Tavern on the Meta - 2011-03-04 (page 1 of 5)

« first day (212 days earlier) ← previous day next day → last day (4803 days later) »

12:01 AM

I have reason to suspect that Rebecca Chernoff is actually TheTXI ...

4

The OutOfStars Exception

@TimPost We're all sockpuppets of Beckers

@TimPost That raises even more questions than it answers!

I thought Rebecca was a sockpuppet of me, who was a sockpuppet of @TimStone

The OutOfStars Exception

@GraceNote Shhhhh. We don't talk about @TimStone like that. He'll replace us all with very small javascripts

@Shog9 None that "Las Vegas" doesn't sufficiently answer.

The OutOfStars Exception

12:04 AM

It might take a few years, but... ;-)

heard he was being talked about

Rebecca Chernoff

@TimPost oh no

Hey - This is @TimPost - I'm in a Net Cafe somewhere in Saigon and I think my OpenID might have been compromised. Has anyone impersonating me said anything lately?

@TimPost Nope, you're cool - back to your game.

@GraceNote That seems like a reasonable sockpuppet hierarchy.

Rebecca Chernoff

12:09 AM

@TimPost we should probably remove that diamond then, eh? (;

@RebeccaChernoff I'll hang on to it for him. We almost have the same name, it'll be less unsettling for people that way.

Wow. Chat is pretty broken on Honeycomb...

Rebecca Chernoff

yeah yeah of course you would @TimStone

Just thinking of the welfare of the community, of course.

Responsible... adding that trait...

12:22 AM

@drachenstern Apple uses Long Long

i->can_haz->threadlock();

I am so sick of maintaining this code that I'm determined to at least make it funny

@TimPost Yeah, that prompted me to use "thingadongdong" in a test case one time.

I'm turning C into lolcode

The happiest of all the codez.

The person I'm supposed to be working with has apparently been over-sleeping for the last 4 1/2 hours. This is what happens when I don't have much else to do.

12:30 AM

@Moshe what comes after int?

Ah, sleep...I'm considering going back to my nap for a bit, heh. Then I think it's time for a little userscripting.

The OutOfStars Exception

@TimStone \o/

@TimStone You should craft a bed.

@GraceNote One of the best new items in Minecraft ;)

Seems like it

12:32 AM

gets three woods and three wools and a workbench

12:57 AM

If you are in this room and haven't read this article Falsehoods Programmers Believe About Names can you tell me why not? Will you rectify that before asap?

badp's sockpuppet

1:09 AM

Out of curiosity

Do web sockets really allow arbitrary communication with any IP and port?

So much for cross-domain restrictions...

@drachenstern That list seems written by a jerk

@badpssockpuppet it probably was :p

badp's sockpuppet

Examples, examples! I require examples.

What culture is it having that problem?

etc.

lol

@drachenstern In iOS?

Long and then LongLong

@Moshe yeah, then I went and looked up a reference online for objc

1:18 AM

@drachenstern heh

yeah, and long is usually numbytes(int)*2 bytes in length

I'm really bad at that apparently.

1 hour later....

Also, it's your Turn WwF

In the Game Kit API we report scores with LongLong

@drachenstern Indeed

whoa, tie game

I realized after playing that I should've played a stronger word, but this should leave me some room to work too

1:20 AM

How do you spell Buddha?

Er, why is that not valid?

Like that

Siddhārtha Gautama (Sanskrit: ; Pali: Siddhattha Gotama) was a spiritual teacher who founded Buddhism. In most Buddhist traditions, he is regarded as the Supreme Buddha (P. sammāsambuddha, S. samyaksaṃbuddha) of our age, "Buddha" meaning "awakened one" or "the enlightened one." The time of his birth and death are uncertain: most early 20th-century historians dated his lifetime as c. 563 BCE to 483 BCE, L. S. Cousins (1996), "[http://indology.info/papers/cousins The dating of the historical Buddha]: a review article", Journal of the Royal Asiatic Society (3)6(1): 57–63. but more recen...

Because it's a name

Yes, it's a title to most people

ugh

well, you know 6/7 of my tiles now

but generally in English when we say Buddha we mean the actual person so it's considered a proper noun

or at least, that's my general understanding

heh

ok

Can you use that triple tile now?

I'm out, cya

Actually... hang on

@drachenstern - Did I tell you how many copies of Nippon went yesterday?

@Moshe no you didn't

1:25 AM

@drachenstern It was featured on Free App Report. So, several hundred.

Look at the Game Center numbers

40 some-odd players

It was free, but it's good exposure

Interesting.

