Re^11: Get the order of HTTP request headers

I meant "denote". But this is getting us nowhere. Your statement ...

There's definitely no attempt to make them random

... is simply wrong and you just won't admit that, will you?

But there is indeed a built-in hash randomization mechanism that is based on the same principle as the RNG in your favorite OS. And this mechanism is clearly documented in perlsec and perlrun.

If you follow those links you will even find the phrases "randomise", "hash randomization" and "random perturbation".

I don't know what else to say.

Comment on Re^11: Get the order of HTTP request headers

Replies are listed 'Best First'.
Re^12: Get the order of HTTP request headers by ikegami (Patriarch) on May 07, 2010 at 16:15 UTC
But there is indeed a built-in hash randomization mechanism We're talking about key orderings. It doesn't matter to this conversation whether bucket selection is random or not. And this mechanism is clearly documented in perlsec and perlrun. I've already refuted that. In fact, perlsec says exactly what I said: Also note that while the order of the hash elements might be randomised, this "pseudoordering" should not be used for applications like shuffling a list randomly. As for perlrun, the passage to which you link doesn't explicitly mention changing key orderings at all. It is not the purpose of the feature to alter the key ordering, much less making it random.	[reply]

Replies are listed 'Best First'.

Re^12: Get the order of HTTP request headers
by ikegami (Patriarch) on May 07, 2010 at 16:15 UTC

But there is indeed a built-in hash randomization mechanism

We're talking about key orderings. It doesn't matter to this conversation whether bucket selection is random or not.

And this mechanism is clearly documented in perlsec and perlrun.

I've already refuted that. In fact, perlsec says exactly what I said: Also note that while the order of the hash elements might be randomised, this "pseudoordering" should not be used for applications like shuffling a list randomly.

As for perlrun, the passage to which you link doesn't explicitly mention changing key orderings at all. It is not the purpose of the feature to alter the key ordering, much less making it random.

[reply]