#pest | 2022-06-17

9976 allows unidurds in messages. Please don't send any uniturds until a reasonable number of people have updated. Available here: http://share.alethepedia.com/blatta/9976-fecal-sprincle.vpatch

↖

01:33

awt

Perhaps phf can't help asciilifeform test that messages from uniturd handles are dropped and logged. This is very difficult to test in blatta, since it's not even possible to register a uniturd peer.

01:33

awt

*can

01:47

shinohai

awt: but muh fuzzing :(

02:01

billymg

ok, i'm on 9976 now, along with bitbot (crawlerbot next)

02:01

billymg

!. uptime

02:01

bitbot

billymg: time since my last reconnect : 0d 0h 1m

02:02

billymg

oh no, what are you doing bitbot

02:07

billymg

^ upgrading to 9976 seems to have dislodged something on bitbot's station

02:08

billymg

(none of the messages are showing up in the logs, but i'm getting them in my console here)

02:08

asciilifeform

and died again, and on restart, again

02:08

billymg

asciilifeform: update to 9976, fixed the issue for me

02:12

*

asciilifeform upped to 9976...

02:13

billymg

!c uptime

02:14

asciilifeform

loox like replays work nao also

02:15

billymg

crawlerbot now on 9976 also -- seems to be a backlog of messages though (i see them replaying in the blatta logs)

02:16

billymg

http://logs.bitdash.io/pest/2022-06-17#1007177 << ty awt, seems to be working

02:16

bitbot

Logged on 2022-06-17 01:31:27 awt[asciilifeform|billymg]: 9976 allows unidurds in messages. Please don't send any uniturds until a reasonable number of people have updated. Available here: http://share.alethepedia.com/blatta/9976-fecal-sprincle.vpatch

02:18

crawlerbot

billymg: last reconnect: 5m ago; last restart: 5m ago

02:18

asciilifeform

lol re ~errybody falling over like bowling pins

02:19

asciilifeform

awt: hm seems like not all replays get marked as such

02:19

asciilifeform

( tho replays nao finally work! a++ )

02:23

asciilifeform

awt et al : http://www.loper-os.org/pub/blatta_replay_screenshit.png

02:26

billymg

updated the downloads mirror with the latest version also

02:26

awt

asciilifeform: you only get one replay notice

02:26

asciilifeform

awt: per spec oughta for each msg that's outta order, neh

02:26

asciilifeform

otherwise how the hell woudlja know it's a replay, if mixed in w/ currents

02:27

asciilifeform

( see screenshit )

02:27

awt

Notice means everything after was received after the timestamp in the notice. Otherwise quite noisy imo

02:28

asciilifeform

noisy, arguably, but old and new msgs shouldn't intermix w/out any mark

02:28

awt

Imagine replay notice for all bazillion lines when syncing

02:28

asciilifeform

( ideally would magically appear upstream in text, but 'teletype' of irc not allows this... )

02:28

asciilifeform

awt: is exactly what asked for in spec tho

02:28

asciilifeform

or you're left guessing

02:29

awt

Could possibly add "end of replay" notice

02:29

asciilifeform

'If the Timestamp of a GetData response (or the earliest message in a chain of responses) pre-dates that of the message most recently displayed to the console, the operator must be informed of this, by prepending said Timestamp to the text emitted to the console.'

↖

02:30

asciilifeform

('3.3.1.6.4. GetData.')

02:30

asciilifeform

awt: 'end of replay' mark still dun tell you precisely which msgs were replays, neh

02:30

asciilifeform

only that some , in a segment, were

02:31

awt

How important is that?

