--- Log opened Sun Mar 27 00:00:50 2011
00:11 -!- XenoPhoenix [~Xeno@cpc13-aztw24-2-0-cust23.aztw.cable.virginmedia.com]
has joined #go-nuts
00:11 -!- crawshaw [~crawshaw@216.239.45.130] has quit [Quit: crawshaw]
00:23 -!- tav [~tav@92.7.141.101] has quit [Quit: Hakuna Matata]
00:24 -!- tav [~tav@92.7.141.101] has joined #go-nuts
00:39 -!- zimsim [~simon@87.72.77.195] has quit [Remote host closed the
connection]
00:48 -!- thebarberian [~Master_Co@216.252.79.105] has joined #go-nuts
00:48 -!- dju_ [dju@fsf/member/dju] has joined #go-nuts
00:48 < thebarberian> Anyone know if gccgo is worth the compilation time?
Is it faster?  What are the benefits?
00:49 -!- dju [dju@fsf/member/dju] has quit [Ping timeout: 246 seconds]
00:49 -!- dju__ [dju@fsf/member/dju] has joined #go-nuts
00:53 -!- dju_ [dju@fsf/member/dju] has quit [Ping timeout: 276 seconds]
00:53 < str1ngs> thebarberian: gccgo makes for smaller binaries
00:55 < str1ngs> and I would guess somethings would be better optimized
00:55 -!- wrtp [~rog@92.17.50.183] has quit [Quit: wrtp]
01:06 -!- rtharper [~tomh@unaffiliated/sioraiocht] has quit [Remote host closed
the connection]
01:06 < thebarberian> interesting...
01:12 < |Craig|> I believe gccgo handles go routines a bit different leading
to some things being slower, and some being faster, but I havent tried it
01:13 < thebarberian> well hopefully in a few hours I will have it compiled
and will try it...
01:13 < thebarberian> will try to compare afterwards and publish some
benchmarks perhaps...
01:21 < str1ngs> thebarberian: I have played with gccgo some, not as much as
I would like.  something I noticed is when compiling multiple *.go files seems to
have trouble.  if you have a monolithic main.go seems ok
01:21 < str1ngs> however I'm sure this is more my issue, but I have not
looked into enough
01:28 -!- Project-2501 [~Marvin@dynamic-adsl-94-36-151-145.clienti.tiscali.it] has
quit [Quit: E se abbasso questa leva che succ...]
01:29 -!- GoBIR [~gobir@c-24-130-224-186.hsd1.ca.comcast.net] has quit [Ping
timeout: 246 seconds]
01:34 < str1ngs> I've made a .tar.gz and .tar.bz2 decompression
https://gist.github.com/888816 . testGz and testBz are basically the same
function.  I would like to abstract this.  but gzip.NewReader returns multiple
where as bzip.NewReader does
01:34 < thebarberian> I can't even get the compiler working yet
01:34 < str1ngs> whould I pass a io.Reader to the function then type assert
it?
01:34 < thebarberian> just failed 30min in the compile
01:35 < str1ngs> thebarberian: did you follow the gccgo instructions?
01:35 < thebarberian> of course!
01:35 < str1ngs> what error did you get
01:36 -!- GoBIR [~gobir@c-24-130-224-186.hsd1.ca.comcast.net] has joined #go-nuts
01:36 < thebarberian> configure-stage1-target-libgcc...
01:36 < str1ngs> if its a long error messge log maybe pastebin it
01:36 < thebarberian> checking for suffix of object files...
01:37 < thebarberian> good one one sec
01:39 < thebarberian> pastebin.com/05f7X4x5
01:39 -!- ajstarks [~ajstarks@pool-98-109-198-180.nwrknj.fios.verizon.net] has
joined #go-nuts
01:40 < thebarberian> pastebin.com/05f7X4x5
01:42 < |Craig|> different error, but same outcome as when I tried to
compile gcc go for mac.  Failed 30 min in or so.  I never did get it to work, but
I didn't try too hard
01:42 < thebarberian> I am very close to taking the same decision branch as
you
01:44 -!- m4dh4tt3r [~Adium@c-69-181-223-245.hsd1.ca.comcast.net] has quit [Quit:
Leaving.]
01:45 < str1ngs> /home/grafa/projects/objdir/x86_64-unknown-linux-gnu/libgcc
should be a config.log file can you pastebin that?
01:46 < thebarberian> absolutely
01:47 -!- GoBIR [~gobir@c-24-130-224-186.hsd1.ca.comcast.net] has quit [Ping
timeout: 252 seconds]
01:47 -!- boscop [~boscop@g230099063.adsl.alicedsl.de] has quit [Ping timeout: 250
seconds]
01:49 < thebarberian> pastebin.com/kHKxXixq
01:50 < thebarberian> could it be something as simple as arch/os definition?
01:55 < str1ngs> one sec let me check that log
01:56 < str1ngs> error while loading shared libraries: libmpc.so.2: cannot
open shared object file: No such file or directory
01:56 < str1ngs> what linux distro are you using?
01:58 < thebarberian> ubuntu
01:58 < thebarberian> that's strange...
01:58 < thebarberian> I definitely installed it
01:59 < thebarberian> gotta see if it's looking in the right place...
01:59 < str1ngs> you might need to install limpc-dev and gpmp-dev
01:59 < str1ngs> err gmp
01:59 < thebarberian> I definitely have libgmp-dev
02:00 < str1ngs> the names I'm just guessing you get the idea though
02:00 < thebarberian> and compiled the new one from scratch just before
5.something
02:00 < thebarberian> yeah..
02:00 < str1ngs> if you still have problems you can overlay them into the
gcc dir
02:01 < thebarberian> actually the configure was failing and then I
installed the all 3 possible libs gmp mpc and some other one
02:01 < str1ngs> but try to avoid that and use your distro's dev packages
02:01 -!- echosystm [~echosystm@unaffiliated/echosystm] has joined #go-nuts
02:01 < str1ngs> also are you building gcc out side of the source tree?
02:01 < thebarberian> I will give it another try in a bit
02:02 < thebarberian> yes
02:02 < str1ngs> kk
02:02 < thebarberian> just like the gccgo_install.html says
02:02 < str1ngs> I would now that you have those headers etc.  clean that
build dir and start again
02:02 < thebarberian> ok thanks
02:03 < str1ngs> if you still have problems find the config.log and see what
errors it fails with
02:03 < thebarberian> yeah good point
02:03 < echosystm> what exactly is the license for go?
02:04 < thebarberian> I gotta dash now but I will try it again in a few
hours...
02:04 < str1ngs> I think there might have been some config cacheing from
before you install mpc headers
02:04 < echosystm> i read "bsd-like", but what does that actually mean?
02:04 < thebarberian> oh....!
02:04 < thebarberian> that makes sense
02:04 < str1ngs> echosystm: bsd simplified iirc
02:04 < thebarberian> I will just delete and start again
02:04 < echosystm> is there a patent grant?
02:04 < thebarberian> thanks str1ngs
02:04 < str1ngs> thebarberian: np
02:05 < str1ngs> echosystm: see PATENTS in the go root
02:06 < echosystm> is there some equivalent of cpan/rubygems/pypi for go?
02:06 < str1ngs> yes goinstall
02:06 < echosystm> ok, thanks
02:07 < str1ngs> its still work in progress but its getting there
02:09 -!- aho [~nya@fuld-590c7df1.pool.mediaWays.net] has joined #go-nuts
02:14 -!- l00t [~i-i3id3r_@189.105.32.98] has joined #go-nuts
02:22 -!- echosystm [~echosystm@unaffiliated/echosystm] has left #go-nuts []
02:31 -!- MX80 [~MX80@cust151.253.117.74.dsl.g3telecom.net] has quit [Read error:
Operation timed out]
02:32 -!- MX80 [~MX80@cust151.253.117.74.dsl.g3telecom.net] has joined #go-nuts
02:32 -!- pothos_ [~pothos@111-240-167-52.dynamic.hinet.net] has joined #go-nuts
02:33 -!- hypertux [~hypertux@vps1.joelegasse.com] has quit [Read error: Operation
timed out]
02:33 -!- hypertux [~hypertux@vps1.joelegasse.com] has joined #go-nuts
02:33 -!- pothos [~pothos@111-240-167-52.dynamic.hinet.net] has quit [Read error:
Operation timed out]
02:38 -!- shvntr [~shvntr@116.26.134.53] has joined #go-nuts
03:11 -!- foocraft [~dsc@78.101.95.201] has quit [Read error: Connection reset by
peer]
03:15 -!- ajstarks [~ajstarks@pool-98-109-198-180.nwrknj.fios.verizon.net] has
left #go-nuts []
03:16 -!- ShadowIce [~pyoro@unaffiliated/shadowice-x841044] has quit [Quit:
Verlassend]
03:23 -!- Stiletto [7f000001@69.195.144.4] has quit [Ping timeout: 240 seconds]
03:24 -!- Stiletto [7f000001@69.195.144.4] has joined #go-nuts
03:34 -!- Stiletto [7f000001@69.195.144.4] has quit [Ping timeout: 264 seconds]
03:34 -!- Stiletto [7f000001@69.195.144.4] has joined #go-nuts
03:39 -!- quotemstr [~quotemstr@173.224.210.52] has joined #go-nuts
03:39 < quotemstr> Err --- how is panic/recover any different from
try/catch?
03:40 < Namegduf> quotemstr: It doesn't require special syntax, and it is
considered equivalent to a crash to let a panic escape a package, so you can rely
on not having to deal with unknown panics in other people's code- they're only in
a package if that package is using them, and only so far as it does use them.
03:41 < quotemstr> Aside from the syntax bit (which reminds me more of
Perl's die mechanism and setjmp), that's still not any different.
03:42 < Namegduf> Yes, aside from the part that IS the language, the
language isn't any different.
03:42 < Namegduf> Also aside from the colour blue is pretty much the same as
red.
03:42 < quotemstr> The semantics are the same though: the stack is unwound
and cleanup functions (derferred ones in Go's case) are run.
03:43 < Namegduf> Yes.
03:43 < quotemstr> So why not use the terminology most other people use?
03:43 < Namegduf> Because Go would be using it differently.
03:43 < Namegduf> The syntax would be different and the idiomatically
correct times and ways to use it are different.
03:44 < quotemstr> So when would you prefer to use panic instead of a return
code?
03:44 < Namegduf> NEVER, in a public API.
03:44 < Namegduf> Internally, you would often use it to abort something
complex, like parsing.
03:44 < Namegduf> And catch it at the top level, where you return an error.
03:44 < quotemstr> Sounds just like the way I use C++ exceptions.
03:45 < quotemstr> Exceptions shall not cross library boundaries.
03:45 < Namegduf> That's reasonable.  Unfortunately, it isn't what
exceptions mean in many other languages.
03:45 < Namegduf> And a lot of people would get confused.
03:45 < Namegduf> It also doesn't have special syntax like most other things
called exceptions, and people would expect the syntax to look similar.
