On Tue, Jan 03, 2017 at 02:17:38PM +0100, Andrea Bolognani wrote:
On Tue, 2017-01-03 at 13:46 +0100, Martin Kletzander wrote:
> > To solve the issue, turn virFirewallAddRule() from a very thin
> > wrapper around virFirewallAddRuleFullV() to a macro that expands
> > to a call to virFirewallAddRuleFull() - itself a very thin wrapper
> > around the aforementioned virFirewallAddRuleFullV() - with no loss
> > of functionality or type safety.
> > ---
> > This only seems to be required on very specific combinations
> > of Clang and host OS, eg. I need it on Clang 3.9 / Fedora
> > rawhide but not on Clang 3.8 or 4.0 / Debian sid.
>
> We sent various patches for this (me, Jan and maybe other people as
> well). I never realized it's not a problem with different versions of
> clang.
Oh, must have missed the previous attempts to fix this.
> I would say it's not a problem for us to solve it in this case,
> however, as I wrote in my solution, it works, but it's undefined from
> the specification point of view. Can it work just because
> virFirewallAddRule() gets optimized into inline function? it shouldn't
> be, though... I don't know.
I lean towards merging this or a comparable solution. It's
true that we aren't currently hitting this on our main
targets, but relying on undefined behavior is definitely
something we want to avoid, plus I don't see any real
drawbacks in changing this to a macro.
Agreed, I think this patch is fine - the original code could easily
have been done this way even ignoring the clang issue.
Regards,
Daniel
--
|:
http://berrange.com -o-
http://www.flickr.com/photos/dberrange/ :|
|:
http://libvirt.org -o-
http://virt-manager.org :|
|:
http://entangle-photo.org -o-
http://search.cpan.org/~danberr/ :|