00:57 |
jonsykkel |
u gota run serpent_init on the 32byte key to expand it |
00:57 |
jonsykkel |
serpent_init(bufer_of_528_bytes,cipher_key,SERPENT_KEYSIZE) |
00:58 |
jonsykkel |
then use that for cipher funcs |
| |
~ 42 minutes ~ |
01:40 |
awt_akris |
jonsykkel: ty. updated. decrypted using both pure python and c serpent and got the same result. |
01:53 |
jonsykkel |
nice |
| |
~ 8 hours 41 minutes ~ |
10:35 |
asciilifeform |
http://logs.bitdash.io/pest/2023-07-17#1028453 << strange; yer at entry does periodically update |
10:35 |
bitbot |
Logged on 2023-07-17 23:01:03 awt_akris[awt|signpost]: asciilifeform: I manually added you at the IP and port I have you at on my other station but I'm not seeing a response to PROD messages going to your station. |
10:35 |
asciilifeform |
i.e. packets are going |
10:36 |
asciilifeform |
last update ~9h ago |
10:49 |
unpx |
http://logs.nosuchlabs.com/log/pest/2023-07-13#1028647 <<< re: no vpatch for docs, in my previous discussions I wanted to bring out this fact, we are still [http://logs.nosuchlabs.com/log/pest/2023-03-30#1025 |
10:49 |
unpx |
788][dreaming] a minimalistic self-maintained software stack |
10:49 |
dulapbot |
Logged on 2023-07-13 14:50:16 asciilifeform: not baked proper vpatch yet for this thing, but will |
10:49 |
dulapbot |
Logged on 2023-06-09 12:43:19 unpx[asciilifeform]: So I guess we are all in a UNIX-like system because it is the most hackable we can have right now without becomeing a system programmer |
10:49 |
dulapbot |
Logged on 2023-07-06 19:06:02 unpx[asciilifeform]: Or at least guide people so they can write they own tools to do all of that the way they want |
10:50 |
unpx |
no vpatch for compiled docs* |
10:53 |
unpx |
[...] we are still dreaming a minimalistic self-maintained software stack. |
10:57 |
dulapbot |
Logged on 2023-03-30 16:44:45 unpx[asciilifeform]: PeterL, asciilifeform: indeed, but what then? I may have a comfy machine for working and then a machine for specific bank/gov stuff, but then what would you have in that work machine? wget, lynx port, Konqueror, Pale Moon? Or would you reduce documents to TXT and forget HT |
| |
~ 44 minutes ~ |
11:41 |
asciilifeform |
unpx: 'if wishes were horses' then neither lynx, nor konkeror, etc. but rather sumthing entirely new and bolixlike, with buncha modular, reusable routines rather than standalone 'apps' |
11:42 |
unpx |
Indeed, the whole "IDE" should be aware |
| |
~ 23 minutes ~ |
12:06 |
unpx |
Moreover, I'm still looking for such a design solution that manages both documentation and code and tie the two together |
12:06 |
dulapbot |
Logged on 2023-05-18 16:37:31 unpx[asciilifeform]: But I cannot see tools that care about both, they are more about the phase one rather then the phase two. |
| |
~ 35 minutes ~ |
12:41 |
signpost |
manages in what sense? |
12:48 |
unpx |
I cannot find the discussion about connecting pdf documents to vpatches, signpost |
12:49 |
asciilifeform |
unpx: vpatches are intended to (borrowing the gpl folxs' terminology) concern 'the form of the document preferred for modification' |
12:49 |
asciilifeform |
so, not pdf, but e.g. its latex src, for instance |
12:51 |
signpost |
yeah, same goes for codegen like the automake cruft currently preserved on pentacle's dissection table. |
12:51 |
unpx |
Yeah, but I recall a discussion where a certain vpatch was mapped to a certain documentation version |
12:51 |
signpost |
generated srcs should in principle not be in there either. |
12:52 |
signpost |
derived items of w/e kind |
12:52 |
unpx |
Probably phf drew a graph |
12:52 |
signpost |
