46300+ entries in 0.035s
mircea_popescu: "we manage
to provide zero cost exceptions
to
the user's environment
through
the simple procedure of just not handling any"
mircea_popescu: fucking lulz of all
time. zcx : zero cost exceptions!!! it's like
the zero cost restaurant : you don't have
to pay anything and htey don't give you any food.
mircea_popescu: a
thing you keep re-serpenting. 64 bit string, w/e, starts as null
diana_coman: I don't quite get
the branches and iterators aka what exactly do you want in
there?
mircea_popescu: 50 + 50 + 50 by
the
time you've used 20 letters you got
the k.
mircea_popescu: right, something like
the above, has buncha branches and iterators, not hard
to write, get it
to run 1k
times or so and lessee.
diana_coman: mircea_popescu, eucrypt's
test on Serpent seem good candidates as one can even adjust how many iterations
to do if you want some specific
time intervals; current full
test of
the serpent module (including i/o because of using
test vectors in file) is reported by
time at ~2.3s without sjlj;
this has no
tasks/exceptions as such;thing is:
time is not extremely precise but I could run I suppose some 1k
times and see
☟︎ mircea_popescu: wrap it in a few for and if clauses, something like like for( a = 1; a < 100; a = a + 1 ){ if a is even
then for( b = 1; b < 100; b = b + 1 ){ ... collector = serpent(collector) }}}
diana_coman: but hm,
the underlying stuff
that
takes most
time is
the mpi c/cpp anyway
diana_coman: hm, it's all about what
the
task does so I suppose it's enough
to plonk in
there some rsa ops
mircea_popescu: diana_coman do you have a non-trivial (non-huge) ada
that you could run once with sjlj and once with zcx in a
timing harness so we can get some data ? ideally something
that
takes 100-1000ms or so.
diana_coman: nope, pragma abort_defer doesn't help (it has
to be outside loop because it's not accepted anywhere other
than immediately after begin and program is stuck just
the same)
mircea_popescu: of course,
that's when
they moved
to unicode. but if memory serves, dwarf2 was kinda braindamaged in some corner cases i now don't recall.
diana_coman: that's how it sounds from adacore's link;
then again
the description of
the pragma in gnat ref manual is "it has
the effect of deferring aborts for
the sequence of statements (but not for
the declarations or handlers, if any, associated with
this statement sequence).
mircea_popescu: (i expect it will stop, yes ; but
the q is, even slower
than good old sjlj perhaps ?)
diana_coman: I read a bit on
that and it seems
to me
that it is meant
to actually add ANOTHER "won't abort here" point so I don't quite see it; but yes, I'll
try it anyway
mircea_popescu: diana_coman what happens if you actually put a pragma Abort_Defer in
the loop ?
diana_coman: it kind of has
to be but really, as you just said: wtf if it breaks
the standard
mircea_popescu: diana_coman possibly because of how
they do
that binder
table bs
mircea_popescu: so basically, gnat runtime needs an hero
to implement dwarf-3 for it, ditch
this entire "zcx" nonsense nobody heard about (and apparently for good fucking reason, it's made for z80s or something)
mircea_popescu: diana_coman
this is so fucking retarded, "our
thing does not implement
the standard". RATIONALE ?
diana_coman: myeah, sjlj is going
to be slower; onth now I don't quite get it WHY is ZCX not able
to support async abort
mircea_popescu: just, i don't
think it can be whipped into a workable exception handling mechanism.
mircea_popescu: but yes, it can't hurt anything
to run it once see what happens.
mircea_popescu: bvt pretty sure
the problem with sjlj is
that it significantly slows down all code execution, even on code
that raises no exceptions.
diana_coman: although indeed, in
this case
the check on
the
tasks' status is
too early so not of much use
diana_coman: confirmed: with --RTS=sjlj, it does
the right
thing and actually aborts
bvt: i expect it will be slower, but it won't hurt
to do
the check.
the impact will depend on how exceptions are used (i don't
think it can have any impact on ffa, for example). but i don't have enough experience with it
to provide any numbers
☟︎ bvt: during
the gnat build,
the sjlj runtime is built, so it should be possible
to switch
to it and
test.
☟︎ a111: Logged on 2019-02-12 03:54 mircea_popescu:
that's
the fucking cost of being a raging asshole.,
mircea_popescu: diana_coman it looks like we'll end up with a patched
tmsr-gnat.
diana_coman: I suppose so far
the options seem
to be 1. maybe
there's a way
to actually make it choose
the sane abort model 2. change
the idiotic "ignore
the signal" part
bvt: mircea_popescu: i will become blog post, but when i get
to
the home machine.
diana_coman: but ftr when a
task is not in
this sort of "can't hear you la la",
there's no need for any wait, no
diana_coman: i.e. if anything, you can say
that my
test code should wait a bit and
then check
bvt: diana_coman: i agree; in
the abort signal handler,
there is a snippet of code
that ignores
the signal (given ZXC exception handling model).
diana_coman: mircea_popescu,
the check is in
the
TEST code
mircea_popescu: you know, it'd be grand if
this were a blog article rather
than a pastebin. odds are ima want
to reference it later, and as a factual matter i want ot reference "More precisely, one of
the issues seems
to be
that abort seems
to be delivered
to
the child
thread after
the check for 'thread aborted' has already been done" specifically even now.
