Why no one should use the AT&T syntax ever | Hacker News Reader

Why no one should use the AT&T syntax ever

elronnd.net

Interesting discussion of some possible pitfalls where AT&T syntax allows you to shoot yourself in the foot.

But to my mind the most important argument is the one mentioned towards the end as a minor one: The official documentation uses Intel syntax. Why would you make your life harder by translating that syntax into a different one?

Only the official documentation for Intel CPU's. Almost nobody uses Intel CPU's writing asm, most asm writers are on arm or mips, in the embedded world.

AT&T is dominant, Intel syntax is for the poor Windows folks.

I do and my assembler is clang.

  .intel_syntax

From what I remember, ARM docs also put the destination register first, use no sigils for marking registers and immediates, and use a not-entirely-obscure syntax for expressing scaling and such in addressing modes. What aspect of ARM's docs do you consider AT&T-like?

I started coding x86 Assembly on the glory days of TASM and MASM and the reality looks quite different on this side.

We are so poor with the hegemony of desktop and game consoles CPUs.

Someone probably put a lot of effort into writing that, but I just had to stop reading.

It's too hard to separate the content from the snark.

That and a 2021 blog post that renders half screen width on my phone

Porting a language runtime from Intel syntax into AT&T just not to add an additional dependency to nasm/yasm when GCC was part of the requirements, was a big mistake.

In retrospective I should just kept nasm/yasm as dependency.

Maybe that could be a curfew project actually, revive the old code just for fun.

Reply on news.ycombinator.com