mircea_popescu: asciilifeform, i honestly can't say what not dead is this ; it's off the web, it's off the web. "oh but it shows up on ipnoje". sooo ?
lobbesbot: asciilifeform: The operation succeeded.
lobbesbot: asciilifeform: The operation succeeded.
ave1: does anyone have a ru to eng or de translation dictionary as a text file?
☟︎ a111: Logged on 2018-03-29 10:22 spyked: phf's ld might add an "static" to "inline"; or not, not sure; can verify by objdump'ing vdiff and seeing if there appear multiple (with local linkage) definitions of xnmalloc in the same file.
spyked: 1. pressed latest vtools; 2. changed all "static inline" to "extern inline __attribute__((gnu_inline))" in
http://btcbase.org/patches/vtools_fixes_static_tohex#L5 ; 3. compiled; 4. diffed the symbols in the generated vdiff against the unmodified version => no changes, xnmalloc et al. are not included in the final binary.
spyked: the __attribute__((gnu_inline)) attr is of course very stupid through its lack of portability, but as per
https://gcc.gnu.org/onlinedocs/gcc/Inline.html , it seems that the wrecking to this compiling/linking behaviour of inline was done pretty early on (C90).
☟︎ spyked: (or that might be C99, actually)
spyked: it might also in a way be interesting to report how I stumbled upon this: I tried to recompile gnupg-1.4.10 on my broken debian system and got the same "multiple definition" linking errors as in vtools' case (though I *did* use gcc<5). so I dug and found the usual kochs "fixing" things to compile gnupg on newer gccs.
☟︎ trinque: doing just fine. watching the rain come down.
mod6: trinque: yeah, I think it's gonna rain here today too.
phf:
http://btcbase.org/log/2018-05-21#1816308 << is there any need for that alternative? from my research static inline is one of the few behavioral commonalities between c99 and pre-c99. it also supposedly works across different compilers, etc.
☝︎☟︎ phf: i'm alive by the way, will start catching up on the mother of logs
esthlos: anyone have a good idea for generating unique identifiers for v's gpg keyring name? I just spent a painful hour debugging because CL's #'get-universal-time is only precise to the second
phf: asciilifeform: that's not going to work cross-invocations though. gensym is incremental across the instance, so same invocations should produce roughly same gensym numbers
phf: (i assume he's basically asking for a lisp-only "mkstemp")
esthlos: phf: was about to mention that
a111: Logged on 2018-03-28 19:32 asciilifeform: funnily enuff, if working rng were standard on pc, 128bits from it would give unique-gensym ( the supposed problem , according to the gcc nitwits, with mktemp , is collision ) without O(N) searching ( as in mkstemp) with probability ~1
esthlos: but under current behaivor, program won't clean up tempdir when it eggogs
☟︎ phf: true, i'd still just go for an ffi to mkstemp. you're dealing with a unix substrate, might as well use unix tooling. otherwise you have to juggle file-probe, create-directory, but restart if there's exception, etc.
esthlos: phf, I assume by ffi you mean a call to run-program, not a use of cffi?
esthlos: oh I see, sbcl has its own ffi
phf: cffi is compatibility layer to native ffis (a rather dodgy one too)
phf: sbcl has both own ffi, and there's also SB-UNIX:SB-MKSTEMP
phf: actually you want mkdtemp, but it's for some reason not included in sb-unix..
esthlos: clozue doesn't look like it has equivalent functions
esthlos: may just have to get dirty with the ffi
phf: clozure doesn't need to, #_mkdtemp
esthlos: is that a reader macro for system calls?
phf: it's a reader macro for ffi
esthlos: gotta run for a bit, thanks for the info phf
phf: ccl has a fancy system, based on hacked up gcc, for generating ffi mappings (called interfaces), which you then can use with (use-interface-dir :foo) (#_foobar ...)
a111: Logged on 2018-05-20 14:58 mircea_popescu: phf, Mocky ^
diana_coman: phf, yes, it does; it's basically version 2.0
deedbot: pipp8 voiced for 30 minutes.
spyked:
http://btcbase.org/log/2018-05-21#1816325 <-- it's not needed, at least not in the case of vtools. it may have been worth it if it had reduced binary size, but from my experiment it makes no difference. gcc seems to remove the definitions of static-inlined functions (after inlining them), but I don't really know if it does that in general.
☝︎ a111: Logged on 2018-05-21 16:34 phf:
http://btcbase.org/log/2018-05-21#1816308 << is there any need for that alternative? from my research static inline is one of the few behavioral commonalities between c99 and pre-c99. it also supposedly works across different compilers, etc.
spyked: but this is easy to verify with readelf -s | grep inlined_func
BingoBoingo: And we have another holiday. Battle of Las Piedras day.
phf: spyked: you kind of have a choice between #define equivalent inline, which is going to be actual inline and increase the binary size, or else i don't think it really matters. supposedly the annotation makes the calls "faster", but the notion of faster is very vague. i suspect it's essentially an equivalent of "register" keyword: things are maybe sometimes reshuffled to make things maybe sometimes faster
spyked: ah. my intuition was that static inline might increase the size by leaving unused definitions in every .o file generated from the respective .c that imported the header.
spyked: I'm pretty sure that happens with functions qualified as static. but the inline qualifier might cause the compiler to eliminate those definitions.
