Hi,
thanks for your thoughts so far, so it seems we just want to get rid
of these conditionals for 2.5 :-)
One thing, though:
On Thu, Aug 16, 2018 at 10:28:05PM +0200, David Sommerseth wrote:
> There is also a related ./configure option: --enable-async-push. This adds
> inotify support, which r
On 16/08/18 13:37, Gert Doering wrote:
[...snip...]
I do see the arguments on both sides here. And in particular where these
warnings halts the compilation (which I think I read some where that newer GCC
compilers would do by default). But ... Gert asks a few reasonable questions
> Anyway. Try
Hi,
On 16/08/18 19:37, Gert Doering wrote:
>> Imho we should just get this patch through and then try to get rid of
>> the whole #ifdef PLUGIN_DEF_AUTH logic at once.
>
> If the idea is to get rid of PLUGIN_DEF_AUTH, then let's do so, not
> add extra #ifdef :-)
Good! Let's focus on that then :)
Hi,
On Thu, Aug 16, 2018 at 07:17:09PM +0800, Antonio Quartulli wrote:
> I totally agree with you that more ifdefs are just ugly and make the
> code worse. However this "new" ifdef is just part of some logic that is
> already there (a few lines above).
It is a new #ifdef, two more lines of code,
Hi,
On 16/08/18 19:08, Matthias Andree wrote:
> Am 15.08.2018 um 08:12 schrieb Antonio Quartulli:
>> The "cleanup" label in ssl_verify.c:verify_user_pass_plugin() is used
>> only when PLUGIN_DEF_AUTH is defined, therefore make the label
>> definition dependent on the same define.
>>
>> Fixes the f
Am 15.08.2018 um 08:12 schrieb Antonio Quartulli:
> The "cleanup" label in ssl_verify.c:verify_user_pass_plugin() is used
> only when PLUGIN_DEF_AUTH is defined, therefore make the label
> definition dependent on the same define.
>
> Fixes the following warning when PLUGIN_DEF_AUTH is not defined: