--- Log opened Sun Jul 19 00:00:26 2020
01:08 < kcalvinalvin> ja I'd have to read the code again... why do you ask?
01:08 < kcalvinalvin> So I'm trying out building proofs for forest and I ran out of disk space
01:09 < kcalvinalvin> For forest in ram for mainnet, I got to ~500,000 before I ran out. Used up 260GB for proofs and ~20GB of memory
02:45 < kcalvinalvin> PR #158 is ready to merge
02:46 < kcalvinalvin> though github actions is failing because variable ```knownTipHeight``` is being passed as 0?
02:46 < kcalvinalvin> bridge just exits with " done pushing blocks to 127.0.0.1:60888" which doesn't happen until it reaches the height that is was passed
02:47 < kcalvinalvin> not sure what's going on..
04:23 -!- dergoegge [uid453889@gateway/web/irccloud.com/x-wqwekxwzcpkzfqdf] has joined #utreexo
04:25 < dergoegge> kcalvinalvin: are you having these problems on master? also mainnet or testnet?
05:25 -!- rjected_ [~dan@pool-71-184-77-198.bstnma.fios.verizon.net] has joined #utreexo
05:27 -!- cfields_ [~cfields@unaffiliated/cfields] has joined #utreexo
05:32 -!- Netsplit *.net <-> *.split quits: rjected, cfields, adiabat
05:39 -!- adiabat [~adiabat@63.209.32.102] has joined #utreexo
06:06 < kcalvinalvin> dergoegge No not on master but on my PR. Locally it works fine but not working on github
06:06 < kcalvinalvin> It's on github so on regtest
06:37 < dergoegge> kcalvinalvin: i left a review on #158 that should fix it.
06:38 < dergoegge> also kinda weird that you didn't have the error locally
07:37 < kcalvinalvin> that fixed it thanks!
11:08 -!- rjected_ [~dan@pool-71-184-77-198.bstnma.fios.verizon.net] has quit [Quit: ZNC 1.7.5+deb4 - https://znc.in]
11:09 -!- rjected [~dan@pool-71-184-77-198.bstnma.fios.verizon.net] has joined #utreexo
13:47 < adiabat> ja: I can explain swap collapses
13:47 < adiabat> it's more complex than I'd like, but it does allow the transform to be reversible
13:48 < ja> adiabat: i got into animations in the last few days, i was thinking of writing a program that animates the operations utreexo does
13:49 < adiabat> oh that would be super cool
13:49 < ja> adiabat: stuff like swapping the children are conceptually simple, but working with the array representation, it is 20 lines of code
13:49 < ja> adiabat: so i have duplicated all algorithms in imperative haskell, now i want to rewrite them using ideomatic haskell
13:49 < ja> adiabat: but the problem is that i don't actually understand what they are doing, what the invariants are
13:49 < adiabat> yeah I will say the (f *Forest) ToString() code was in some way the most important
13:50 < adiabat> because that let me print out what was going on much better
13:50 < adiabat> "collapses" are swaps that can't immediately take place
13:50 < adiabat> there can only be one collapse per row
13:50 < adiabat> and it's the last thing created for that row.  It's either a leaf that's becoming a root or a root that's becoming a leaf
13:51 < adiabat> or... I think a root that stays a root but moves also counts as a collapse
13:51 < adiabat> the only difference between swaps and collapses is that swaps can happen immediately as you're going from the bottom row to the top
13:52 < adiabat> collapses though have to be applied after all the swaps, including swaps on higher rows
13:52 < ja> ok, can we take an example i generated?
13:52 < adiabat> but the collapses themselves can be affected by higher up swaps
13:52 < adiabat> sure
13:52 < ja> if swaps is [[(35,28),(51,47)],[(152,145)],[(210,204)],[],[],[],[]]
13:53 < ja> and collapses is [Just (76,94),Just (167,174),Just (216,214),Just (232,234),Just (245,244),Nothing,Nothing]
13:53 < adiabat> heh that's big, can't write that on paper...
13:53 < ja> the resulting collapse is [Just (76,76),Just (167,166),Just (216,210),Just (232,232),Just (245,244),Nothing,Nothing]
13:53 < ja> ah, well i generated the examples randomly, computer doesn't know they should be small
13:53 < adiabat> right I just mean
13:53 < ja> maybe it would be better with a smaller example
13:54 < adiabat> if it's small I can trace it out on paper
13:54 < adiabat> with those I'm not sure what's going on
13:54 < ja> ok, well, any way i could understand the intuition would be good
13:55 < adiabat> ok a smaller one is like
13:55 < adiabat> printout.txt in the accumulator folder
13:55 < adiabat> line 21 has the 15 leaves forest
13:55 < adiabat> so an example would be
13:56 < adiabat> delete 4,5,6,7,8,9
13:56 < adiabat> on the bottom row, no swaps
13:56 < adiabat> but, 14 needs to collapse, it can't stay there
13:56 < adiabat> we know that 14 will end up at 8, because there were 15 leaves, and we deleted 6
13:57 < adiabat> but it's a collapse, not a swap
13:57 < adiabat> we can't swap 14 and 8 right now
13:57 < adiabat> because other swaps will move those around
13:58 < adiabat> so for row 0, we say, OK, no swaps, collapse: (14, 8)
13:58 < adiabat> then row 2, deletions are 18, 19, 20
13:58 < ja> is this work going on before calling swapCollapses? since you started with the list of nodes to delete
13:59 < adiabat> this is before yes
13:59 < adiabat> we get the (14, 8) collapse
14:00 < ja> aah ok, so this is the result of makeCollapse called in remTrans2?
14:00 < adiabat> and it will be transformed by swapCollpases into ... 14, 12?
14:00 < adiabat> 1 sec
14:01 < adiabat> no no 14, 4
14:01 < adiabat> in this case swapCollapses will change 14, 8 into 14, 4
14:01 < adiabat> because of the 25, 26 swap that happens on the 3rd row
14:01 < adiabat> (row 2)
14:02 < ja> the swapCollapses will never change the amount of collapses, right?
14:02 < adiabat> that 25, 26 swap affects the 14, 8 collapse because 8 is under 26
14:02 < adiabat> right the amount of collapses will not change
14:02 < adiabat> sometimes it will make collapses become like "null"
14:02 < adiabat> like can change things into 14, 14
14:02 < adiabat> or from a 14, 14 to a 14, 12 or something
14:02 < adiabat> but actual number will not change
14:03 < adiabat> once swap collapses has happened, we can do everything row by row
14:03 < adiabat> we know 14 needs to end up at 8
14:03 < adiabat> but we can't swap 14 and 8 right at the bottom row
14:03 < adiabat> but if we swap 14 with 4, it will end up at 8 after all the rows of swaps have happened
14:04 < adiabat> maybe a better way to call it is ... turn collapses into swaps?
14:04 < adiabat> I don't even think the name "collapse" is that great... it's like a movement that is temporarily on hold
14:05 < ja> names are hard, but i think i almost got the intuition, that is most important :)
14:05 < ja> thank you for the walkthrough!
14:05 < adiabat> sure!
14:05 < ja> ok, i think i need to read your walkthrough a few times, but i think it will help a lot. i'll report back with questions, gotta run! thank you so much!
14:05 < adiabat> ok sure see you
14:05 < adiabat> also I think there will be a "release" in tuesday
14:05 < adiabat> just have a binary and server up
14:06 < adiabat> maybe people will be interested :)
14:06 < ja> no doubt about that :D
--- Log closed Mon Jul 20 00:00:26 2020