dan <[EMAIL PROTECTED]> writes:
> Simon Josefsson wrote:
>> dan <[EMAIL PROTECTED]> writes:
>>
>>>Good afternoon, all -
>>>
>>> I'm new to the list, I'd like to first thank the developers of
>>> GNUTLS for their fine work. We appreciate it.
>>>
>>> We're developing a threaded application in our development
>>> department, and it's going alright so far. However, we keep
>>> getting an error which we believe comes from generate_dh_params().
>>> The error is as follows:
>>>
>>>Fatal: no entropy gathering module detected
>>>
>>> The most we've been able to gather using google.com and assorted
>>> resources is that entropy cannot be established. The only solution
>>> that we had found that was even remotely related to ours just
>>> suggested making sure that /dev/{u}random had world-readable
>>> permissions. Aside from that, we don't know what to go by.
>>>
>>>Any pointers?
>> That is a libgcrypt error. Check how you built libgcrypt. If the
>> Win32/Linux/Unix RNG can't be used, you might need to run EGD.
>> You could also try building GnuTLS with --with-nettle.
>> Hope this helps,
>> Simon
>>
>
> Simon -
>
> Thanks for the response.
>
> The only argument we're compiling libgcrypt with is '--disable-asm'. I
> took the origional RPM for libgcrypt-1.2.0-1, and modified it slightly
> for -1.2.1, just in case. We get this error regardless of which version
> we're using.
>
> Why wouldn't we able to use the RNG on this Linux platform? What would
> prevent us?
I don't know, perhaps you can quote relevant parts of the output from
libgcrypt's "configure" output? Does 'make check' in libgcrypt work?
Perhaps some libgcrypt expert can help.
Regards,
Simon
_______________________________________________
Help-gnutls mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/help-gnutls
