| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Comments:
<0> through the wire
<1> anyway, I'm quitting (only the channel) as the disconnects/reconnects seems to bother some people
<2> phy: you can attach a signal handler for SIGINT
<1> zanella, what could deteriorate more than that?
<2> phy: os.signal(signal.SIGINT, somefunction) or something. help(os.signal), help(signal), man 7 signal
<1> zanella, I'm sshing to my PC now, it looks ok
<3> Jerub: okay, thanks. i'll give that a try.
<4> do I import twisted?
<2> bulio: http://twistedmatrix.com/
<4> oh, I have to dl it?
<2> bulio: unless you're using debian or something that has it included
<2> also, docs are here http://twistedmatrix.com/projects/words/documentation/
<4> ok
<5> Hello all, What would be the best strategy to create a hierarchy objects of the same cl***, where each child may or may not overide their parent's attributes?
<3> Jerub: the SIGINT is only caught when handler is installed by the C extension.
<6> gojab, dynamic mixins with __cl***__ tricks.
<2> phy: huh? no, it should be caught at any time.
<2> gojab: why would you do that? what's your use-case?
<3> example: if you have:
<3> import myext signal.signal(signal.SIGINT, x) myext.blocking_call()
<3> that signal will not be caught after execution enters myext.blocking_call() (at least on OSX)
<2> phy: .. that can't be right.
<3> trying on freebsd now
<5> I'm just thinking but I have an XML file (wurfl.sf.net) that has elements defined with subelements that are different from a referenced parent element - but each element is really the same so I don't really think I can use a subcl***.
<3> Jerub: nope the ext _must_ install it's own signal handler on (osx and freebsd)
<5> Kinda like defining a parent/child relationship in real terms where the generic cl*** is human.
<7> Question: the 'finally: keyword' if there is _no_ exception does this still get executed?
<7> andycat, if there IS an exception it still gets excuted?
<7> er 'and'
<5> hari`: Googling, thanks for the tip
<8> shawn_work, finally executes in both cases
<8> (else, why would you need a third state?)
<7> ok, since im trying to debug a thread race condition in my code :-(
<7> Yango: true
<9> Yango: I though except and finally do not play well together
<4> I can use python irclib
<9> "There are two forms of try statement: try...except and try...finally. These forms cannot be mixed "
<9> http://docs.python.org/ref/try.html
<2> roderyk: until python2.5
<7> this is on Python 2.3
<9> Jerub: well, I'm ***uming shawn_work wants to use it now :)
<2> shawn_work: finally: will always be executed.
<7> you can do a try without an except I ***ume
<2> shawn_work: always always always.
<7> try...finally. These forms cannot be mixed (but they can be nested in each other).
<7> right im ok on that
<8> shawn_work, yup, if you use finally
<8> :)
<7> its just a 'try: .... finally: ...'
<7> ok that's fine
<8> thought that ****s
<8> though
<9> shawn_work: I understood you were doing try....except... except....finally
<9> my mistake
<7> for some reason my threads in python are in:
<7> futex(0x802b2c, FUTEX_WAIT, 1, NULL
<7> ^^^
<7> and i dont know why yet.. :/
<2> shawn_work: that's a lock.
<7> yes
<2> shawn_work: a futex is a really efficient kind of threading lock.
<7> when i create 64 pool theads, and use only 3 of them, I have the rest held
<8> is futex short for fast mutex?
<2> Yango: ask rusty
<7> I think for some reason, those threads that are locked aren't being unlocked somewhere else even if they are not in use
<7> since the constructor initially holds each thread
<8> shawn_work, the most probable reason is a mistake in your locking code :)
<7> yeah, in this case, non-used threads aren't being unlocked properly
<10> morning
<11> hi hi hi
<9> parks: evening ^^
<8> morning? are you in japan, parks?
<7> Yango: ie the parent ID is 31928 and 2 child threads have 31928 and are all in a FUTEX_WAIT
<8> but you don't do any locking call?