03:45 < quotemstr> You could have used setjmp and longjmp then.
03:46 < str1ngs> Namegduf: you are saying that all public api's should
handle panics?  where need
03:46 < Namegduf> It doesn't behave like those, either.
03:46 < Namegduf> Those don't unwind the stack.
03:46 < Namegduf> recover() also allows catching things which aren't
manually initiated with a panic()
03:46 < quotemstr> Namegduf: They do on some platforms.  :) But that's
beside the point.  I see what you mean.
03:46 < Namegduf> You can catch nil pointer dereferences
03:46 < Namegduf> Or array indexing faults
03:47 < Namegduf> THe other case to use recover() is where it isn't
expressly paired with a panic(), and that's to catch errors resulting from
programmer errors and stop them going any further.
03:48 < Namegduf> Similar to the way you'd have something catching "any
exception" in Java or Python or whatever.
03:48 < Namegduf> Perhaps in a client connection handler or something.
03:49 < Namegduf> Defensive; if they do figure out how to crash your code
somehow, all it does is abort their thing unless they manage to corrupt state
elsewhere.
03:50 < quotemstr> COM does something similar by default for RPC servers.
03:50 < quotemstr> If some component dies horribly, COM catches it and just
returns E_FAIL or somesuch.
03:50 < Namegduf> Yeah, you can do that.
03:50 < Namegduf> recover() is nice and general enough to cover that case,
too.
03:50 * quotemstr is not in favor of that approach.
03:50 < Namegduf> And the name makes sense.
03:50 < quotemstr> It puts components into states they don't expect.
03:51 < Namegduf> Yeah, it's not perfect..
03:51 < quotemstr> Does Go have lexical closures?
03:51 < Namegduf> Closures which can access variables in the outside
function?  Yes.
03:53 < quotemstr> e.g., in JavaScript, function() { var x = 4; return
function() { return ++x; }; };
03:53 < Namegduf> str1ngs: http://golang.org/doc/effective_go.html#panic
03:53 < Namegduf> str1ngs: It emphasises that panic() is "taking down the
whole program"
03:53 < Namegduf> And discusses various cases (init issues) where that's
appropriate
03:54 < Namegduf> The next section discusses recover() and has an example
function.
03:54 < Namegduf> Which uses it internally.
03:54 < Namegduf> Then goes on to say: "Useful though this pattern is, it
should be used only within a package.  Parse turns its internal panic calls into
os.Error values; it does not expose panics to its client.  That is a good rule to
follow."
03:55 < str1ngs> ah ok
03:55 < str1ngs> so the idea is a package will never stop another calling
cmd or pkg
03:55 < Namegduf> When you are writing a package
03:55 < Namegduf> You only ever have to worry about panics you put there
03:56 < Namegduf> In *that* package
03:56 < str1ngs> ok gotcha
03:56 < quotemstr> Assuming everyone follows the convention.
03:56 -!- st-6308 [foobar@92.250.11.202] has joined #go-nuts
03:56 < Namegduf> If they break the convention, "their code crashes".
03:57 < Namegduf> People can write crashy code if they like, but it won't be
very liked
03:57 < Namegduf> I don't think it's a real issue.
03:57 < Namegduf> It is a valid point, though.
03:58 < str1ngs> I really need to look into this some more.  I have one cgo
package.  that I should make sure handles this
03:58 < Namegduf> Most of the time, you get the "unwind stack" ability to
deal with throwing things up many levels, but don't get the horrible cost of
dealing with random exceptions being thrown by everything.
03:58 < Namegduf> Well, if you don't panic() explicitly
03:58 < str1ngs> even more so due to the C pointers
03:59 < quotemstr> Well, a library author could just declade that errors are
propagated via panic.
03:59 < quotemstr> Sure, that might give some people the vapors, but it'd
work, yes?
03:59 < Namegduf> Then a panic() which goes up through your package is
actually a programmer error and a crash
03:59 < Namegduf> They could, but it'd be bad.
03:59 < quotemstr> I never bought the argument that exceptions were bad
things to be avoided at all costs.
03:59 < Namegduf> Then their users would need to deal with tracking pancis
originating from foreign code.
03:59 < quotemstr> If throwing is expensive, fix throwing.  Don't blame the
feature.
04:00 < Namegduf> It isn't a matter of cost, it's a matter of safety.
04:00 < cbeck> It's not as simple as 'fix' throwing.  There are engineering
tradeoffs to be made
04:01 < Namegduf> You can add checked exceptions ala Java
04:01 < Namegduf> But those are painful
04:01 < cbeck> I mean the underlying mechanism
04:01 < Namegduf> quotemstr: The problem is that panics() are not part of
the signature, and documentation on panics() it can generate is not compile-time
checkd
04:02 < Namegduf> A mistake in ensuring all panics thrown up past an API
will result in a crash and none of it is possible to verify.
04:02 < Namegduf> *past an API are handled
04:03 < Namegduf> Thus the trivial solution is to say that APIs do not
panic() on purpose, ever
04:04 < Namegduf> Which cuts the problem of tracking all possible cases
where a panic() is manually generated which should be caught, and a recover()
exists to catch them, into a single package
04:04 < Namegduf> Which should be small enough for a human brain to handle
okay.
04:05 < Namegduf> You *can* ignore it, but it's like taking interface{}
parameters and then using reflect to figure out what they are, except there's
actually uses for that.
04:05 < Namegduf> You're losing compile-time checking and a bunch of safety.
04:05 < Namegduf> Regardless of speed considerations.
04:06 -!- nsf [~nsf@jiss.convex.ru] has joined #go-nuts
04:07 -!- bpalmer` [~user@adsl-69-232-221-158.dsl.pltn13.pacbell.net] has joined
#go-nuts
04:12 -!- rejb [~rejb@unaffiliated/rejb] has quit [Ping timeout: 246 seconds]
04:18 -!- nettok [~quassel@200.119.190.184] has joined #go-nuts
04:23 -!- nsf [~nsf@jiss.convex.ru] has quit [Quit: WeeChat 0.3.4]
04:23 -!- bugQ [~bug@c-67-171-127-76.hsd1.ut.comcast.net] has quit [Quit: but
don't be sad.]
04:30 -!- st-6308 [foobar@92.250.11.202] has quit [Remote host closed the
connection]
04:33 -!- st-6519 [foobar@92.250.11.202] has joined #go-nuts
04:36 -!- st-6519 [foobar@92.250.11.202] has quit [Remote host closed the
connection]
04:40 -!- kanru [~kanru@kanru-1-pt.tunnel.tserv15.lax1.ipv6.he.net] has joined
#go-nuts
04:47 -!- nettok [~quassel@200.119.190.184] has quit [Ping timeout: 248 seconds]
04:48 -!- iant [~iant@216.239.45.130] has joined #go-nuts
04:48 -!- mode/#go-nuts [+v iant] by ChanServ
04:55 < steven> <3
04:56 -!- yihuang [~yihuang@183.17.204.27] has joined #go-nuts
04:56 -!- yihuang [~yihuang@183.17.204.27] has left #go-nuts []
05:01 -!- foocraft [~dsc@dyn-86-36-41-37.wv.qatar.cmu.edu] has joined #go-nuts
05:04 < str1ngs> my awesome tarball extractor https://gist.github.com/888927
05:07 < quotemstr> Namegduf: Fair enough.  Not that it matters --- these
days, I'm writing a lot of elisp and C, both of which are primitive in their own
way.
05:07 < quotemstr> Night.
05:10 -!- pilgrum [~pilgrum@cpe-67-49-71-222.socal.res.rr.com] has joined #go-nuts
05:30 -!- krutcha1 [~krutcha@S010600045a27676a.vs.shawcable.net] has joined
#go-nuts
05:31 -!- quotemstr [~quotemstr@173.224.210.52] has left #go-nuts []
05:37 -!- foocraft [~dsc@dyn-86-36-41-37.wv.qatar.cmu.edu] has quit [Remote host
closed the connection]
06:02 -!- pphalen [~patrick@66.92.11.149] has joined #go-nuts
06:09 -!- pphalen [~patrick@66.92.11.149] has left #go-nuts []
06:15 -!- ExtraSpice [XtraSpice@88.118.35.153] has joined #go-nuts
06:26 -!- nsf [~nsf@jiss.convex.ru] has joined #go-nuts
06:31 -!- photron [~photron@port-92-201-133-223.dynamic.qsc.de] has joined
#go-nuts
06:37 -!- aho [~nya@fuld-590c7df1.pool.mediaWays.net] has quit [Quit:
EXEC_over.METHOD_SUBLIMATION]
06:39 -!- iant [~iant@216.239.45.130] has quit [Ping timeout: 246 seconds]
06:41 -!- [Pete_27] [~noname@110-174-103-31.static.tpgi.com.au] has joined
#go-nuts
06:47 -!- pilgrum [~pilgrum@cpe-67-49-71-222.socal.res.rr.com] has quit [Ping
timeout: 240 seconds]
06:57 -!- iant [~iant@adsl-71-133-8-30.dsl.pltn13.pacbell.net] has joined #go-nuts
06:57 -!- mode/#go-nuts [+v iant] by ChanServ
07:03 -!- |Craig| [~|Craig|@panda3d/entropy] has quit [Quit: |Craig|]
07:29 -!- tensorpudding [~user@99.148.205.193] has quit [Remote host closed the
connection]
07:34 -!- ronnyy [~quassel@p4FF1C0A2.dip0.t-ipconnect.de] has joined #go-nuts
07:51 -!- huin [~huin@91.85.185.181] has joined #go-nuts
07:52 -!- awidegreen [~quassel@c-eacae555.08-2-73746f39.cust.bredbandsbolaget.se]
has quit [Ping timeout: 240 seconds]
07:55 -!- awidegreen [~quassel@c-eacae555.08-2-73746f39.cust.bredbandsbolaget.se]
has joined #go-nuts
07:55 -!- tensai_cirno [~cirno@77.232.15.216] has quit [Ping timeout: 246 seconds]
08:02 -!- awidegreen [~quassel@c-eacae555.08-2-73746f39.cust.bredbandsbolaget.se]
has quit [Quit: No Ping reply in 180 seconds.]
