[Pkg-clamav-devel] tomfastmath is ready (take 2)

[Andreas Cadhalpun]

Hi Sebastian,

On 04.11.2015 09:15, Sebastian Andrzej Siewior wrote:
> On 2015-11-03 23:43:08 [+0100], Andreas Cadhalpun wrote:
>> Regarding the armhf problem: Have you filed a gcc bug report?
> No. I will narrow it down first. Right now I am not sure what the problem
> is. There were bugs in C-code which were patched later (and I backported
> them but the bug remained). So it could be both. Once I know I will either
> fix the code or report a gcc bug. But right now I will first try to get
> new clamav into exp before looking at this again.

OK.

>> Regarding the other patch: Making function parameters const is technically
>> an API change, which is not really nice to do in a Debian-only patch.
> Okay.
> 
>> On the other hand these const problems can be worked around with casts,
>> so maybe it would be better to add some casts to clamav?
> No. You don't cast const away and have a good night sleep. Here in our
> case it wouldn't matter. However it is a bad example. I would prefer to
> remove the const in clamav until it gets changed upstream.

I consider it OK to cast const away, when one has manually checked that
it's not a problem. But removing the const in clamav should be fine, too.

>> It would also be useful to check what other programs using tomsfastmath
>> in Debian expect.
> Okay, I can do it.

Anyway, if it's not only clamav that expects the const, patching tomsfastmath
might still be the better approach.

Best regards,
Andreas