| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Comments:
<0> time to breka out venkman i think
<1> parks: or typeof(obj)
<2> anyone here using the xmpp library?
<0> god i hate ajax
<1> I like javascript, actually, it's a little clunky, but overall quite nice.
<0> i like JS
<0> i hate ajax
<3> does there happen to exist a good python module for dealing with opendocument files, or do I have to do it all by hand with a xml module?
<1> parks: Ajax is a really tiny thing. I'm surprised it's causing you special problems.
<0> Joao i am living that nightmare
<0> CSWookie a particular library
<4> heh that is so silly that people think that web apps can replace all the desktop apps
<1> If browsers didn't **** so very horribly, I'd be ok with that,
<2> right now I'm having problems with the xmpp library, when I send the client.auth(user,p***word) thing, it hangs
<1> But they do.
<4> I have even seen some of the web based office type apps and god they ****!
<4> the problem is browsers can not do everything
<2> but if I'm in ipython and I send a cl.auth, cancel the call and call cl.auth again it works
<1> kosh: I know.
<5> and the more they do, the more they're going to fall over and lose your work.
<1> kosh: And what they can do, they do half-***edly.
<4> people want them to do just about everything possible
<4> CSWookie: it is not really on purpose though
<4> no other framework I can think of is expected to deal with such crap data
<1> Tue.
<1> true, rather.
<4> if you handed the same level of crap to mfc, qt, .net, gtk etc etc it would never compile
<4> much less run
<4> and nobody would expect it to
<5> and javascript isn't just crap, it's astoundingly crap.
<4> so a web browser is supposed to deal with every crap people come up with and still manage to run it the way it was intended by the brain damaged person that did it
<5> any ajax page that doesn't reload a page every 5 minutes is broken.
<0> prehaps im mis understanding something
<1> Javascript isn't bad. It could be much better.
<5> because the js will leak memory all over the floor and explode.
<1> But it could be a lot worse, too.
<5> CSWookie: uh, i'm talking about the implementation, not the sepc.
<1> Ah.
<5> the spec is fine. js is a p***able language and mochikit makes it manageable.
<5> I'm talking about the VM that runs js.
<1> Yeah. JS interpreters should be a lot tighter.
<4> if you push ajax very far in ANY browser you can trivially kill it
<0> lynx
<1> I know.
<4> the problem is that when you make a js interpreter tighter it breaks on more code
<0> anyway, prehaps someone can explain this
<6> mh, no way to get an object by a specific id?
<4> kjs used to be completely strict, they had to loosen it up to run the crap code out there
<4> and with loosening it up you end up with more bugs
<7> Scraper, not by id alone, that I know of.
<6> k thx
<0> [o for o in globals() if id(o) = 'id']
<1> I was so proud of myself when I made it so that my JS did not leak memory all over the place, but at a slow, acceptable rate.
<4> browsers **** because they have to do deal with utter crap
<4> ajax pretty much screws it to an extreme, I have not seen any ajax stuff that does not leak
<0> the only place ive seen JS leak memory all over the floor is with IE and event handlers
<0> apparently it doesnt release its references
<4> ie will leak with xmlhttprequest also
<1> parks: IE gets into cyclic memory problems between the dom and js.
<1> parks: You have to take special precautions to make sure that you never leave any cycles behind.
<0> yeah
<2> I wonder why I have to call an authentication function twice for it to work
<0> Joa tell your users to click harder
<8> I'm trying to write a function that dynamically loads a module and returns a new instance of a cl*** based on information in a database. Code: http://deadbeefbabe.org/paste/3035 (Hopefully all the information needed is there). Only I'm getting an error message 'No frozen submodule named Services.TestServ' on imp.find_module. Can anybody tell me what I'm doing wrong?
<1> But now that I know what those precautions are, I am superman, and event handlers don't **** up memory.
<0> CSWookie yeah, ive gotten around it too
<4> IE and opera have other really nasty issues
<0> opera has come a long way recently
<4> god I want to shoot the person in opera that copied the IE name id confusion
<1> Thankfully we don't support opera for our main project... OH ****
<9> kosh: Opera doesn't do that any more.