Congrats

Thank you.

@EduardoMolteni I'm gonna assume you don't sweat the small stuff

Going to push a beta build soon.

No time to fix the Long bug now, but let me look into it real quick

Wow, tough crowd ...

Figured he would be quick with a:

> What, me worry?

1:31 AM

er, what is a LongLong in NSInteger-land\?

looks at docs

/facepalm

Longer than a ShortLong, shorter than a LongerLong

user image

@TimPost yes, you deserve something for that. I don't think it's a star tho

@drachenstern There is a strong correlation between me clicking that link and my computer crashing.

1:32 AM

@mootinator there should not be, but ok

It's down to 0.5 now.

0

Q: How should I declare a long in Objective-C? Is NSInteger appropriate?

I see NSInteger is used quite often and the typedef for it on the iPhone is a long, so technically I could use it when I am expect int(64) values. But should I be more explicit and use something like int64_t or long directly? What would be the downside of just using long?

iphone objective-c

0

Q: Convert int64_t to NSInteger

Hi all, How can i convert int64_t to NSInteger in Objective-C ? This method returns into score an int64_t* and I need to convert it to NSInteger: [OFHighScoreService getPreviousHighScoreLocal:score forLeaderboard:leaderboardId]; Thank you.

objective-c convert cast nsinteger

@mootinator I'm glad to hear that

Right, int64t

Which probably means rewriting a fair dinkum chunk of code

@drachenstern Just a bunch of replacements

1:35 AM

@Moshe and more testing

Before you go much further, can I suggest a proposal?

It's [NSNumber setInteger:"" forKey:@""] becomes [NSNumber setLongLong for key:]

What's the maximum expected cartsize? 512?

@drachenstern Yea, you want me to cap it?

What's the maximum expected sell price/cost of SZ?

@drachenstern sure

1:36 AM

25,000,000?

@Moshe no

Keep in mind these are algo's you've written that you can manipulate or artificially limit.

(Source code removed)

So what's the maximum number of times you can buy/sell something at 512 cart size?

@Moshe that's obviously not right if I'm selling it at 16mil+

I've changed it.

@drachenstern It was 16mil+ before

It takes like 8 upgrades to get to the max cart size

Put it back, that wasn't a terrible thing

Right, so...

1:38 AM

it only hits it like 4 times a game, on average

so you figure what's 30 - (8/2) + 1 * 16mil * 512

@Moshe ineffective edit

@drachenstern heh, but humerous

True

You now need to be able to store a minimum maximum score of 221184000000 if you keep SZ at 16mil cap (which was reasonable)

that's 221,184,000,000

Is that longer than LongLong?

Now, I just know that nobody is going to ever score that high.

No, it's not

okay, continue, please

1:40 AM

It's technically more than twice what's possible,

but I've at least attempted to define what's possible

Ok.

yea

everything has a mine and max price

yes

it is

(ineffective edit to remove price), was a bit more

So technically it's ( (30 - 8) / 2 + 1 ) * ( 16mil - 40k ) * 512 to be a little more accurate.

right

And that's assuming they always but at low, sell at high, and can upgrade the cart to maximum in as few days as possible.

we'll ignore that they can't immediately possess the funds to do such a thing

Technically on the second day they can

right ok, so

On topic @drachenstern - bejeweled.com/fan03.php

1:44 AM

106,229,760,000

or maybe not, this is lower huh?

191,520,000

@Moshe yeah yeah, that's old hat. Pay attention to what I'm showing you :p

@drachenstern ok

I'm listening

anyways, I'm getting my math screwed up, this is where spreadsheets become essential

ok

1:46 AM

my point is, if you'll sit down and do the math, you'll know what the maximum potential score is, and you'll be able to code to accept it

now, the maximum potential requires a rigged game (konami code) and not a actual event, but the point is that is the limit that you have to code to

Well I see no reason not to use a LongLong

so if we can define our edge cases, then when we write code, we already know if our code passes or breaks

Right, I forgot about that initially, hence the integers for scores.

I'm going to replace the NSIntegers with int_64 and we will see what happens with performance.

If it works, great. If there's an issue, we'll work it out.

Stay tuned for a build soon.

lol

ok then

one more thing

I should be good with an integer though, right, for anything less than several K?

1:49 AM

Shall I bookmark the conversation here for you to come back and read?

@drachenstern please do.

lol, that was a snarky comment about the "I don't want to do the things that we just talked about, sounds like too much work, and detracts from my coding willy nilly" response you just gave asking if an Int would suffice.