02:31

asciilifeform

imho quite

02:31

asciilifeform

so you have some way of knowing each of'em wasn't just nao sent

02:31

asciilifeform

picture, somebody logs in after 2wks at sea

02:31

awt

Incidentally jonsykkel unaffected

02:32

asciilifeform

starts seeing ancient msgs intermixed w/ current convo

02:32

asciilifeform

awt: he prolly uses his client, which evidently not suffered from given eggog

02:32

awt

yes

02:32

jonsykkel

smalpest chads just cant stop winning

02:32

asciilifeform

lolyes

02:33

awt

asciilifeform: in most cases you won't see later messages until you've got the oldies

02:34

asciilifeform

awt: asciilifeform expected so, but in screenshit can see that apparently not

02:34

asciilifeform

likely asciilifeform even created a bit of fork, too, by speaking into a pestron which was waiting for 'head' of current msgs

02:35

asciilifeform

( tho afaik forkage not handled presently )

02:35

awt

asciilifeform: which is the message not from you that got intermixed in the replay?

02:36

*

asciilifeform loox..

02:36

asciilifeform

awt: seems like only mine, aha

02:36

asciilifeform

intermixed

02:37

*

asciilifeform wonders what will happen when these are getdata'd by the folx currently offline

02:38

asciilifeform

will they end up 'in the past' on acct of netchain ? or in 'correct' order by selfchain

02:39

asciilifeform

( is there anyone to whom this q not makes sense? )

02:41

awt

asciilifeform: will have to reload netchain/selfchain logic into my head before I can contemplate

02:41

asciilifeform

awt: no particular hurry

02:41

asciilifeform

but imho interesting q, and prolly needs elaboration in spec too

02:42

asciilifeform

( what happens to msgs spoken into a pestron which is 'behind' or fulla chain gaps )

02:42

asciilifeform

or rather, what oughta happen

02:42

asciilifeform

( 2 separate q, lol )

02:44

asciilifeform

( imho 'ideal' pestron would in fact display msg by netchain, reordering as req'd (by default; could toggle and order by timestamp, say )

02:44

asciilifeform

naturally impossible w/ irc frontend tho

↖ ↖

02:45

asciilifeform

and , apropos, when reconnect, oughta 'prod' peers, and quickly see if yer on 'same page' with'em

02:46

awt

I had been working on this update over several days. Fortunate timing that it came out of the oven tonight. Or perhaps almost fortunate.

02:46

asciilifeform

( iirc atm you have to actually speak , to learn that yer 'behind' )

02:46

asciilifeform

awt: heh

02:46

awt

asciilifeform: someone has to speak to you

02:47

asciilifeform

a

02:47

asciilifeform

possibly the 'ignores' send out to keep nat ports alive oughta be 'prods'

02:47

asciilifeform

*sent

02:48

asciilifeform

( or at least some % of'em )

02:49

asciilifeform

if done this way, would be virtually impossible to end up speaking into a desynced pestron

02:49

asciilifeform

(tho not ~entirely~ impossible)

02:50

awt

asciilifeform: true but would reduce the entropy of IGNORE command bodies.

02:50

asciilifeform

imho would suffice to send prod at a $prodinterval in place of an 'ignore' there

02:51

asciilifeform

where suggested interval is, say, 1hz

02:52

asciilifeform

let operator decide which he prefers, moar entropic chaff or smoother sync

02:53

asciilifeform

( overall flavour of pestronics, in asciilifeform's pov -- that which can be made to individual taste of $operator -- oughta be, rather than to force agreement where none is req'd )

02:55

asciilifeform

observe that per spec, (other than in 'rekey') nobody ever ~has~ to send an 'ignore'

02:55

awt

你好

03:08

*

signpost blows on mic

03:14

asciilifeform

wb signpost

~ 1 hours 47 minutes ~

05:02

phf

гуртом і батька легше бити

05:10

awt

"herd and dad is easier to beat"? google fails me

~ 8 hours 2 minutes ~

13:12

PeterL

Hi, I have updated to 9976

~ 2 hours 20 minutes ~

15:33

phf

awt: together with frens it's even easier to beat up dad

15:37

asciilifeform

wb phf

15:38

asciilifeform

phf: are you using own cl pestron nao ?

15:40

phf

asciilifeform: i've always been using own cl pestron, but i switched from reading logs to hand unpacking packets. the decoder is there, but the pipeline is manual :)

15:41

asciilifeform

a++

15:41

asciilifeform

phf: consider posting , even if thing aint half done just yet

15:42

*

asciilifeform will add to links , then, and we can say 'we've 3 prototypes!'