<0> i use opera
<4> document.getElementById is wrong in opera and IE
<0> kosh, how?
<1> Our new project is going to be more loose... I wonder what we will support.
<4> sanxiyn: unless they fixed it in a more recent beta then I tested the problem still exists
<9> kosh: IE User-Agent thing?
<4> parks it puts the name and id namespace together on that if you do a request for something with an id of foo if there is something with a name of foo earlier in the document it will grab that object instead
<9> Opera 9 uses Opera/9.0 (Windows; etc).
<4> it can create some very nasty problems
<9> The most sensible.
<4> that problem exists with opera 9 also
<4> this has nothing to do with user agent strings
<4> this is javascript and it is a nasty reported bug and they have not fixed it
<9> kosh: Ah, name/id confusion?
<10> oh well, is it so hard to avoid, though?
<9> I just won't use name attribute... :(
<4> there are times that it is
<9> Except in form of course.
<4> in forms you have to use name and sometimes you want to use that name to control an id somewhere else
<9> Yes, indeed.
<4> so in ie and opera have to do name mangling to achieve that
<0> name="f_formname"
<9> parks: Yes that's the point.
<4> instead since we only use it for backend code what I did instead is just say that only khtml and gecko are supported
<4> screw ie and opera
<0> you can blame document.all['rofflez'] for that
<4> opera has no marketshare, and our customers have no problem using firefox, safari etc since many times they have to use it already for security issues
<4> I could hack it but screw it
<9> parks: They could support legacy document.all while supporting valid getElementById!
<9> Does anyone here understand Python 2.5 "with" syntax?
<4> when you are dealing with lots of autogenerated code that is used in many ways, setting up name mangling for an ie,opera issue can end up being a lot of work
<7> No, but other languages have had 'with' before, it's pretty simple.
<0> sanxiyn i pray its not based on VB's 'with' statement
<9> hari`: Python's "with" is very different.
<11> parks: lisp's
<2> hrm is the xmpp library broken?
<9> hari`: It looks more close to unwind-protect.
<11> parks: or more to say, based on lisp's with-open-file and other with-foo forms
<9> parks: It's completely unlike Pascal's.
<9> sproingie: Which is implemented with unwind-protect IIRC.
<11> i sincerely doubt it's as comprehensive as unwind-protect
<11> it's more like try/finally syntax sugar
<9> It isn't. Just like generator isn't coroutine (and not continuation)
<12> unwind-protect?
<11> *snort* ... probably wouldn't be politic of me to opine about the python language design process
<9> So, the idea seems to be
<9> @contextmanager
<9> def closing(it):
<9> try: yield it
<9> finally: it.close()
<11> Yango: unwind-protect is lisp's version of try/finally ... or as lispers would say, try/finally is other languages versions of unwind-protect
<9> with closing(open(filename)) as f:
<9> # do something with f
<11> http://www.lisp.org/HyperSpec/Body/speope_unwind-protect.html
<12> help(yield) => invalid syntax
<9> sproingie: Or setjmp/longjmp (yuck)
<11> unwind-protect isn't extremely interesting. handler-case is what makes lisp so freakin cool
<11> scheme has dynamic-wind which is also nice, since it works in both directions
<11> probably nothing python needs
<12> i was taught scheme at first year of university, but we were all too young and too amused and annoyed by all those ('s to really learn it
<9> def f():
<9> try:
<9> a = 1
<9> return a
<9> finally:
<9> return a + 1
<9> (Guess what does it do)
<13> interesting
<9> (It returns 2, of course.)
<1> I was never taught a Lisp dialect, and I feel that I have been cheated.
<14> CSWookie: so learn one?
<4> well I would never write that cod and it should probably be fixed somehow in the language
<4> that is just nasty to read
<1> shric: It's part of the plan.
<11> nothing to fix, that's try/finally working correctly
<11> it's definitely not too smart
<11> to write that is
Return to
#python
or
Go to some related
logs:
synaptic ubuntu xfce update
ubuntu jstest
kylixen
Failed to load shadow images ubuntu
debian sis_sata
#gentoo
perl tohex
define scripitng
vertilo
+fglrx +mouse +garbled +ubuntu
|
|