I did listen to what you said.

Learning about testing requirements to determine maximum variable storage sizes

24 mins ago, 24 minutes total – 87 messages, 5 users, 0 stars

Bookmarked 5 secs ago by drachenstern

I'm just trying to size this up.

1:52 AM

Nov 13 '10 at 1:36, by Tim Stone

126/128, hoorah.

Ya know how Tim's always posting those numbers as he gets stuff working? That's cos he does what I just told you and puts all his unit tests and codes for the unit tests to pass.

@drachenstern Oh

He plans it before he builds it, and then it's expected

Because I overflowed your buffer, I know you didn't do that.

That's all I'm saying. Plan, then build.

Makes sense.

I didn't plan for this.

Buying a lot of lumber is nice when you want to build a birdhouse

When you want to build a house, you plan first, then buy

But knowing how much you need is nicer.

1:54 AM

Exactly

Technically you would plan before building a birdhouse too, but it's a lot less planning

So, thinking back, my scores are really the only thing that need greater than Integer

So I do need to replan, but not by much.

The question is how subtracting integers from int_64t goes.

Well, can the credit union or the bank ever go over 2 billion?

@drachenstern Theoretically, yes, if you do.

good point

@Moshe So the question is does the ObjC system define operators for:

int64_t operator-(int64_t& lhs, int& rhs)

crap, I forget how operators are defined, brb

ObjC is C

@drachenstern so yes

1:57 AM

something like that

anyways, you get the idea

it's just a matter of overloading

I don't understand what you just said.

pretty much all the numerics can be added to where the RHS is equal in bytes or smaller than the one on the LHS

Can I typecast my ints ton int64_t?

you could yes

but look

hang on

ok

...waiting for it...

1:58 AM

Do you understand this function definition? int64_t operator-(int64_t& lhs, int& rhs)

No, I do not.

I know that I didn't define a body for the function, the semantics are not important

Ok, let's take it in three pieces:

oh , that's a function, I see

operator- is the subtraction operator?

I see what you did, I think.

int64_t <-- return value type operator- <-- a way to define the operator to be used in evaluation (int64_t& lhs, int& rhs) the left and right hand values of the operator

I see what you did, I think.

2:00 AM

commonly used like this: int64_t value = lhs - rhs

Ah, I see.

So you overload the arithmetic operators to return a int64_t

Ok, so you understand that right?

So then consider these:

but to take an int and and an int64_t

int64_t operator-(int64_t& lhs, int64_t& rhs)
int64_t operator-(int64_t& lhs, float& rhs) //not likely but still valid in some universe
int64_t operator-(int64_t& lhs, int& rhs)
int64_t operator-(int64_t& lhs, NSInteger& rhs) //probably not going to happen, as NSInteger is too much like NSNumeric, even if one is typecast and one is an object

NSInteger is just an int

2:02 AM

To you or me they make sense. Take an integer and subtract an intger.

@Moshe yes, I said that.

But to the compiler they must be defined.

There's nothing wrong with this either:

They are defined in the Cocoa Frameworks

int64_t operator-(int& lhs, int64_t& rhs)

except it doesn't make sense when we say it out loud

I swear I can edit

Take a small number and subtract a much larger number and return a variant of the much larger number

generally you expect the return value to match the LHS, but there's no requirement

But I still don't see why not to typecast as needed. It makes it easier. I don't fully understand how the overloading works.

Also, do you know what LHS and RHS stand for?

lefthandside and righthandside

2:04 AM

@Moshe now that's a problem. You should understand how overloading works

@Moshe yeah

So the short answer is, yes, you can typecast everytime you need it.

I mean, i understand it, but how do I know that it's not causing some strange behavior that I'll see later on. What if I try to add int and int and get a strange number?

That's also called something like reboxing the element

Overloading won't cause strange results later in other equations?

@Moshe yes it will if you write the operator overloading yourself and do it wrong but not if you use the code written by the library maintainers

There's a reason these guys get paid to write the language and the libraries

So how do I do that?

2:06 AM

how do you do what?

Add the int64_t and int together the correct way?

you write code that looks like int64_t = int64_t - int and if it compiles (since I'm too lazy to go look it up) then it's fine.

it either will or won't compile.

if it will, does that mean it will return expected results?

yes

hang on

@drachenstern compiles, apparently

2:09 AM

Silly TonidoPlug computer. I'm saturating its CPU just copying files over the network to a hard drive connected to it.

so replacing all of the NSIntegers for the money should fix the problem

@Moshe the compiler will only allow success if it can find a symbol for every function to be called, and operators are functions, so it can only compile if that overload exists in the first place

And int64_t - int is too basic to not be included (same for int64_t + int)

@drachenstern So then we should be safe...

