Re: apr_wait_for_io_or_timeout() decl in apr_support.h which is in public include directory

[Brad Nicholes <bn...@novell.com>]

It sounds like we need to expose it correctly now but immediately mark
it as deprecated.  Then in the next release of APR, hide it again.  It
just seems wrong and will continue to cause problems if we leave it half
hidden as it is now.

Brad

>>> Joe Orton <jo...@redhat.com>  >>>
On Sun, Nov 13, 2005 at 08:25:42AM -0500, Jeff Trawick wrote:
> apr_support.h has these comments...
> 
> /**
>  * @defgroup apr_support Internal APR support functions
>  * @ingroup APR
>  * @{
>  */
> 
> It must be for pragmatic reasons that apr_support.h is in the apr
> include directory instead of include/arch.  Perhaps
> include/arch/apr_private_common.h is more appropriate?

Ouch, yes, definitely.  I wonder whether we need to have an argument 
about whether unintentially exposed functions constitute part of the ABI

or not first before hiding this again...

joe

Re: apr_wait_for_io_or_timeout() decl in apr_support.h which is in public include directory

[Jeff Trawick <tr...@gmail.com>]

On 11/26/05, Brad Nicholes <bn...@novell.com> wrote:
> It sounds like we need to expose it correctly now but immediately mark
> it as deprecated.  Then in the next release of APR, hide it again.  It
> just seems wrong and will continue to cause problems if we leave it half
> hidden as it is now.

somebody stumbled on a header file, missed or ignored the "Internal"
comment, used it on a platform that doesn't require it to be
explicitly exported...  they're not going to care about a "deprecated"
comment either...  better they fix their code sooner than later since
it isn't portable anyway