159100+ entries in 0.099s
phf: (the old, not "modern") emacs approach was
that you provide non-ambiguous, predictable building block operations
that ~specific operator~ can build dwim
tools out of. but
those dwim
tools shouldn't really leave your own
toolbox
mircea_popescu: people gotta get used
to
the whole alphabet schtick already, "POSSIBLE LESS, DO MORE!"
mircea_popescu: phf possibly. in any case,
the issue is
the "DWIM" symbols in
thought, let alone language.
phf: shaping a file list is not necessarily simple op, but i'm not sure if ascii's response is even meant
to make sense..
trinque: mircea_popescu: it is, just
thread leapt from meaning of *
to omg my keyboard?
☟︎ mircea_popescu: trinque it is a valid point of how
to make a proper kbd
tho.
trinque: how does
that enter into how
to properly represent
the
thought
phf: well, why not:
tar cvfz foo.tar.gz `tree bar/baz|grep .txt` ?
mircea_popescu: we're discussing how
things should be or how
they are ?
mircea_popescu: thgat's
the fucking point. | is wel lformed. * is malformed.
mircea_popescu: because
there's no latin character named *. it's an utfism.
spyked: asciilifeform yeah, but "rm -rf *" doesn't mean "all files reachable in
this path". it means "expand *
to arbitrarily long sausage,
then pass result as args
to rm".
mircea_popescu: asciilifeform because * as a semantic notation is part of a class
that goes a*.txt
too.
mircea_popescu: asciilifeform stop an
think of what i'm saying rather
than just chasing your strand!
phf: asciilifeform: my point is what spyked wrote is not even complete in his own
terms
a111: Logged on 2017-09-10 19:50 mircea_popescu: i do not want results of computation
to be dependent on parameters supplied at
tiem fo computation. let me do cat all | grep what i want, not fucking bullshit cat all -flag -otherflag nonsense.
phf: note
that plan9/rc
tried actually making
the last command work, rm `{du|awk '{print $1}'|grep
txt} almost works as expected out of
the box. (note
the somewhat questionable decision
to "replace" find with du, but
then you have
to awk out first column..)
phf: the fact
that you have
to do
this demonstrates a dozen of unix retardations. correct method could just be rm `find .|grep
txt`, but
this fails in a number of interesting ways
mod6: i'll hold off, but
this is a good discussion of design goals.
a111: Logged on 2017-09-10 19:50 spyked: mircea_popescu
took
the words right from my mouth. find . -name '*.txt' | xargs rm is IMHO elegant enough (decomposes into stream of files and repeated removal).
then again, find is another beast on its own
mircea_popescu: as it is, now he's going
to hack
together an undesigned, ad-hoc "sane human-readable form", which is nogood.jpg and also how you get dedicated, good folk in
trouble.
mircea_popescu: (which is what happens when people-other-than-mp say what
to9 be done!!!)
a111: Logged on 2017-09-10 19:26 mod6: oh.
that'd be useful huh. basically just pass in "" or "all" and it'll dump every priv key and its associated pub key?
mircea_popescu: golden standard is 1.
take ye olde wallet.dat ; 2. dump all
txn in it ; 3. dump all addresses involved in
them ; 4. get privkeys for all addresses in
that list (most of which fail).
a111: Logged on 2017-09-10 19:45 mircea_popescu:
http://btcbase.org/log/2017-09-10#1711759 << folks generally do not use
the "wallet" abstraction/idiocy anymore
than
they use
the "accounts" idiocy baked in. 1 address per wallet sorta deal.
mircea_popescu: asciilifeform i was explaining why pywallet has
the dependency.
a111: Logged on 2017-09-10 19:43 mircea_popescu:
http://btcbase.org/log/2017-09-10#1711757 << you don't understand :
the deprecated lib is part of
th ewallet, because
the mike hearn / power ranger imbeciles "updated"
the db scheme of prb in order
to INCLUDE a new db dependency, and one made by google, but DID NOT remove
the old one.
spyked: eh. shell is by definition glue/tape macgyver. and "shell" is just a convenient extension of "unix",
the same way 3 is a "conveniently chosen approximation" for pi.