<7> Yango: yes
<7> I think this is happening because my queue of threads state never ends
<7> so we have a stack of items in a queue
<7> while self.__tasks != [] <-----
<7> != [] is also != None ?
<7> aka, the list is null/empty
<12> hello
<12> is there any way to find the number of the arguments p***ed to a python program?
<13> len(sys.argv)
<12> :)
<12> but isnt it a list of strings?
<13> yes
<12> so len works with that too.eh?
<13> sure
<13> len() works on any sequence or mapping
<12> thanks
<14> hello... is there a precompied pyopengl for python2.4 on debian ? i've been trying to compile it myself but it's a nightmare...
<15> hi guys
<15> where can i get access to the method called by print?
<15> im trying to ***ign system = print for debugging
<16> jmg: sys.stdout.write
<17> When using __add__ in cl***es, one's function is def __add__(self, other), where self is the first object and other is the second, correct?
<6> Yes.
<17> I've just discovered OO, and I think I like it. :P
<6> And __radd__ is called on the second if there is no __add__ on the first.
<2> phreq: operator overloading isn't OO, it's just operator overloading.
<6> That's not OO. Well, it's (O)perator (O)verloading, but that's not OO.
<17> Oh okay. I was wondering what overloading was.
<17> Thanks.
<6> More generally, overloading is making one thing mean more than one thing based on context.
<17> Second question... I'm trying to write a cl*** for unit manipulation, because I can't find any modules already written for it. Am I overlooking a module tucked away on the net somewhere?
<10> of course
<10> you can overload .__add__ to do whatever you want
<10> roderyk g'day
<10> shawn None and [] are not the same
<18> let's beer with music!
<19> hi, i opened an existing file on windows (python 2.4) with mode 'r+b' - while writing to it all occurrences of hex 0A get preceeded by 0D, which together is the windows linesep combo
<19> does anyone know how to suppress that?
<4> http://rafb.net/paste/results/jsuAOP20.html
<4> anyone know why that doesn't connect?
<20> cmd: use the os module perhaps
<19> thanks, i'll give it a try :)
<18> cmd: does the existing file have \x0d in it?
<10> bulio what error does it give...
<20> os.write() it appears
<4> no error
<17> Am I confused? I thought open(,rb) was binary read?
<4> parks: it just doesn't connect
<10> phreq what platform?
<18> phreq: "r+" is read/write
<10> theres no such thing as "binary" mode unix
<18> parks: there's still "binary" mode, and you should still use it on unix.. it's just that binary mode = text mode on unix
<19> mcmillen: yes it has
<10> bulio does it throw an exception...
<4> parks: nothing
<21> hiya
<4> Enter the real name of the bot: bulio-bot
<4> after that it just waits
<19> mcmillen: i think i got you wrong, its not that i've mistaken an existing 0D for the one located before the 0A, it really gets there on writing
<4> this is annoying
<17> parks, mcmillen: os module w/ Ubutnu Linux is the platform, and thanks mcmillen. I didn't know that.
<10> bulio sure its not stuck in irc.process_for_ever() ?
<17> I love Python.
<10> phreq warts and all?
<17> parks: We're past warty. Now it's the Breezy badger. ;)
<17> And it gets better, next it'll be Dapper Drake. :P
<6> I have to say, the names are the worst part of ubuntu.
<17> Bah, I love 'em.
<10> phreq pyNumeric ?
<17> Computing should be fun.
<10> as for units ... metric baby
<22> Is there a way to have python return a list of available methods for a user defined cl***?
<10> help(cl***)
<17> pyNumeric looks helpful for vectors, at least.
Return to
#python
or
Go to some related
logs:
#awk
mp3 won't play in sterio
#fluxbox
#xorg
turbogear connection pool
ubuntu install kde
emerge-xdm
#perl
mencoder 3g2 sync
#gentoo
|
|