15:43

*

asciilifeform at one time started on a cl pestron, but not got far before sank into salt mines again

15:50

*

PeterL looking forward to having prod implemented for when sation starts, seems silly that it sits there potentially indefinately until somebody says something and then I get pages of replays

15:50

signpost

phf: cool!

15:51

PeterL

also, I agree with asciilifeform, it is not clear which messages are new and which are still replays

15:52

signpost

yup, what's the current notion, turn them into notices?

15:52

asciilifeform

signpost: per spec

15:52

bitbot

Logged on 2022-06-17 02:29:40 asciilifeform: 'If the Timestamp of a GetData response (or the earliest message in a chain of responses) pre-dates that of the message most recently displayed to the console, the operator must be informed of this, by prepending said Timestamp to the text emitted to the console.'

15:53

signpost

yeah that works

15:53

asciilifeform

would be a++ incidentally to have a www logotron that reorders correctly (i.e. eats from pest db, rather than irc frontend)

15:53

signpost

eventually a bespoke UI can handle more clearly too

15:53

asciilifeform

aha

15:54

*

asciilifeform doesn't expect anyone to bake a gui any time soon, given the epic gnarl of anyffin to do with guis

15:57

signpost

pest-aware logotron's a fine step closer.

↖

15:58

signpost

weechat plugin might also be capable of rewriting buffer history and scrolling up to the oldest new message, and rendering all previously unseen with a highlight or similar.

15:58

signpost

none of these things are urgent, just throwing ideas out.

15:58

*

asciilifeform currently pesting on weechat, not overly fond of it on acct of difficulty of clean cut'n'paste

15:59

asciilifeform

is the reason wai used xchat & similar historically

15:59

*

signpost only deals in leastbad in software selection

16:00

signpost

dunno how extensible erc is, but would assume "very".

16:02

asciilifeform

signpost: seems that you'd have to either implement entire pestron in elisp (ugh) or use some kinda glue (and then simply anuther irc frontend?)

16:02

asciilifeform

ideally imho gui pestron would be monolithic

16:03

asciilifeform

( and support threading, etc. and other knobs absent from irc frontend )

16:04

signpost

curious what phf has in mind

16:05

*

signpost found mcclim to be a hobbyist joke (fielded a product that used this for a minimal restaurant kitchen UI)

↖

16:05

signpost

actually moved to mcclim from w/e the cl Qt binding is called.

16:06

awt

I attempted prepending timestamps at one point but ran into some issue.

16:06

signpost

message length?

16:07

awt

signpost: that's a concern in certain cases but not what I ran into. Could have been something dumb like the colon in the timestamp confusing the IRC client.

16:09

awt

Also, as I said above, a notice per line makes it really unreadable.

16:09

asciilifeform

prepending is potentially quite gnarly (may need to re-chunk)

16:13

awt

Bouncers pull it off somehow. IIRC the text is formatted in italic.

16:18

awt

http://logs.bitdash.io/pest/2022-06-17#1007287 << indeed or even just a script you could run to dump the logs into less or whatever reader you prefer.

16:18

bitbot

Logged on 2022-06-17 15:57:13 signpost: pest-aware logotron's a fine step closer.

16:18

asciilifeform

awt: znc in particular uses some kinda oddball extension to protocol, which some clients support

~ 25 minutes ~

16:44

asciilifeform

http://logs.bitdash.io/pest/2022-06-17#1007298 << afaik all attempts at gui glue are doomed to be either intimately os-specific (and straightjacketed firmly into whatever knobs $os supports) or 'hobbyist joke'; or, often enuff, ~both~ of these

↖

16:44

bitbot

Logged on 2022-06-17 16:05:05 signpost: found mcclim to be a hobbyist joke (fielded a product that used this for a minimal restaurant kitchen UI)

~ 2 hours 35 minutes ~

19:19

PeterL

billymg: do I remember correctly that there is some way to get your logs as black text on white? the white text on black background hurts my eyes if I try reading more than a minute or two

~ 1 hours 1 minutes ~

20:21

billymg

PeterL: it's coming, not there yet