08:02 -!- awidegreen [~quassel@c-eacae555.08-2-73746f39.cust.bredbandsbolaget.se]
has joined #go-nuts
08:06 -!- tobier [~tobier@c-1e9de055.712-1-64736c11.cust.bredbandsbolaget.se] has
quit [Read error: Operation timed out]
08:07 -!- awidegreen [~quassel@c-eacae555.08-2-73746f39.cust.bredbandsbolaget.se]
has quit [Ping timeout: 250 seconds]
08:08 -!- tobier [~tobier@c-1e9de055.712-1-64736c11.cust.bredbandsbolaget.se] has
joined #go-nuts
08:10 -!- rutkowski [~adrian@178235051224.walbrzych.vectranet.pl] has joined
#go-nuts
08:13 -!- pilgrum [~pilgrum@cpe-67-49-71-222.socal.res.rr.com] has joined #go-nuts
08:15 -!- femtoo [~femto@95-89-249-242-dynip.superkabel.de] has joined #go-nuts
08:20 -!- ljs
[~user@customer4523.pool1.Croydon-GLN2000-BAS0001.orangehomedsl.co.uk] has quit
[Ping timeout: 264 seconds]
08:25 -!- pilgrum [~pilgrum@cpe-67-49-71-222.socal.res.rr.com] has quit [Quit:
Leaving]
08:25 -!- tvw [~tv@e176000013.adsl.alicedsl.de] has joined #go-nuts
08:34 -!- krutcha1 [~krutcha@S010600045a27676a.vs.shawcable.net] has quit [Quit:
Leaving]
08:36 -!- tensai_cirno [~cirno@77.232.15.216] has joined #go-nuts
08:40 -!- ronnyy [~quassel@p4FF1C0A2.dip0.t-ipconnect.de] has quit [Remote host
closed the connection]
08:40 -!- littlebobby [~bob@unaffiliated/littlebobby] has joined #go-nuts
08:40 -!- ronnyy [~quassel@p4FF1C0A2.dip0.t-ipconnect.de] has joined #go-nuts
08:43 -!- ronnyy [~quassel@p4FF1C0A2.dip0.t-ipconnect.de] has quit [Remote host
closed the connection]
08:46 -!- femtoo [~femto@95-89-249-242-dynip.superkabel.de] has quit [Read error:
Connection reset by peer]
08:46 -!- ronnyy [~quassel@p4FF1C0A2.dip0.t-ipconnect.de] has joined #go-nuts
08:47 < huin> what's generally regarded as more efficient: create a
goroutine + syncronous channel to make a python-like generator, or allocate a
slice and return, in terms of generating on the order of 400 items?
08:47 -!- ronnyy [~quassel@p4FF1C0A2.dip0.t-ipconnect.de] has quit [Remote host
closed the connection]
08:49 -!- rutkowski [~adrian@178235051224.walbrzych.vectranet.pl] has quit [Quit:
WeeChat 0.3.3-dev]
08:57 -!- rlab [~Miranda@91.200.158.34] has joined #go-nuts
09:00 < edsrzf> huin: Returning a slice is more efficient, but you'll use
more memory.
09:00 -!- napsy [~luka@tm.213.143.73.175.lc.telemach.net] has quit [Quit: Lost
terminal]
09:01 -!- ronnyy [~quassel@p4FF1C0A2.dip0.t-ipconnect.de] has joined #go-nuts
09:01 < huin> thanks :)
09:02 < huin> i guess i have to decide on CPU vs memory efficiency, but at
this stage it's probably premature optimisation to do so
09:03 < edsrzf> Yeah, at 400 objects you can probably spare the memory,
unless the objects are huge.
09:03 < huin> i think i can just code this "the nice way" and decide later
what's best.  changing this later would be easy as it's a confined section of code
09:04 < huin> given that goroutines probably use 4k or so each on startup
(is that right?), then memory savings would likely be negated anyway
09:05 < edsrzf> No, I don't believe it's that much
09:11 -!- edsrzf [~chickench@122-61-221-144.jetstream.xtra.co.nz] has quit [Remote
host closed the connection]
09:14 -!- tobier [~tobier@c-1e9de055.712-1-64736c11.cust.bredbandsbolaget.se] has
quit [Read error: Operation timed out]
09:15 -!- tobier [~tobier@c-1e9de055.712-1-64736c11.cust.bredbandsbolaget.se] has
joined #go-nuts
09:21 -!- Count_Niedar [~bleh@ip68-99-166-222.hr.hr.cox.net] has joined #go-nuts
09:23 -!- cw [~anticw@parsec.stupidest.org] has quit [Ping timeout: 248 seconds]
09:23 -!- cw [~anticw@parsec.stupidest.org] has joined #go-nuts
09:23 -!- Niedar [~bleh@ip68-99-166-222.hr.hr.cox.net] has quit [Ping timeout: 248
seconds]
09:29 -!- ronnyy [~quassel@p4FF1C0A2.dip0.t-ipconnect.de] has quit [Remote host
closed the connection]
09:39 -!- tvw [~tv@e176000013.adsl.alicedsl.de] has quit [Remote host closed the
connection]
09:40 -!- Project_2501 [~Marvin@dynamic-adsl-94-36-166-18.clienti.tiscali.it] has
joined #go-nuts
09:49 -!- rtharper [~tomh@unaffiliated/sioraiocht] has joined #go-nuts
10:16 -!- tensai_cirno [~cirno@77.232.15.216] has quit [Quit: Leaving]
10:16 -!- saturnfive [~saturnfiv@219.144.238.246] has joined #go-nuts
10:17 -!- saturnfive [~saturnfiv@219.144.238.246] has left #go-nuts []
10:27 -!- jumzi [~none@c-89-233-234-125.cust.bredband2.com] has joined #go-nuts
10:42 -!- nsf [~nsf@jiss.convex.ru] has quit [Quit: WeeChat 0.3.4]
10:45 -!- espeed [~espeed@63.246.231.57] has joined #go-nuts
10:59 -!- boscop [~boscop@g230087148.adsl.alicedsl.de] has joined #go-nuts
11:01 -!- waqas [~waqas@jaim.at] has joined #go-nuts
11:02 -!- Niedar [~bleh@ip68-99-166-222.hr.hr.cox.net] has joined #go-nuts
11:04 -!- shvntr [~shvntr@116.26.134.53] has quit [Ping timeout: 246 seconds]
11:06 -!- Count_Niedar [~bleh@ip68-99-166-222.hr.hr.cox.net] has quit [Ping
timeout: 276 seconds]
11:43 -!- femtoo [~femto@95-89-249-242-dynip.superkabel.de] has joined #go-nuts
11:45 -!- shvntr [~shvntr@116.26.134.53] has joined #go-nuts
11:47 -!- littlebobby [~bob@unaffiliated/littlebobby] has quit [Quit: Ex-Chat]
11:58 -!- rejb [~rejb@unaffiliated/rejb] has joined #go-nuts
12:11 < xyproto> str1ngs: nice bzip2/gzip snippet.  How about .xz as well?
It is quickly becoming more widespread.  :)
12:14 -!- ShadowIce
[~pyoro@HSI-KBW-109-193-120-162.hsi7.kabel-badenwuerttemberg.de] has joined
#go-nuts
12:14 -!- ShadowIce
[~pyoro@HSI-KBW-109-193-120-162.hsi7.kabel-badenwuerttemberg.de] has quit
[Changing host]
12:14 -!- ShadowIce [~pyoro@unaffiliated/shadowice-x841044] has joined #go-nuts
12:18 < str1ngs> xyproto: thats next on my list
12:19 < str1ngs> I could use a better test for file type something like
/usr/bin/file
12:20 < str1ngs> but that might be overkill for now
12:25 -!- nsf [~nsf@jiss.convex.ru] has joined #go-nuts
12:26 < xyproto> str1ngs: maybe it's possible to just look at the two first
bytes of the file?  All tar.bz2 files I have lying around starts with 5a42, for
example
12:27 < xyproto> str1ngs: and tar.gz files starts with 8b1f
12:27 < str1ngs> not sure
12:27 < xyproto> str1ngs: I think that's how file does it.
12:27 < xyproto> str1ngs: I've looked just briefly at the file sourcecode,
and I am on the file mailinglist, but I'm not a file-expert
12:27 < str1ngs> I also need to figure out away to peek at the tar, get how
many dir it contains
12:28 < str1ngs> most contain one dir.  but I cant assume that.
12:29 -!- unofficialmvp [~dev@94-62-164-227.b.ipv4ilink.net] has joined #go-nuts
12:29 < str1ngs> actually what I might do is just do tmp/wget/wget-build
tmp/wget/wget-stage untar to wget-stage and the get dir list
12:30 < str1ngs> thats saner.  and I dont have to peek at the tar if I've
already staged and just doing a rebuild
12:31 -!- unofficialmvp [~dev@94-62-164-227.b.ipv4ilink.net] has left #go-nuts []
12:31 -!- wrtp [~rog@92.17.50.183] has joined #go-nuts
12:31 < xyproto> str1ngs: sounds like a plan :)
12:32 < xyproto> hi wrtp
12:32 < wrtp> xyproto: hi
12:32 < str1ngs> xyproto: but I was thinking about making a gotar.  and if I
do that then ya it should have better filetype handling
12:34 < str1ngs> xyproto: if you want to play with it I can convert it to a
pkg
12:35 < xyproto> str1ngs: sure, I'll try it out
12:35 < str1ngs> xyproto: ok I'll stick it on githup.
12:36 < str1ngs> something like libarchive I guess
12:36 < str1ngs> but to be honest I didnt do much work just pieced what
stdlib together
12:37 < xyproto> it's fine, I'll give it a go :)
12:38 < nsf> iant: oh, I wanted to ask you..  in your constant eval code
there is a thing for bit shifts (left shift op): "shift > 0x100000" is it just
a protection from memory bloat?
12:39 -!- wrtp [~rog@92.17.50.183] has quit [Quit: wrtp]
12:40 -!- wrtp [~rog@92.17.50.183] has joined #go-nuts
12:40 < nsf> if user tries to shift more than 0x100000 bits, it's a shift
overflow in your code
12:41 < aiju> Intel CPUs do something funny when you shift too much
12:41 < aiju> x <<= 32; is a nop
12:41 < nsf> aiju: it's in const eval code :)
12:41 < nsf> arbitrary precision math here
12:41 < aiju> sure, i'm just mentioning something related ;P
12:41 < nsf> and well, shifting MAX UINT bits doesn't make sense
12:41 < nsf> the resulting int will be 500 megs
12:42 < nsf> if no compression is used :)
12:42 < aiju> gzip in memory
12:42 < nsf> simple run length encoding is fine here
12:50 < str1ngs> xyproto: https://github.com/str1ngs/goarchive I reused the
go license.  it might not work without some tweaking.  and the test framework is
there but needs to be filled in
12:54 -!- yihuang [~yihuang@183.17.189.72] has joined #go-nuts
12:56 -!- gr0gmint [~quassel@87.61.162.99] has joined #go-nuts
13:08 -!- ccc_ [62cb062b@gateway/web/freenode/ip.98.203.6.43] has joined #go-nuts
13:14 -!- Project-2501 [~Marvin@82.84.70.6] has joined #go-nuts
13:15 < ww> rofl.  these android people.
13:15 < ww> "The main point is, an insecure/public API bridge should only be
13:15 < ww> launch-able via direct user action.  "
13:15 < aiju> wtf?
13:15 < ww> well...  i can launch one with a shell command
13:15 < aiju> wtf is an "API bridge" after all?
13:15 < aiju> newspeak for wrapper?
