The samples are in the osce course which is not free , But there are many samples on the net about this

please ive been ctfing for a year and making money off it and you’re telling me im overcomplicating all pwn challenges

I really want to know which weed people around here take

The good kind

If you need a low-level scripting language LLVM would be your best bet tbh. This C thing you're doing isn't really scripting

You can use it as a backend for a JIT tho

Making a language that transpiles to C then compiling that code in memory and running it

Making a language that transpiles to C then compiling that code in memory and running it

transpiling adds another layer making the jit ing slower tho

transpiling adds another layer making the jit ing slower tho

You only do so once

You don't transpile 150 times while executing

once everytime i run

but its still slower than no overhead

once everytime i run

That is why JIT are only suitable for long running programs anyway

Sure, but the less overhead you have the better

But making a JIT using TCC is not that worse compared to another way of doing so

it is worse though

especially on older archs

it is worse though

Worse than an interpreted script?

I doubt it

Hello I looking for a ready project with data base can someone send me please?

especially on older archs

Mind you, on Atari, a linked list is more than 50 times faster to explore than an array

Worse than an interpreted script?

worse than a direct compiler

worse than a direct compiler

Not an apple to apples comparison

Mind you, on Atari, a linked list is more than 50 times faster to explore than an array

i doubt that lol

Not an apple to apples comparison

What?

i doubt that lol

Doubt it all you want, it is a peculiarity of the platform

Lol

What?

So if you want an embedded language in a software, the embedded language should be entirely compiled before running, this is slower and wasteful and require complex tooling

While using a JIT here is a compromise

And a good one at it

JIT compiles too

both do the samr

Which is why people have gone the way of making JIT for Javascript and libraries like LibJIT and LibLightning

except if you jit to cpp and let tcc jit the output of the first jit you lose time

JIT compiles too

Then why are you saying it is worse?

except if you jit to cpp and let tcc jit the output of the first jit you lose time

this is what i said

i didnt say that jit is worse than compilation

well it is since you lose enrgy

but shrug.

except if you jit to cpp and let tcc jit the output of the first jit you lose time

You will always have intermediate representations, it is a good way to make a JIT with it because it provides you a readable intermediate representation, and one that you can output for exclusion of that allow you to include C libraries in you embedded programming language without overcomplicating its design

intermediate representations slow down stuff

intermediate representations slow down stuff

No

intermediate representations slow down stuff

Strange that your compiler use two of them right?

No

two compilers are slower than one

two compilers are slower than one

No

Strange that your compiler use two of them right?

shouldnt a good compiler output a binarh and the linker then links it against shit?

Also, note that TCC is actually pretty fast, about 250 times faster than GCC

two compilers are slower than one

clang use two compiler

Also, note that TCC is actually pretty fast, about 250 times faster than GCC

thats its purpose

gcc use two compiler

gcc use two compiler

both conpilers you mentioned are rather slow though

shouldnt a good compiler output a binarh and the linker then links it against shit?

Except you transform the source in an abstract syntax tree, optimize said tree, then make a output syntax tree, optimize that, them make a binary that you will link

Except you transform the source in an abstract syntax tree, optimize said tree, then make a output syntax tree, optimize that, them make a binary that you will link

Not to mention the preprocessor

Except you transform the source in an abstract syntax tree, optimize said tree, then make a output syntax tree, optimize that, them make a binary that you will link

Except you want to use twice that

both conpilers you mentioned are rather slow though

faster than their old "direct" compilers

Except you want to use twice that

TCC is not that complex, it only have one tree intermediate form and no optimization

faster than their old "direct" compilers

😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂

TCC is not that complex, it only have one tree intermediate form and no optimization

Because it is designed exactly to be simple and fast even if it have to generate imperfect code

😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂

Another example is bytecode. Intermediate step for increase performance

Another example is bytecode. Intermediate step for increase performance

right, byte code is an intermediary step

Another example is bytecode. Intermediate step for increase performance

Bytecode is generally the binary form of the abstract syntax tree

Except you want to use twice that

nope. cfe only converts your C/C++ code to a .bc (no optimizations) file containing LLVM bytecode. llvm-opt then optimizes that and llc compiles it

TCC is not that complex, it only have one tree intermediate form and no optimization

Well, thats still worse than instantly output bytecode

Bytecode is generally the binary form of the abstract syntax tree

That is the case in Java and C# for ezample

Well, thats still worse than instantly output bytecode

But not that worse, you may have a loss of a ten of percents or two

But look at the price of complexity you actually eliminate

But not that worse, you may have a loss of a ten of percents or two

If your first compiler is as good as tcc, you might have a slowdown of 100%

If your first compiler is as good as tcc, you might have a slowdown of 100%

But your "first compiler" will have to be hand rolled and support about 10 to 20 architectures

Are you ready to have a handrolled complete compiler to accelerate filters on your spreadsheet application?

But your "first compiler" will have to be hand rolled and support about 10 to 20 architectures

werent you the one saying that yacc is bad?

werent you the one saying that yacc is bad?

Yacc is just a parser generator

A compiler compiler ;-)

It is very limited if you ask me and that is the issue I have with it

A compiler compiler ;-)

It doesn't make a compiler, just a parser

Flex does that

Flex does that

No, it is a syntax checking tool

then bitching about it maybe isn't the best idea?

I do FreeBSD too, and bitching about Linux is what you can do in that platform. I am sorry but thats true.

I do FreeBSD too, and bitching about Linux is what you can do in that platform. I am sorry but thats true.

Of all OS I used, OpenBSD was the most friendly to code in

Agree

Flex does that

Also, flex and bison and yacc do not work in lots of cases

I do FreeBSD too, and bitching about Linux is what you can do in that platform. I am sorry but thats true.

I get it Linux isn't good, that's true, but now every second message from Maxi is about that and it's quite annoying

I get it Linux isn't good, that's true, but now every second message from Maxi is about that and it's quite annoying

Lol. xD

btw why couldn't you build the llvm toolchain?