[Click] Two quick patches regarding ELEMENT_LIBS() and long long vs int64
Christian Kreibich
christian at icir.org
Fri Aug 31 14:04:40 EDT 2007
Hi Eddie,
On Fri, 2007-08-31 at 08:29 -0700, Eddie Kohler wrote:
> Hi Christian,
>
> Thanks and thanks!!
>
> First patch: The normal way to do this, for example in gcc's __attribute__
> syntax, is with an extra pair of parentheses, and this is what I have checked
> in -- preferable because you won't have problems with linker arguments that
> contain quotes.
Ah, much nicer. Thanks!
> Second patch: Why isn't HAVE_LONG_LONG defined on your system?!?!?!?!??!?!
> Can you look at config.log and tell me? In the meantime your patch is totally
> fine anyway and is checked in.
I looked into it a little more. It's my fault. I have a library that
parses a certain kind of configuration files, and I'm using it across a
few elements. The library uses Click classes, but is not a Click element
itself. It turns I erroneously used this:
#include <click/config-userlevel.h>
#include <click/config.h>
Changing it to
#include <click/config.h>
#include <click/config-userlevel.h>
solved the problem since HAVE_LONG_LONG is in fact alive and well --
sorry.
ps: I initially tried --disable-int64, but that caused the Click build
to break:
CXX ../lib/confparse.cc
../lib/confparse.cc: In function ‘bool cp_file_offset(const String&, off_t*)’:
../lib/confparse.cc:924: error: no matching function for call to ‘cp_integer(const String&, uint64_t*)’
../lib/confparse.cc:763: note: candidates are: const char* cp_integer(const char*, const char*, int, int32_t*)
[...]
--
Cheers,
Christian
More information about the click
mailing list