leaving aside for a moment that if you need codegen you're probably already in a bad place. |
12:57 |
unpx |
>markdown docs |
12:59 |
unpx |
something something dependency.png |
13:05 |
unpx |
aye, it was jonsykkel |
13:05 |
dulapbot |
Logged on 2023-03-25 21:46:41 jonsykkel[asciilifeform]: example http://zzz.st/up/olNWGD1z/example.png |
13:06 |
unpx |
Will ponder a bit with phf tools too |
| |
~ 46 minutes ~ |
13:52 |
awt_akris |
Alright now running with jonsykkel's serpent (calling it c_serpent). CPU usage looking good. |
13:56 |
awt_akris |
Also flushing the toilet is not resulting in poo in the freezer. |
13:57 |
asciilifeform |
awt_akris: a++ |
13:58 |
* |
asciilifeform still puzzled wai direct peering not worx. given that some packets are getting through |
14:02 |
awt_akris |
asciilifeform: this is the prod I'm sending you that's not getting a reply: http://paste.deedbot.org/?id=nr9x |
14:03 |
asciilifeform |
awt_akris: indeed not appears in log here |
14:04 |
awt_akris |
I am behind a nat here, but everyone else is replying so maybe not my nat. |
14:04 |
asciilifeform |
awt_akris: hmm wai same ip ( asciilifeform's) on both sides of the quoted line ? |
14:04 |
asciilifeform |
did log notation change from blatta's ? |
14:04 |
signpost |
lol, think how quickly you can unpack the groceries when you line 'em up JUST SO |
14:04 |
signpost |
oh god, impaled the cat with a pineapple |
14:04 |
asciilifeform |
signpost: hm? |
14:04 |
signpost |
bad joke about reading into a struct |
14:05 |
asciilifeform |
aa |
14:05 |
signpost |
slow on the draw |
14:05 |
awt_akris |
asciilifeform: first IP is the address the prod is being sent to. second IP is the last known address of the intended recipient. |
14:05 |
asciilifeform |
a! got it |
14:08 |
awt_akris |
signpost: not sure which struct you're referring to. Desperately want to get the joke. |
14:10 |
awt_akris |
asciilifeform: double checked the key and seems right. On blatta, it is possible to accidentally add spaces to the key when entering it. |
14:11 |
asciilifeform |
key defo right (or wouldn't be seeing at updates here, neh) |
14:11 |
awt_akris |
What address do you have for me? |
14:11 |
signpost |
awt_akris: it's common to point a struct pointer directly at where you recv'd a packet to chunk into fields |
14:12 |
signpost |
making a joke about how this goes wrong with field misalignment |
14:12 |
asciilifeform |
awt_akris: 99.14.101.191:1024 |
14:12 |
signpost |
it was funnier in my head |
14:13 |
awt_akris |
asciilifeform: yep that's correct. |
14:16 |
awt_akris |
signpost: speaking of impaling cats, should be a lot easier to integrate ocpy now that I've got a working example going with c_serpent. |
14:16 |
signpost |
turd's pretty slow still, but it's a starting point |
14:16 |
signpost |
c one's fast but unfinished |
14:17 |
awt_akris |
signpost: sorry I was referring to the c version. Can't remember what it's called. |
14:17 |
asciilifeform |
will also work a++ in asciilifeform's (sadly stalled atm) cpp pestron, come to think of it |
14:17 |
signpost |
still might be worthwhile to try the thing out with ocpy |
14:17 |
signpost |
occ |
14:17 |
awt_akris |
oh lol ok makes sense. |
14:18 |
* |
signpost did some work on it yesterday evening, probably not too far off |
| |
~ 1 hours 49 minutes ~ |
16:07 |
* |
unpx doesn't know if asciilifeform and signpost understood what I did mean here by ``manages both documentation and code'' http://logs.nosuchlabs.com/log/pest/2023-07-18#1028928 |
16:07 |
dulapbot |
Logged on 2023-07-18 13:03:29 unpx[asciilifeform]: aye, it was jonsykkel |
16:09 |
awt_akris |
