[gs-devel] feature request : better seeding the random number generator

[Didier LINK]

Hello all,

I face on a problem of randomness in gs.

Some users used to pre-seed by just doing "realtime srand", but now they
see that realtime has a very NONrandom value at all. The same suite of
numbers appeared in short time intervals.

I've see some bugs reported in gs bugzilla (version 7.X) that talk about
the non-portable use of /dev/(u)random devices but they disappeared in
the code since.

I think a better initial value in the rand functions can be implemented
on unix plateforms, at least, with using a configure time check.

There is some reasons to have removed the /dev/urandom read in the gs
rand functions ?

Can have you a stronger method to initialize the rand functions ?

Thanks for your comments.

Best regards.

Didier Link

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Ceci est une partie de message
	=?ISO-8859-1?Q?num=E9riquement?= =?ISO-8859-1?Q?_sign=E9e?=
Url : http://ghostscript.com/pipermail/gs-devel/attachments/20061101/57af6da6/attachment-0002.pgp