BingoBoingo: <asciilifeform> 'AmRRON Dark Labs, a type of Skunk Works for the Patriot community, has built a machine...' blah << LOL, FBI-otron in the wild!
BingoBoingo: asciilifeform: Well, the existence of FUCKGOATS raises the bar
a111: Logged on 2018-01-09 20:29 asciilifeform: in other oldies, seiko's 'DPU414' thermal printer, can be had for <fiddybux, and not only prints 80cols x infinity (continuous tape) but knows how to eat rs232 and... print hexdump (up to 19200 baud)
phf: asciilifeform: transliteration code failing, give me a sec
phf: asciilifeform: fixed now
a111: Logged on 2018-01-09 18:56 a111: Logged on 2017-01-15 23:05 a111: Logged on 2017-01-15 21:47 indiancandy: !!register <だ댧卾ᅟ룱ࢤ�ᫎ䂳䯫证慢ቊ콥贐较本덁㡚࠘譂ꘪ蜈出ᜄ粘锝⛥箥戴ᷜꔐ俟组․ഒ㻒㌕낌ᥥ款ړ忲䧚뾨շ脵됾쓰娊嫥៏䙿槥晥获᷑룜巵歛踈>
a111: Logged on 2018-01-09 18:57 mircea_popescu: unicode is the solution!
mimisbrunnr: Logged on 2018-05-17 14:40 mircea_popescu: what is behind comments like
http://btcbase.org/log/2018-05-16#1814037, "I moved more recently than this fellow and I'll get my id after he does" ? "I don't, properly speaking, think bb is a person, at least not in the same sense / to the same degree I am" ?
mircea_popescu: there's two important advantages of reasoned decisionmaking. first and foremost it gathers your people around you, this is what loyalty is made of, at its best. secondly and more importantly it permits history, which is to say humanity.
ben_vulpes: mircea_popescu: it doesn't make sense to put delicate items acquired with much suffering like BingoBoingo's cedula immediately into the statists hopper (localbitcoins demands "an ID" to post ads), when one can pick from the sale offers and operate on the side of the transaction where id is not asked for
☟︎ lobbesbot: ben_vulpes: Sent 22 hours and 13 minutes ago: <asciilifeform> is there a problem with the invoice ? ( aside from the fact that i forgot the 36 usd fee , in it ) ? how come i'm still waiting ?
mircea_popescu: ben_vulpes, but you can also talk to ads already posted.
ben_vulpes: asciilifeform BingoBoingo: i await a transaction to clear through deedbot, then i shall settle your invoices
a111: Logged on 2018-05-21 00:01 asciilifeform: by ~that~ measure it was dead 3+y ago , when it turned to openly infecting the stored tarballs
a111: Logged on 2018-05-21 10:12 ave1: does anyone have a ru to eng or de translation dictionary as a text file?
a111: Logged on 2018-05-21 11:40 spyked: the __attribute__((gnu_inline)) attr is of course very stupid through its lack of portability, but as per
https://gcc.gnu.org/onlinedocs/gcc/Inline.html , it seems that the wrecking to this compiling/linking behaviour of inline was done pretty early on (C90).
a111: Logged on 2018-05-21 16:38 esthlos: but under current behaivor, program won't clean up tempdir when it eggogs
mircea_popescu: usefully, too, if anyoen wants to keep these they're autosorted by run
a111: Logged on 2018-05-21 16:55 phf:
http://btcbase.org/log/2018-05-20#1815927 << i'll take a look, i had further comments about diana_coman's original spec and the responses, but does this new spec replaces the other one entirely?
a111: Logged on 2018-05-21 19:23 asciilifeform: 'AmRRON Dark Labs, a type of Skunk Works for the Patriot community, has built a machine...' blah
ben_vulpes: mod6 asciilifeform: a sales incentive scheme proposal: we kick back 5% of revenue to affiliates who've recruited up to 10 paying customers, 10% up to 25 paying customers, 15% to 50 customers and then 20% for recruiting over 50 customers, which'd be a tidy .03 btc/mo, way better than any other "while you sleep" scheme out there.
☟︎☟︎☟︎ mircea_popescu: aand in today's aphorisms, "Licking the power socket will never kill the licking of power sockets, even though it may kill all lickers of power sockets."
a111: Logged on 2018-05-21 22:47 ben_vulpes: mod6 asciilifeform: a sales incentive scheme proposal: we kick back 5% of revenue to affiliates who've recruited up to 10 paying customers, 10% up to 25 paying customers, 15% to 50 customers and then 20% for recruiting over 50 customers, which'd be a tidy .03 btc/mo, way better than any other "while you sleep" scheme out there.
lobbes: ben_vulpes: Here's my initial q's: 1) Is that in perpetuity? 2) I assume this is a percentage of the revenue of the set of the customers the sales person recruits and not of the entire set of customers?
ben_vulpes: lobbes: for so long as a customer is paying pizarro is the intention
ben_vulpes: and yes, of the affiliate's recruited set, not of all customers.
ben_vulpes: this incentivizes the hunt for long term customers
lobbes: and would this be for both shared hosting customers and rockchip (and anything else)?
ben_vulpes: lobbes: shared shell alone, i haven't examined the implications of this scheme for rockchips yet
lobbes: well, sounds good to me.
lobbes: in other news, lobbesnode1 just passed block height 300000!