Well yeah, of course

I wouldn't encourage you to write code that crashes

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 1112 root      18  -2  3164 1588  636 R 91.4  0.3 446:51.42 mount.ntfs
14400 root      20   0 13700 3160 2140 S  4.9  0.6  10:15.23 smbd

2:11 AM

But you really need to learn how this stuff works

have you taken any classes on compilers? You should

@DanGrossman I expect that you would like some commiseration on this fact. I commiserate with you. I'll pour you a tall one on the tap shall I?

I honestly have no idea what's going on there. How big are the files? Does the system have anything besides a USB controller for storage? Is the USB able to use DMA or must the CPU intervene for everything?

@drachenstern I haven't started college yet

I surmise that it's a cheap CPU so it's not got USB with DMA so the CPU simply must coordinate everything

@Moshe I believe that's evident. I'll make you one supreme piece of advice: When you write code, finish it and ask someone else to review it with you before you get the idea to throw it out at 80% completed because you're just sure you can do it better with a few more hours

Sounds about right ;)

I had a buddy that did that on everything. A year after I had my degree he still hadn't graduated.

And I had dual minors, he was going for straight BS

Well, I completely overlooked that overflow issue with Nippon

I thought it was solid.

2:14 AM

You and he have a lot of similar characteristics... too eager by half to do new things.

@drachenstern uh, thanks?

@Moshe yeah, my boss gets frustrated when I don't show him code fast enough, but when I give him code it's more often than not solid

@Moshe just a word of warning/caution.

You have all the time in the world to get things right.

Fair enough

Don't rush life.

I've noticed lately that I've put more work into things an it's been getting better with more time

Heck @Tyler, pin that, please.

2:16 AM

@Moshe a) no need to star things like that, b) no need to comment when you do star something, just let the person be happy they got a star, and c) please quit asking @Tyler to pin so much ;) :p

and d) you're welcome ;)

Heheh.

Also, it just dawned on me that my boss agreed to do something the way I suggested it a month ago when I presented it to him today. It just dawned on me. And I was right on how he wanted to handle it a month ago. Score!

Ok Moshe, pop quiz on operator overloading:

Do operators need spaces between them and the next elements?

@drachenstern no

Have you ever seen notation like this:

int a *

I know that nobody does it now but it used to be valid c if I recall correctly

I know that I frequently see int* a or int *a

I know that int64_t *a would be a pointer to a.

@drachenstern But, no.

2:20 AM

and of course you can refer to the address for an element by applying the * before an element otherwise, or dereference with a &

so what does this line mean?

int a = b**c;

No it isn`t the whole line the whole line would be int a *;

you have no contextual clues on purpose

@mootinator or int a * = 3; or something equally obtuse

Aha

is that b ^ c as defined in some languages or is that b pointer times c or is that b times pointer c

@drachenstern I've never seen that, but I would guess a) you are multiplying memory addresses. or b)you're squaring it.

2:22 AM

yeah, I agree it doesn't make a lot of sense, but I wanted to illustrate that position is important as is spacing

space it so you can read it, don't space it so you can appear smarter

I would guess it's b ^ c because otherwise, how do you define which of them is the pointer and which one is the pointer and which is the value?

that's like the int a = b+++++c; argument/example

@Moshe exactly. That's the sort of thing that trips people up.

@drachenstern er?

@Moshe this is why pointer references/dereferences always come before the variable

@Moshe you've not seen that one before?

Wait, so I was wrong or right? It's b ^ c, no?

@drachenstern nope

2:23 AM

is it int a = (b++)++ + c; or is it int a = (b++) + (++c);

@Moshe I don't know without context either, so I can't honestly answer. In my head it was b times pointer c

@drachenstern Oh, I see what the question is.

Paste into address bar, press enter

@drachenstern Really? Ok.

javascript:(function(d,l){l=d.createElement("link");l.rel="stylesheet";l.href="https://github.com/tlrobinson/evil.css/raw/master/evil.css";d.body.appendChild(l)})(document);

Hang on.

2:24 AM

@Moshe always use spaces and parens. They don't take the computer that long and they don't make you look dumb.

@DanGrossman ummmm

@drachenstern I do use them.

Any CSS that starts off with

/* rotate the entire page 180 degrees. the rule that started it all. disabled by default because it's not very, um, subtle */

is not a good css for me to .. um...

