[PATCH] Duplicate detection in fail over did not work.

[Michal Židek]

This should fix the duplicate detection. Patch is attached. Please 
review with caution, i was in hurry.

https://fedorahosted.org/sssd/ticket/1472

Thanks
Michal

[Jakub Hrozek]

On Fri, Aug 10, 2012 at 05:13:48PM +0200, Michal Zidek wrote:
> This should fix the duplicate detection. Patch is attached. Please
> review with caution, i was in hurry.
>
> https://fedorahosted.org/sssd/ticket/1472
>
> Thanks
> Michal

I think the general approach with compare function is good, but it
should be a property of the fo_service. I don't think there is any
scenario where we'd like to treat two servers with same user_data
contents as the same.

The default of the compare function should be NULL, which would
translate to "just compare the pointers", as we did until now.

[Jakub Hrozek]

On Fri, Aug 10, 2012 at 07:09:13PM +0200, Jakub Hrozek wrote:
> On Fri, Aug 10, 2012 at 05:13:48PM +0200, Michal Zidek wrote:
> > This should fix the duplicate detection. Patch is attached. Please
> > review with caution, i was in hurry.
> >
> > https://fedorahosted.org/sssd/ticket/1472
> >
> > Thanks
> > Michal
>
> I think the general approach with compare function is good, but it
> should be a property of the fo_service. I don't think there is any
> scenario where we'd like to treat two servers with same user_data
> contents as the same.

^^^ Sorry, this should have read "two servers from the same service".

[Jakub Hrozek]

On Fri, Aug 10, 2012 at 07:12:38PM +0200, Jakub Hrozek wrote:
> On Fri, Aug 10, 2012 at 07:09:13PM +0200, Jakub Hrozek wrote:
> > On Fri, Aug 10, 2012 at 05:13:48PM +0200, Michal Zidek wrote:
> > > This should fix the duplicate detection. Patch is attached. Please
> > > review with caution, i was in hurry.
> > >
> > > https://fedorahosted.org/sssd/ticket/1472
> > >
> > > Thanks
> > > Michal
> >
> > I think the general approach with compare function is good, but it
> > should be a property of the fo_service. I don't think there is any
> > scenario where we'd like to treat two servers with same user_data
> > contents as the same.
>
> ^^^ Sorry, this should have read "two servers from the same service".

You also need to amend the testsuite as well when changing the fo_
interface, otherwise the test suite doesn't work:

CC     src/resolv/fail_over_tests-async_resolv.o
src/tests/fail_over-tests.c: In function ‘test_fo_resolve_service’:
src/tests/fail_over-tests.c:233:5: warning: passing argument 5 of ‘fo_add_server’ makes pointer from integer without a cast [enabled by default]
In file included from ./src/providers/dp_backend.h:26:0,
                 from ./src/providers/ldap/sdap.h:25,
                 from ./src/tests/common.h:31,
                 from src/tests/fail_over-tests.c:35:
./src/providers/fail_over.h:120:5: note: expected ‘int (*)(void *, void *)’ but argument is of type ‘int’
src/tests/fail_over-tests.c:233:5: error: too few arguments to function ‘fo_add_server’

[Michal Židek]

On 08/13/2012 11:13 AM, Jakub Hrozek wrote:
> On Fri, Aug 10, 2012 at 07:12:38PM +0200, Jakub Hrozek wrote:
> > On Fri, Aug 10, 2012 at 07:09:13PM +0200, Jakub Hrozek wrote:
> > > On Fri, Aug 10, 2012 at 05:13:48PM +0200, Michal Zidek wrote:
> > > > This should fix the duplicate detection. Patch is attached. Please
> > > > review with caution, i was in hurry.
> > > >
> > > > https://fedorahosted.org/sssd/ticket/1472
> > > >
> > > > Thanks
> > > > Michal
> > >
> > > I think the general approach with compare function is good, but it
> > > should be a property of the fo_service. I don't think there is any
> > > scenario where we'd like to treat two servers with same user_data
> > > contents as the same.
> >
> > ^^^ Sorry, this should have read "two servers from the same service".
>
> You also need to amend the testsuite as well when changing the fo_
> interface, otherwise the test suite doesn't work:
>
> CC     src/resolv/fail_over_tests-async_resolv.o
> src/tests/fail_over-tests.c: In function ‘test_fo_resolve_service’:
> src/tests/fail_over-tests.c:233:5: warning: passing argument 5 of ‘fo_add_server’ makes pointer from integer without a cast [enabled by default]
> In file included from ./src/providers/dp_backend.h:26:0,
>                   from ./src/providers/ldap/sdap.h:25,
>                   from ./src/tests/common.h:31,
>                   from src/tests/fail_over-tests.c:35:
> ./src/providers/fail_over.h:120:5: note: expected ‘int (*)(void *, void *)’ but argument is of type ‘int’
> src/tests/fail_over-tests.c:233:5: error: too few arguments to function ‘fo_add_server’