13:16 < ww> and then look at netstat to find the port in a brute force way
(ther are only a handful)
13:16 < ww> aiju: more or less.  json over tcp <-> java stuff
13:16 < str1ngs> ww: andriod is really not bad iphone is way worst
13:16 < aiju> Hitler is worse than android, ergo android is good
13:16 < ww> there's some funny kool-aid thing going on though
13:17 * ww wishes openmoko had gotten more funding
13:17 < str1ngs> I'd still take the lesser of 2 evils every time :P
13:17 -!- Project_2501 [~Marvin@dynamic-adsl-94-36-166-18.clienti.tiscali.it] has
quit [Ping timeout: 250 seconds]
13:18 < ww> google just looked at openmoko and said, hey we can take that
and make it shiny on the surface and screw it up underneath
13:18 < aiju> i remember openmoko to be fucked up as well
13:18 < str1ngs> ww: it might get worst before it get's better.  if you
really look at what apple is doing its scary
13:19 < ww> aiju: openmoko had the right idea, just not enough resources to
do the work and clout with the manufacturers and carriers
13:19 < aiju> what is the "right idea"?
13:19 -!- napsy [~luka@88.200.96.18] has joined #go-nuts
13:19 -!- shvntr [~shvntr@116.26.134.53] has quit [Quit: leaving]
13:19 < str1ngs> apple axed opera on iphone, stating it was not suitable for
children under 18
13:19 < aiju> when all you have is a Linux, everything looks like a server
13:19 < ww> str1ngs: what?
13:20 < str1ngs> ww: because opera doesnt have some parental guidance thing.
its not suitable for children under 18.  according to apple
13:21 < ww> obviously they don't have children
13:21 < ww> if they did, they'd know that the first thing children learn how
to do is subvert parental controls
13:22 < str1ngs> but I need to find this source.  I saw it somewhere in
passing
13:24 < xyproto> aiju: Godwin's law :P
13:24 < xyproto> str1ngs: I'll look at it :)
13:24 < aiju> Hitler would send you to concentration camp for mentioning
Godwin's law
13:24 < xyproto> aiju: oh no, I have no defense for recurse Godwin!  :D
13:24 < xyproto> *recursive
13:26 < xyproto> str1ngs: it compiles fine and the test says PASS here
13:26 < str1ngs> xyproto: test will pass because there is not logic
13:26 < str1ngs> no*
13:26 < xyproto> oh, ok
13:27 < str1ngs> I just put the framework together.  I'll add a test to
Untar
13:27 < str1ngs> might be easier to add a Tar then use that for testing
Untar
13:28 -!- jokoon [~eio@feu30-1-82-242-58-229.fbx.proxad.net] has joined #go-nuts
13:28 < str1ngs> if you plan on doing alot of hacking on it I'll add you to
github so you can just push and dont have to do pull requests.  and I'll do some
testing as you push
13:31 < str1ngs> also I dont Stat for dest dir.  I need to add that.  since
my calling app does that
13:31 < str1ngs> thats kind important I guess ..
13:32 -!- tvw [~tv@e176000013.adsl.alicedsl.de] has joined #go-nuts
13:32 < nsf> iant: hehe, found the comment regarding that 0x100000,
strangely enough not in a hg repo, but somewhere in the nowhere:
http://patchwork.ozlabs.org/patch/83905/
13:32 < xyproto> str1ngs: I could add the detection of .gz and .bz2 at
least, if you want to.  Feel free to add me, it would be fun.  :)
13:33 < str1ngs> xyproto: sure no pressure though.  xyproto on github?
13:34 < xyproto> str1ngs: yes
13:34 < xyproto> str1ngs: I haven't been that active on github, but I want
use it more
13:35 < str1ngs> xyproto: like I said no pressure.
13:35 < nsf> iant: or I guess hg repo of the gofrontend was created from the
gcc's svn repo, right?  :)
13:35 < nsf> and that's the place where I should look for comments
13:35 < xyproto> str1ngs: no worries
13:37 < str1ngs> xyproto: ok added
13:39 < str1ngs> does gccgo require libffi.  the maintainer of gcc on
archlinux was asking about ti
13:39 < str1ngs> it*
13:40 < str1ngs> my guess is that gccgo uses libffi to interface with either
C or C++ but I'm not sure
13:43 < nsf> str1ngs: I think it's very unlikely
13:43 < nsf> gccgo uses just gcc
13:43 < nsf> afaik
13:43 < nsf> although, I don't know
13:43 < nsf> ask iant
13:44 < str1ngs> readelf -a libgo.so | grep ffi ..  seems to say yes
13:45 < nsf> hm..
13:47 < str1ngs> meh probably not important he can ask on the mailing list
if he needs to
13:52 -!- fabled [~fabled@mail.fi.jw.org] has quit [Quit: Ex-Chat]
13:53 < skelterjohn> morning
13:53 -!- schmrkc [~marcus@sxemacs/devel/schme] has quit [Ping timeout: 260
seconds]
14:10 < nsf> [nsf @ crawl]$ echo 'const X = 3.14 * 4 > 12.3 && (4*8)/24
== 1;' | ./crawl -no-ast
14:10 < nsf> const 'X': 1
14:10 < nsf> :P
14:10 < aiju> > is shift?
14:11 < nsf> no
14:11 < nsf> comparison
14:11 < aiju> shouldn't X be true then?
14:11 < nsf> greater than
14:11 < nsf> yes
14:11 < nsf> I use int for booleans
14:11 < aiju> oh you have integers as booleans?
14:11 < aiju> oic
14:11 < nsf> 1 and 0
14:11 < aiju> 2 should cause a segfault ;P
14:11 < nsf> but yeah, I should print type as well :)
14:13 < nsf> [nsf @ crawl]$ echo 'const X = (1 << 100) >> 98;' |
./crawl -no-ast
14:13 < nsf> const 'X': 4
14:14 < nsf> Go's shift example :)
14:14 < aiju> what's it written in?
14:14 < aiju> Go or C?
14:14 < nsf> uhm, crawl?
14:14 < nsf> in C++
14:14 < aiju> why C++?  D:
14:14 < nsf> uhm, for future tight integration with LLVM
14:14 < nsf> and libclang
14:15 < nsf> also a good bootstrap target as well
14:15 < aiju> doesn't llvm has a C api?
14:15 < nsf> also has std::vector and std::unordered_map
14:15 < aiju> *have
14:15 < nsf> it has, but it's incomplete
14:16 < nsf> [nsf @ crawl]$ echo 'const X = (7+8) / (4+16-20);' | ./crawl
-no-ast
14:16 < nsf> stdin:1:17: error: division by zero
14:16 < nsf> const X = (7+8) / (4+16-20);
14:16 < nsf> ~~~~~ ^ ~~~~~~~~~
14:16 < nsf> I have nice error messages as well :)
14:17 < nsf> or should I say as clang?  lol
14:17 -!- yihuang [~yihuang@183.17.189.72] has left #go-nuts []
14:17 < aiju> the UNIX assembler had the best error messages
14:17 < aiju> N0059
14:17 < aiju> no need to be any more verbose ;P
14:18 < exch> you had a lookup table printed out and stuck to the wall
behind your desk?  :p
14:18 < aiju> no, you knew them by heart
14:18 < exch> hardcore
14:18 < aiju> the number is the line number
14:18 < exch> ah
14:18 < aiju> most errors were "unknown label" and such
14:19 < aiju> http://man.cat-v.org/unix-6th/1/as
14:19 < aiju> oh it even printed the file name
14:19 < aiju> luxury
14:20 < aiju> the reason for those cryptic errors is probably the same as
for ed: long error messages are just annoying on 10 char/sec terminals
14:21 < nsf> lol
14:42 -!- nettok [~quassel@200.119.188.116] has joined #go-nuts
15:09 < xyproto> I just discovered the package-name can not be "main" when
having a something.go and something_test.go test.  Why is that, I wonder?
15:10 < xyproto> make test fails when the first line of both is "package
main"
15:11 < xyproto> strangeness, it works in a different directory.  Maybe I've
done something strange.
15:12 < str1ngs> I dont think there is a need to have main as a package name
15:12 < xyproto> when I have "package main" in both files, I get:
_testmain.go:9: undefined: __main__
15:12 < xyproto> and also "main redeclared in this block"
15:12 < str1ngs> is it a command package?
15:13 < xyproto> str1ngs: it's a little test application.  The Makefile
includes Make.pkg
15:13 < str1ngs> I'm not sure if testing works with command projects
15:13 < xyproto> str1ngs: it works if I change "package main" to something
else, though
15:14 < str1ngs> whats happening is its creating as main() which conflicts
with main from foo.go
15:14 < xyproto> Ah, since foo.go is not a package, it does have a main,
hence the conflict.  I see.
15:15 < xyproto> It just surprised me that the package name would cause a
conflict, that's all.
15:15 < str1ngs> right thats why foot_test.go is used for testing packages
since there is no main
15:15 < str1ngs> make test creates a main() etc
15:15 * ww has to have a look at how to test commands...
15:16 < ww> maybe it doesn't really matter that much
15:16 < str1ngs> I just make a test: target in the Make file and run make
test.
15:16 < exch> If I'm not mistaken, gotest is set to work with commands as
well as packages at some point.  Not sure when that will occur or if it already
has in the latest versions
15:16 < xyproto> "make test" works here without making a test target, just
by including the go makefile includes
15:16 -!- jokoon [~eio@feu30-1-82-242-58-229.fbx.proxad.net] has quit [Quit:
Leaving]
15:16 < skelterjohn> xyproto: gb can do testing for package main :)
15:16 < xyproto> skelterjohn: what is gb?  :)
15:17 < skelterjohn> go-gb.googlecode.com
15:17 < xyproto> *checking*
15:17 < skelterjohn> build tool i created
15:17 < nsf> D has a special syntax construction: unittest {}
15:17 < nsf> what do you think about it?
15:17 < xyproto> nsf: sounds nice, but maybe a bit too helpful?  :)
15:17 < exch> Not sure if a dedicated language construct is really necessary
tbh
15:18 < nsf> well, why can't we have it?  it's pretty much an orthogonal
feature
15:18 < nsf> you can even use the language without knowing it exists
15:18 < nsf> but it makes testing simpler
15:18 < Namegduf> I don't think you can
15:18 < nsf> because every tester tool can recognize unittest clauses
15:18 < nsf> and execute the code
15:18 < Namegduf> A programmer needs to know all the stuff he can encounter
and deal with when dealing with other people's code
15:18 < nsf> Namegduf: but then it's very simple
15:19 < nsf> even without reading the doc :)
15:19 < Namegduf> It's not a great complexity add but does it really provide
a great deal of gain to efficency of writing code?
15:19 < Namegduf> *efficiency
15:19 < nsf> it affects culture a bit
15:20 < nsf> by showing that language supports unit testing as a known
technique
15:20 < nsf> other than that
15:20 < nsf> I don't think there are any gains
15:20 -!- tvw [~tv@e176000013.adsl.alicedsl.de] has quit [Remote host closed the
connection]
15:20 < ww> i guess it might be nice to be able to arrange to have unit
tests near the code that they test...
