On Jul 30, 2008, at 5:50 AM, Ralf S. Engelschall wrote:
On Wed, Jul 30, 2008, Jeff Johnson wrote:
This "band-aid" prevents installing argv.h in /usr/include/rpm w/o
messy
AutoFu.
Really? If no Autoconf #defines are present, _nothing_ happens. The
stuff degrades gracefully to just no #include at all. So, this should
not cause any problems.
All rpm include files should __NOT__ carry AutoFu #defines into /usr/
include.
Automagically including <stdint.h> is far worse than just forcing
applications to
include <stdint.h> (or inttypes.h or using their own typedefs's or
whatever)
Is it __REALLY_ that hard to get application developers to add
#include <stdint.h>
where needed?
Well, it was not my idea to include <stdint.h> there. I just wanted to
fix the build-time problems the new addition of <stdint.h> caused. I'm
happy to leave out everything (the status-quo as before <stdint.h> was
added), but I can also live with the current band-aid as it at least
degrades gracefully IMHO.
Instead of arguing abt <stdint.h> (which I will start ripping after
my 2nd cup of coffee
today, sigh), let's try the forward looking goal.
What arekm wants is __ALL__ rpm include files to be self consistent,
dragging
in everything necessary for the file to be included by itself.
No matter that rpm has __NEVER__ promised self-consistency.
The traditional promise for using the rpm API was
#include <rpmlib.h>
and nothing else.
In order to start refactoring code, I had to break that traditional
promise.
What is there instead is
#include <rpmcli.h>
which drags in every necessary rpm include file.
Refactoring the self consistency to __EVERY__ rpm include file has not
been attempted.
And its premature to just charge into the task of making __EVERY__ rpm
include file self consistent. I also think its pointless, a forked
project cannot
possibly pretend to an API, see the incompatible prototypes for
headcerGet()
between rpm.org and rpm5.org for at least one instance of insanity.
If you want an API (other than #include <rpmcli.h> dragging in all
other necessary
includes), then I will add private typedef's for ints to remove all
C99isms and otherwise
aggressively start refactoring rpm includes today.
I most definitely do __NOT__ want HAVE_SILLY_H defines in __ANY__ rpm
include files.
73 de Jeff
Ralf S. Engelschall
[EMAIL PROTECTED]
www.engelschall.com
______________________________________________________________________
RPM Package Manager http://rpm5.org
Developer Communication List rpm-devel@rpm5.org
______________________________________________________________________
RPM Package Manager http://rpm5.org
Developer Communication List rpm-devel@rpm5.org