Ok. I rewrote it, so the user_data_cmp is now part of fo_service.
If it is set to NULL, user_data is not used as part of duplicity detection.
Also testsuite is now updated.

New patch is attached.

Thanks
Michal

[Ondrej Kos]

Nack

Working, but hard to read, use typedef for those strcmp casts.

O.

On 08/13/2012 05:25 PM, Michal Zidek wrote:
> On 08/13/2012 11:13 AM, Jakub Hrozek wrote:
> > On Fri, Aug 10, 2012 at 07:12:38PM +0200, Jakub Hrozek wrote:
> > > On Fri, Aug 10, 2012 at 07:09:13PM +0200, Jakub Hrozek wrote:
> > > > On Fri, Aug 10, 2012 at 05:13:48PM +0200, Michal Zidek wrote:
> > > > > This should fix the duplicate detection. Patch is attached. Please
> > > > > review with caution, i was in hurry.
> > > > >
> > > > > https://fedorahosted.org/sssd/ticket/1472
> > > > >
> > > > > Thanks
> > > > > Michal
> > > >
> > > > I think the general approach with compare function is good, but it
> > > > should be a property of the fo_service. I don't think there is any
> > > > scenario where we'd like to treat two servers with same user_data
> > > > contents as the same.
> > >
> > > ^^^ Sorry, this should have read "two servers from the same service".
> >
> > You also need to amend the testsuite as well when changing the fo_
> > interface, otherwise the test suite doesn't work:
> >
> > CC     src/resolv/fail_over_tests-async_resolv.o
> > src/tests/fail_over-tests.c: In function ‘test_fo_resolve_service’:
> > src/tests/fail_over-tests.c:233:5: warning: passing argument 5 of
> > ‘fo_add_server’ makes pointer from integer without a cast [enabled by
> > default]
> > In file included from ./src/providers/dp_backend.h:26:0,
> >                   from ./src/providers/ldap/sdap.h:25,
> >                   from ./src/tests/common.h:31,
> >                   from src/tests/fail_over-tests.c:35:
> > ./src/providers/fail_over.h:120:5: note: expected ‘int (*)(void *,
> > void *)’ but argument is of type ‘int’
> > src/tests/fail_over-tests.c:233:5: error: too few arguments to
> > function ‘fo_add_server’
>
> Ok. I rewrote it, so the user_data_cmp is now part of fo_service.
> If it is set to NULL, user_data is not used as part of duplicity detection.
> Also testsuite is now updated.
>
> New patch is attached.
>
> Thanks
> Michal
>
> sssd-devel mailing list
> [e-mail]
> https://lists.fedorahosted.org/mailman/listinfo/sssd-devel

[Dmitri Pal]

On 08/14/2012 04:54 AM, Ondrej Kos wrote:
> Nack
>
> Working, but hard to read, use typedef for those strcmp casts.