15:20 < ww> they also serve as examples...
15:21 < ww> like python's doctests (that are underused in favour of much
heavier test harneses)
15:21 < Namegduf> *_test.go accomplishes that as well, though.
15:21 < Namegduf> "near", if not next to.
15:21 < nsf> I think *_test.go is worse than unittest
15:21 < Namegduf> They are unit tests
15:21 < nsf> because it's an undocumented collection of some kind of
conventions
15:21 < ww> Namegduf: sure.  a bit farther than doctests, but still near
15:21 < nsf> I mean D's 'unittest'
15:21 < nsf> syntax
15:22 < Namegduf> I'm not sure trying to represent all conventions through
syntax is a good approach for a simple language
15:22 < Namegduf> Or one that's feasible without stopping to be a simple
language
15:22 < Namegduf> That's my opinion, anyway.
15:23 < ww> i guess it wouldn't be exceptionally difficult to extract tests
from comments and generate *_test.go
15:23 < ww> that way examples would flourish in the documentation...
15:23 < Namegduf> Or visa versa
15:24 -!- tvw [~tv@e176000013.adsl.alicedsl.de] has joined #go-nuts
15:24 < Namegduf> Just search *_test.go for function calls and you have
examples.
15:24 < xyproto> I would like quicktests, where calls and expected results
are written in a short comment above the function declaration
15:24 < exch> tests will not always be a good example representation.
Sometimes you want tests to fail deliberately
15:24 < Namegduf> I'm dubious of those providing enough completeness.
15:24 < Namegduf> Ah, fair point.
15:25 < exch> or rather, the tests should succeed in failing
15:25 < exch> :P
15:25 < Namegduf> Yeah.
15:25 * Namegduf is lazy about writing unit tests quickly.  :(
15:26 < Namegduf> It's because I like to change semantics and provided
functionality a lot while still in the "get version 0.1" phase.
15:26 < Namegduf> Try to look at the whole thing and come up with ways to
provide what the actual code using a package is doing simpler and blah.
15:27 -!- xyproto [~alexander@77.40.159.131] has quit [Quit: WeeChat 0.3.4]
15:28 -!- xyproto [~alexander@cm-84.208.177.43.getinternet.no] has joined #go-nuts
15:28 -!- Fish- [~Fish@9fans.fr] has joined #go-nuts
15:28 -!- waqas [~waqas@jaim.at] has left #go-nuts []
15:31 < ww> ok...  homebrew is quickly devolving into me just downloading
stuff and doing ./configure && make && make install
15:32 < aiju> homebrew?
15:32 < ww> someone said ditch macports for homebrew
15:32 < ww> and homebrew devolves much more quickly into roll-your-own
15:32 < aiju> wtfi homebrew?  are you playing with some console?
15:32 < ww> aiju: get a mac :p
15:33 < aiju> isn't there ports?
15:33 < aiju> homebrew is based on fucking ruby
15:34 < ww> yeah...  i think this started when i was making a cross compiler
and macports got in the way...
15:34 -!- Fish- [~Fish@9fans.fr] has quit [Ping timeout: 276 seconds]
15:34 * ww is feeling very anti-package-management now
15:35 < aiju> a friend of mine ported portage to OS X
15:35 < skelterjohn> I'm pretty sure the language a package manage is
written is not its most important feature
15:35 < aiju> you actually have a sort of Gentoo subsystem with that
15:36 < aiju> yeah, because a package manager eating dozens of megs is such
a good idea
15:37 < Namegduf> Dozens?
15:37 < aiju> Namegduf: being optimistic
15:37 < Namegduf> Don't be ridiculous.
15:37 < Namegduf> Hundreds!
15:37 < aiju> not to mention the speed
15:37 -!- nettok [~quassel@200.119.188.116] has quit [Ping timeout: 240 seconds]
15:37 < Namegduf> I don't really care, though
15:37 < aiju> portage dependency resolution can take longer than actual
compilation <.<
15:38 * Namegduf understands macports to be about as reliable as portage and is
all for homebrew
15:38 * ww rolls own
15:38 < Namegduf> Ah, OS X.
15:38 < ww> they all do the wrong thing
15:39 < aiju> are records a planned feature for Lion yet?
15:39 < Namegduf> Records?
15:39 < aiju> file records
15:39 < ww> you know, vinyl
15:39 < Namegduf> Lion has an App Store
15:40 < aiju> in the old days, everyone KNEW that files can't be byte
streams
15:40 < ww> show your 33 1/3s to the camera and play them with itunes
15:40 < Namegduf> It's like package management but only for OS X programs
without dependencies
15:40 < aiju> they have to be split into records
15:40 -!- femtoo [~femto@95-89-249-242-dynip.superkabel.de] has quit [Read error:
Connection reset by peer]
15:40 < Namegduf> aiju: That's what multiple files are for
15:40 < aiju> Namegduf: a record is more like a line
15:40 -!- Fish- [~Fish@9fans.fr] has joined #go-nuts
15:40 < aiju> files aren't streams of bytes, they are streams of lines
15:41 < Namegduf> No they aren't
15:41 < aiju> UNIX got rid of that in favour of a more simple model
15:41 < Namegduf> Text files might be streams of lines
15:41 < Namegduf> But everything can be anything at all
15:41 < aiju> Namegduf: i'm telling you the situation on non-Unices
15:41 < aiju> another decision in UNIX was to allow only one stream per file
15:42 < aiju> OS X reintroduced multiple streams and hence i expect them to
provide records, sooner or later ;P
15:42 -!- zimsim [~simon@87.72.77.195] has joined #go-nuts
15:42 < Namegduf> Ah.
15:49 -!- boscop_ [~boscop@g230087148.adsl.alicedsl.de] has joined #go-nuts
15:51 -!- boscop [~boscop@g230087148.adsl.alicedsl.de] has quit [Ping timeout: 240
seconds]
15:52 -!- boscop__ [~boscop@g230087148.adsl.alicedsl.de] has joined #go-nuts
15:53 -!- nettok [~quassel@200.119.188.116] has joined #go-nuts
15:54 -!- boscop_ [~boscop@g230087148.adsl.alicedsl.de] has quit [Ping timeout:
264 seconds]
15:55 -!- Fish- [~Fish@9fans.fr] has quit [Ping timeout: 250 seconds]
15:55 -!- skejoe [~skejoe@188.114.142.162] has joined #go-nuts
15:56 -!- jrslepak [~jrslepak@24-247-40-178.dhcp.mrqt.mi.charter.com] has joined
#go-nuts
15:56 -!- DerHorst [~Horst@e176098106.adsl.alicedsl.de] has joined #go-nuts
15:56 -!- jrslepak [~jrslepak@24-247-40-178.dhcp.mrqt.mi.charter.com] has quit
[Client Quit]
15:58 -!- boscop__ [~boscop@g230087148.adsl.alicedsl.de] has quit [Read error:
Connection reset by peer]
16:03 -!- jrslepak [~jrslepak@24-247-40-178.dhcp.mrqt.mi.charter.com] has joined
#go-nuts
16:05 -!- jrslepak [~jrslepak@24-247-40-178.dhcp.mrqt.mi.charter.com] has quit
[Client Quit]
16:06 -!- boscop [~boscop@g230087148.adsl.alicedsl.de] has joined #go-nuts
16:07 * aiju found a really nice use of channels
16:07 < aiju> iterating through the results of a recursive process like
calculating all permutations
16:08 -!- boscop [~boscop@g230087148.adsl.alicedsl.de] has quit [Read error:
Connection reset by peer]
16:09 -!- XenoPhoenix [~Xeno@cpc13-aztw24-2-0-cust23.aztw.cable.virginmedia.com]
has quit [Read error: Connection reset by peer]
16:09 < exch> like so?
https://github.com/jteeuwen/calculon/blob/master/mods/permute/word.go#L103
16:09 -!- XenoPhoenix [~Xeno@cpc13-aztw24-2-0-cust23.aztw.cable.virginmedia.com]
has joined #go-nuts
16:09 < exch> it's not recursive though
16:09 -!- boscop [~boscop@g230087148.adsl.alicedsl.de] has joined #go-nuts
16:10 -!- Fish- [~Fish@9fans.fr] has joined #go-nuts
16:10 < aiju> i have no clue how that one works ;P
16:10 < exch> black magic
16:11 < aiju> i calculated all cycle structures recursively
16:11 < aiju> and every new result is pushed down a channel
16:11 < exch> yea mine does that to
16:11 < aiju> http://aiju.phicode.de/up/a.html
16:11 < aiju> this table was calculated at basically constant memory
16:12 < skelterjohn> i have no idea what that is
16:12 < aiju> math
16:12 < exch> factorials galore
16:13 < skelterjohn> i'm ok at math, but your data isn't well labeled
16:13 < aiju> the numbers are only remotely related to factorials
16:13 < aiju> skelterjohn: "well labelled" would require several pages of
introduction
16:14 < aiju> it lists all cycle structures occuring in S35
16:14 < aiju> with their counts, order and parity
16:14 < skelterjohn> what's S35
16:14 < aiju> the 35th symmetric group
16:14 -!- boscop [~boscop@g230087148.adsl.alicedsl.de] has quit [Ping timeout: 240
seconds]
16:15 -!- boscop [~boscop@g230087148.adsl.alicedsl.de] has joined #go-nuts
16:20 -!- boscop [~boscop@g230087148.adsl.alicedsl.de] has quit [Ping timeout: 250
seconds]
16:22 -!- femtoo [~femto@95-89-249-242-dynip.superkabel.de] has joined #go-nuts
16:27 -!- nettok [~quassel@200.119.188.116] has quit [Ping timeout: 276 seconds]
16:29 -!- zimsim [~simon@87.72.77.195] has quit [Remote host closed the
connection]
16:47 -!- xyproto [~alexander@cm-84.208.177.43.getinternet.no] has quit [Quit:
WeeChat 0.3.5-dev]
16:48 -!- xyproto [~alexander@77.40.159.131] has joined #go-nuts
16:49 -!- toyoshim_ [~toyoshim@y099101.dynamic.ppp.asahi-net.or.jp] has joined
#go-nuts
16:51 -!- ronnyy [~quassel@p4FF1C0A2.dip0.t-ipconnect.de] has joined #go-nuts
16:56 -!- jumzi [~none@c-89-233-234-125.cust.bredband2.com] has quit [Remote host
closed the connection]
17:01 -!- bortzmeyer [~stephane@2a01:e35:8bd9:8bb0:d963:394:1870:b27d] has joined
#go-nuts
17:09 -!- ronnyy [~quassel@p4FF1C0A2.dip0.t-ipconnect.de] has quit [Quit:
http://quassel-irc.org - Chat comfortably.  Anywhere.]
