My sons bedtime is in 30 minutes

Lol

ahaha

I don't remember having seen the chat quiet for that much time!

No chat Tuesdays

I don't recall ever seeing "12 hours later..."

Nor 17, if we count the FIP out.. :P

oh shit you're right

Today smells like summer!

Hot girl posted on Instagram that she beat the new Zelda in 40 - 50 hours

Decent.

Hot girl plays through Zelda in 2 days. Marriage material right there man.

HOLY FUCK

UNITY BOUGHT TEXT MESH PRO. IT IS NOW FREE.

Not sure what that is, but it sounds useful!

It is what @PearsonArtPhoto just paid 100$ for.

Star that shit man. People here are gonna wanna know this shit.

can some1 help me with mysql need a query to generate 100-200 new rows name password can be same the only thing is important is playRegion randomaly 1-5

@yahoo5000 you can look that up on google pretty easily/quickly

keyword ?

how to generate new rows using a query?

Not really surprised they bought TMP.

I only spent $60 for it, I think.

@PearsonArtPhoto kinda sucks for you though. you just bought it not too long ago, no?

Yeah, especially as I didn't really need it yet. Oh well...

Ouch. Well, you should be able to make that back quickly, right? :P

Hopefully.

One good thing is, I understand TMP, so...

@dcfyj would that work ?

Truth be told, I used money I essentially got for free to buy it, so...

INSERT INTO user_acc(username, password,playRegion, firsttimelogin) VALUES ([test],[test],[1-5],[1])

On the plus side, it wasn't more expensive than $60 ^^

@yahoo5000 um, if you're using [1-5] as a random generator, I think you'll be disapointed.

live very disapointed ?

@yahoo5000 this might come in handy. Although a little google work is all that was needed to find it

@yahoo5000 the easiest way to call a query inserting a single row 100-200 times

yea i am doing it now

or a single query could do...

what about this ?

i also got syntax error on line 1

gamedev.stackexchange.com/a/138026/10408 <- Enjoy this physics simulation!

DELIMITER //
DECLARE x INT DEFAULT 0;
 SET x = 1;

 WHILE x  < 200 DO
INSERT INTO `user_acc`(`username`, `password`,`playRegion`, `firsttimelogin`) VALUES ([test],[test],[RAND() * (6)+1],[1])
SET x = x +1;
 END WHILE;
DELIMITER ;

not working no error nothing but he doest do anything

MSSQL D:

mysql :P

Huh? Never seen the [] notation in MySQL

i remove them but still there is

i got them by default

Never had to work with loops in MySQL, sorry

looks scary

Sounds like you're doing something the wrong way

hmm i was askign this question on stackoverflow , havent got any good answer yet onde dude said should using mysqlslap because the current code is not supported or something

loops on data are not how to use SQL

you're using the wrong tool

how should i do ?

you would generate your random data somewhere else then use an SQL query to enter the data into the db

don't do the logic in SQL

SQL is notoriously difficult to debug.

i could do that in c++ but was hoping easy way

doing logic in SQL is not "easy way"

it would be kind of like using C++ to make a website

wrong tool

I'd create a python script that generates the SQL queries, output that in a file and run that file..

yea but as far as i know i can use c++ to make website :D

oh cool, free text mesh pro

I guess they will fix it now so that it actually works, ffs

You can, but that's not the point.

yea i am trying to run shell on mysql and then use mysqlslap but all passwords wrong

yea some people are doing it :D

Or anything in Java, but you really don't want to

@yahoo5000 you can. but it would be really stupid to do it that way.

yea did knew that was hoping would be faster then write in c++ but the time i spend already coudl right it many times so far

@jgallant It has this huge downside with Unity UI that it breaks the UI instancing, our UI takes more than half of our drawcall budget in our mobile game

instancing, or batching

@Lasse Tried to split up UI to different canvas?