FYI
http://www.freeipa.org/page/Coding_Style#Use_of_Typedefs
> O.
>
> On 08/13/2012 05:25 PM, Michal Zidek wrote:
> > On 08/13/2012 11:13 AM, Jakub Hrozek wrote:
> > > On Fri, Aug 10, 2012 at 07:12:38PM +0200, Jakub Hrozek wrote:
> > > > On Fri, Aug 10, 2012 at 07:09:13PM +0200, Jakub Hrozek wrote:
> > > > > On Fri, Aug 10, 2012 at 05:13:48PM +0200, Michal Zidek wrote:
> > > > > > This should fix the duplicate detection. Patch is attached. Please
> > > > > > review with caution, i was in hurry.
> > > > > >
> > > > > > https://fedorahosted.org/sssd/ticket/1472
> > > > > >
> > > > > > Thanks
> > > > > > Michal
> > > > >
> > > > > I think the general approach with compare function is good, but it
> > > > > should be a property of the fo_service. I don't think there is any
> > > > > scenario where we'd like to treat two servers with same user_data
> > > > > contents as the same.
> > > >
> > > > ^^^ Sorry, this should have read "two servers from the same service".
> > >
> > > You also need to amend the testsuite as well when changing the fo_
> > > interface, otherwise the test suite doesn't work:
> > >
> > > CC     src/resolv/fail_over_tests-async_resolv.o
> > > src/tests/fail_over-tests.c: In function ‘test_fo_resolve_service’:
> > > src/tests/fail_over-tests.c:233:5: warning: passing argument 5 of
> > > ‘fo_add_server’ makes pointer from integer without a cast [enabled by
> > > default]
> > > In file included from ./src/providers/dp_backend.h:26:0,
> > >                   from ./src/providers/ldap/sdap.h:25,
> > >                   from ./src/tests/common.h:31,
> > >                   from src/tests/fail_over-tests.c:35:
> > > ./src/providers/fail_over.h:120:5: note: expected ‘int (*)(void *,
> > > void *)’ but argument is of type ‘int’
> > > src/tests/fail_over-tests.c:233:5: error: too few arguments to
> > > function ‘fo_add_server’
> >
> > Ok. I rewrote it, so the user_data_cmp is now part of fo_service.
> > If it is set to NULL, user_data is not used as part of duplicity
> > detection.
> > Also testsuite is now updated.
> >
> > New patch is attached.
> >
> > Thanks
> > Michal
> >
> > sssd-devel mailing list
> > [e-mail]
> > https://lists.fedorahosted.org/mailman/listinfo/sssd-devel

[Jakub Hrozek]

On Tue, Aug 14, 2012 at 08:24:08AM -0400, Dmitri Pal wrote:
> On 08/14/2012 04:54 AM, Ondrej Kos wrote:
> > Nack
> >
> > Working, but hard to read, use typedef for those strcmp casts.

Ondrej's explanation on list was quite short, more details were provided in
person, along with a reminder that the explanation needs to be more
verbose next time so that it's easily consumable just by following the
list.
> FYI
> http://www.freeipa.org/page/Coding_Style#Use_of_Typedefs

As per the typedefs, we forbid the use of typedefs for data types, which
would obscure the real type underneath. However, the patch used a
pointer to a function to provide a generic comparison mechanism.

When using a function pointer, we use typedefs, copying the full
function header would be ugly.

Something like:
    typedef int (*cmp_fn)(void *, void *);

Is allowed, because consumers of the compare functions can use it quite
easily:
    int do_compare(void *p1, void *p2, cmp_fn fn);

I think our coding style needs to me amended to explicitly state that we
only forbid typedefs when declaring data types.

[Michal Židek]

On 08/14/2012 10:54 AM, Ondrej Kos wrote:
> Nack
>
> Working, but hard to read, use typedef for those strcmp casts.
>
> O.

Ok. I now use typedef for the function type to make it easier to read. I 
also created separated functions
to compare user_data for each provider, so it will be easy to change it 
without affecting other providers
(right now all compare functions are only wrappers around strcmp).
> On 08/13/2012 05:25 PM, Michal Zidek wrote:
> > On 08/13/2012 11:13 AM, Jakub Hrozek wrote:
> > > On Fri, Aug 10, 2012 at 07:12:38PM +0200, Jakub Hrozek wrote:
> > > > On Fri, Aug 10, 2012 at 07:09:13PM +0200, Jakub Hrozek wrote:
> > > > > On Fri, Aug 10, 2012 at 05:13:48PM +0200, Michal Zidek wrote:
> > > > > > This should fix the duplicate detection. Patch is attached. Please
> > > > > > review with caution, i was in hurry.
> > > > > >
> > > > > > https://fedorahosted.org/sssd/ticket/1472
> > > > > >
> > > > > > Thanks
> > > > > > Michal
> > > > >
> > > > > I think the general approach with compare function is good, but it
> > > > > should be a property of the fo_service. I don't think there is any
> > > > > scenario where we'd like to treat two servers with same user_data
> > > > > contents as the same.
> > > >
> > > > ^^^ Sorry, this should have read "two servers from the same service".
> > >
> > > You also need to amend the testsuite as well when changing the fo_
> > > interface, otherwise the test suite doesn't work:
> > >
> > > CC     src/resolv/fail_over_tests-async_resolv.o
> > > src/tests/fail_over-tests.c: In function ‘test_fo_resolve_service’:
> > > src/tests/fail_over-tests.c:233:5: warning: passing argument 5 of
> > > ‘fo_add_server’ makes pointer from integer without a cast [enabled by
> > > default]
> > > In file included from ./src/providers/dp_backend.h:26:0,
> > >                   from ./src/providers/ldap/sdap.h:25,
> > >                   from ./src/tests/common.h:31,
> > >                   from src/tests/fail_over-tests.c:35:
> > > ./src/providers/fail_over.h:120:5: note: expected ‘int (*)(void *,
> > > void *)’ but argument is of type ‘int’
> > > src/tests/fail_over-tests.c:233:5: error: too few arguments to
> > > function ‘fo_add_server’
> >
> > Ok. I rewrote it, so the user_data_cmp is now part of fo_service.
> > If it is set to NULL, user_data is not used as part of duplicity 
> > detection.
> > Also testsuite is now updated.
> >
> > New patch is attached.
> >
> > Thanks
> > Michal
> >
> > sssd-devel mailing list
> > [e-mail]
> > https://lists.fedorahosted.org/mailman/listinfo/sssd-devel

