tmsr - Log 2018-10-16

asciilifeform: in oooother noose, 'pixelqi' lcd ( thread http://btcbase.org/log/2018-10-11#1860677 ) is ~not~ compat with 'lenovo s10-3' -- screw holes in wrong places , and connector dun reach . ☝︎

a111: Logged on 2018-10-11 17:08 asciilifeform: !Q later tell phf i recently found that lcd panel made by (defunkt) 'pixelqi co', pq3qi-01 , is still sold by chinese ( i have nfi if clones, or old stock ), ~100 $ ea; the thing toggles from 1024x600 colour/backlit into a 3072x600 greyscale reflector thing, worx with various lappies ( sadly not x60, it's a 16:9 )

asciilifeform will try again with 'samsung n150' box, with sadder kbd, but same chipset, and reportedly fitting ☟︎☟︎

lobbes: http://btcbase.org/log/2018-10-15#1862662 << yeah, I think I must've still been waking up that morning. my unexamined line of thought at the time was that since it doesn't list in here, there's no need to have it here. Then I remembered there's still use in, e.g. being able to bid, view orders, etc. ☝︎

a111: Logged on 2018-10-15 03:47 mircea_popescu: http://btcbase.org/log/2018-10-14#1862346 << wait, what ?

lobbes: I'll start working to get the voicing figured out for auctionbot and bring it in here.

lobbes: http://btcbase.org/log/2018-10-15#1862788 << this is a good idea for a feature. Once I get logbot-voicing figured out I'm thinking of setting up a general purpose 'miscbot' for such things as the later-tell and the duckduckgo search. I'll try and work this into the later-tell feature ☝︎

a111: Logged on 2018-10-15 10:20 mircea_popescu: http://btcbase.org/log/2018-10-15#1862617 << speaking of which, hey lobbes, feel like adding a param to !Q later tell, whereby it can tell next it sees guy but not earlier than x hours ?

lobbes: well, my ultimate aim is to eventually rewrite ALL of lobbesbot's functionality to stand on top of logbot versus the heathen limnoria. auctionbot was the first victory in this regard.

lobbes: but there's still the !Qlater tell, the #eulora log-quoting, #eulora log search, etc. I guess I figured why not take the opportunity to organize vastly different functionalities into separate bots. (e.g. all the logger-related functionality in one bot, auction/ticker-announce/price-history in auctionbot, everything-else in 'miscbot')

mircea_popescu: and if you're putting them all on the same box what's "separate" mean anyway ?

lobbes: but maybe you have a point. I gotta noodle on it. there's a balance somewhere

mircea_popescu: is it that your mental picture of the bot gets too gnarly because of all these diff functions ?

lobbes: there's that. then there's also the "command namespace collisions" issue. For example http://btcbase.org/log/2018-01-15#1770696 ☝︎

a111: Logged on 2018-01-15 11:44 lobbes: btw "!Qsarchive" (without the "-") also works to call the command. ("!Qs" was already in-use to search logs.minigame.bz)

mod6: http://btcbase.org/log/2018-10-15#1862873 << Thanks for reporting. I'll add this to the list to take a look @ & replicate. ☝︎

a111: Logged on 2018-10-15 15:44 asciilifeform: in unrelated noose, mod6 et al : loox like trb doesn't remove banned peers from its addr table ! i.e. offers their addrs up to others, as if they were proper noades!! i never noticed this previously, it had to wait to be found empirically. really oughta be fixed, and pretty simple patch.

lobbes: yeah, that makes sense. Ol' causes and purposes. The only cause I can see right now for separate bot is if wanting to keep it on separate box

mircea_popescu: meanwhile in french class, "Je sens sur ta lèvre une odeur de fièvre de gosse mal nourri ; et sous ta caresse je sens une ivresse qui m'anéantit"

ave1: http://btcbase.org/log/2018-10-15#1863087, look at the pasted code. Yes, you move it from a dynamic allocation on the heap to a "dynamic allocation" on the stack. With the major advantage that the stack is a well defined mechanism with predictable results where dynamic allocations are not. ☝︎

a111: Logged on 2018-10-15 19:13 bvt: ave1, can you expand on getting rid of string allocations? I see a way to get rid of it for Mode, but not for the Path argument. i can move its c-string from heap to stack, but that would be it. anyhow, i will revise this aspect.

