| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Comments:
<0> un-bitpacking the chmod string
<1> Patrick`: depends on what you're doing, and what you want to do.
<2> nay.array(iter([[1,2,3],[4,5,6]])) <- fails
<0> Yhg1s: if a file is world-writeable, panic
<0> I stat() it and then parse that
<1> Patrick`: 'if os.stat(filename).st_mode & 02: ...'
<3> >>> import numarray as nay; nay.array(list(iter([[1,2,3],[4,5,6]])))
<3> array([[1, 2, 3],
<3> [4, 5, 6]])
<0> Yhg1s: is a bitwise and?
<2> right but the reason that i'm constructing an array *directly* from an iterator
<2> is because the Python list takes up too much memory
<1> Patrick`: & is a bitwise and, yes, like it is in most languages.
<3> Connelly: What's your iterator, exactly?
<2> uhm
<1> Pythy: too big, apparently :)
<1> Connelly: I suggest posting to the numarray/numpy/scipy/whatevertheycallitnow lists.
<0> Yhg1s: you'll have to excuse me, I don't understand how that works but I'll look it up
<0> the syntax, I grok binary
<0> ish
<4> anyone here using SPE?
<5> Connelly: I just got in here a few seconds ago, but why dont you just yield up values
<1> Patrick`: it takes 'os', retrieves the 'stat' function, calls that on 'filename', takes the 'st_mode' argument of the result (which is the file's mode as an integer), and AND's that with 02, which evaluates to either 2 or 0, depending on whether bit 2 is set. if bit 2 is set, the file is world-writable.
<1> oneeyedelf1: numarray doesn't consume iterators, it requires an object with a length.
<0> yep, I get everything except "why 02" ?
<1> Patrick`: because that's the 'world writable' bit.
<0> no, no
<0> to my eyes, that will return non-zero only if the last 2 bits are "10"
<0> yet it works for 10 and 11
<2> Pythy: it's an iterator of 1352078 lists of form similar to [range(12),range(12,24)]
<1> Connelly: alternatively, you may be able to wrap your iterator in an object that adds a sensible __len__ method; that requires you know how large the resultset will be, though.
<0> oh, I get it ...
<0> I'm an idiot, disregard me
<0> and I was doing bloody logic gates last week
<1> Patrick`: it's bitwise AND. it returns the bits in the mask that are set.
<2> um, my iter->array code works fine
<0> 01 & 10 = 00, 11 & 10 = 19
<0> *10
<0> and so on
<0> dur dur dur
<2> i'll post it as a recipe
<3> Connelly: The inner lists were created using range()?
<0> but why prepend it with a zero? force of habit (octal, and can test for, say, 0755)
<1> Patrick`: yes.
<2> um, they're regular lists
<2> but they're created from some set partitioning math
<6> whoaaaaa
<6> new website
<6> shinnnnnyyyyy
<7> what is the best resource for learning py from the beginning
<8> ?>tutorials@Zamolxe
<9> $!>cheeseburger_and_fries@Zamolxe
<7> ok ok i got it :))
<9> the cheeseburger?
<10> mmm, McBot
<7> before i'll do some rtfm, can someone please tell me what is the main purpose of py
<7> can i use it like php
<6> yes.
<7> i mean i;ve worked with zope before
<6> unlike php, py doesn't have a main purpose; it is general purpose. however, it comes with batteries included, and there are libraries available for just about any purpose.
<7> i've been into parsing html, crawling and databases. the problem with php is that is not so fast as i espected, and i ran in a lot of seg faults using regular expressions (thou i runned some patches)
<7> is py faster?
<10> Python is almost always fast enough, and there are things you can do for when it's not
<1> maybe. maybe not. it depends on the code you write. also, parsing HTML with large regular expressions is neither fast nor easy.
<1> Python has quite a few excellent HTML parsers, though.
<9> PHP, ew
<8> Zalamander, it's all in the Big Oh. ;-)
<6> Yes, you really should use a real parser to parse HTML. Fortunately, there are plenty to choose from.
<11> beautiful soup is neat :)
<6> I'm looking for a list of built-in decorators, but I can't find it.
<7> davidmccabe i don;t understand, do you mean parsing/processing HTML with regular expressions is bad?
<6> Zamolxe: yes.
<8> davidmccabe, staticmethod and cl***method, basically. Decoraters are just functions, after all.
<7> hm...
<1> with a signle one, yes.
<6> Zamolxe: HTML is complex enough to warrant a lex-based parser.