New patch is attached.

Thanks
Michal

[Ondrej Kos]

Ack

On 08/14/2012 03:40 PM, Michal Zidek wrote:
> On 08/14/2012 10:54 AM, Ondrej Kos wrote:
> > Nack
> >
> > Working, but hard to read, use typedef for those strcmp casts.
> >
> > O.
>
> Ok. I now use typedef for the function type to make it easier to read. I
> also created separated functions
> to compare user_data for each provider, so it will be easy to change it
> without affecting other providers
> (right now all compare functions are only wrappers around strcmp).
> > On 08/13/2012 05:25 PM, Michal Zidek wrote:
> > > On 08/13/2012 11:13 AM, Jakub Hrozek wrote:
> > > > On Fri, Aug 10, 2012 at 07:12:38PM +0200, Jakub Hrozek wrote:
> > > > > On Fri, Aug 10, 2012 at 07:09:13PM +0200, Jakub Hrozek wrote:
> > > > > > On Fri, Aug 10, 2012 at 05:13:48PM +0200, Michal Zidek wrote:
> > > > > > > This should fix the duplicate detection. Patch is attached. Please
> > > > > > > review with caution, i was in hurry.
> > > > > > >
> > > > > > > https://fedorahosted.org/sssd/ticket/1472
> > > > > > >
> > > > > > > Thanks
> > > > > > > Michal
> > > > > > I think the general approach with compare function is good, but it
> > > > > > should be a property of the fo_service. I don't think there is any
> > > > > > scenario where we'd like to treat two servers with same user_data
> > > > > > contents as the same.
> > > > >
> > > > > ^^^ Sorry, this should have read "two servers from the same service".
> > > >
> > > > You also need to amend the testsuite as well when changing the fo_
> > > > interface, otherwise the test suite doesn't work:
> > > >
> > > > CC     src/resolv/fail_over_tests-async_resolv.o
> > > > src/tests/fail_over-tests.c: In function ‘test_fo_resolve_service’:
> > > > src/tests/fail_over-tests.c:233:5: warning: passing argument 5 of
> > > > ‘fo_add_server’ makes pointer from integer without a cast [enabled by
> > > > default]
> > > > In file included from ./src/providers/dp_backend.h:26:0,
> > > >                   from ./src/providers/ldap/sdap.h:25,
> > > >                   from ./src/tests/common.h:31,
> > > >                   from src/tests/fail_over-tests.c:35:
> > > > ./src/providers/fail_over.h:120:5: note: expected ‘int (*)(void *,
> > > > void *)’ but argument is of type ‘int’
> > > > src/tests/fail_over-tests.c:233:5: error: too few arguments to
> > > > function ‘fo_add_server’
> > >
> > > Ok. I rewrote it, so the user_data_cmp is now part of fo_service.
> > > If it is set to NULL, user_data is not used as part of duplicity
> > > detection.
> > > Also testsuite is now updated.
> > >
> > > New patch is attached.
> > >
> > > Thanks
> > > Michal
> > >
> > > sssd-devel mailing list
> > > [e-mail]
> > > https://lists.fedorahosted.org/mailman/listinfo/sssd-devel
>
> New patch is attached.
>
> Thanks
> Michal
>
> sssd-devel mailing list
> [e-mail]
> https://lists.fedorahosted.org/mailman/listinfo/sssd-devel