ave1: Using the stack makes code less error prone and a lot more readable, please read the FFA code; http://www.loper-os.org/?cat=49.

bvt: thanks, i see what you mean. the Interfaces.C.Strings.New_String() comments explicitly say that the heap allocation is necessary for the large strings, which is not the case for vpatch temporary file names.

bvt: i have already reworked the code based on the following code from GNAT: http://p.bvulpes.com/pastes/xjjec/?raw=true (g-locfil.adb, lockfile implementation). i.e. System.Address of an aliased String as argument to C function.

deedbot: http://ossasepia.com/2018/10/16/smg-comms-chapter-2-raw-types/ << Ossasepia - SMG Comms Chapter 2: Raw Types ☟︎

mircea_popescu: yknow it occurs to me, one could make a likely very popular film, call it "captive" or w/e, simply through filming the male seminal canal from teste to ligature point in one of those tube-tied doods. 4 minuites of voice-over at the beginning explaining what's what ("this is joe. joe is a sperm." bla bla) and then the rest of an hour and a half just... let people watch the random walk.

mircea_popescu: http://btcbase.org/log/2018-10-16#1863187 << this'll take some rereading ; ima take it with me on trip. ☝︎

a111: Logged on 2018-10-16 11:20 deedbot: http://ossasepia.com/2018/10/16/smg-comms-chapter-2-raw-types/ << Ossasepia - SMG Comms Chapter 2: Raw Types

mircea_popescu: http://btcbase.org/log/2018-10-15#1863111 << thinking about this in the interim, it is EVIDENTLY the case. "files" as in ~the fucking thing REFERENCED~ in the "file"system names aren't even first class abstractions! the drive's a BYTE device! there's no fucking files ANYWHERE except in the imagination. ☝︎☟︎

a111: Logged on 2018-10-15 19:24 mircea_popescu: s a file system AT ALL, and the "all things are files" is just a thin wraper on a turd sausage made out of "we have no data model beyond ram and our disk doesn't actually work".

mircea_popescu: what passes for a "file system" currently is this attempt to force a byte-device (sometimes with some half-baked geographical allocation bolted on) through a half-baked database model. wtf is a "FAT" or a "journal" besides poorly implemented mysql/postgres dben ?

mircea_popescu: in any case there's no such thing as a filename anywhere at all, it's not even 2nd class, it's a sort of indian.

mircea_popescu: and obviously most of the reason the sql flavour is broken is because "oh, structure could never be aforeknown" and so "must have maximal hallucinatory freedoms".

mircea_popescu: "then you want to put these '''maximal freedom''' elements in a tree ?" "it's not a tree, it's a '''directory structure'''." "wtf is that ?" "a half-baked tree." "you're really dedicated to this half-baked thing beyond all else, aren't you ?"

mircea_popescu: man-drive dialogue flows always like : "i want file X" "wtf is that ?" "tell me what i told you before file X is supposed to be" "bytes 4978459876 through 50987498098" "i want bytes 4978459876 through 50987498098 please" "here you go" "are these file X ?" "fuck you" "does file X even exist ?" "little bitch..." "grep -a -C 500 'known pattern' /dev/sda | tee /tmp/recover" ☟︎

asciilifeform: http://btcbase.org/log/2018-10-16#1863191 << bonus: nobody actually sells a hdd that's 'byte device', they're addressable in 512byte blox ( and internally with 4096byte flash sectors ) and it actually makes a diff , not only in re speed but the life of the iron ( erry sector can be written ~2000 times ) ☝︎

a111: Logged on 2018-10-16 11:44 mircea_popescu: http://btcbase.org/log/2018-10-15#1863111 << thinking about this in the interim, it is EVIDENTLY the case. "files" as in ~the fucking thing REFERENCED~ in the "file"system names aren't even first class abstractions! the drive's a BYTE device! there's no fucking files ANYWHERE except in the imagination.

lobbesbot: asciilifeform: The operation succeeded.

asciilifeform: diana_coman: if my nitpick isn't clear, i can post an example later today.

a111: Logged on 2018-10-16 11:51 mircea_popescu: man-drive dialogue flows always like : "i want file X" "wtf is that ?" "tell me what i told you before file X is supposed to be" "bytes 4978459876 through 50987498098" "i want bytes 4978459876 through 50987498098 please" "here you go" "are these file X ?" "fuck you" "does file X even exist ?" "little bitch..." "grep -a -C 500 'known pattern' /dev/sda | tee /tmp/recover"

asciilifeform: a http://btcbase.org/log/2018-10-15#1863136 - machine needs no predefine fs, you can use whatever data structures you like and they'll still be there when the power is cycled, per http://www.loper-os.org/?p=231 formula. ☝︎

a111: Logged on 2018-10-15 19:34 asciilifeform: for instance 1 possible answer is 'no files, errything is an s-expression, and names live in gns'

asciilifeform: the fs as we know it is a classic example of 'plugging funnel at wrong end'.

lobbesbot: diana_coman: Sent 14 minutes ago: <asciilifeform> http://ossasepia.com/2018/10/16/smg-comms-chapter-2-raw-types/comment-page-1/#comment-4339

asciilifeform: when disk was refrigerator-sized, its write speed was 100kB/sec, and ram cost 1 toyota per 128kB, in 1980, it made sense to ask operator to manually move data structures to/from disk/ram. imho today it does not. run in ram, persist it to flash when power cuts out , running from supercap.

asciilifeform: diana_coman: dun worry, it will make sense very quickly when you bake the record

diana_coman: I expect /hope so; meanwhile it's very good to have that observation there, thank you!

asciilifeform: diana_coman: didja ever look at my 'nqb' example ? it has buncha records , that get serialized/deserialized correctly

asciilifeform: ( some components -- 'manually', as shitoshi used wildly variable sizes; some -- cleanly )

diana_coman: at any rate when I get to the records I might come and whine loudly if I get stuck ☟︎

asciilifeform: diana_coman: you won't need to do any of the streams horror that i did there, because those of your subfields that are arrays, have strictly bounded sizes ☟︎

diana_coman: thing is: there are lots of this "int8 number of keys" followed by "n*int8 key ids" and this in turn is just part of the wider thing

asciilifeform: diana_coman: if you get stuck i'ma show you how to make it go, cleanly, i have a working example of nearly the same thing you're doing ( gossip prototype thing ) that simply hasn't been cleaned for publication of yet

asciilifeform: of all the various items asciilifeform has used over the yrs, ada ( and not even 'modern', but even back in ada-83 ) has the absolute sanest support for strict record layout that actually worx as described

asciilifeform: kinda like cl's 'binary-types' thing, but as part of the language, rather than add-on hackolade

diana_coman: and added for clarification re converters since it struck me that it might be unclear there what's all about: http://ossasepia.com/2018/10/16/smg-comms-chapter-2-raw-types/comment-page-1/#comment-4341

diana_coman: that was my whole point of the separation between layer 0 and layer 1 there: at 0 everything is just an array of octets, nothing more

asciilifeform: diana_coman: aha; you dun actually need to shovel raw octets anywhere other than the udp transceiver

asciilifeform: back when i wrote in c, i also serialized/deserialized 'just once', when casting packet to the correct struct, but to precisely nail down the latter required gcc-specific pragmas, as c per se does not give this ability. but in ada the required knobs are standard.

deedbot: http://bimbo.club/?p=54 << Bimbo.Club - TMSR Log Summary - 10/13/2018

a111: Logged on 2018-10-16 13:56 asciilifeform: diana_coman: you won't need to do any of the streams horror that i did there, because those of your subfields that are arrays, have strictly bounded sizes

mircea_popescu: things are kinda deliberately made to dovetail with ada, so i hope they do and if they don't we fix.

diana_coman: kk; I'll keep up this approach of published-even-if-stumbling small steps so on one hand there is something concrete to shoot down if needed and on the other hand we can correct any trouble early on, hopefully

mircea_popescu: anyway, /me shall be back ~thurs, though might say hi on the road. but will certainly read teh logs!

deedbot: diana_coman updated rating of phf from 2 to 4 << writes at http://barksinthewind.com/

deedbot: diana_coman unrated EDLionX.

deedbot: diana_coman updated rating of hanbot from 2 to 5 << writes at http://thewhet.net/; met irl.

deedbot: diana_coman updated rating of mircea_popescu from 5 to 7 << writes at trilema.com; at the centre of it all, like it or not; met irl.

deedbot: diana_coman updated rating of mod6 from 3 to 4 << writes at http://mod6.net/

deedbot: diana_coman unrated wyrdmantis.

deedbot: diana_coman updated rating of lobbes from 2 to 3 << writes at http://blog.lobbesblog.com/

deedbot: diana_coman unrated andreicon.

deedbot: diana_coman updated rating of esthlos from 1 to 2 << writes at http://blog.esthlos.com/

deedbot: diana_coman updated rating of BingoBoingo from 3 to 4 << writes at bingology.net and qntra.net; moved to Uruguay and started Pizarro ISP.

deedbot: diana_coman updated rating of trinque from 3 to 4 << writes at trinque.org; deedbot and cuntoo; he makes me want again to visit Texas.

deedbot: diana_coman rated bvt 1 << introduced himself with well formed question on my blog

deedbot: diana_coman rated deedbot 10 << Bot: voice, deeds, wallet; http://deedbot.org/help.html

BingoBoingo back from paying the rent, pretty sure they eyes managed to get a bit of sunburn in spite of sunglasses

asciilifeform: ( possibly the 1 win from wearing glasses, is the uv filter coat... )

BingoBoingo: Well, in this southern hell afternoon walks may lead to a bit of a headache. It increases in intensity and seens to be centered around the eyes.

asciilifeform: mno then not this, no headache from cornea burn, just 1 hell of an itch and quicker cataract

diana_coman: BingoBoingo, second asciilifeform there: it does sound like heat stroke really; cover your head

diana_coman: I can say the same, but honestly, it was never enough protection against the sun (and fwiw I have thick hair anyway)

asciilifeform: when i was in BingoBoingostan, found myself on numerous occasions wishing for cork helmet.

asciilifeform: but interestingly 0 sunburn, in asciilifeform's experience sunburn seems to exist only in usa..

diana_coman: ahahah; well, I did get sunburn on my shoulders in Costa Rica but mainly because of being in the water most of time while at the beach

BingoBoingo has met a number of not entirely white Brazillians who managed to pick up profound beach sunburn

deedbot: http://pizarroisp.net/2018/10/16/pizarro-september-2018-provisional-statement/ << PizarroISP - Pizarro September 2018 Provisional Statement

diana_coman: I can easily picture that really; at any rate, I got otherwise at least 2 shades darker after the few days by the ocean and even now (so...2 months later) I'm still darker than my natural shade

BingoBoingo: My arms and legs manage to tan to darker shades with sun exposure, but my face pretty much constantly stays in a state of more or less accute sunburn regardless of the quantity and frequency of "bloqueador de sol" applied

Mocky: BingoBoingo, usually by the end of winter I look like a coffee with too much cream, but slightly too much summer sun and I turn into kunta kinte for 6 months ☟︎

BingoBoingo: Mocky: That sounds rather healthy. My face and ears have taken to the color red as a seemingly permanent part of their color palette

BingoBoingo: ATTN: Over the next few weeks I will be making some fairly big changes to the Pizarro website (theme chopping, pictures, etc) If the pizarro site is broken and stays broken please let me know.