mircea_popescu: bvt
that's a good point actually,
the check comes
too late in
the current implementation.
bvt: not really, just writing from
the salt mine.
bvt: hello,
thanks for voice
diana_coman: "More precisely, one of
the issues seems
to be
that abort seems
to be delivered
to
the child
thread after
the check for 'thread aborted' has already been done:" ->
this ftr is NOT an issue; if only it were delivered and worked afterwards but it's not;
that check is just
testing stuff,
the more important check is
that program hangs
mircea_popescu: is
there anything like c's "signal(value, pointer)" in ada ?
diana_coman: mircea_popescu, my hope is
that maybe
there is some flag or option
that sets it; I'll have
to dig deeper
to find out exactly
mircea_popescu: (needless
to say, literatly chick also fat as all get-out, but by
then one pretty much expects
the whole
triad.)
mircea_popescu: evidently florrie gave our pal jack keefe some daughters,
too. left-handed,
to boot.
mircea_popescu: meanwhile at
the ~angry~ retard farms, moedirtcircle 26F Primal Prey 4h It Literatly only says "not even kidding"
thats it. And I'm
the one
thats failing 😒😒😒😒 get
that stick from up your ass and be nicer
to people YOU message. Fuck off dipshit. moedirtcircle 26F Primal Prey 4h Damn I finally just seen your profile. You Literatly have no looks going for you. And not intelligence either, or compassion. Have fun wit
mircea_popescu: diana_coman is
this abort-handler something you can define from within your proggy ?
diana_coman: hm, asciilifeform from what I see at a quick glance in
that s-taprop-linux__adb it actually calls pthread_kill ONLY if "abort_handler_installed" and
that in
turn seems set only if not some default state ? might be a clue
to follow perhaps, I haven't dug much deeper yet
a111: Logged on 2019-02-12 01:31 asciilifeform: relatedly
to entire subj,
http://nosuchlabs.com/pub/warez/burns.djvu is
the most readable ref i've found on
the subj (of how
the
thing is supposed
to work,
that is , when actually worx )
a111: Logged on 2019-02-10 16:03 diana_coman: I also
tried asynchronous
transfer i.e. supposedly "try
this and if
timeout
then do
that" but apparently it's in fact still "oh, but ONLY if abortable"
phf: if any log reader is getting a 404 not found, even if intermittent, please let me know. i was just getting it for a couple of minutes, but i can't figure out if
that's
the actual situation on
the machine or some russian dns problem
phf: asciilifeform:
there's also sgi machines with
their famously well built x11 implementation. i believe
there's no source for
that one
though
☟︎ mircea_popescu: meanwhile in
trilema roulette, fell on
http://trilema.com/2010/ce-i-de-facut/ ; would propose it
to
teh rotaku club as perhaps
the foremost use of
that language found on my blog, stylistically perfect, elegant without equal, i doubt
there's much romanian
that can stand up
to
that romanian
to be found anywhere.
mircea_popescu: dmca requires someone
to register a claim, and
they can actually be sued if it's fraudulent.
mircea_popescu: asciilifeform it's dumber, schmuck's like "hi
there, i r anon please delete my name". how fucking stupid does he need
to be, "anon" is not listed
there.
mod6:
http://p.bvulpes.com/pastes/ht6ey/?raw=true << Ok, inflating /dev/sdb with cuntoo. Will report back more
tomorrow after install is complete, and I update
the lilo.conf via LiveCD chroot mount of cuntoo. Will post lilo.conf changes, and results.
a111: Logged on 2019-02-12 02:01 asciilifeform: it's how
the
thing fans out
to N jacks
a111: Logged on 2019-02-12 01:34 asciilifeform: for
total
thread-completeness,
there is a workaround
that works, but it is imho ugly ( ada 'interrupts' package ), i'ma post example if mircea_popescu et al ask.
a111: Logged on 2019-02-12 01:25 asciilifeform: i suspect
this is gonna be a '1-line' patch,
to fix. q is where..
a111: Logged on 2018-10-26 02:14 asciilifeform: meanwhile, in gnat bugs : apparently ( and
this is documented or mentioned nowhere ) : it is impossible
to have a Ada.Finalization.Limited_Controlled
type ANYWHERE inside a static library, unless it is generic all
the way down (i.e. if
the lib package is generic, any sub-packages must also be instantiated as generics )
mircea_popescu: wanna be a man,
then gotta be a man. for example, wanna be a restaurant maitre d',
then you're stuck apologizing
to
the people
that ordered 19 oz ny steak and got 1.5 oz cafeteria hoof cut. and so on. posturing brings with it
the cost of sucking all
the dick you earn
the sucking of in
the usual course of
the business you're posturing with.
mircea_popescu: because EVERY SINGLE PERSON EVER INVOLVED WITH GNAT AT ALL ~~MUST~~~ apologize for
this,
to me, personally./
mircea_popescu: then when
they fail
to respond,
they get a default negrate.
mircea_popescu: but -- email. must email
them. all of
them.
their fucking support list, and every individual involved.