[Jakub Hrozek]

On Wed, Aug 15, 2012 at 11:02:15AM +0200, Ondrej Kos wrote:
> Ack

Two more nitpicks, please use strcasecmp for comparing host names and
make functions that are not exported in a header file static.
> On 08/14/2012 03:40 PM, Michal Zidek wrote:
> > On 08/14/2012 10:54 AM, Ondrej Kos wrote:
> > > Nack
> > >
> > > Working, but hard to read, use typedef for those strcmp casts.
> > >
> > > O.
> >
> > Ok. I now use typedef for the function type to make it easier to read. I
> > also created separated functions
> > to compare user_data for each provider, so it will be easy to change it
> > without affecting other providers
> > (right now all compare functions are only wrappers around strcmp).
> > > On 08/13/2012 05:25 PM, Michal Zidek wrote:
> > > > On 08/13/2012 11:13 AM, Jakub Hrozek wrote:
> > > > > On Fri, Aug 10, 2012 at 07:12:38PM +0200, Jakub Hrozek wrote:
> > > > > > On Fri, Aug 10, 2012 at 07:09:13PM +0200, Jakub Hrozek wrote:
> > > > > > >On Fri, Aug 10, 2012 at 05:13:48PM +0200, Michal Zidek wrote:
> > > > > > >>This should fix the duplicate detection. Patch is attached. Please
> > > > > > >>review with caution, i was in hurry.
> > > > > > >>
> > > > > > >>https://fedorahosted.org/sssd/ticket/1472
> > > > > > >>
> > > > > > >>Thanks
> > > > > > >>Michal
> > > > > > >I think the general approach with compare function is good, but it
> > > > > > >should be a property of the fo_service. I don't think there is any
> > > > > > >scenario where we'd like to treat two servers with same user_data
> > > > > > >contents as the same.
> > > > > > ^^^ Sorry, this should have read "two servers from the same service".
> > > > >
> > > > > You also need to amend the testsuite as well when changing the fo_
> > > > > interface, otherwise the test suite doesn't work:
> > > > >
> > > > > CC     src/resolv/fail_over_tests-async_resolv.o
> > > > > src/tests/fail_over-tests.c: In function ‘test_fo_resolve_service’:
> > > > > src/tests/fail_over-tests.c:233:5: warning: passing argument 5 of
> > > > > ‘fo_add_server’ makes pointer from integer without a cast [enabled by
> > > > > default]
> > > > > In file included from ./src/providers/dp_backend.h:26:0,
> > > > >                  from ./src/providers/ldap/sdap.h:25,
> > > > >                  from ./src/tests/common.h:31,
> > > > >                  from src/tests/fail_over-tests.c:35:
> > > > > ./src/providers/fail_over.h:120:5: note: expected ‘int (*)(void *,
> > > > > void *)’ but argument is of type ‘int’
> > > > > src/tests/fail_over-tests.c:233:5: error: too few arguments to
> > > > > function ‘fo_add_server’
> > > >
> > > > Ok. I rewrote it, so the user_data_cmp is now part of fo_service.
> > > > If it is set to NULL, user_data is not used as part of duplicity
> > > > detection.
> > > > Also testsuite is now updated.
> > > >
> > > > New patch is attached.
> > > >
> > > > Thanks
> > > > Michal
> > > >
> > > > sssd-devel mailing list
> > > > [e-mail]
> > > > https://lists.fedorahosted.org/mailman/listinfo/sssd-devel
> >
> > New patch is attached.
> >
> > Thanks
> > Michal
> >
> > sssd-devel mailing list
> > [e-mail]
> > https://lists.fedorahosted.org/mailman/listinfo/sssd-devel
>
> -- 
> Ondrej Kos
> Associate Software Engineer
> Identity Management
> Red Hat Czech
>
> cell:  [phone]
> phone: [phone]
> ext.:  82-62558
> irc:   okos @ #brno
> _______________________________________________
> sssd-devel mailing list
> [e-mail]
> https://lists.fedorahosted.org/mailman/listinfo/sssd-devel

[Michal Židek]

On 08/15/2012 01:17 PM, Jakub Hrozek wrote:
> On Wed, Aug 15, 2012 at 11:02:15AM +0200, Ondrej Kos wrote:
> > Ack
>
> Two more nitpicks, please use strcasecmp for comparing host names and
> make functions that are not exported in a header file static.

