00:05 |
awt |
Pretty cool watching the log view get updated in realtime during an initial sync. |
| |
~ 10 hours 25 minutes ~ |
10:30 |
awt |
!!ticker btc usd |
10:30 |
deedbot |
$30,118 |
10:33 |
awt |
First round of syncing got me back to: http://logs.bitdash.io/pest/2021-11-16#1000669 |
10:33 |
bitbot |
Logged on 2021-11-16 13:25:16 PeterL: hey everybody |
10:33 |
awt |
If I reset the failed GetDatas I could probably get more. |
| |
~ 1 hours 43 minutes ~ |
12:17 |
signpost |
!!ticker btc usd |
12:17 |
deedbot |
$31,129 |
| |
~ 1 hours 14 minutes ~ |
13:31 |
phf |
http://logs.nosuchlabs.com/log/pest/2023-06-23#1027659 << i think the correct method is to trigger a cluster of getdatas on a periodic schedule, maybe keeping track of their counts/failures. so you start by spooling a moving window of getdatas into the past, eventually settling on |
13:31 |
phf |
infrequent getdatas for packets that never returned |
13:31 |
dulapbot |
Logged on 2023-06-23 10:36:06 awt[asciilifeform]: If I reset the failed GetDatas I could probably get more. |
13:34 |
* |
asciilifeform suspects that in all but pathological cases ( msg sent by ancient buggy pestrons ) getdata will return in <1s |
13:34 |
asciilifeform |
for broadcasts, that is |
13:35 |
asciilifeform |
( for directs, may have to wait until the pertinent peer is online, if he's cold ) |
13:36 |
* |
asciilifeform could be wrong tho, will be interesting to see what awt finds re the failed getdatas -- do they ever succeed on retry ? (i.e. is it worth to have the moving parts phf was referring to) |
13:38 |
asciilifeform |
then again when using questionable modem, 'sliding' getdata oughta win |
13:38 |
asciilifeform |
so scratch all of above, lol |
13:43 |
awt |
Can confirm that they succeed on retry. Resetting attempts to 0 after initial sync putters out results in more responses. This is afaict due to blatta getting overloaded. (Shouldn't happen with multithreaded station) |
13:45 |
awt |
Current approach is perhaps similar to "sliding window". Turns out to be expanding sliding window tho. |
13:46 |
awt |
Also, currently only connected to 1 station. Should see better results when connected to multiple. |
13:47 |
awt |
"multi-process approach" |
| |
~ 1 hours 59 minutes ~ |
15:46 |
awt |
phf: I'm running into issues displaying one of your messages. I these two characters: \udbff\udffd - they seem to be "surrogate pair codepoints used by UTF-16". Any idea how these originated or should be dealt with? They're also a possibly byproduct of my hack to remove multipart data from messages marked as non-multipa |
| |
↖ ↖ |
15:46 |
awt |
rt. |
| |
~ 47 minutes ~ |
16:34 |
awt_akris |
test |
16:37 |
awt_akris |
This time without the newline |
16:40 |
awt_akris |
!!ticker btc usd |
16:40 |
deedbot |
$30,881 |
| |
~ 34 minutes ~ |
17:14 |
asciilifeform |
http://logs.bitdash.io/pest/2023-06-23#1027265 << iirc phf had implemented the prev. version of asciilifeform's 'pre-draft', where these chars marked multipart text msgs ( later, per phf's suggestion, asciilifeform revised spec to make'em their own msg type ) |
17:14 |
bitbot |
Logged on 2023-06-23 15:46:25 awt: phf: I'm running into issues displaying one of your messages. I these two characters: \udbff\udffd - they seem to be "surrogate pair codepoints used by UTF-16". Any idea how these originated or should be dealt with? They're also a possibly byproduct of my hack to remove multipart data from messages marked as non-multipa |
17:15 |
* |
asciilifeform had picked those particular uniturds, having determined that they don't seem to produce catastrophic effects in afaik any existing displayer |
17:16 |
asciilifeform |
( nor occur 'naturally' in any extant text likely to be pasted ) |
17:21 |
asciilifeform |
thread |
17:21 |
dulapbot |
Logged on 2022-11-28 11:06:10 phf[asciilifeform]: unfortunately until there's a great switchover, can't really do multipart reliably, since it'll spit noise into log |
17:22 |
asciilifeform |
( but! oddly not == to what awt saw in the messages tho, so possibly red herring?? ) |
17:23 |
* |
asciilifeform evidently not correctly recalled the uniturd he had used |
| |
~ 2 hours 58 minutes ~ |
20:22 |
signpost |
https://zerohedge.com/geopolitical/wagner-chief-declares-full-coup-against-russian-defense-ministry << wild, did the west flip prigozhin |
20:25 |
asciilifeform |
signpost: 'everyone has his price' |
20:25 |
signpost |
guess so eh. |
20:25 |
signpost |
guy has serious pills and booze physiognomy |
20:33 |
asciilifeform |
eh, that's just ordinary 'orc >40' lol |
20:33 |
signpost |
haw |
20:33 |
asciilifeform |
re upstack, see also. |
| |
~ 1 hours 54 minutes ~ |
22:28 |
awt_akris |
tk widgets are practically useless. Can't even sent the font inside a Listbox. |
22:28 |
awt_akris |
everything must be a Text |
22:33 |
signpost |
this magical teletype box is vanity anyway. we must return to the quill. |
22:34 |
* |
awt strokes remarkable2 |
22:34 |
* |
signpost has also, fantastic device. |
22:35 |
awt |
been using to to reference the pest spec quite a bit |