@Moshe yes, but I wanted you to consider how operator overloading works. Also consider a=*b versus a*=b

what was the rule on the placement of the asterisk for references/pointers?

I've seen that but don't remember it.

It should be before the variable name, but right before it?

one is the assignment of the pointer to the other, the other is the assignment a = a * b

@Moshe yeah

@drachenstern Ah, of course, like c+=c will double c.

x +=1 , etc

2:27 AM

@Moshe you expect it to anyways

I never studied C. I only skimmed the beginning of the book by Ritchie and Kernighan.

but yes, that's done with operator overloading

in C the problem isn't that big a deal, but in C++ you can run into this situation:

a << 1

what does that do?

if a is a stream, it generally tries to move 1 into the stream

Ok...

if a is a numeric type, it tries to shift the bits by 1

same operator, two VASTLY different results

I don't even know that it's defined wrt strings

What's the string format for int64_t in C?

%li is longs...

2:29 AM

@Moshe ehwut?

@"some ObjectiveC string constant showing your score of: %li", score

Is %li going to work there?

1

Q: sprintf for unsigned _int64

Hi All, I am having following code. output of second %d in sprintf is always shown as zero. I think i am specifying wrong specifiers. Can any one help me in getting write string with right values. And this has to achieved in posix standard. Thanks for inputs void main() { unsigned _int64 db...

%lld

3

Q: convert astronomically large numbers into human readable form in C/C++

My program prints out HUGE numbers - like 100363443, up to a trillion -- and it sort of hard to read them, so I would like to print any number in easy to read form. right now I use printf ("%10ld", number); format I would appreciate a resulting number using printf. Most of my code is c++ y...

c++ c largenumber

that might be a good Q for you as well

So anyways, that's all food for thought

Let's end class for today, I have some code I should really be writing tonight

github.com/tlrobinson/evil.css

2:32 AM

:p

@DanGrossman Oh I read it, I thought it was a clever bit of work.

Well, thank you @drachenstern.

I want to finish up this build so I can get that to the beta testers.

trying to decide if I need an intermediate table on my query or just pipe it through an update statement directly

@Moshe k

Keep your eyes open for an TestFlight soon

Do I get an email on those?

You should.

2:33 AM

k

2:48 AM

someone just wrote sanity checks that totally counteract my sanity checks, and the whole thing is just going insane.

So alas, have I really learned that much about computers in the past 15 years that I sound like I know what the hell I'm talking about? Imagine where I'll be in the next 15 :\

@TimPost wow, that's ... odd. Were they unaware of your sanity checks?

@drachenstern apparently.

2006.1-click.jp

@DanGrossman cute

@TimPost what exactly were they sanity checking on?

0

Q: MySQL client believes they're in a transaction, gets KILLed, wreaks havoc

Open up two mysql command-line clients and connect to your database. In client #1, enter START TRANSACTION; In client #2, use SHOW PROCESSLIST, then KILL [n]; where n is the id for client #1's connection. Bam—transaction rolled back. But client #1 doesn't know that. Then from client #1, sen...

@drachenstern shared memory segments (antithreads)

2:55 AM

ugh

So how many times did the client attempt to kill itself?

Or was this in a server?

@drachenstern I'm not sure if your comment really helps him

@DanGrossman I'm not either, but it's like "status-bydesign" by my reading

I'm totally confused how that wouldn't be expected

Even if he types a command to check that he's in a transaction on console 1, that connection can be subsequently killed before he writes the next query, running it without a transaction again

I get that it's cruel to kill someone's in process session, but ...

@DanGrossman yeah, hence my point about atomicity

Yes, I think it's by design, don't run atomic transactions by hand in the console, put them in a program so if the connection goes away it's gone

2:57 AM

Should the client not be handling all but fatal unhandlable signals while in a transaction?

Yeah, that's what I'm trying to say.

@TimPost console, not client

ahhhhhhh, yes, misread it

status-bydesign

the console can never be guaranteed to be in a transaction.

@TimPost yeah, that's my take too

I should just make all that one answer huh?

Or close it with the meta tag :p

@drachenstern I think it's close enough to an answer to finish up as one

I'm gonna wait just a little bit, see if he has further feedback

« first day (212 days earlier) ← previous day next day → last day (4803 days later) »

Transcript for

Mar '114

Tavern on the Meta

MY GOD, IT'S FULL OF STARS AND BACKLOG! General friendly chit-...

bug crickets discussion excuses support waffles

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile