adiabat> except not exactly first, there are a couple before that, but they have blank pubkeys
21:01 < kcalvinalvin> yeah.. ok I kinda see what you're saying
21:02 < kcalvinalvin> I'll look at it and maybe push a PR today
21:02 < adiabat> ok if not I'll look at it tomorrow
21:03 < adiabat> close though!  if the rev stuff works, then genAddDel is done-ish
21:03 < adiabat> and with no extra database thankfully
21:03 < kcalvinalvin> yay
21:04 < kcalvinalvin> ⊂(◉‿◉)つ
--- Log closed Thu Apr 02 00:00:41 2020