[3900] in linux-net channel archive
Re: EVEN MORE 2.0 vs 1.2 and INN config.
daemon@ATHENA.MIT.EDU (Thomas Quinot)
Mon Jul 29 17:37:24 1996
To: submit-linux-dev-net@ratatosk.yggdrasil.com
From: thomas@cuivre.fdn.fr (Thomas Quinot)
Date: 29 Jul 1996 13:47:25 GMT
Miquel van Smoorenburg (miquels@drinkel.cistron.nl) =E9crit :
> > 3. for file locking , LOCK_STYLE select NONE, FLOCK, FCNTL or LO=
CKF
> > you have FCNTL, Debian also has FCNTL
> > while the FAQ suggests using FLOCK .... wonder which is best i=
f it
> >matters <?>
> flock() used to be emulated in the C library since Linux didn't have =
it.
Yup, this is why I picked FCNTL historically...
> > 4. for use waitpid instead of wait3 <?>
> > you have HAVE_WAITPID DONT
> > both the FAQ and Debian sets this to DO..... is DO the corre=
ct choice
>=20
> Again it doesn't matter. Chances are waitpid and wait3 are the same i=
nternally
> anyway (both emulated in the libc with a call to wait4).
They are.
> Linux doesn't have vfork, and in the BSD compat library headers there=
's just a
> #define vfork fork
Indeed. I ould change this one...
> sysconf() is just a libc wrapper that calls getrlimit as far as I kno=
w,
> so it's reasonable to just call getrlimit directly.
Actually sysconf(3) does not call getrlimit(2) (as of libc-5.3.12 that =
is),
but gives a hard-coded value (the value of OPEN_MAX at libc compile tim=
e),
so sysconf() is just not as good a solution as getrlimit IMHO...
> Yeah. I don't think it matters, and I can't remember why I used
> (isascii((c)) && ixXXXXX((c))).. perhaps it was needed with an olde=
r
> version of libc. I could take it out but it doesn't hurt I guess.
The current is* man page has the provision that "c" shall be an 8-bit
ASCII character or EOF, but it is not assumed that it shall be a 7-bit
ASCII character, so isacii() is not necessary.
--=20
Thomas.Quinot@Cuivre.FdN.FR <URL:http://Web.FdN.FR/~tquinot=
/>