<6> (well, you know what I mean)
<1> or a parser built ontop of many small regular expressions.
<10> right, BeautifulSoup can be good for "wild" HTML, things like ElementTree are good for XHTML
<7> Yhg1s of course
<6> thanks TFK.
<11> ...and xmltramp for parsing XML :)
<7> thanks guys i apreciate. one more thing, every time i read some paper about web documents, parsing, etc, i hear about python. i heard that at google to
<7> why?
<11> why not?
<1> Zamolxe: I'm not sure what you are asking.
<7> i mean, why not use php or perl
<2> Numeric/numarray/numpy array from iterator:
<2> http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/476217
<1> Zamolxe: because Python is easier to write, easier to get right and easier to extend.
<7> is there something different about python, that i must consider
<7> ok
<12> to put it really vaguely, both PHP and Perl have a lot of nastiness of various kinds that Python seems to have avoided quite well.
<2> Zamolxe: i find that Python is similar to my mental pseudocode
<2> so it's really easy to "think" in Python
<13> Zamolxe: python has been called 'executable pseudocode'
<7> Connelly i see
<7> oh that's great
<14> XSLT is also worth a look if you are doing XML processing - it's not that easy to learn though
<2> i bet some people "think" in Perl
<15> poor guys
<2> :)
<15> I've even heard of some actually thinking in that stuff sendmail is configured with
<8> Hmm. Think in J, communicate in Python, execute in C...
<15> but I live almost next to a mental hospital, so that might be the case...
<16> hi. if i want to check whether a value is a tuple, should i just go with "type(x) == type(())"? or is something more straightforward that i should be using?
<0> python makes itself look nice
<17> How freaking long does it take to get a project renamed on sourceforge. It's been over a week now.
<8> amrit|wrk, ?.isinstance@amrit|wrk
<0> I don't care for sourceforge, they rejected my project request
<0> I was a bit flamey
<2> amrit|wrk: isinstance((1,2,3),tuple)
<10> amrit|wrk, you can use isinstance, but you probably shouldn't
<0> "unlike 90% of the project requests you're likely to approve today, I have actual files ready!"
<17> Patrick`, I know. it's sick
<8> Why do I keep mentioning the asker's nicks when using the bot? :-/
<0> and "this program exists because all other alternatives are wank"
<0> "please write a 10,000 word egowank about your project"
<0> so, yes, I'm going to try for freshmeat
<17> Patrick`, What about berlios.de ?
<16> TFK: comcor benji : thanks.. but why shouldn't i use it?
<0> exonic: I'm not german
<0> (this is the first time I've heard of it)
<17> Patrick`,You have to be german to use it?
<0> I have no idea
<8> amrit|wrk, type-checking is usually bad practice.
<0> but it's .de and that's all I know about it
<18> what about launchpad (no idea myself)
<17> Patrick`, I don't think u do.
<0> exonic: exactly
<8> amrit|wrk, depends on what you need it for. You are welcome to elaborate :-)
<16> TFK: ah. well, i'm writing some stuff where i want functions to have a nice and clean design in general - they just return a bool. but in a few special cases, i want to be able to return diagnostic messages as well, along with the boolean. the functions are only called from one place, so i was just going to have that spot check whether the value returned is a tuple, and if so, handle it appropriately.
<8> amrit|wrk, either use exceptions, or ALWAYS return a tuple of (result, err_code)
<1> a nice and clean design doesn't return different types of things from the same function.
<8> result, err_code = myfunc(); if err_code == OK: ...; else: ...
<1> exceptions are preferable.
<16> yeah, i suppose you're right
<16> thanks
<19> Ok this is a little more complicated than my situation but let's say I have an if loop and it will do something if either of the cl***'s two variables is greater than ten.
<19> However not all of the cl*** instances will have the second variable.
<19> How would I set it up?
<10> eugman, that doesn't sound like a very good design, perhaps if the second variable doesn't apply it can be set to None instead?
<20> eugman: You could use getattr with a default value
<21> eugman: create a method ab_more_than_ten () and use a try
<10> then you can check if foo is not None and foo > 10:
<10> s/foo/self.foo/g
<17> eugman, program state!
<17> heh
<19> Benji's seems to be the smartest thing to do. thanks for the help.
Return to
#python
or
Go to some related
logs:
Metadata file does not match checksum yumex
kde button Malformed URL system:/
#physics
NX host not found
lvm create RAID1 fedora
fluxbox overlay snap
#php
invalid module format i810
java-void(0)
qemu xscale
|
|