--- Day changed Tue Nov 05 2019
03:15 < kcalvinalvin> Ok so I have some things that I've been working on locally on branch add-resume-for-parseblock. Hopefully I can push them before the meeting.
05:00 < stevenroose> So it's in an hour right? If my timezone calculation was correct?
05:00 < stevenroose> Right at lunch time for me tbh.. Will try to make it. I'm very curious of the state of affairs :)
05:12 < kcalvinalvin> Yes it's in 48 minutes 
05:45 < adiabat> yup... maybe a couple minutes after 9 as I've got to set up the computer at work & linux audio always works the first time :)
06:01 < kcalvinalvin> Hello. Tell me when you're good to go
06:05 < adiabat> hey i'm on meet.jit.si/utreexo
08:20 < jamesob> ah dammit, missed it again. so we're doing these every tues at 9AM EST?
10:29 < adiabat> jamesob: yeah 9AM EST seems ok for people (& I have more reliable internet at work :)
10:38 < jamesob> cool, will throw it on the calendar
12:54 < adiabat> kcalvinalvin: about the add-resume branch & what we were talking about
12:55 < adiabat> talking to cory / fanquake found a maybe simpler way to do it
12:55 < adiabat> also a little ugly, but ... ok:
12:55 < adiabat> first pass parse through all the blocks files, but only read the first 80 bytes of each block -- just get the headers
12:56 < adiabat> then, in-ram, do the whole nextMap thing, piece everything together, and write the following to a file
12:56 < adiabat> just a bunch of 8 bytes for each block: [b1 fnum][b1 offset][b2 fnum][b2 offset][b3...]
12:57 < adiabat> where fnum is the file number, offset is byte offset in block
12:57 < adiabat> should take only... 30 seconds? to build that file.  Depends how fast your disks are but it's quick.
12:57 < adiabat> then once it builds that file, it can just go through linearly and it's super easy to build the db/ txos file
12:58 < adiabat> you wouldn't even need a progress file at all; just look at the last line of the testnet.txos text fiile and see what block you're on
12:58 < adiabat> then seek to 8*blockheight in the offset file, and read where to look next
12:59 < adiabat> anyway, I think that's probably fastest / easiest; I can implement that like... tomorrow ish.  or you can try if you like, let me know
21:03 < kcalvinalvin> Ah ok I see. One more step 
21:04 < kcalvinalvin> I feel like implementing core to write the files in order wouldn't be hard
21:05 < kcalvinalvin> Ah nvm getting the consensus for that is probably not so easy?
21:07 < kcalvinalvin> I'll give it a go while you're sleeping :)