p-lang thread: (now (have you (problems two)))

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > p-lang thread: (now (have you (problems two)))

«‹›1784 »

Volte: Oct 4, 2004; woosh woosh

getting mercilessly destroyed as a programmer and as a human being is the number two reason i would not want to work on the linux kernel. having to write mission-critical C code is numero uno

# ? Oct 5, 2015 16:37

Adbot: ADBOT LOVES YOU

# ? Jun 8, 2024 20:11

Sinestro: Oct 31, 2010; The perfect day needs the perfect set of wheels.

my feelings are clearly the most important thing

# ? Oct 5, 2015 16:45

Luigi Thirty: Apr 30, 2006; Emergency confection port.

edgy

# ? Oct 5, 2015 16:50

Carthag Tuek: Oct 15, 2005; Tider skal komme,
tider skal henrulle,
sl�gt skal f�lge sl�gters gang

if you think making GBS threads on people creates better work, you should kill yourself

# ? Oct 5, 2015 16:55

Sinestro: Oct 31, 2010; The perfect day needs the perfect set of wheels.

well, my parents were emotionally abusive and i turned out perfect, so clearly

# ? Oct 5, 2015 16:59

Volte: Oct 4, 2004; woosh woosh

has anyone started a coding collective to rewrite the linux kernel in rust yet? i'll make the code of conduct

# ? Oct 5, 2015 17:02

Luigi Thirty: Apr 30, 2006; Emergency confection port.

I didn't quote the bit at the bottom where she says she's going to stay off the internet for the next few days because you know how angry nerds are going to respond

# ? Oct 5, 2015 17:06

Wheany: Mar 17, 2006; Spinya^{ha^{ha^haha}ha}ha_{ha_{ha_haha}ha}ha!; Doctor Rope

Luigi Thirty posted:

I didn't quote the bit at the bottom where she says she's going to stay off the internet for the next few days because you know how angry nerds are going to respond

good call

# ? Oct 5, 2015 17:11

tef: May 30, 2004; -> some l-system crap ->

look, i dunno about you put standing up on the breakfast table and making GBS threads everywhere is a normal process. look, the guys are just blowing off steam, i know it's a bit lovely, but this is how the big boys work in the big world.

we've tried to accommodate your feelings, heck, piss-only-thursdays has been a great success, but we all came here to do make a good product, and you can't make an omelette without cracking a few eggs, and you can't make good software without telling other people that they're wrong, loudly, and scoring points. meritocracy is not a team game.

you've being far too sensitive and you're bringing down morale with your complains and negative attitude. you need to join in the fun, roll with the punches, and poo poo on the breakfast table.

# ? Oct 5, 2015 18:01

tef: May 30, 2004; -> some l-system crap ->

being a jerk doesn't make you good at programming, but you do get to feel awfully smug when you tell people they're wrong

i mean, how else is everyone going to know you're right? i mean you didn't write any documentation

# ? Oct 5, 2015 18:03

gonadic io: Feb 16, 2011; >>=

tef posted:

look, i dunno about you put standing up on the breakfast table and making GBS threads everywhere is a normal process. look, the guys are just blowing off steam, i know it's a bit lovely, but this is how the big boys work in the big world.

we've tried to accommodate your feelings, heck, piss-only-thursdays has been a great success, but we all came here to do make a good product, and you can't make an omelette without cracking a few eggs, and you can't make good software without telling other people that they're wrong, loudly, and scoring points. meritocracy is not a team game.

you've being far too sensitive and you're bringing down morale with your complains and negative attitude. you need to join in the fun, roll with the punches, and poo poo on the breakfast table.

https://www.youtube.com/watch?v=urSubeOywus

# ? Oct 5, 2015 18:06

Blotto Skorzany: Nov 7, 2008; He's a PSoC, loose and runnin'
came the whisper from each lip
And he's here to do some business with
the bad ADC on his chip
bad ADC on his chiiiiip

tef posted:

being a jerk doesn't make you good at programming, but you do get to feel awfully smug when you tell people they're wrong

cf. thomas figg's php talk from some edinburgh programmer meetup

# ? Oct 5, 2015 18:11

tef: May 30, 2004; -> some l-system crap ->

gonadic io posted:

https://www.youtube.com/watch?v=urSubeOywus

https://www.youtube.com/watch?v=co_DNpTMKXk

# ? Oct 5, 2015 18:11

Sapozhnik: Jan 2, 2005; Nap Ghost

# ? Oct 5, 2015 23:44

Bloody: Mar 3, 2013

Mr Dog posted:

.

same, Mr Dog. same

# ? Oct 6, 2015 01:05

suffix: Jul 27, 2013; Wheeee!

i had to concatenate a lot of strings in javascript today

normally with immutable strings you'd either want to use some sort of string builder or save them up in an array for a single concatenation call, so a 'correct' solution might be:

code:

var ButtArray = [];
for (var i = 0; i < 1000000; i++) {
  ButtArray.push("butt");
}
var Butts = ButtArray.join("");

but javascript programmers don't do that poo poo, so browsers engines use lazy concatenation so the naive code is actually faster:

code:

var Butts = "";
for (var i = 0; i < 1000000; i++) {
  Butts += "butt";
}

but there's no language guarantee that's non-quadratic, it just happens to work on the major browsers, except old versions of internet explorer

so you can either do it the 'correct' way and be some hundred milliseconds slower on good browsers, or do it the naive way and take several minutes on ie8 and other rando browsers

or more likely figure out a way not to do this in javascript

# ? Oct 6, 2015 02:55

fritz: Jul 26, 2003

suffix posted:

i had to concatenate a lot of strings in javascript today

normally with immutable strings you'd either want to use some sort of string builder or save them up in an array for a single concatenation call, so a 'correct' solution might be:
code:
var ButtArray = [];
for (var i = 0; i < 1000000; i++) {
  ButtArray.push("butt");
}
var Butts = ButtArray.join("");
but javascript programmers don't do that poo poo, so browsers engines use lazy concatenation so the naive code is actually faster:
code:
var Butts = "";
for (var i = 0; i < 1000000; i++) {
  Butts += "butt";
}
but there's no language guarantee that's non-quadratic, it just happens to work on the major browsers, except old versions of internet explorer

so you can either do it the 'correct' way and be some hundred milliseconds slower on good browsers, or do it the naive way and take several minutes on ie8 and other rando browsers

or more likely figure out a way not to do this in javascript

doesn't JavaScript have promises, do it both ways and use whichever returns first

# ? Oct 6, 2015 03:04

AWWNAW: Dec 30, 2008

heads you lose. tails you lose

# ? Oct 6, 2015 03:23

DONT THREAD ON ME: Oct 1, 2002; by Nyc_Tattoo; Floss Finder

fritz posted:

doesn't JavaScript have promises, do it both ways and use whichever returns first

oooommmmmgggg lol

# ? Oct 6, 2015 03:28

tef: May 30, 2004; -> some l-system crap ->

Blotto Skorzany posted:

cf. thomas figg's php talk from some edinburgh programmer meetup

yep i'm a big jerk

# ? Oct 6, 2015 03:31

Max Facetime: Apr 18, 2009

the trouble with high-level languages is you need to write an AST rewriter for your code to be able to reason about what it will actually do

# ? Oct 6, 2015 03:35

Suspicious Dish: Sep 24, 2011; 2020 is the year of linux on the desktop, bro; Fun Shoe

the reddit peanut gallery is the worst

https://www.reddit.com/r/linux/comments/3nksp0/closing_a_door_the_geekess/cvp0eso?context=1

# ? Oct 6, 2015 03:38

pepito sanchez: Apr 3, 2004; I'm not mexican

Max Facetime posted:

the trouble with high-level languages is you need to write an AST rewriter for your code to be able to reason about what it will actually do

http://asmjs.org/spec/latest/

HEHEHEHEHEHHE

# ? Oct 6, 2015 03:43

sarehu: Apr 20, 2007; (call/cc call/cc)

Suspicious Dish posted:

the reddit peanut gallery is the worst

Reddit posts belong in the HN thread.

