[Bug libc/3662] Implementation bugs in random_r and friends

[archie.cobbs at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=3662

Archie Cobbs <archie.cobbs at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |archie.cobbs at gmail dot com

--- Comment #10 from Archie Cobbs <archie.cobbs at gmail dot com> ---
Through a long and painful process of debugging and hair-pulling I also arrived
at this issue.

I am now in a state of weary disbelief.

> I already said no.  The interfaces are used internally and that's really
> their only purpose.  Design and implement your own library with randomization
> functions if you must.  If you don't want to provide any more patches to
> the documentation the bug might as well be closed.

WTF are you talking about? These functions are exposed and documented as normal
library functions just like all the others:

    NAME
       random_r, srandom_r, initstate_r, setstate_r - reentrant
       random number generator
    ...
    DESCRIPTION
       These functions are the reentrant equivalents of the functions
       described in random(3).  They are suitable for use in multithreaded
       programs where each thread needs to obtain an independent,
       reproducible sequence of random numbers.

This is a malevolent segfault booby-trap!

I can't believe the maintainers of glibc don't care.

-- 
You are receiving this mail because:
You are on the CC list for the bug.