@# Quotes DB     useful, funny, interesting





Google
 
Web www.quotesdb.info
Undernet  |  EFnet  |  Quakenet  |  Freenode  |  Dalnet  |  Ircnet  |  Galaxynet
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16



Comments:

<0> I suspect -d val does too.
<1> I keep getting confused when to use PyFoo and _PyFoo and Py_Foo
<2> optparse? is that a standard module?
<0> http://www.python.org/doc/2.4.2/ext/backToExample.html
<0> timonator: Yes
<0> Use it
<0> :p
<2> okay
<0> It's awesome-ish
<0> cout: I use the Py_RETURN_NONE macro.
<0> That's why I never saw this.
<0> If you have a C function that returns no useful argument (a function returning void), the corresponding Python function must return None. You need this idiom to do so (which is implemented by the Py_RETURN_NONE macro):
<0> Mention of the macro is the only was this 2.4 doc differs from the 2.3 doc.
<0> I suspect this is done simply to be technically correct.
<0> And that is doesn't really matter.
<0> But, it's in a macro, so I just do it.



<1> cosmint: ah, I didn't know about that
<2> youre right, it IS awesome-ish
<0> parser = OptionParser(usage='usage: %prog [options] srcpath1 srcpath2 ..etc..',version="060126.1")
<0> parser.add_option('-n','--network',action='store_true',dest='network',help='Specify when transferring over network. (default)')
<0> parser.add_option('-l','--local',action='store_false',dest='network',help='Specify when transferring files locally.')
<0> parser.add_option('--broot',dest="broot",help="Specify explicit backup root.",default="k:/")
<3> hi does anyone here know a python chat server / chat client (very simple one), that is _not_ based on any standards like irc, jabber, msn, ect.
<0> kristalino: maybe something in twisted
<3> cosmint, ok, i check that. But anything that is in twisted frightens me a little, i don't know why ;)
<0> It should.
<0> I would write my own.
<1> so any ideas why my extension crashes? http://rafb.net/paste/results/abKYfq60.html
<4> cout, wouldn't the ref count be one even if the data was just freed ? and the mem manager didn't write something over it yet
<1> eblis: I thought the object was freed when the ref count dropped to zero
<0> No
<0> It is marked for freeing
<4> i don't know if python decreases the count to 0 and then check or checks to see if it's 1 and then just frees it knowing it would have been 0 after a --
<0> You don't know when it frees
<1> that's what I meant
<4> *check or checks = checks :)
<1> but it won't free if the refcount is 1 unless it detects a cycle, which I think is impossible with a refcount of 1
<1> this is my first extension, so if I'm getting the details wrong, please let me know
<0> It probably crashes because you are using gtk.
<1> hehe
<0> Seriously.
<0> Toolkits always do weird stuff.
<0> main threads, other threads, main loops...
<5> In metacl***es, should the arguments for __new__ be (cls, name, bases, dct) or (cls, name, bases, attrs)? I've seen both used.
<0> Putting that into a python extension probably takes care.
<0> veracon_: I think dct is attrs.
<0> veracon_: "dict" is in the examples, but it's just a variable NAME.
<5> cosmint, that's what Django does, but a guide to metacl***es I'm reading says dct
<5> (I'm looking at Django for some examples)
<4> is Py_BuildValue("()") faster than PyDict_New() ?
<0> veracon_: Doesn't matter, it's just a name. And "dict" is the name in the Python docs.
<1> cosmint: so I'm not doing anything obviously wrong other than that?
<0> cout: I didn't look.
<5> cosmint, dict will cause a collision (or what you call it), wouldn't it? dct wouldn't
<4> oh, i just saw it's a touple, not a dict
<0> veracon_: No
<0> >>> def foo(dict):
<0> ... print dict
<0> ...
<0> >>> foo('s')
<0> s
<5> cosmint, won't it collide with the dict type?
<5> I see
<0> It's just in the examples.
<0> I wouldn't recommend it.
<0> That's why someone picked dct.
<0> And someone else picked attrs.
<5> I see
<0> Since attrs might be a better variable name.
<5> The style guide didn't say anything on this
<0> As I said, it's just a variable name.
<5> So I ***ume it's not something like self or cls -- that's pretty strict?
<6> it will only hide it inside that function
<0> veracon_: If we are talking about it, then it's probably not recommended.
<5> I see, I will use attrs, thanks :)
<0> I hope django doesn't make people close-minded to my much better web application development solution. :p
<7> It probably will



