The Big Programming Thread - Page 882

On May 17 2017 04:58 supereddie wrote:
Also, maybe you don't need a database at all and just use Excel's data tables? Just put the 'source data' on a seperate sheet
.
Also also, you can enable 'Filter' on rows and columns (just select the rows/columns in Excel and press Filter): possible filter values will be available in a dropdown.

On May 17 2017 05:22 tofucake wrote:
I've never even heard of them. What language are they for?

On May 16 2017 16:17 Manit0u wrote:


It's C89/90:
https://software.intel.com/en-us/articles/old-style-c-function-argument-declarations-are-supported

Also, the code is formatted really bad even in the PDF. I guess someone has been teaching C for a long time and never bothered to actually track changes to the language or educate themselves past the initial point... Doesn't bode well.

Edit:

Here's the code brought up to a bit more modern standards (but it still doesn't make much sense):
+ Show Spoiler +

#include <stdio.h>

int setjmp(int v);
int longjmp(int v);
int fun1(void); 
int fun2(void); 
int *add; //using global, not sure if best idea
int main(void)
{
    int r = setjmp(r);
	if(r!=0)
	{
		printf("error\n");
		return 2;
	}
	
	fun1();
	return 0;
}
int setjmp(int v)
{
  add = &v; //used a global variable
  return 0;
}
int longjmp(int v)
{
  int *p; //pointer
  p = &v; //save argument address
  *(p-1) = *add; //return address = address in main
  return 1;
}
int fun1(void)
{
    printf("hi1");//testing
    fun2();
    printf("hi2");//testing
    return 0;
}
int fun2(void)
{
    int a;      
    longjmp(a);
    return 0;
}//output SHOULD be "hi1"
 //output is currently "hi1" "hi2"

Showing 90 changed files with 98 additions and 517 deletions

On May 17 2017 05:40 Manit0u wrote:


They're for Elixir

Overall, this looks like slightly worse Rails. If you want simple and performant, you could always go with something like Kemal which is built on top of Crystal. Something I've been wanting to get into lately. We've recently created a simple script in Crystal to populate our database for performance testing and it did it so good that it killed our server (weren't expecting it to be this good).

On May 17 2017 05:37 Artesimo wrote:


Yeah, after checking out "ribbontab" (never heard that term before, thought it was something special lol), I dulged in a lot of helpful stuff that most likely covers all my needs. Feel a bit stupid now, I just googled waaaaaay to complicated. I might ditch the database since they already distribute their by hand sorted tables per dropbox.

On May 17 2017 20:28 Nesserev wrote:


So, setjmp and longjmp's goal is to basically unwind the stack instantly to a certain prior state (when setjmp is called; stack frame of calling function still has to be on the stack when calling longjmp). Normally, this is done by saving the content of all registers to a buffer, and then using longjmp later to overwrite all current register values.

Problem is, you can't access the register values without assembly... so yeah, throw the idea of implementing the actual behaviour of setjmp and longjmp out of the boat, cause that's not gonna happen.

(An added problem is that, because during the execution of a function not all variables are necessarily stored within registers, and variables can be changed or pushed/popped from the stack during function execution, it becomes clear that setjmp and longjmp can lead to some nasty shit.

However, this probably won't be a problem with such simple functions, so let's ignore that.)

--

So, let's go back to the ultimate goal of using setjmp/longjmp, which is stack unwinding. Can you unwind the stack without directly accessing any registers?

Well, you could try manipulating registers by manipulating the return addresses on the stack, but then you end up again in "implementation land" with call conventions, compiler implementations, etc. which is probably way worse than the relatively clean "default" solution of setjmp and longjmp.

So, uh... I think this is a lost battle.

I'd be glad if anyone could correct me :/

On May 17 2017 23:28 Manit0u wrote:
Wait a second. Your professor is giving you examples of code written in style that was used 30 years ago, that is compiled for embedded system/router architecture which introduced 32 years ago?

This is madness...

On May 17 2017 23:13 dsyxelic wrote:
I asked on stackoverflow and got similar responses (cant be done or awful assignment forcing bad workarounds which defeats purpose of assignment). I gave pretty much all the directions I was given...

Dunno if this helps but this was what the professor showed and talked about using the stack. He said he basically gave us the answer to the project with this:
+ Show Spoiler +

On May 17 2017 23:38 dsyxelic wrote:


Yeah idk this is easily the worst cs class ive taken. My data structures class was in pascal as well but wasnt as bad since we were allowed to choose from python or c/c++ for any submissions. This prof is ridiculously lazy. Used same slides for years (not even made by him and is out of date) and the random midterm uploaded online by a student 5 years ago was pretty much identical to the midterm we just took.

        addiu   $sp,$sp,40

On May 17 2017 23:41 Acrofales wrote:


Image links are broken.