17:10 -!- tvw [~tv@e176000013.adsl.alicedsl.de] has quit [Read error: Connection
reset by peer]
17:16 -!- nsf [~nsf@jiss.convex.ru] has quit [Quit: WeeChat 0.3.4]
17:30 -!- bugQ [~bug@c-67-171-127-76.hsd1.ut.comcast.net] has joined #go-nuts
17:43 -!- pphalen [~pphalen@66.92.11.149] has joined #go-nuts
17:57 -!- zozoR [~Morten@56344e3e.rev.stofanet.dk] has joined #go-nuts
18:01 -!- bpalmer` [~user@adsl-69-232-221-158.dsl.pltn13.pacbell.net] has quit
[Ping timeout: 252 seconds]
18:01 -!- bpalmer` [~user@adsl-69-232-221-158.dsl.pltn13.pacbell.net] has joined
#go-nuts
18:09 < thebarberian> str1ngs: Deleting the objdir and running it again
helped
18:09 -!- femtoo [~femto@95-89-249-242-dynip.superkabel.de] has quit [Read error:
Connection reset by peer]
18:10 -!- TheMue [~TheMue@p5DDF60EA.dip.t-dialin.net] has joined #go-nuts
18:12 -!- |Craig| [~|Craig|@panda3d/entropy] has joined #go-nuts
18:23 -!- Project_2501 [~Marvin@dynamic-adsl-94-36-166-222.clienti.tiscali.it] has
joined #go-nuts
18:25 -!- Project-2501 [~Marvin@82.84.70.6] has quit [Ping timeout: 250 seconds]
18:27 -!- ccc_ [62cb062b@gateway/web/freenode/ip.98.203.6.43] has quit [Quit: Page
closed]
18:32 -!- bugQ [~bug@c-67-171-127-76.hsd1.ut.comcast.net] has quit [Ping timeout:
248 seconds]
18:46 -!- littlebobby [~bob@unaffiliated/littlebobby] has joined #go-nuts
18:53 -!- rlab [~Miranda@91.200.158.34] has quit [Read error: Connection reset by
peer]
19:04 < str1ngs> thebarberian: good stuff
19:08 -!- Tuller [~tuller@c-68-33-63-208.hsd1.va.comcast.net] has joined #go-nuts
19:09 -!- rlab [~Miranda@91.200.158.34] has joined #go-nuts
19:11 -!- zozoR [~Morten@56344e3e.rev.stofanet.dk] has quit [Remote host closed
the connection]
19:13 -!- foocraft [~dsc@89.211.195.128] has joined #go-nuts
19:15 -!- zimsim [~simon@87.72.77.195] has joined #go-nuts
19:19 -!- lmoura [~lauromour@187.113.121.19] has quit [Read error: Connection
reset by peer]
19:20 -!- lmoura [~lauromour@187.113.125.71] has joined #go-nuts
19:26 < steven> does Go give us control over endianness?
19:26 -!- toyoshim_ [~toyoshim@y099101.dynamic.ppp.asahi-net.or.jp] has quit
[Remote host closed the connection]
19:29 -!- gr0gmint [~quassel@87.61.162.99] has quit [Read error: Connection reset
by peer]
19:29 < str1ngs> ww: you need via its written in go :P
19:30 -!- nettok [~quassel@200.119.151.202] has joined #go-nuts
19:33 -!- skejoe [~skejoe@188.114.142.162] has quit [Quit: Lost terminal]
19:34 -!- m4dh4tt3r [~Adium@c-69-181-223-245.hsd1.ca.comcast.net] has joined
#go-nuts
19:34 * ww looks at vIa just now
19:35 < str1ngs> ww: still alot of work to be donw
19:35 < str1ngs> done*
19:35 < steven> HELLO HOW ARE YOU DOING GOOD IM GLAD TO HEAR IT
19:36 < str1ngs> oh hi steven :P
19:38 < steven> :)
19:38 < steven> bbl
19:40 -!- femtoo [~femto@95-89-249-242-dynip.superkabel.de] has joined #go-nuts
20:02 < ww> str1ngs: where is via again?
20:02 < str1ngs> ww: github..  not very useable
20:03 < str1ngs> ww: but its alot like homebrew.  would need some hacking.
I plan on abstracing the OS ARCH stuff just not done yet
20:05 < str1ngs> ww: if you want to look at it give me some idea's might be
helpful
20:10 -!- ljs
[~user@customer4523.pool1.Croydon-GLN2000-BAS0001.orangehomedsl.co.uk] has joined
#go-nuts
20:11 < steven> does Go give us control over endianness?
20:15 < str1ngs> ww: what problems are you haveing with homebrew?
20:16 -!- araujo [~araujo@gentoo/developer/araujo] has quit [Quit: Leaving]
20:17 < skelterjohn> steven: what do you mean "control"
20:17 < skelterjohn> it uses whichever endianness the architecture uses
20:17 < skelterjohn> though when you encode for network transfer, you can
choose
20:18 < ww> oh, homebrew built python differently, anydbm using gdbm instead
of bdb, they're not quite equivalent, different locking behaviour.  some
irritating loss of a time sorting it out...  not very go related...
20:18 -!- Scorchin [~Scorchin@host109-156-219-102.range109-156.btcentralplus.com]
has joined #go-nuts
20:19 < str1ngs> hmm have not heard of gdbm before
20:20 -!- araujo [~araujo@gentoo/developer/araujo] has joined #go-nuts
20:20 < str1ngs> I was hopeing to add osx support to via but its ways off.
its kinda linux centric right now
20:21 -!- aho [~nya@fuld-590c7951.pool.mediaWays.net] has joined #go-nuts
20:21 < dforsyth> so in via you have to build all of the plans?
20:22 < ww> hmmm.  so a package description is actually compiled into the
binary?
20:22 < str1ngs> dforsyth: right now yes but I'm adding packaging eventually
right now it installs to a chroot
20:22 < str1ngs> ww: yes which is not optimal but has some benefits
20:23 < dforsyth> it might be interesting to roll a dpkg clone complete with
libdpkg in go
20:23 < ww> one of the more common things i do with package managers is
change the build flags or configure arguments
20:23 < ww> would be harder if that meant i needed tp rebuild the package
manager...
20:23 < str1ngs> dforsyth: I'd like to write a package manager in go that be
great
20:24 < str1ngs> ww: let me push my changes so you can see how I handle
flags
20:24 < dforsyth> i was thinking about rolling freebsd package tools in go
20:24 * ww is more of a bsd person as well
20:24 -!- bortzmeyer [~stephane@2a01:e35:8bd9:8bb0:d963:394:1870:b27d] has quit
[Quit: Leaving.]
20:24 < dforsyth> it wouldnt be very difficult.  maybe ill do it next
weekend if i dont feel like working on my editor
20:25 < str1ngs> ww:
https://github.com/str1ngs/via/blob/devel/plans/ncurses.go
20:26 < str1ngs> ww: I have default gnuflags ie --prefix.  but you can also
add your own
20:27 -!- Tuller_ [~tuller@c-68-33-63-208.hsd1.va.comcast.net] has joined #go-nuts
20:27 -!- nsf [~nsf@jiss.convex.ru] has joined #go-nuts
20:28 < str1ngs> I'm worred though that compiling the meta data is not very
scaleable
20:28 -!- krutcha1 [~krutcha@S010600045a27676a.vs.shawcable.net] has joined
#go-nuts
20:30 -!- Tuller__ [~tuller@c-68-33-63-208.hsd1.va.comcast.net] has joined
#go-nuts
20:30 -!- Tuller [~tuller@c-68-33-63-208.hsd1.va.comcast.net] has quit [Ping
timeout: 246 seconds]
20:31 -!- femtoo [~femto@95-89-249-242-dynip.superkabel.de] has quit [Read error:
Connection reset by peer]
20:32 -!- bpalmer` [~user@adsl-69-232-221-158.dsl.pltn13.pacbell.net] has quit
[Ping timeout: 264 seconds]
20:32 < ww> would tend to thnk that metadata should live elsewhere
20:32 < str1ngs> if I can figure out how to attach hooks and logic
20:32 < str1ngs> there is so many corner cases that you have to have some
logic in the meta data
20:33 -!- Tuller_ [~tuller@c-68-33-63-208.hsd1.va.comcast.net] has quit [Ping
timeout: 252 seconds]
20:33 < ww> that's why makefile based systems like ports are nice ...
20:33 < str1ngs> I agree but they have there limitations to
20:33 < ww> also provides clear separation between build system and package
management
20:34 < str1ngs> this wont be a package manager
20:34 < str1ngs> if I use something it will be like pacman
20:35 < ww> i see
20:35 < str1ngs> with this all you need is make and gcc.  nothing else.  and
away you go
20:35 < str1ngs> no I'm not saying this is perfect.  its has aways to go.
it more proof on concept right now
20:37 < str1ngs> if you can think of a meta format that is scriptable or I
can attach hooks to . that might help
20:37 -!- krutcha1 [~krutcha@S010600045a27676a.vs.shawcable.net] has quit [Quit:
Leaving]
20:38 < ww> so one way you could not build in all the build instructions is
to make plans a bdb
20:38 -!- foocraft [~dsc@89.211.195.128] has quit [Quit: Leaving]
20:38 < ww> use gob and you have equivalent to python's shelve
20:38 < str1ngs> can gob handle functions though?
20:39 < ww> not directly
20:39 < str1ngs> and I"m worried about a db format.  since you need an
interface with it
20:39 < ww> but if you make methods on Plan...
20:39 < ww> could get ugly quickly though
20:39 < str1ngs> https://github.com/str1ngs/via/blob/master/plans/lsof.go
here is a corner case for example
20:40 < str1ngs> this will change to . since I"m making helper functions to
make corner cases easier to handly stuff like func ln()
20:40 < ww> right.  so you'd have to resort to some registry of special
functions...
20:41 < str1ngs> its not often most time's they fall under gnuBuild
20:41 -!- bpalmer` [~user@adsl-69-232-221-158.dsl.pltn13.pacbell.net] has joined
#go-nuts
20:42 < ww> you should have a plan for go
20:42 < str1ngs> I had one.  but I had to take it out.  because I need to
work on the gitDownload and hgDownload
20:42 < str1ngs> plus hgStage gitStage
20:43 -!- nighty__ [~nighty@tin51-1-82-226-147-104.fbx.proxad.net] has joined
#go-nuts
20:43 -!- st-5426 [~st-5426@a89-154-147-132.cpe.netcabo.pt] has joined #go-nuts
20:44 < str1ngs> also maps is not a good idea.  map keys are not unique lol
20:45 < str1ngs> well you change the value of one
20:45 < str1ngs> I guess I can just use a function and check for existing
key
20:46 -!- fabled [~fabled@83.145.235.194] has joined #go-nuts
20:49 -!- krutcha1 [~krutcha@S010600045a27676a.vs.shawcable.net] has joined
#go-nuts
20:53 -!- virtualsue [~chatzilla@nat/cisco/x-btxvoreqdosakzsd] has joined #go-nuts
20:54 -!- ildorn [~ildorn@dslb-188-099-199-131.pools.arcor-ip.net] has joined
#go-nuts
20:58 -!- ljs
[~user@customer4523.pool1.Croydon-GLN2000-BAS0001.orangehomedsl.co.uk] has left
#go-nuts ["ERC Version 5.3 (IRC client for Emacs)"]
21:01 -!- ildorn [~ildorn@dslb-188-099-199-131.pools.arcor-ip.net] has quit [Quit:
Leaving.]