The text has to be separated to it's own tree of objects so it can batch properly

Never used it.. Does it break all canvas?

it's horrible manual job / spaghetti code if done by code

it has to change material in between drawing the UI elements and the text, so yes

Whaaat

@yahoo5000 mySQL is still the wrong tool to use.

mySQL stores data, and retreives it.

it should not be used for more than that.

it's possible to do so. But that's like emptying your bathtub with a spoon instead of just pulling out the plug.

@Almo I agree'); DROP TABLES;

I should note that "just" storing data and retreiving it is immensely complicated, when considering that multiple processes can be accessing the same data for read and/or write at the same time.

ok fuck mysql doing in c++

and it took 5minutes

@Lasse I think you are doing something wrong, they are suppose to be way faster than the built in Unity Text renderers

@jgallant well they are

It's even more horrible with the Unity UI own text

haha

Because you usually dont want that

how can i do this then ?

why cant i copy 2 bytes into two bytes i dont see any conversion

std::copy(p.buffer[0], p.buffer[2], &(uint8_t*)type);

I think

Also Im lazy and using C-style cast, you shouldn't

still error doing this way

u mean memcpy ?

yes

what about this ? (int16_t)&p.buffer

no

wtf is going on here lol

are you guys trying to memcpy(p.buffer, p.buffer + 2, 2) or something?

yea

i did with memcpy cuz std::copy was giving conversion error

actually std::copy works differently

std::copy(int first , int last , dest);

std::copy uses iterators

yea i know

I personally hate the living fuck out of the iterator paradigm completely, and have never in real life seen a single example where it made any program design more efficient or convenient. Actually every single time ever period I've found that either it's tedious to type out, or introduces an auto which is convenient but somewhat ambiguous leading to bugs when less experienced devs touch the code.

with that said though, the error with that std::copy above is trivial, put the & in the right place

std::copy(p.buffer[0], p.buffer[2], (uint8_t*)&type);

std::copy(&p.buffer.front(), &p.buffer.front() + 2,& type);

i have tried (uint8_t*)&type was giving error

well I think the first 2 arguments are broken also

tbh this entire thing is frustratingly horrible

this works

std::copy(&p.buffer[0], &p.buffer[2], (uint8_t*)&type);

Beat me to it..

In actual reality, the obviously correct solution is type = *(short*)p.buffer

But that's not standard compliant C

(Though it's used in real life extensively, and works efficiently and reliable on fucking every desktop ever)

Obviously that works lol

But holy fucking god that's ugly code, you know?

std::copy(&p.buffer[0], &p.buffer[2], &type);

this works too

still ugly as fucking hell

yea but this is better then memcpy

and i was need just for debugging

type = p.buffer[0] + p.buffer[1] * 256;

which is still ugly as fuck but at least standards compliant

if it's for debugging why are you screwing with standards when you can just do type = *(short*)p.buffer.data();

i dont like bitshifting :D

i was doing this (int16_t)&p.buffer

Well that's severely broken

why its seems same as (short)p.buffer.data();

You're converting the pointer to the std::vector to a ushort

@yahoo5000 Because it is

That's basically just a random number

You want to dereference the pointer, to get the data it points to

But before you do that, you want to get the right pointer

The pointer to the std::vector structure is probably not the same as the pointer to its data :P

but its same as .front ?

or [0]

I'm not sure if those are defined explicitly as giving a suitable pointer to the entire buffer

Though it'll work in real life, but if you're trying to use standards you're supposed to be using .data() for that

And even that might be relying on C++0x

.data() is C++11 I think

hmm i see

So if you want it to be "technically correct" under vanilla C++, then type = p.buffer[0] + p.buffer[1] * 256; is the "correct" solution anyways

actually... type = p.buffer[0] + p.buffer[1] << CHAR_BIT; is technically correct

But since fuck standards. real life. Why use the long ugly code?

yea but thats looks shit :D i rather do std::copy or just as u mentioned *(int16_t *)p.buffer.data()

Why would that be more correct than just memcpy?

