6
ddephor
7y

ATTENTION PLEASE! Important announcement following:

Please check your interface implementations for correct byteorder according specification BEFORE YOU START COMPLAINING ABOUT DATA FAILURES ON EXCHANGING DATA.

Freakin hell, if I'd get some money for every byte order mismatch on testing interfaces, I'd be a be a billionaire.

And why are all those highlevel I-know-every-fucking-framework developer incapable of checking the real memory content of a datatype, and the real data content on the interface even if you tell them that their byte order is obviously wrong?
No, your system is not the centre of the universe and I don't care how you get your less-than-32bit-datatypes-are-for-assembler-usage-frameworks to change byteorder. It's not rocket science, if there's no ready-to-use-function then write those 4 lines yourself.

Next time I get to specify an interface I'll go for mixed-endian, just to make sure everybody involved knows the concepts of endianess afterwards.

Comments
  • 0
    I get the struggle, I really do.

    BUT any framework that does not abstract away endianess is utterly and completely useless. Leaky abstractions are toxic!
  • 0
    Most devs don't care about that because they always stay in their microcosm.
    And what the framework doesn't provide, does not exist for them.

    What annoys me so much is that so many devs don't even know about endianess. They take their frameworks output as granted and often insist on it being correct without checking it.
Add Comment