21:06 < ww> str1ngs: what about buIld depends?
21:08 < str1ngs> ww: have not started on that yet
21:09 -!- rtharper [~tomh@unaffiliated/sioraiocht] has quit [Read error:
Connection reset by peer]
21:09 -!- rtharper [~tomh@unaffiliated/sioraiocht] has joined #go-nuts
21:09 < str1ngs> my plan so far was to use a shallow direct dependancy list
of some kind
21:09 -!- crazy2be [~crazy2be@d209-89-248-73.abhsia.telus.net] has joined #go-nuts
21:10 < crazy2be> what is the format of envv?
21:10 < crazy2be> used when you call exec.Run or os.StartProcess
21:10 < str1ngs> []string iirc
21:11 < str1ngs> let me check though
21:11 < crazy2be> yeah but what does each string look like
21:11 < crazy2be> like []string{"FOO=BAR", "BLAR=FOO"}?
21:11 < str1ngs> /usr/bin/ls, []string{"ls","-l"}
21:12 < str1ngs> you need to pass the full binary PATH in then the
arguements usually with the binary name
21:12 < ww> crazy2be: yes lIke that
21:12 < crazy2be> ww: ok, thanks!
21:12 -!- TheMue [~TheMue@p5DDF60EA.dip.t-dialin.net] has quit [Quit: TheMue]
21:12 < crazy2be> any reason they don't use a map[string]string?
21:13 < str1ngs> because arguments are not alwasy foo=bar see my ls example
21:13 < crazy2be> i mean for the enviroment
21:13 < crazy2be> argv is the args, envv is the enviroment :P
21:13 < str1ngs> gah sorry confusing the two
21:14 < ww> there is os.Getenv which is kind of like environ[key]
21:14 < ww> not sure why it isn't just a map...
21:14 < ww> it directly mirrors what the system call wants...
21:14 < str1ngs> you can use os.Environ() to inherit the existing enviroment
21:14 < str1ngs> most times thats all you need
21:15 < ww> i wonder what happens if you put something not of the form k=v
in the environment...
21:16 < ww> it probably gets passed around happily...
21:16 < str1ngs> I normal use os.Setenv its more strict
21:16 < str1ngs> even more so if you use os.Clearenv
21:16 < ww> i think k=v is more of a widespread convention
21:16 < ww> but not entailed by the semantics of fork etc
21:17 < crazy2be> ww: If you don't do it, then what is the value of the
variables?
21:17 < crazy2be> just that they are set?
21:17 < ww> they aren't variables, just strings
21:17 < ww> you have a string in the environment i guess
21:17 < ww> you can iterate over the array and see if your string is
there....
21:17 < crazy2be> wierd
21:17 < ww> just speculating...
21:18 < str1ngs> you can try nil which might give 0 enviroment variables
21:18 < crazy2be> but it's usually used as variables
21:18 < ww> crazy2be: yes, it is
21:19 < crazy2be> i do wish i didn't have to do exec.PassThrough,
exec.PassThrough, exec.PassThrough tho
21:20 < crazy2be> it's kinda a pain to specify it three times when i don't
want to customize it
21:20 < str1ngs> I made my own general Exec funtion with formating.  ie
Exec("ls -l %v","/usr/")
21:21 < crazy2be> str1ngs: that would run ls -l /usr/ ?
21:21 < str1ngs> crazy2be: right but.  if you need to pipe and get data back
you need access to the pipes
21:22 < str1ngs> crazy2be:
https://github.com/str1ngs/via/blob/master/fork.go
21:22 < str1ngs> that wont be something you can just plug in and use though.
21:23 < str1ngs> goinstall also has another good example of a gen exec.  see
genRun
21:23 < str1ngs> which inspired me
21:25 -!- Tuller__ [~tuller@c-68-33-63-208.hsd1.va.comcast.net] has quit [Quit:
Computer has gone to sleep.]
21:25 < crazy2be> huh
21:25 < crazy2be> i don't use it enough to justify a function like that
21:25 -!- DerHorst [~Horst@e176098106.adsl.alicedsl.de] has quit [Remote host
closed the connection]
21:25 < crazy2be> but that's cool
21:25 < str1ngs> fair enough.  just trying to be helpful
21:26 < dforsyth> my editor has a file browser \o/
21:26 < crazy2be> just out of curiousity, why not accept a string and have
the caller do fmt.Sprintf("format", args)?
21:26 < crazy2be> i suppose that's what you do internally anyway
21:27 < str1ngs> I guess this way I can format long commands
21:27 < crazy2be> dforsyth: so does mine :) kate is nice
21:27 < dforsyth> mine was also written by me :D
21:28 < crazy2be> is there any way to detect a particular error?
21:28 < crazy2be> it seems that the os.Error interface simply defines
String()
21:28 < str1ngs> some times ie.  err == os.EOF
21:28 -!- tvw [~tv@e176000013.adsl.alicedsl.de] has joined #go-nuts
21:30 < crazy2be> hmm
21:32 < crazy2be> what error corresponds to a non existant file?
21:32 -!- thebarberian [~Master_Co@216.252.79.105] has quit [Read error:
Connection reset by peer]
21:32 < crazy2be> EBADF?
21:33 < str1ngs> check with os.Stat see what it returns
21:34 -!- thebarberian [~Master_Co@222-74-252-216.dsl.colba.net] has joined
#go-nuts
21:34 < crazy2be> str1ngs: Err...  how do i do that?
21:35 < str1ngs> _,err := os.Stat("/bogus/")
21:35 < crazy2be> yes but
21:35 < crazy2be> how do i check what that error code is?
21:35 < str1ngs> fmt.Printf("%T",err)
21:36 < crazy2be> oh it's a os.PathError
21:37 < crazy2be> but that still doesn't help me, it just has path, op, and
error fields
21:37 < dforsyth> os.ReadDir doesnt give you '..'?
21:37 < crazy2be> hmm
21:38 < str1ngs> if err == os.PathError { blah }
21:38 -!- preflex [~preflex@unaffiliated/mauke/bot/preflex] has quit [Ping
timeout: 252 seconds]
21:38 < str1ngs> dforsyth: I dont think so
21:39 < crazy2be> str1ngs: Would be nice, but PathError is a type :/
21:39 < crazy2be> and i could check that it is of that type
21:39 < crazy2be> but that doesn't really help me
21:39 < crazy2be> since os.Stat could return errors of that type for all
kinds of reasons
21:39 < str1ngs> hmm os.EOF is compariable
21:41 < str1ngs> crazy2be: I guess you have to compare .String() or use
general nil
21:41 -!- tvw [~tv@e176000013.adsl.alicedsl.de] has quit [Remote host closed the
connection]
21:41 < crazy2be> :/
21:41 < crazy2be> i'll just assume that all errors are does not exist errors
21:42 < str1ngs> well if its not nil then there is an error handle it either
way
21:43 -!- preflex [~preflex@unaffiliated/mauke/bot/preflex] has joined #go-nuts
21:45 < str1ngs> crazy2be: I think you can type assert with a switch
21:45 < crazy2be> yeah but the type won't help me
21:45 < crazy2be> i don't think
21:46 < crazy2be> hmm but Stat() uses Errno as well
21:46 < str1ngs> crazy2be: blind leading the lind here :P
21:46 < crazy2be> so it might be comparable
21:46 < crazy2be> the question is, to what?
21:46 < str1ngs> I saw that but it passed it of PathError
21:48 -!- photron [~photron@port-92-201-133-223.dynamic.qsc.de] has quit [Ping
timeout: 240 seconds]
21:48 < str1ngs> http://golang.org/pkg/os/#PathError
21:48 -!- GoBIR [~gobir@c-24-130-224-186.hsd1.ca.comcast.net] has joined #go-nuts
21:49 -!- fabled [~fabled@83.145.235.194] has quit [Read error: Operation timed
out]
21:50 < crazy2be> ah ENOENT
21:50 < str1ngs> from Op?
21:51 < dforsyth> wow all the dir readers explicitly skip "." and ".."
21:51 < str1ngs> yes why do you need them?
21:51 < str1ngs> if you need them just add them :P
21:51 < str1ngs> . is always redundant
21:52 < dforsyth> i will need to add them, but it would be convinient to
already have them stat'd from a single ReadDir call
21:52 < dforsyth> its not redundant if i havent stat'd it yet
21:52 < str1ngs> how so though?  9/10 times you have to remove them
21:52 < crazy2be> :/
21:52 < dforsyth> and 1/10 i need it
21:52 < crazy2be> comparing err to os.ENOENT didn't exist
21:52 < dforsyth> so why not let me skip it myself?
21:52 < crazy2be> *didn't work
21:52 < str1ngs> dforsyth: there is also the walker
21:52 < dforsyth> str1ngs: fn name?
21:53 < str1ngs> dforsyth: forget one sec
21:53 < crazy2be> dforsyth: path/filepaths/Walker
21:53 < crazy2be> or something
21:53 < crazy2be> i assume that's what he's talking about
21:53 < crazy2be> er, filepath
21:53 < dforsyth> oh thats like fts
21:54 < str1ngs> find it?
21:54 < dforsyth> yeah
21:54 < str1ngs> dunno if that will help I'm confused why you need .
21:54 < str1ngs> ..  is handy but I"m sure all you need to know if you are
at the root path or not
21:54 < dforsyth> i need ..
21:55 < dforsyth> i dont think either should be removed though
21:55 < str1ngs> sorry if I come of strong I find . and ..  anoying
21:55 < str1ngs> for my shell thats another story
21:56 < crazy2be> it's mostly a political battle...  but . and ..  don't
actually exist
21:56 < crazy2be> they are just created by the system
21:56 < crazy2be> what are you trying to do, anywya?
21:56 < crazy2be> *anyway
21:56 < str1ngs> there useful in a shell for relative movement.  but for
codeing 9/10 times I have to filter it out
21:56 < crazy2be> yeah
21:56 < dforsyth> it would just be convinient to get ..  when i list this
directory
21:56 < dforsyth> i agree its hardly useful
21:57 < dforsyth> but i think that should be my decision
21:57 < dforsyth> system language -> distort the system/
21:57 < str1ngs> ok so stat the parent dir
21:57 < crazy2be> dforsyth: Call the syscall interface yourself?
21:58 < dforsyth> ?
21:58 < str1ngs> dforsyth: parent, child os.Split(thisPath)
21:58 < str1ngs> err path.Split my bad
21:59 < dforsyth> yes, i can stat is myself
21:59 < str1ngs> you need to becareful with that thisPath = / will cause
issues
21:59 < dforsyth> thats not really the issue
22:00 < str1ngs> \o/ I give up :P
22:00 < crazy2be> pathWithDots := make([]string, len(pathWithoutDots)+2)
22:01 < crazy2be> pathWidthDots[len(pathWithDots)-2] = ".."
