|
00:13
stmuk joined
01:48
ilbot3 joined
02:22
lizmat joined
06:12
domidumont joined
06:17
domidumont joined
|
|||
| nwc10 | ianozsvald.com/2016/06/20/results-f...tists-use/ -- interesting. if you read it, please read the "Bias" bit at the end | 07:46 | |
| moritz | if I'd start a new python project, it'd certainly be in 3.4+ | 08:00 | |
| but legacy, legacy, legacy | 08:01 | ||
| nwc10 | me too*. and yes, legacy | ||
| * I wanted something not-perl to validate XML I generated. But I found a command-line thing that seemed to be C, that was already installed | |||
|
08:23
rurban_ joined
08:34
cognominal joined
|
|||
| arnsholt | Yeah, Python 3 is a lot nicer | 09:03 | |
| When I looked at the differences in semantics and such between 2 and 3 when I started Snake, I basically looked at old-style vs. new-style classes and went NOPE! | |||
| All of the nopes, in fact | |||
| timotimo | yeah, never do old-style classes | 09:04 | |
| arnsholt | Yeah, I wanted to do something moderatly clever in a Python 2 project, involving descriptors and couldn't figure out why the descriptors weren't called | 09:05 | |
| Then I realized: it's old-style unless you explicitly inherit from object | |||
|
10:54
domidumont joined
11:36
domidumont joined
11:46
brrt joined
|
|||
| brrt | nwc10: interesting | 11:46 | |
| otoh, descriptors are such a dumb design | 11:47 | ||
| timotimo | what's that again? | ||
| brrt | its... | 11:48 | |
| an object | |||
| which takes the place of an attribute on another object | |||
| and which has three methods: get,set,delete | |||
| timotimo | oh | ||
| brrt | e.g. | ||
| moritz | a layer of indirection over an attribute | 11:49 | |
| much like returning a Proxy in Perl 6 | |||
| python makes heavy use of them | |||
| brrt | class Foo(object): bar = descriptor(); f = foo(); f.bar = 3; (calls f.bar.__setitem__(f.bar, f, 3) | ||
| perl6 has containers, which (imho) is more sane | 11:50 | ||
| timotimo | we also have NativeRefs which basically are those descriptors :P | ||
| those that are such a dumb design | |||
| brrt | not sure i agree | ||
| moritz | the weird thing in python land (IMHO) is that from the outside, you have to know if something is an attribute or a method | ||
| timotimo | maybe :) | ||
| brrt | that what moritz says | 11:51 | |
| moritz | but if you access it as an attribute, there might still be a method hidden behind it (trhough such a descriptor) | ||
| timotimo | right | ||
| NativeRef works in terms of assign and decont | |||
| brrt | on the inside, you somehow have to store something in the object hash, without conflicts pretty please | ||
| on the inside, you don't necessarily know what 'name' your attribute has | |||
| timotimo | hm | 11:52 | |
| brrt | i'm very confident that descriptor are the wrong way to approach the question of 'magic attributes' | ||
| timotimo | magibutes | ||
| brrt | but yeah, proxies are also like that | 11:53 | |
| timotimo | proxies don't let you set a name without mixing in an anon role with a .^name on it, eh? | ||
| brrt | except that you probably create a proxy per object (or per invocation), whereas the descriptor is a class-global instance | 11:54 | |
| timotimo | ah, yeah, the proxy needs to close over the instance you want to access | 11:55 | |
|
11:56
lizmat joined
|
|||
| brrt | \o lizmat | 11:59 | |
| lizmat | brrt o/ | 12:00 | |
| timotimo | <o | 12:03 | |
| jnthn | o/ lizmat | 12:05 | |
| Enjoying YAPC? | |||
| Or TPC as I heard it's now called...there's one less FAQ :) | |||
| nine | New FAQ: "What happened to "YAPC"?" | 12:06 | |
| brrt | and will we be calling it TPC::EU next year? | ||
| jnthn | And if it's TPC::EU, will the UK be allowed to host it again? :P | 12:08 | |
| moritz | well, Kiev wasn't really "EU" either :-) | 12:09 | |
| brrt | we'll host it in switzerland or norway | 12:10 | |
| moritz | in both! | ||
| jnthn | Goodness, the beer prices... :P | ||
| moritz | Poland, maybe? :-) | ||
| closer to me, cheaper to eat and drink | |||
| brrt | we can also do YAMC for a moarvm conference | 12:11 | |
| timotimo | yam yam | 12:12 | |
| jnthn doesn't mention how nice it'd be to have it in Prague at some point, 'cus he knows that's a case of "well volunteered" :P | |||
| brrt | which is just us 5 or 10 guys hanging out | ||
| prague is an excellent idea | |||
| it's settled then | |||
| timotimo | i need to offer brrt hugs | ||
| brrt | YAMC::EU 2016 in prague | ||
| timotimo | this year, already? | ||
| we only have have a year to plan it, absolute-maximum | 12:13 | ||
| brrt | :-D | ||
| moritz | well, a smaller conference needs less planning | 12:14 | |
| also, please spell it MOARCONF | |||
| brrt | awesome | 12:15 | |
| nine | +1 for MOARCONF | 12:16 | |
| timotimo | COANF | ||
| jnthn | heh, a MOARCONF would be fun | ||
| lizmat | jnthn: sorry, was a way for a bit having breakfast | 12:21 | |
| TPC is nice, hope the videos of yesterday will be up today | 12:22 | ||
| the guys were working on it last night already | |||
| MOARCONF is an excellent idea, BTW | 12:23 | ||
| perhaps after Alpine Workshop ? | |||
| nwc10 | I assume it's more likely to be a Moar Unconference (come hackathon) | ||
| lizmat | or maybe *at* the Alpine Workshop ? | ||
| timotimo | i'm not sure we want to pull in the whole UN for this | 12:25 | |
| brrt | when is the APW? | 12:27 | |
| 2-3 september | 12:32 | ||
| hmmmm | |||
| oh well, i can always try to get free time for that :-) | 12:35 | ||
|
13:28
lizmat joined
14:08
domidumont joined
14:46
rurban_ joined
16:01
cognominal joined
16:59
domidumont joined
17:30
lizmat joined
18:57
FROGGS joined
19:01
lizmat joined
19:32
mst___ joined,
japhb joined
20:05
lizmat joined
20:10
lizmat joined
20:38
cognominal joined
21:45
lizmat joined
|
|||