# ? Oct 6, 2015 03:44

Suspicious Dish: Sep 24, 2011; 2020 is the year of linux on the desktop, bro; Fun Shoe

sarehu posted:

Reddit posts belong in the HN thread.

why would i read the hn thread

# ? Oct 6, 2015 03:49

Cybernetic Vermin: Apr 18, 2005

string concatenation being loving expensive in java is p. stupid tbqh, only some kind of stockholm syndrome that makes the explicit stringbuilder poo poo seem acceptable

# ? Oct 6, 2015 09:38

Soricidus: Oct 21, 2010; freedom-hating statist shill

Cybernetic Vermin posted:

string concatenation being loving expensive in java is p. stupid tbqh, only some kind of stockholm syndrome that makes the explicit stringbuilder poo poo seem acceptable

this isn't the terrible programmer thread

# ? Oct 6, 2015 09:58

sarehu: Apr 20, 2007; (call/cc call/cc)

Assume we're on a 32-bit platform. Maybe "size" should be a 31-bit unsigned integer type ranging from [0, 2^31). Then, when you subtract two sizes, you get a signed 32-bit integer. There are no implicit integer conversions.

Discuss.

# ? Oct 6, 2015 11:28

Volte: Oct 4, 2004; woosh woosh

sarehu posted:

Assume we're on a 32-bit platform. Maybe "size" should be a 31-bit unsigned integer type ranging from [0, 2^31). Then, when you subtract two sizes, you get a signed 32-bit integer. There are no implicit integer conversions.

Discuss.

Restricting the very notion of a size to be a positive integer seems like an unnecessarily heavy thing to do. It's one thing for a negative size to be invalid semantically, but its a whole other ballgame for it to be invalid syntactically. Negative sizes as arguments to an operation could easily be interpreted as an intention to perform the opposite operation (eg, shrink instead of grow), so why eliminate that option for seemingly no benefit?

edit: also you wouldn't be able to subtract two sizes and then use the result as a size, even though the (absolute) difference between two sizes is still intuitively a size

# ? Oct 6, 2015 12:01

sarehu: Apr 20, 2007; (call/cc call/cc)

Volte posted:

It's one thing for a negative size to be invalid semantically, but its a whole other ballgame for it to be invalid syntactically.

What does this have to do with syntax?

Volte posted:

Negative sizes as arguments to an operation could easily be interpreted as an intention to perform the opposite operation (eg, shrink instead of grow), so why eliminate that option for seemingly no benefit?

For that, you would use the signed 32-bit type that you get when you subtract two sizes. Let's say it's named ptrdiff_t or diff, or ssize_t or something else that isn't int32_t.

Volte posted:

edit: also you wouldn't be able to subtract two sizes and then use the result as a size, even though the (absolute) difference between two sizes is still intuitively a size

You could, by converting the value back to a size.

# ? Oct 6, 2015 12:34

Volte: Oct 4, 2004; woosh woosh

sarehu posted:

What does this have to do with syntax?

A negative size is inexpressible in the syntax of the language.

sarehu posted:

For that, you would use the signed 32-bit type that you get when you subtract two sizes. Let's say it's named ptrdiff_t or diff, or ssize_t or something else that isn't int32_t.

But then you aren't using the size type at all and its existence (or non-existence) is moot.

sarehu posted:

You could, by converting the value back to a size.

If the conversion has to be done by the user and cannot be validated by the type system (i.e., what if you try to convert a negative size?) then I don't see what this new type offers at all, other than a bunch of new syntactic overhead.

# ? Oct 6, 2015 12:51

sarehu: Apr 20, 2007; (call/cc call/cc)

Volte posted:

A negative size is inexpressible in the syntax of the language.

I don't know what this even means. Whether you can make a negative size or not has nothing to do with syntax at all.

Volte posted:

But then you aren't using the size type at all and its existence (or non-existence) is moot.

Yes you are, because values of type size can't be negative, and values of type ptrdiff can be.

Volte posted:

If the conversion has to be done by the user and cannot be validated by the type system (i.e., what if you try to convert a negative size?) then I don't see what this new type offers at all, other than a bunch of new syntactic overhead.

If you try to convert a negative value to a size, you get an error.

# ? Oct 6, 2015 13:39

distortion park: Apr 25, 2011

Cybernetic Vermin posted:

string concatenation being loving expensive in java is p. stupid tbqh, only some kind of stockholm syndrome that makes the explicit stringbuilder poo poo seem acceptable

I thought the '+' operator gets optimised into string builder now

# ? Oct 6, 2015 13:51

Shaggar: Apr 26, 2006

Cybernetic Vermin posted:

string concatenation being loving expensive in java is p. stupid tbqh, only some kind of stockholm syndrome that makes the explicit stringbuilder poo poo seem acceptable

strings are immutable in java so that's why its a thing. but if you want to be lazy just use like java 1.6+ cause the complier will detect some kinds of string concatenation in loops and replace it w/ string builder.