22:01 < crazy2be> etc
22:01 < dforsyth> why would you hide somethign like this with a system
language though?
22:01 -!- ShadowIce [~pyoro@unaffiliated/shadowice-x841044] has quit [Quit:
Verlassend]
22:01 < str1ngs> its not hiding anything
22:01 < str1ngs> ..  . are not real dir
22:02 < str1ngs> there just realative
22:02 < ww> str1ngs: actually they are real
22:02 < ww> you can infer that from the link count on the directories
22:03 < str1ngs> sure but useful to what end.
22:03 < crazy2be> it all depends on how you define real
22:03 < str1ngs> cd ../ <- useful
22:03 < str1ngs> os.Stat("../") I dont see the point
22:03 -!- wrtp [~rog@92.17.50.183] has quit [Quit: wrtp]
22:03 < crazy2be> i do
22:03 < ww> crazy2be: real meaning present in the directory structure in the
filesystem
22:03 < crazy2be> but getting that in the output would be silly imo
22:04 < crazy2be> ww: Are they actually stored on disk?
22:04 < ww> yes
22:04 < steven> probaby as a doubly linked list or something
22:04 < ww> well, ..  is . isn't
22:04 < str1ngs> lol steven
22:04 -!- fabled [~fabled@83.145.235.194] has joined #go-nuts
22:04 < str1ngs> they need to add ...
22:04 < str1ngs> j/k
22:05 < steven> triply linked list.
22:06 < steven> . == self, ..  == parent, child = child
22:06 < str1ngs> dforsyth: have you tried os.Stat("../" ?
22:06 < dforsyth> no, again stating the directory is not the problem
22:06 < str1ngs> because I'm pretty sure relative paths work
22:06 < str1ngs> ok whats the problem?
22:07 < dforsyth> it would be convinient to me if that entry was already in
the results from os.ReadDir
22:07 < Rennex> yeah, this time
22:07 < dforsyth> its not like i dont know theyre there, or that theyre
unreachable
22:07 < dforsyth> Rennex: yes, this time
22:07 < str1ngs> 9/10 times its not though is my point
22:07 < Rennex> what's convenient for you in one occasion isn't necessarily
the wise thing to implement in a library ;)
22:08 < dforsyth> Rennex: its a system language, this shouldnt be something
that is skipped
22:08 < Rennex> i don't agree
22:08 < dforsyth> okay
22:08 < Rennex> because they are not files in that dir
22:09 < crazy2be> here's why i don't agree
22:09 < Rennex> they are names for that dir and its parent
22:09 < crazy2be> because they always exist
22:09 < crazy2be> they are implicit
22:09 < str1ngs> yes be we all know they exist
22:09 -!- nighty__ [~nighty@tin51-1-82-226-147-104.fbx.proxad.net] has quit [Ping
timeout: 260 seconds]
22:09 < crazy2be> so getting them from a directory listing function has no
use
22:09 < crazy2be> since they are static
22:09 < crazy2be> known at the time of writing
22:10 < crazy2be> returning them in the map would be a waste of memory
22:10 < crazy2be> and a pain most of the time
22:10 < str1ngs> not only that but you have to filter them out
22:10 < dforsyth> theyre getting filtered anyway
22:10 < str1ngs> which would be more work in the long run for people.  then
you adding them to your file browser
22:11 < Rennex> even in a shell, i don't think "." and ".." should ever be
included in a file listing, and especially the pattern ".*" shouldn't match ".."
as well
22:11 < steven> ive seen ..  used a lot in production code.
22:11 < steven> path = dirname(__FILE__) + "/../../../somefile"
22:12 < Rennex> that is a completely different matter though
22:12 < steven> and i have to tell ya, im not a fan.
22:12 < str1ngs> steven: or inclusion type languages that hard to get around
ie ruby
22:13 < str1ngs> if they use absolute paths and you move the app it breaks
22:13 < skelterjohn> what you should do is have a APP_ROOT global variable
22:13 < skelterjohn> and do things like APP_ROOT+"/something else"
22:13 < skelterjohn> rather than ..s
22:13 < str1ngs> aye that cleaner
22:14 < skelterjohn> initialize APP_ROOT correctly at the beginning
22:14 < Rennex> steven: used like that they are doing what they're meant
for, i.e.  to signify "parent dir".  When a file listing says "this dir contains
the file 'foo' and the dir 'bar' and also itself and its parent", anyone would say
"wtf?" if they hadn't gotten used to it :)
22:15 < dforsyth> you're assuming i want this done on the directory my app
is in
22:15 < skelterjohn> just because they're using ".." as intended doesn't
make it a good thing:)
22:16 < Rennex> that's true, they're quite unreadable.  especially when you
have more than one ".." - one i'm just about fine with
22:16 < str1ngs> filepath.Glob and profit
22:17 < str1ngs> Rennex: I was * would match .* thats annoying
22:22 -!- huin [~huin@91.85.185.181] has quit [Quit: leaving]
22:26 -!- dfr|mac [~dfr|work@ool-182e3fca.dyn.optonline.net] has joined #go-nuts
22:30 -!- zimsim [~simon@87.72.77.195] has quit [Remote host closed the
connection]
22:37 -!- kamaji [~kamaji@cpc2-aztw22-2-0-cust775.aztw.cable.virginmedia.com] has
quit [Quit: leaving]
22:37 -!- rlab [~Miranda@91.200.158.34] has quit [Quit: Miranda IM! Smaller,
Faster, Easier.  http://miranda-im.org]
22:41 -!- boscop [~boscop@f055112013.adsl.alicedsl.de] has joined #go-nuts
22:42 -!- ukai [~ukai@nat/google/x-iyfodgmovgxvavgn] has quit [Ping timeout: 240
seconds]
22:45 -!- nsf [~nsf@jiss.convex.ru] has quit [Quit: WeeChat 0.3.4]
22:45 -!- krutcha1 [~krutcha@S010600045a27676a.vs.shawcable.net] has quit [Quit:
Leaving]
22:45 -!- dju__ [dju@fsf/member/dju] has quit [Quit: Quitte]
22:56 -!- Tuller [~tuller@c-68-33-63-208.hsd1.va.comcast.net] has joined #go-nuts
22:59 -!- virtualsue [~chatzilla@nat/cisco/x-btxvoreqdosakzsd] has quit [Ping
timeout: 252 seconds]
23:01 -!- werdan7 [~w7@freenode/staff/wikimedia.werdan7] has quit [Excess Flood]
23:02 -!- Tuller [~tuller@c-68-33-63-208.hsd1.va.comcast.net] has quit [Quit:
muffins]
23:07 < steven> skelterjohn: we have both.  its kidna gross the way it
works.
23:08 -!- rtharper [~tomh@unaffiliated/sioraiocht] has quit [Remote host closed
the connection]
23:08 -!- dfc [~dfc@eth59-167-133-99.static.internode.on.net] has joined #go-nuts
23:09 -!- ukai [~ukai@nat/google/x-txmogypglguivovp] has joined #go-nuts
23:10 -!- werdan7 [~w7@freenode/staff/wikimedia.werdan7] has joined #go-nuts
23:18 -!- Project_2501 [~Marvin@dynamic-adsl-94-36-166-222.clienti.tiscali.it] has
quit [Quit: E se abbasso questa leva che succ...]
23:18 -!- _Eko_ [~eko@c-24-130-224-186.hsd1.ca.comcast.net] has joined #go-nuts
23:19 -!- ExtraSpice [XtraSpice@88.118.35.153] has quit [Ping timeout: 276
seconds]
23:21 -!- GoBIR [~gobir@c-24-130-224-186.hsd1.ca.comcast.net] has quit [Ping
timeout: 276 seconds]
23:21 -!- Eko [~eko@c-24-130-224-186.hsd1.ca.comcast.net] has quit [Ping timeout:
246 seconds]
23:23 -!- GoBIR [~gobir@c-24-130-224-186.hsd1.ca.comcast.net] has joined #go-nuts
23:33 -!- pphalen [~pphalen@66.92.11.149] has left #go-nuts []
23:35 < steven> guys gusy gusy gsuyg usygsu gysu gusy guys guys!!!!
23:35 < steven> !!!!
23:35 < steven> quick!!!!!!!
23:35 < steven> i have a quesiton!!
23:35 < steven> oh wait nm.
23:36 < crazy2be> lol
23:37 < crazy2be> putting that many !!!!!  makes me want to ignore your
question :P
23:41 -!- tensorpudding [~user@99.148.205.193] has joined #go-nuts
23:42 < steven> btw, how do you recommend iterating over a directory's
contents (to only one level, ie without delving into any subdirectories) in Go?
23:43 < crazy2be> os.Readdirnames?
23:43 < dforsyth> this FileInfo has no Name...
23:43 < steven> sure it does.  just try harder.
23:46 -!- boscop [~boscop@f055112013.adsl.alicedsl.de] has quit [Ping timeout: 240
seconds]
23:47 < steven> what the heck!
23:48 < steven> os.Open's second 2 parameters are not documented.  i dont
know what valid values are.
23:48 < crazy2be> what?
23:48 < crazy2be> durp
23:48 < steven> name string, flag int, perm uint32
23:48 < crazy2be> flag is os.O_SOMEFLAG
23:48 < steven> oh
23:48 < crazy2be> perm is standard unix permissions
23:49 < steven> i do NOT like typing 0666 #sigh
23:49 < crazy2be> 0666?
23:49 < crazy2be> why?
23:49 < steven> ...
23:49 < crazy2be> you almost always want 0644, don't you?
23:49 < steven> havent read Revelation, have you?
23:49 < steven> probably.
23:50 < crazy2be> well i would assume you don't like it because it means
it's world-writable
23:50 * steven can never remember how to translate unix permissions into those
octal numbers
23:50 -!- Scorchin [~Scorchin@host109-156-219-102.range109-156.btcentralplus.com]
has quit [Quit: Scorchin]
23:50 < crazy2be> which is enough to give anyone shivvers
23:50 < steven> ha
23:50 < dforsyth> is an empty name ever expected behavior from os.Stat?
23:52 < crazy2be> dforsyth: never happened to me...
23:52 < crazy2be> anyone know if there is an alternative to os/inotify for
macs?
23:53 < steven> yes, there is.  i dont think it's in Go though
23:53 < steven> but it exists in C on mac
23:53 < steven> its called,
23:53 < dforsyth> kqueue
23:53 < dforsyth> ?
23:53 < steven> noooo
23:54 < steven> FSEvents i think
23:54 < crazy2be> hmmm
23:54 < steven> yeah, FSEvents.  ive used it in the past to be notified when
a directory has changed.
23:54 < steven> (when its contents have changed rahter)
23:54 < steven> good C api
23:55 < steven> i think its the same api spotlight `mds` uses to recache for
spotlight
23:55 < steven> (as opposed to kqueue)
--- Log closed Mon Mar 28 00:00:50 2011