Because the memcpy approach assumes that std::vector keeps all elements in a contiguous buffer, which I think is only guaranteed after some revision of C++0x or C++11

I mean I don't think there's any STL insane enough to do anything else, but we're being pedantic here, not realistic :P

Right. The bool template kinda ruins that assumption :P

So it's guaranteed for all others?

@JoshPetrie sweet thanks

@JoshPetrie So what approach would you chose in Yahoo's place?

Merge two uint8 into a uint16

(where the two bytes are at the front of a vector)

copy 2 bytes from std::vector<uint8_t> to int16_t

Booooom son!

:D

i will stick with Mick way anyway look nicer :D *(int16_t *)p.buffer.data()

I like that way because it's more obvious what's going on, and would use it for debugging without a second thought, but I gotta emphasize it's technically incorrect lol

@yahoo5000 Did you just defy the advice of our almighty C++ god?

i did not :D but i like nice looking code :D not some ulgy :D

btw Josh's version is actually the most ideal so far, because | (instead of +) could possibly be more efficient on some processors

Though I'd be horrified if any complier produced anything other than mov reg16, [ptr]

@jgallant just booted gorb I love the :o on they sound sliders

hmmm any possible reason why winsock send() could stuck on sending 22 bytes ? and never finished never returns from send() function

Wasn't that your suggestion?

@yahoo5000 You're using a blocking socket and the other side stopped sending ACKs?

my server using blocking socket client using nonblicking

If the output buffer gets full, send will block

output buffer whats the size of it ?

Don't worry about that

I mean, come up with a solution that doesn't depend on the output buffer size, since the user can change it arbitrarily

hmm i believe this is same problem i had before , i just skip it for now . Switching to non blocking socket would solve everything right

?

Probably not, a non blocking socket will just throw away data in the same case a blocking socket would halt

You have to check when you send on a non-blocking socket, to make sure it'll actually send

i am checking but its never gives me returncheck

bool Server::SendAllBytes(std::shared_ptr<Connection> connection, char * data, int totalbytes)
{
	int totalBytesSent = 0;
	while (totalBytesSent < totalbytes) // while full buffer not send keep sending
	{
		int ReturnCheck = send(connection->socket, data + totalBytesSent, totalbytes-totalBytesSent, NULL);
		if (ReturnCheck == SOCKET_ERROR || ReturnCheck == 0)
		{
			return false;
		}
		totalBytesSent += ReturnCheck;
	}
	return true;
}

so if send success then returns true if socket error or send 0 will fail

I think that should "work" but it'll smash the CPU with 100% load if you hit the limit but it's non-blocking

That code is clearly meant for a blocking socket

yep

Also the real problem is that your output buffer is full

output buffer what is that ? i mean i am sending 22 bytes , what does output buffer storing ?

and when does he gets cleaned

i had similar problem when i was sending 70k bytes it might fill output buffer not sure

if it really "never" returns, then somehow the client is not receiving the messages and winsock is stuck waiting and retrying until the clients sends an ACK for the first thing in the output buffer

So did you kill the client quietly after it accepted the connection?

once i kill client i got failed to send packet

until then he gives nothing in return

just stuck in send function

Ok then are you actually reading the data on the client side?

i also was debugging and when step into send function it just stuck

Or is the client also stuck waiting for a send?

its reading

output buffer gets cleaned once client reads data right ?