# ? Oct 6, 2015 14:13

Cybernetic Vermin: Apr 18, 2005

Shaggar posted:

strings are immutable in java so that's why its a thing. but if you want to be lazy just use like java 1.6+ cause the complier will detect some kinds of string concatenation in loops and replace it w/ string builder.

it is a pretty weak reason though, just being a bit lazy in general about it would be easy to do and could be done fairly safely, just keeping a collection of strings until some linear operation both motivates and hides the effort of realizing the concatenation. an argument could possibly be made that deferring it is a bit dangerous to legacy software which does a *lot* of concatenations and then performs an indexof/charat in a fantastically performance-sensitive part of the system, which is the only O(1) operation on a string as it stands, since some more work would need to be performed at that point, but such software is effectively already making assumptions about gc so w/e

i rather suspect that this is mostly a matter of not invalidating a bunch of old hotspot optimizations which rely on the internals of the current string implementation (as well as general inertia), but it is pretty clearly a bit of an overly primitive gotcha detail that java otherwise retains rather few of

# ? Oct 6, 2015 14:45

Volte: Oct 4, 2004; woosh woosh

sarehu posted:

I don't know what this even means. Whether you can make a negative size or not has nothing to do with syntax at all.

Yes it does. Writing -1 where a size is expected becomes syntactically invalid. Developers are not given the freedom to define semantics for a negative size because the syntax does not abide it.

sarehu posted:

Yes you are, because values of type size can't be negative, and values of type ptrdiff can be.

Begging the question.

sarehu posted:

If you try to convert a negative value to a size, you get an error.

So in other words, the whole thing is an overreaching runtime assertion that over-defines the meaning of "size".

# ? Oct 6, 2015 15:11

Blotto Skorzany: Nov 7, 2008; He's a PSoC, loose and runnin'
came the whisper from each lip
And he's here to do some business with
the bad ADC on his chip
bad ADC on his chiiiiip

It doesn't overdefine size, negative sizes are absurd and your proposed use for them to reverse operations is a dopey footgun

# ? Oct 6, 2015 15:22

sarehu: Apr 20, 2007; (call/cc call/cc)

Hrm. I was just wondering how overflow-checked subtraction of type (u32, u32) -> i32 could be implemented. It turns out overflow happens if SF != CF (those are the sign flag and the carry flag). Unfortunately there's no condition code for that with Intel.

# ? Oct 6, 2015 15:28

sarehu: Apr 20, 2007; (call/cc call/cc)

Volte posted:

Yes it does. Writing -1 where a size is expected becomes syntactically invalid.

That's perfectly valid syntax. It just won't typecheck or will fail at runtime.

Volte posted:

Begging the question.
So in other words, the whole thing is an overreaching runtime assertion that over-defines the meaning of "size".

This is nonsense posting.

Edit:

Volte posted:

Restricting the very notion of a size to be a positive integer seems like an unnecessarily heavy thing to do. It's one thing for a negative size to be invalid semantically, but its a whole other ballgame for it to be invalid syntactically. Negative sizes as arguments to an operation could easily be interpreted as an intention to perform the opposite operation (eg, shrink instead of grow), so why eliminate that option for seemingly no benefit?

edit: also you wouldn't be able to subtract two sizes and then use the result as a size, even though the (absolute) difference between two sizes is still intuitively a size

Do you even program? Sizes are already unsigned in C or C++ or other languages with size types.

sarehu fucked around with this message at 15:36 on Oct 6, 2015

# ? Oct 6, 2015 15:32

Adbot: ADBOT LOVES YOU

# ? Jun 8, 2024 20:11

Volte: Oct 4, 2004; woosh woosh

Blotto Skorzany posted:

It doesn't overdefine size, negative sizes are absurd and your proposed use for them to reverse operations is a dopey footgun

Depends on the language and the use case. Growing by a negative size being the same as shrinking is hardly a footgun...it's more like the definition of shrinking, or at least an identity. Another use would be to pass the size -1 as a sentinel, which you cannot deny is a common convention in languages that lack tagged unions.

# ? Oct 6, 2015 15:36

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > p-lang thread: (now (have you (problems two)))

«‹›1784 »