<whitequark>
pause after printing a screenful of text
<devyn>
hmm, must be really old because I always remember DOS having 'more'
<devyn>
dir | more
<whitequark>
what
<whitequark>
dos doesn't have pipes
<devyn>
er
<devyn>
pretty sure it does
<whitequark>
no
<whitequark>
never had
<whitequark>
it can't even launch two regular processes in parallel
<devyn>
oh good point
<devyn>
hmm
<devyn>
whitequark: "Because DOS is a single-tasking operating system, piping is achieved by running commands sequentially, redirecting to and from a temporary file. COMMAND.COM makes no provision for redirecting the standard error channel."
<devyn>
so uh
<devyn>
yeah
<whitequark>
I wonder if that's post-6.22
<devyn>
I searched and it is in an old MS-DOS 6.22 book
<devyn>
let me go earlier
<whitequark>
eh don't bother, I'm wrong then
<devyn>
ok
<devyn>
well, it's a pretty dumb way to do pipes anyway
<devyn>
but it does work in the case of 'dir | more' haha
mcc has quit [Quit: This computer has gone to sleep]
prophile has joined #elliottcable
alexgordon has joined #elliottcable
Sgeo has quit [Read error: Connection reset by peer]
yorick has joined #elliottcable
<vil>
ELLIOTTCABLE: I do not partake in the snapping of chats
sharkbot has quit [Remote host closed the connection]
sharkbot has joined #elliottcable
prophile has quit [Quit: The Game]
eligrey has joined #elliottcable
Sorella_ has joined #elliottcable
gozala has joined #elliottcable
prophile has joined #elliottcable
Sorella_ is now known as Sorella
Sorella has quit [Changing host]
Sorella has joined #elliottcable
<alexgordon>
glowcoil: swift has no \v muahaha
<alexgordon>
someone was listening
<whitequark>
Chris Lattner, duh
<whitequark>
if my life was hinging on that guy's software design skills, I'd live to 150
<alexgordon>
haha
<prophile>
Lattner afraids of very few things
<joelteon>
who's good at python here
<nicksergeant>
joelteon: i know me some python
<nicksergeant>
i wouldn't consider myself an expert though
<nicksergeant>
just enough to know that i'm a shitty python person
<katlogic>
its how bitcoin and tor win32 binaries are proven that they are not backdoored
<katlogic>
if 50 respectable people attach their gpg saying this is sha sum indeed corresponds to this git head, it is hard to dispute that
<devyn>
well, unless someone comes along and manages to pay them all off, but it's unlikely
<devyn>
:p
<devyn>
kidding. seems really neat
<katlogic>
Yeah, it's dangerous if all are not anonymous
<katlogic>
They can be NSL'd all at the same time, though there is usually predetermined signal to tell others they've been NSLd
<devyn>
well, I think ideally don't even worry if they're all "reputable" — just allow people to sign up anonymously to do it
<devyn>
just anyone
<katlogic>
mixed bag of anon + non-anon works best
<devyn>
yeah I suppose
<katlogic>
if non-anons are NSLd, anons will can freely speak up
<devyn>
how flexible is that toolchain? can you choose to use clang?
<joelteon>
with what
<devyn>
and how about building for Windows; how does that work
<devyn>
gitian
<joelteon>
i'll tell you this
<katlogic>
devyn: It's pretty crappy actually :/
<joelteon>
you can't use clang with nix
<katlogic>
just a vm with preinstalled ubuntu
<katlogic>
usually the point is to do binaries for win32, which is ok, its just default ubuntu mingw cross compiling
<devyn>
hmm ok
<devyn>
is it possible to cross compile for OS X? probably not, because any of the headers past plain Darwin are copyrighted right
<katlogic>
Indeed cross compiling for os x sucks hard
<katlogic>
would need hackintosh vm
<devyn>
yeah
<katlogic>
On the up side, the architecture is stable
<katlogic>
which means one would not actually even need a VM
<katlogic>
just that all signers would need mac
<katlogic>
(afaik clang does not do silly things like include timestamps in binaries)
<joelteon>
so the binaries would have to be the same
<devyn>
I can test that right now
<joelteon>
identical
<joelteon>
i don't know if that's reliable
<katlogic>
if there are two people in channel with osx lion
<katlogic>
just compile printf("hello world") and compare sha sums
<katlogic>
chances are high that sums will match
<joelteon>
afaik gcc uses hardware sampling when it bootstraps itself so even two gccs built from identical source trees with identical libcs might be different
<devyn>
I'm on 10.9.3
<devyn>
how about you joelteon
<joelteon>
10.10
<devyn>
ah damn
<katlogic>
ah, far from stable :/
<joelteon>
neigh
<katlogic>
i thought mac is rolling-version or something
<devyn>
no, they have big-ish releases
<devyn>
current stable is Mavericks (10.9)
<devyn>
next is Yosemite (10.10)
<katlogic>
joelteon: Indeed. That is why the VM don't actually build gcc. The gcc is simply pulled from debian / ubuntu repo.
<devyn>
[devyn@DevBook ~]$ clang test.c -o test
<devyn>
[devyn@DevBook ~]$ shasum -a 256 test
<devyn>
[devyn@DevBook ~]$ clang test.c -o test
<devyn>
[devyn@DevBook ~]$ shasum -a 256 test
<devyn>
4f31a4572b96ed9d6b2ae67f74fb13cf3100e09ae102b6548693bf9cd2935e53 test
<devyn>
4f31a4572b96ed9d6b2ae67f74fb13cf3100e09ae102b6548693bf9cd2935e53 test
Sgeo has joined #elliottcable
<devyn>
interestingly because there's no vectorization to be done, clang -O2 and clang -O3 provide identical binaries on that little program
<devyn>
but -O1 and no flag produce different binaries
<devyn>
and -Os is different, too, of course
<katlogic>
hmm, what differs with O1/Os ?
<katlogic>
You mean same program, each time with O1 is different?
<katlogic>
(difference between O1 and O2 is obvious ..)
<devyn>
no, I mean there's a difference between nothing, -O1, -O2, and -Os
<devyn>
but -O2 and -O3 produce identical binaries
<devyn>
in this case
<katlogic>
thats pretty normal :)
<devyn>
no, the interesting part is that -O2 and -O3 were identical
<devyn>
meaning that there's no storage of flags or anything in the binary