asciilifeform: which is better? Annoy operators whose clock has drifted back, or use the time the message was received at as the timestamp to order messages by when displaying them? Or something else? |
| |
↖ |
16:10 |
unpx |
http://logs.nosuchlabs.com/log/pest/2023-07-18#1028926 <<< w/o proper tools to render it, markdown is not worth `compiling' to html, since you then need a browser |
16:10 |
dulapbot |
Logged on 2023-07-18 12:55:57 unpx[asciilifeform]: >markdown docs |
16:10 |
awt_akris |
IIRC ntp is not a desirable dependency. |
16:12 |
awt_akris |
I'm finding frequent instances of out of sync clocks and it does disrupt log reading during a hot thread. |
| |
~ 19 minutes ~ |
16:32 |
signpost |
why not order by netchain with a recursive query? (leaving aside indexing or w/e to speed up, might not be needed) |
16:32 |
signpost |
imho the ordering I care about is the one the speaker intended, resolving ties in w/e sensible way. |
16:39 |
awt_akris |
signpost: the goal is ultimately to order by netchain - currently blatta's handling of netchain is unspeakably broken, so until we have a reliably contiguous netchain, gotta order by timestamp... |
16:39 |
awt_akris |
With what I have right now, I could attempt to use netchain as a heuristic when timestamp < netchain tip's timestamp. Even so, the spec disallows such messages - technically akris should be discarding them. |
| |
↖ |
| |
~ 45 minutes ~ |
17:25 |
signpost |
makes sense |
| |
~ 23 minutes ~ |
17:49 |
asciilifeform |
http://logs.bitdash.io/pest/2023-07-18#1028527 << track the deltas (while warning operator) and estimate 'troo' time by weighted pull of peers with historically low drift. |
17:49 |
bitbot |
Logged on 2023-07-18 16:09:56 awt_akris[awt|signpost]: asciilifeform: which is better? Annoy operators whose clock has drifted back, or use the time the message was received at as the timestamp to order messages by when displaying them? Or something else? |
17:50 |
asciilifeform |
( if they're all drifting by a ~constant, the br0ken 1 is.. you ) |
| |
↖ |
17:51 |
asciilifeform |
the original conceit, however, was that pestrons only need to be 'city tower bell' accurate (i.e. +/- 15m) to work reliably. |
17:51 |
asciilifeform |
( vs. e.g. ntpism ) |
17:51 |
asciilifeform |
conceivably can sync with sundial. |
17:51 |
asciilifeform |
'worx 100% with sundial time' was explicit goal of asciilifeform's orig design. |
17:52 |
asciilifeform |
sundial being the only apolitical clock avail. on planet3. |
18:04 |
* |
signpost has long wanted to do an optical clock which recalibrates using the night sky |
18:06 |
awt_akris |
http://logs.bitdash.io/pest/2023-07-18#1028539 << On this pestnet, this heuristic could turn out to be an "insanity check" |
18:06 |
bitbot |
Logged on 2023-07-18 17:50:28 asciilifeform[5]: ( if they're all drifting by a ~constant, the br0ken 1 is.. you ) |
18:08 |
awt_akris |
asciilifeform's clock consistently matches my own, fwiw. |
| |
~ 1 hours 55 minutes ~ |
20:04 |
asciilifeform |
iirc mine synced from router which syncs from isp which from hitler |
| |
~ 1 hours 35 minutes ~ |
21:40 |
awt_akris |
Testing out using recieved_at - "worx" but would not work well during initial sync when received at could be years off from the packet timestamp. |
21:43 |
awt_akris |
Testing out using recieved_at - "worx" but would not work well during initial sync when received at could be years off from the packet timestamp. |
21:56 |
awt_akris |
Not a dupe ^ an exception was preventing the message from displaying and I sent it twice. |
| |
~ 28 minutes ~ |
22:24 |
awt_akris |
nm can just use originator timestamp for getdata responses. |