mircea_popescu: spyked it's not elegant except in
the sense
that in
the "dara ea-i mai mare muta, muci-i marg, din gura-i pute" world a washed face seems stylish and classy.
mircea_popescu: this latter is
the java way, and it should die with fucking java
spyked: mircea_popescu
took
the words right from my mouth. find . -name '*.txt' | xargs rm is IMHO elegant enough (decomposes into stream of files and repeated removal).
then again, find is another beast on its own
☟︎ mircea_popescu: i do not want results of computation
to be dependent on parameters supplied at
tiem fo computation. let me do cat all | grep what i want, not fucking bullshit cat all -flag -otherflag nonsense.
☟︎ a111: Logged on 2017-09-10 19:26 mod6: oh.
that'd be useful huh. basically just pass in "" or "all" and it'll dump every priv key and its associated pub key?
mircea_popescu: alternatively iirc xargs
takes a param count limit, you keep re-running it
mircea_popescu: i guess
THE LINUX WAY, inherited from
the c way, inherited from new jersey is
to go rm -rf a*.txt and so following.
a111: Logged on 2017-09-10 18:42 phf:
http://btcbase.org/log/2017-09-10#1711748 <<
this is a kernel level limit on number of arguments you can pass
through argv (documented in e.g. man execve). i
think recent linux kernels actually have
this work somewhat dynamically.
a111: Logged on 2017-09-10 18:09 asciilifeform: so what do folx with coldwallets use ?
their
teeth and a pair of wires ?
a111: Logged on 2017-09-10 18:07 asciilifeform: ( not a single one of
the boxes i have here now, will eat
the 'deprecated' bdb lib for python,
that
the
thing insists on )
mircea_popescu:
http://btcbase.org/log/2017-09-10#1711757 << you don't understand :
the deprecated lib is part of
th ewallet, because
the mike hearn / power ranger imbeciles "updated"
the db scheme of prb in order
to INCLUDE a new db dependency, and one made by google, but DID NOT remove
the old one.
☝︎☟︎ a111: Logged on 2017-09-10 13:13 phf: i need
to figure out how
to persist it better. for example i
think
that
the flags should be passed on
through <- and -> page changes. or perhaps you should be able
to store
them in cookie
a111: Logged on 2017-09-10 13:10 phf:
this is a btcbase.org psa: if you add ?nick=yournick
to
the end of a message listing url you're going
to get highlight for messages from yournick and messages
that have yournick in
their
text. handy for when you want
to see recent messages related
to yourself
mircea_popescu: and speaking of "i've been around for 8 years and i imagine
this means something notwithstanding i got nothing for show for it", hey kanzure are ye busy reading
the logs and reforming yourself into republican shape, or are ye back on
the old bandwagon of idiocy, sucking from
the same old bottle of stupid ?
mircea_popescu: there are many proprietary shits, but
they "publish" summaries in
the vein of us scientism, "summaries" and coloring book pdfs.
mircea_popescu: would actually be iunteresting, as a phuctor sort of
thing. "here's
the mempool pile"
mircea_popescu: it's basically a bdb dump of keys and
txn, first part (but not 2nd) aesencrypted.
mircea_popescu: asciilifeform> i.e. complete disasm of wallet.dat << it also holds
txns
mod6: i'll have
to look at
that.
mod6: oh.
that'd be useful huh. basically just pass in "" or "all" and it'll dump every priv key and its associated pub key?
☟︎☟︎ mod6: <+asciilifeform> it also doesn't dump WHOLE keypool << what is meant here? you want
to see every private key dumped?
phf: hyperspec mandates it
to be no less
than 50, which basically means you can't rely on #'apply working on unbound lists in a portable way
phf: but
there's also CALL-ARGUMENTS-LIMIT so
this particular failure point is not limited
to unix..
phf:
TSTKSIZ controls
this.
phf: list,
typically around 409,600 bytes.
The kernel constant
phf: There is a large but finite limit on
the size of an argment
phf: heh, plan9 lists
this in "bugs" section.