Done.
> > On 08/14/2012 03:40 PM, Michal Zidek wrote:
> > > On 08/14/2012 10:54 AM, Ondrej Kos wrote:
> > > > Nack
> > > >
> > > > Working, but hard to read, use typedef for those strcmp casts.
> > > >
> > > > O.
> > >
> > > Ok. I now use typedef for the function type to make it easier to read. I
> > > also created separated functions
> > > to compare user_data for each provider, so it will be easy to change it
> > > without affecting other providers
> > > (right now all compare functions are only wrappers around strcmp).
> > > > On 08/13/2012 05:25 PM, Michal Zidek wrote:
> > > > > On 08/13/2012 11:13 AM, Jakub Hrozek wrote:
> > > > > > On Fri, Aug 10, 2012 at 07:12:38PM +0200, Jakub Hrozek wrote:
> > > > > > > On Fri, Aug 10, 2012 at 07:09:13PM +0200, Jakub Hrozek wrote:
> > > > > > >> On Fri, Aug 10, 2012 at 05:13:48PM +0200, Michal Zidek wrote:
> > > > > > >>> This should fix the duplicate detection. Patch is attached. Please
> > > > > > >>> review with caution, i was in hurry.
> > > > > > >>>
> > > > > > >>> https://fedorahosted.org/sssd/ticket/1472
> > > > > > >>>
> > > > > > >>> Thanks
> > > > > > >>> Michal
> > > > > > >> I think the general approach with compare function is good, but it
> > > > > > >> should be a property of the fo_service. I don't think there is any
> > > > > > >> scenario where we'd like to treat two servers with same user_data
> > > > > > >> contents as the same.
> > > > > > > ^^^ Sorry, this should have read "two servers from the same service".
> > > > > > You also need to amend the testsuite as well when changing the fo_
> > > > > > interface, otherwise the test suite doesn't work:
> > > > > >
> > > > > > CC     src/resolv/fail_over_tests-async_resolv.o
> > > > > > src/tests/fail_over-tests.c: In function ‘test_fo_resolve_service’:
> > > > > > src/tests/fail_over-tests.c:233:5: warning: passing argument 5 of
> > > > > > ‘fo_add_server’ makes pointer from integer without a cast [enabled by
> > > > > > default]
> > > > > > In file included from ./src/providers/dp_backend.h:26:0,
> > > > > >                   from ./src/providers/ldap/sdap.h:25,
> > > > > >                   from ./src/tests/common.h:31,
> > > > > >                   from src/tests/fail_over-tests.c:35:
> > > > > > ./src/providers/fail_over.h:120:5: note: expected ‘int (*)(void *,
> > > > > > void *)’ but argument is of type ‘int’
> > > > > > src/tests/fail_over-tests.c:233:5: error: too few arguments to
> > > > > > function ‘fo_add_server’
> > > > >
> > > > > Ok. I rewrote it, so the user_data_cmp is now part of fo_service.
> > > > > If it is set to NULL, user_data is not used as part of duplicity
> > > > > detection.
> > > > > Also testsuite is now updated.
> > > > >
> > > > > New patch is attached.
> > > > >
> > > > > Thanks
> > > > > Michal
> > >
> > > New patch is attached.
> > >
> > > Thanks
> > > Michal

New patch is attached.

Thanks
Michal

[Jakub Hrozek]

On Wed, Aug 15, 2012 at 01:38:03PM +0200, Michal Zidek wrote:
> On 08/15/2012 01:17 PM, Jakub Hrozek wrote:
> > On Wed, Aug 15, 2012 at 11:02:15AM +0200, Ondrej Kos wrote:
> > > Ack
> >
> > Two more nitpicks, please use strcasecmp for comparing host names and
> > make functions that are not exported in a header file static.
>
> Done.

Ack

[Jakub Hrozek]

On Wed, Aug 15, 2012 at 01:49:57PM +0200, Jakub Hrozek wrote:
> On Wed, Aug 15, 2012 at 01:38:03PM +0200, Michal Zidek wrote:
> > On 08/15/2012 01:17 PM, Jakub Hrozek wrote:
> > > On Wed, Aug 15, 2012 at 11:02:15AM +0200, Ondrej Kos wrote:
> > > > Ack
> > >
> > > Two more nitpicks, please use strcasecmp for comparing host names and
> > > make functions that are not exported in a header file static.
> >
> > Done.
>
> Ack

Pushed to master.