** UnhandledExceptionEventHandler :: OFFICIAL LC / CHATTER THREAD **
05-09-2018
, 05:43 PM
Quote:
have you ever had this common mistake wreck absolute havoc at runtime and not understood why?
str += "foo";
this is fine for small strings. but if your string is say, the size of pride and prejudice (which happened to me recently) then it has to destroy and recreate the entire string just to append. you need to use stringbuilder to concat to a large string.
now, every java dev knows this and will avoid this mistake, but it's quite unintuitive that every method that modifies a string actually destroys the entire object and recreates it.
str += "foo";
this is fine for small strings. but if your string is say, the size of pride and prejudice (which happened to me recently) then it has to destroy and recreate the entire string just to append. you need to use stringbuilder to concat to a large string.
now, every java dev knows this and will avoid this mistake, but it's quite unintuitive that every method that modifies a string actually destroys the entire object and recreates it.
Lol, we had a stream process that built up a string to be written to a few diff places. We spent like 2 hours wondering why it ran so slow and it ended up being we could prepend rather than append strings.
I disagree with the wording that its about something being immutable though. Immutability is the bees knees and is a huge issue i have with a lot of java (things like passing a hash map into a function that returns void and modifying the hash map are terrible programatic practices that are encouraged in java)
Re: java 8, i have somewhat limited experience, but cramming a feature in (their "streams" that arent streams) 10 years after every other language (including php) is likely not enough to help me appreciate it.
All hate aside though, you are dead on with the libs. Especially Springs maturity. You can have a full api up in 10 seconds with dependency injection with 0 configuration files
05-09-2018
, 05:46 PM
05-09-2018
, 07:42 PM
Quote:
I just dont like when a languages "best practices" seem to go against the grain for well defined practices in general programming. Mutating stuff goes against the grain for me, and Java very much encourages that. Rails is much worse in that regard. Setting instance variables and randomly mutating them in methods because "thats the rails way". Ugh.
(I know gaming_mouse agrees with you too
)
05-09-2018
, 09:34 PM
05-09-2018
, 09:48 PM
05-09-2018
, 10:37 PM
05-09-2018
, 10:50 PM
Common pattern I dont care for
Code:
class FooService
def initialize
@foo = Foo.new()
end
def method_100_lines_later
# Do something with @foo here, mutating or just reading
end
end
Code:
class FooService
def method_100_lines_later(foo)
# Do something with foo here, mutating or just reading
end
end
And obviously there are considerations for a rule. I tried doing 100% immutable but that was just stupid annoying and slows you down on anything with a slight bit of logic to it.
Im more butthurt by the private or self modifier being on one random line in bigger files than any of this stuff btw
05-09-2018
, 11:26 PM
Quote:
On the extremely rare occasion I want to mutate parameters in C#, I always declare the parameter with the "ref" keyword, which makes a pointer reference pass, even if I don't actually need a pointer reference, just to alert any users of the function that their parameter is not safe from mutation.
05-09-2018
, 11:33 PM
When I say "I always do this", I mean I've probably done it like twice in my career, because there's virtually always a better way to write the code.
Edit: I suppose people might just think they'll be safe making a copy of the pointer before passing, so maybe it's a bad idea. There's unfortunately no good way to deal with it in the language.
Quote:
C++ and "const" for the win
05-09-2018
, 11:41 PM