12:28 |
phf |
http://logs.bitdash.io/asciilifeform/2022-08-06#1113069 << you need books on marine chronometers, the great ships and the lesser known wwii aerial battles. also books on lathe work, and biographies of men you respect despite their flaws. |
| |
↖ |
12:28 |
bitbot |
(asciilifeform) 2022-08-06 scoopbot: New article on 0xcdd.com: Books |
| |
~ 54 minutes ~ |
13:23 |
shinohai |
$ticker btc usd |
13:23 |
busybot |
Current BTC price in USD: $23186.09 |
13:26 |
awt |
Prod doesn't quite fit in with the order buffer mechanism. When you receive a prod with a tip you don't have, there's no message to add to the order buffer. But... |
13:27 |
awt |
Gotta add something to the order buffer so the staleness check can be avoided for the GetData response. |
13:31 |
phf |
so when a packet comes in you know that it's a getdata response by checking your existing messages again? |
| |
↖ |
13:33 |
phf |
(as opposed to e.g. having a special list of pending getdata messages) |
| |
~ 17 minutes ~ |
13:51 |
awt |
http://logs.bitdash.io/pest/2022-08-06#1011030 << yeah blatta keeps a list of messages that are "out of order" in the order buffer. When ancestors come in via a GetData response, they avoid the timestamp check if a direct descendent is in the order buffer. |
13:51 |
bitbot |
Logged on 2022-08-06 13:31:54 phf[awt]: so when a packet comes in you know that it's a getdata response by checking your existing messages again? |
13:52 |
awt |
jonsykkel and I also ended up with a datastructure to keep track of outgoing GetData messages to ensure messages aren't requested multiple times. |
13:52 |
awt |
* data structure |
| |
~ 16 minutes ~ |
14:09 |
phf |
apparently i've been ignoring netchain messages (or rather not producing right ones) |
| |
↖ |
14:12 |
phf |
here's how my packet log looks, http://glyf.org/tmp/pest-packets.pdf red lines are selfhchains, black lines are netchains. messages that are just hashes are missing from my backlog, and if either netchain or selfchain is 0 i don't render it |
| |
↖ |
14:13 |
phf |
i like how there's chains of my messages just kind of floating in the aether, until e.g. asciilifeform responds with a proper netchain to give it placement |
| |
↖ |
14:26 |
phf |
this is test |
14:26 |
phf |
hello, world |
| |
↖ |
14:26 |
phf |
pestbot: ping |
14:27 |
phf |
http://logs.bitdash.io/pest/2022-08-06#1011040 |
14:27 |
bitbot |
Logged on 2022-08-06 14:26:35 phf[awt]: hello, world |
14:27 |
phf |
blah |
14:42 |
phf |
http://glyf.org/tmp/pest-packets-fixed-netchain.pdf |
| |
~ 36 minutes ~ |
15:18 |
awt |
wow nice I've been meaning to write a tool to display chains somehow |
| |
~ 1 hours 34 minutes ~ |
16:53 |
awt |
Red is net chain? |
16:56 |
phf |
red is self chain |
16:58 |
phf |
i don't have any preferences, if you're building your own at the moment, we can standardize on any other colors, as long as they are shared, so that it's easier to read them moving forward |
17:03 |
awt |
pestbot: ping |
17:04 |
awt |
I think empty messages crash pestbot |
17:04 |
awt |
pestbot: ping |
17:04 |
pestbot |
pong |
17:08 |
awt |
phf: what did you use to generate the pdf? |
17:08 |
phf |
graphviz, it's just a simple dot graph |
| |
↖ |
17:08 |
phf |
one sec |
| |
~ 16 minutes ~ |
17:25 |
phf |
http://logs.bitdash.io/pest/2022-08-06#1011055 << http://paste.deedbot.org/?id=3kT3 |
17:25 |
bitbot |
Logged on 2022-08-06 17:08:26 phf[awt]: graphviz, it's just a simple dot graph |
17:38 |
awt |
thanks |
| |
~ 34 minutes ~ |
18:13 |
jonsykkel |
http://zzz.st/up/iT8fbjF8/giraf.png |
| |
↖ ↖ |
18:13 |
jonsykkel |
neat tool |
18:19 |
awt |
lol how crude |
18:20 |
jonsykkel |
u mean comprehensive |
18:21 |
jonsykkel |
spent like 30min on this |
18:22 |
phf |
coks stands supreme, the giraf don't lie |
18:23 |
jonsykkel |
didnt even need to tell program to do that |
18:23 |
jonsykkel |
very good software |
| |
~ 36 minutes ~ |
19:00 |
awt |
I think some sort of horizontal grouping by time increment would make phf's graph a bit more clear |
19:08 |
awt |
actually after staring at it for a while my thread seems to make sense. Just phf's long thread with no netchain is weird |
19:16 |
phf |
awt, also no time is intentional. i want to understand the packet graph's behavior outside of time domain |
19:17 |
phf |
e.g. good old topological sort should give you mostly correct packet order (of course when netchain/selfchain is valid) |
| |
~ 1 hours 1 minutes ~ |
20:19 |
awt |
net chain should zigzag between your selfchain thread and mine. But the beginning of your thread doesn't do that. There should be a net chain edge from "* data" to "apparently", but there is none. |
| |
~ 41 minutes ~ |
21:01 |
awt |
http://logs.bitdash.io/pest/2022-08-06#1011036 << ah I see you are already aware, just noticed rereading logs |
| |
↖ |
21:01 |
bitbot |
Logged on 2022-08-06 14:09:48 phf[awt]: apparently i've been ignoring netchain messages (or rather not producing right ones) |
| |
~ 42 minutes ~ |
21:43 |
awt |
Cool. Got a prototype of Prod working that actually sends out and handles GetData requests. |