Yes, it should stay in the buffer until the client sends an ACK packet to tell you they got it

I'm assuming you're using TCP

hmm then output buffer should by clean , because client have read all data so far i send to him only those last 22 bytes

yes

and client still trying to recv those 22 bytes

it could be the last problem i skiped have to do with it . because i was sending 70k bytes and was having issue tu recv them all it might still by on output buffer :/

Well it's not enough information, you're leaving an important detail out somewhere and I can't guess where

thats all i know my self

The only other guess I can make is that you need to flush somewhere because of Nagle's algorithm, but that's a stretch and I don't think it's what's going on here

Make a tiny program that causes the same glitch and share it

Just copy paste the broken part and not much else, only the minimum to re-create the situation

u mean on pastebin ?

sscce.org

ok will do

will just clean my server and client to make them tiny

@yahoo5000 I just got the feeling I should mention that when I said should "work" earlier, I put scare quotes on it because it'll return false even when things are fine, if you send anything bigger than the output buffer, since send() will return 0 as soon as that's full

not that it matters so much, I wanted to gloss over it to stay on topic, but you should notice still

ok

OMG Putin wears his watch on his right wrist! WTF man.

I'm starting to get fed up with stupid fucken users of picross Touch

hahah

what now

Like I'm ready to start telling people to fuck off

No auto save. 1 star

If you turn off the game in the middle of a puzzle it doesn't save the progress

Are you fucking with me?

haha

I wish I could revoke access for certain accounts

heheh

I haven't been sleeping properly for the past week and quite frankly I have a short fuse right now

:(

And I just got a 5 star rating

Thanks random guy

I'm not saying you should make the official complaint process "Step 1. Blow me. Step 2. Please wait." but I am saying those words

I have his google account and name. Let's wreck his life. Lol

Jj

Jk

jk - just killing

Jk Rowling

She tweeted that pewdiepie is a racist

I used to have no opinion of her, I can forgive her for harry potter, but recently I saw her making some kind of ignorant noise

yeah that's what it was!

And the context didn't matter

She stated context does not matter.

Like what? Really?

Take a really fucked up line from one of her books and attribute the quote to her

Context doesn't matter.

This girl called out idubbz for saying the N word

He found footage of her saying it on 3 occasions

lel

His video ripped her apart

@jgallant lmao

He went to her show and waited in line for a meet and greet

Hilarious

youtube.com/shared?ci=8MbJjt3Jsto

Stupid link but video us hilarious

That looked somewhat like gay porn intro

Idubbz is by far the best YouTuber

The frame rate though, is silky smooth

And yeah his content cop intros are pretty strange.

Well he is pretty strange

Everything he does is strange

@MickLH i think i made exactly same story just tiny client and server but this time instead of stucking he returns me -1 return check and failed to send same packet .

witch error code 10054

@yahoo5000 Then that tells you that you have to look at something you didn't expect

like what ?

And that's pretty common, actually one of the reasons I ask for an SSCCE is exactly because it forces people to find out for sure if they are looking at the right piece of code at all

i got same error code on my normal server when i close client

@yahoo5000 Well it means that whatever causes it to hang, is not part of what you thought should go into the SCCEE

So you should look at other things, start with the most related bits of code

hmm i am checking it now , removing mysql part and will run once more

It could be very simple, maybe the size of the send data is really big unexpectedly

It could also be very complex, maybe you have memory corruption and the debugger is lying about where you've crashed

You might be able to find the issue easier by making a copy of the entire source tree, and then removing as much as you can from it until it stops hanging

That should give you a good hint at what's going on, once you find that point

yea thast great idea i am currently doing it now tnx Mick ;)

how to find errors like this ?

@yahoo5000 Are you linking against user32.dll?

ok i was fucked up my project settings was missing %(AdditionalDependencies)

Took me a while to figure it out was looking problem not where she was :D

Girlfriend's phone died in middle of picross puzzle.

It didn't save.

hahahahahaha

fucking lolll

For crying out loud

I'm still laughing ty

Lol same

Man, I should probably ask my doctor why it hurts to laugh so much

Herpes

♫ ... the pope's got it and so do youuu!... ♫

I guess I could ask WebMD what it mi-CANCER.

Don't worry about it

that's mostly what I do, I always ignore the issue until it's absolutely unbearable lol

why do anything when it'll probably be magically fixed tomorrow morning

(or in this case, a moment after I stop laughing)