44300+ entries in 0.374s

ben_vulpes: yeah i recently tried
a 'modern' asdf and recoiled in horror at how much it broke
phf: from xach's quicklisp update "This update was created with an older version of SBCL. The latest SBCL includes ASDF 3.3.1, which breaks
a handful of projects in ways that have not been resolved yet."
mod6: lol, you're giving away
a herse?
pete_dushenski: just read about the 5451-piece richard mille "jura" clock in quebec city, as it happens, and i'll definitely be checking it out next time i'm there. apparently the thing took 6 years to build, which isn't so hard to believe when you see that incorporates
a perpetual calendar (which knows leap years from non-leap years, unlike an annual calendar), rementoire d'egalite (for improving accuracy), and equation of
pete_dushenski:
http://btcbase.org/log/2017-12-23#1757448 << it just so happens that i saw one in vancouver's gastown neighbourhood last month. not the most tuneful steam whistle and it looks far older than the 40 years of age it actually is, but the plaque did indicate that it was something of
a technical innovation. looked like
a simplified rube goldberg machine to me.
☝︎ a111: Logged on 2017-12-29 20:50 asciilifeform: ( asciilifeform , like complete idiot, went and thought... 'i can make
a useful diagram! with svg ! which exists!' )
a111: Logged on 2016-09-18 22:53 phf: yes, scp, screen, asdf. i had
a v based deployment but i wasn't happy with it, so i'm trying to rethink it
mircea_popescu: nor is it reasonable thereby to expect "this lisp instance runs while
a new press gets pulled in"
mircea_popescu: nor is there such
a thing as triple bypass while running
a marathon
mircea_popescu: the point is that item either
a) operates continuously, provided it is THIS item, ie, this specific collection of parts or else b) is stopped while beingt serviced, ie having
a cog replaced. there's no clock that works while having its cogs switched around.
mircea_popescu: i can understand the fascination with "this orrery has been in clickety-clacking continuously since 1625", but let's point out that it relies on
a) THIS orrery, as opposed to "constantly changing randomly pile of cogs" and b) it's
a discrete mechanism, like the human heart. it takes
a break every beat. essentially the problem has been hidden, by these, not resolved.
mircea_popescu: but from
a sustainability pow : eulora server currently restarts weekly ; not necessarily because it absolutely must, but because i deliberately did not want to provide
a "continuous" item in this sense.
a111: Logged on 2017-12-29 19:35 phf: lisps deal with freshly-pressed instances as well as anything else: by doing
a clean restart. my point is not lisps, my point is that current solution assumes lack of state and delegates the problem to
a non-v-tronic build system
mircea_popescu: is the entirety of this fiddling you going "i'm curious how this thing could work/break at the edges?" or is it rather "i wonder how i could run
a v tree as an infrastructure node without reboots" ?
a111: Logged on 2017-12-29 19:18 phf: for example one thing that i tried with my lisp workflow is to have the system automatically compile/load touched files in their order of appearance in
a v patch. but the order here is explicitly linear, requires fiddling with patch order, and is definitely not how we use it now
a111: Logged on 2017-12-29 19:07 asciilifeform: v dun have
a sat-solver!!
mircea_popescu: asciilifeform nothing wrong with summary as is, so it'll change, still better an idea than
a blank.
a111: Logged on 2017-12-28 02:58 shinohai: ah mircea_popescu lookup has been there
a while, but i think i forgot to have pete_dushenski push it to his bots page updates.
mircea_popescu: phf and if you don't share with heathens, make it so !!shelf returns
a pile encrypted to my key for instance, nothing wrong with that.
☟︎ mircea_popescu: here's
a problem i perceive phf : you could guess about log(n) of my understanding of various things that interest me on the basis of reading trilema ; i could not guess epsilon of thge say your understanding of sbcl on the basis of reading whatever you provide voluntarily. i could glean it from this kind of interaction, but here's what that means :
http://btcbase.org/log/2017-12-29#1760839 ☝︎☟︎☟︎ a111: Logged on 2017-12-27 02:01 asciilifeform: picture
a kind of 'multiverse ada', where you dun call foo(bar), but instead foo:somepatchid(bar:somepatchid) etc, explicitly conforming to 'multiversism'...
phf: lisps deal with freshly-pressed instances as well as anything else: by doing
a clean restart. my point is not lisps, my point is that current solution assumes lack of state and delegates the problem to
a non-v-tronic build system
☟︎ phf: obviously not designed for that, but the assumption here then is that you're running
a fresh instance for each fresh press, which is suboptimal for
a lisp (this particular detail will likewise come up when v is run on an always-on system of any kind, be that
a cuntoo or
a lispm)
phf: also v doesn't recognize state. in
a tree
A->B pressing to
A, then loading, then pressing to B, then loading is different than pressing to B right away and loading the result
phf: for example one thing that i tried with my lisp workflow is to have the system automatically compile/load touched files in their order of appearance in
a v patch. but the order here is explicitly linear, requires fiddling with patch order, and is definitely not how we use it now
☟︎ phf: v runs with
a special case of sat solver. i don't understand which part of "asdf doesn't have
a sat solver" and "sat is never mentioned" is not clear. i'm saying that whatever dependency resolution is always
a special case of sat. you put some contraints on it (no cirlces, etc.) and then you can have special case solutions.
phf: well, dependency resolution is
a SAT problem at its core
phf: fare claims somewhere that he solved
a gnarly dependency resolution bug in asdf, but the nature of bug is never revealed, and SAT is never mentioned (i suspect most bugs in asdf would be SAT related)
phf: there's actually some extensions that are not necessarily evil (like :email or :author in
a defsystem), bulk of fail comes from test-op, that's not supported in old version, but that nobody uses test-op consistently at all
ben_vulpes: while useful in some contexts, once phf said 'asdf is
a pig with lipstick', could not unsee.
phf: (fwiw ircbot is already sbcl specific, since it relies on
a bunch of sbcl threading extensions, so pulling compat layer for cl-irc is
a waste)
phf: one solution could be to, e.g. post cl-irc, but hard coded to
a specific sbcl implementation. whoever wants to port it elsewhere (like cmucl or whatever) can just post
a #-sbcl #+cmucl vpatch, etc.
phf: it would be perhaps worthwhile to triangulate sbcl towards
a common deployment platform
phf: so instead of addressing the core machinery, they put these hooks all over the place, where some things are part of serve event, and some are not. anyway, you ran into the fact that 1.0.42 introduced
a flag to socket-make-stream :serve-events, which is true by default. it was later changed to false
phf: asciilifeform: serve event is
a cmucl way of tracking streams of every kind (since no threading, repl mainloop is
a reactor pattern). sbcl has just lifted the whole thing, but then never really redid the architecture
mircea_popescu: please tell me hylaea is not
a reference to the fucking futurists
mircea_popescu: so give her
a copy, correct the item she produces and there you go, she got
a blog now.
BingoBoingo: The ordering is more one of prioritizing. No need to stop. At any number of languages, but the focus can shift. As time makes the spanish more comfortable, Portuguese is
a natural next point of focus.
BingoBoingo: asciilifeform: It turns out that Portuguese is making
a strong case for third lang
mircea_popescu: (for les autres : defloration is another word for losing one's virginity. the woman's name being lorena, and her disposition deeply pubescent anxiotic (you should know what this is) which is beyond ridiculous in
a 40yo not to mention hysterical given the amount of compensatory covers-for-impotence she engages in, pretending entirely without basis to be an actress, poetess and whatnot -- the nedeflorena name immediately clicke
a111: Logged on 2017-12-29 15:47 mircea_popescu:
http://btcbase.org/log/2017-12-29#1760654 << fucking annoying this imperial summarization style. "it had
a bug" costs as much to say as "it had so and so bug". but nooooo, god forbid anything effectual ever occurs.
a111: Logged on 2017-03-02 18:10 asciilifeform:
a 'secure prng' is fundamentally THE SAME animal as the 'secure hash' and the 'secure blockcipher'.
BingoBoingo: And plenty of pretty girls need to know who to ask for
a light
mircea_popescu: but not on "byte" meaning anything other than what it does, which is to say... "how much the machine takes in at
a time"
☟︎ mircea_popescu: and why the fuck would you maintain it in
a place which has no need to maintain it. ffa is free from said flesh.
mircea_popescu: asciilifeform nibble was
a HALF BYTE! and
a byte was REGISTER WIDTH. so no, it wasn't customarily 4, it was customarily half of 8 for as long as registers were 8. nibble today should be 32 bits.
a111: Logged on 2017-12-29 02:59 phf: i mean, if you read norvig's python snippets you can clearly see they are written by
a very experienced lispers. you literally never see python like that in the wild, but yet there it is.
a111: Logged on 2017-12-29 15:30 mircea_popescu:
http://btcbase.org/log/2017-12-29#1760601 << "Releases prior to release 1.0.2 had
a bug that could lead to input being ignored and hence wrong message digests being calculated for input with repeated calls to sha3-update (see README file for details). Please upgrade as soon as possible, since this bug can lead to unintended collisions in generated message digests." and of course "The library should be portable across nearly
a mircea_popescu: "PMSF IT Consulting is proud to announce its acceptance as
a member of ASAM e.V.. As
a long time user of ASAM standards, like XCP and A2L, we see the ongoing central role of ASAM as
a provider for standards in the embedded domain and beyond, with
a special focus on automotive applications."
mircea_popescu:
http://btcbase.org/log/2017-12-29#1760601 << "Releases prior to release 1.0.2 had
a bug that could lead to input being ignored and hence wrong message digests being calculated for input with repeated calls to sha3-update (see README file for details). Please upgrade as soon as possible, since this bug can lead to unintended collisions in generated message digests." and of course "The library should be portable across nearly
a ☝︎☟︎ BingoBoingo: Apple's own, choosing the letter "J" with
a bite taken out of the side."
BingoBoingo: And guess who the wops beat in court: "Brothers Vincenzo and Giacomo Barbato named their clothing brand "Steve Jobs" in 2012 after learning that Apple had not trademarked his name. "We did our market research and we noticed that Apple, one of the best known companies in the world, never thought about registering its founder's brand, so we decided to do it," the two told la Repubblica Napoli. The Barbatos designed
a logo that resembles