Hi Richie,
...
> > I wonder Petr, is it still necessary to define IFF_ECHO now only
> > <linux/if.h> is included? Or do they somehow symlink linux/if.h ->
> > net/if.h?
> No.
> > Indeed it seems the current version of uclibc-ng doesn't include
> > IFF_ECHO in <net/if.h>. OTOH musl does define it.
> Yes => please keep that definition in can_common.h (enough here, we don't have
> to bother with if.h when we don't have it yet).
> And I'll send tonight patch to uclibc-ng.
> The more mature musl is the less relevant uclibc-ng is.
Just for a record, I was wrong. Using <linux/if.h> is enough, no need to
keep definition on can_common.h. Sorry for wasting your time.
I mixed two problems: <linux/if.h> and <net/if.h> conflict (there are more headers
which conflict) [1] and sometimes missing definition on uclibc-ng.
*But* musl defines IFF_ECHO in <net/if.h> as they try to allow people not having
to depend on <linux/*.h>, which is IMHO better than blindly relying on
<linux/*.h> which glibc and uclibc{,-ng} (which follows glibc) does much more
than musl:
$ git grep '^#include <linux/.*\.h>' |wc -l
43 # glibc
37 # uclibc-ng
3 # musl
Kind regards,
Petr
[1] https://sourceware.org/glibc/wiki/Synchronizing_Headers#Known_Pairs_of_Headers_that_Conflict
