Random number generation about toxcore

[irungentoo]

I'm sorry, for some reason I was under the mistaken assumption that random() produced better random numbers than the rand() function.

If you check the code however you will realize that the only place the random_int() function is used for crypto is for generating the nonces.

The key generation is left to NaCl or sodium.

This will however be fixed.

Thank you for looking over the code.

[benwaffle]

Just seed both random() and rand() in general, and there should be no troubles.
On Jul 13, 2013, at 6:34 PM, irungentoo [email protected] wrote:

I'm sorry, for some reason I was under the mistaken assumption that random() produced better random numbers than the rand() function.

If you check the code however you will realize that the only place the random_int() function is used for crypto is for generating the nonces.

The key generation is left to NaCl or sodium.

This will however be fixed.

Thank you for looking over the code.

—
Reply to this email directly or view it on GitHub.

[gmorenz]

Your impression of random() and rand() isn't completely mistaken, there are libraries that have them separate, and in those random() is usually better at generating properly distributed random numbers. I mostly mentioned that in case it mattered to someone debugging who expected certain random numbers or something, personally I would leave a comment in the code about it and then ignore the issue.