<0> hehe
<7> Popular things do that. Look at XML.
<0> Yeah, I think I knew that.
<7> I want to dispatch data to different functions depending on the type of the data. Yesterday someone said you could use decorators to do it but I can't think how.
<8> " my much better web application development solution" <- reinventing SQL are we cosmint?
<0> kousu: Ehh, yeah..
<0> I want cl*** decorators already. Anyone know if they are in 2.5?
<0> cout: I'd take out the gtk stuff.
<0> cout: And learn to make your first extension run before adding it back.
<1> cosmint: that's sort of integral to what I need to do
<0> You need to do something else first though.
<0> Make this a skeleton module.
<0> I bet you will find gtk is crashing it.
<1> I've written C extensions for both perl and ruby.
<0> "was" crashing it.
<0> Then learn to add gtk back in, and initialize and bridge it correctly.
<1> I can
<0> 't
<1> I know that taking gtk out will stop if from crashing, because it happens to be crashing in a callback from gdk
<0> Those last 6 lines are, for me, extremely suspicious.
<0> cout: Ok, thanks for telling us.
<9> is there a "python way" to document C code for an extension module?
<4> hmm, don't you have to define a sentinel while enumerating the functions
<0> Though, you don't really know a crashed is really caused by the place it crashes..
<0> Anyway, if you think you know it's gtk, then you simply have to learn to initialize and bridge it into Python correctly.
<0> You may find the relevant information in topics detailing putting gtk in shared libraries.
<0> Or gtk components in shared libraries.
<4> the static PyMethodDef Methods[] ... don't you have to have an entry with NULL, NULL, 0, NULL ?
<0> ie. Not specifically related to Python extensions, but probably directly related to your problem.
<1> I don't understand what you mean. gtk is already in a shared library.
<0> Calling gtk from shared libraries.
<0> Using gtk from shared libraries.
<1> hmm.
<9> Eblis: an entry with just {NULL} works
<0> Maybe you first need to understand why this is even a concern.
<0> It has all to do with threads, main and loops.
<4> bmonty: if in c++, right ? anyway, there's no such entry either
<0> And, inside a dll/so, all perspective on those things changes.
<1> I only have one thread at the moment.
<0> Couple this with a toolkit, which is very particular about such things, and you have the current dilemma.
<9> Eblis: I'm using it with C, and my understanding is that it is required but unfortunately I haven't figured out why yet
<2> is there a more convenient way to get an absolute correct etc path to a folder than os.path.abspath(os.path.expanduser(os.path.expand(......)))?
<7> Can't os.normalize() replace abspath and expanduser and expand?
<2> dunno, ill have a look
<1> cosmint: if your goal is to frustrate me through subtle insults, you are succeeding.
<2> kousu: seems like i dont have that function
<7> Hmm
<7> It must be called something else
<7> I'll look
<2> you mean normpath?
<7> That might be it
<2> seems like that doesnt expand ~...
<2> i wonder if i need that expanding stuff anyways...
<9> timonator: you could create a wrapper to simplify things for yourself
<2> hmm... that might be overkill right now :)
<2> os.path.expanduser(os.path.normpath("~/.ut2004/Cache/..")) <-- this seems to be just fine
<7> How would you detect a recursive structure like a={}; b={}; c={}; a['b']=b; b['c']=c; c['a']=a ?
<10> anyone know the correct/safe/secure way to add a user to the system (linux) from a python script?
<7> I know with OpenBSD there's a OpenBSD.py module that gives access to system services like that
<7> http://docs.python.org/lib/allos.html doesn't list anything like what you want. It might not be in the standard lib
<11> kousu, traverse your data structure and remember objects you see.
<10> what about disabling an account (not deleting, but preventing the user from connecting with ssh,rsync,ftp,etc until renabled)?
<7> You'd edit /etc/users or something
<7> I think if you're just looking for a quick script you can get away with tweaking the files "by hand"
<12> or set the p***word to an invalid hash of the appropriate type
<7> soyt: and then something like "if obj in remembered_objs"? But that only compares with ==, not 'is'.
<7> soyt: would the only way be to loop through the entire list each time and use 'is' myself?
<2> cachedir = os.path.join(os.path.expanduser(os.path.normpath(options['directory'])),"Cache")
<2> AttributeError: Values instance has no attribute '__getitem__'
<2> what the hell? :(
<11> kousu, remember objects ids in a set


Name:

Comments:

Please enter the result of the sum 63 + 46 (to avoid spam):






Return to #python
or
Go to some related logs:

#linux
#math
etherape quit unexpectedly
livinded
ntpd.conf fedora core 4
both webcam and tv-card linux udev
#math
#bash
#linux
ubuntu timestamp too far in future



Home  |  disclaimer  |  contact  |  submit quotes