Re: Remaining Interfaces
Here're my comments regarding glibc. Some functions are GNU
extensions - do we want to add them?
Andrew Josey <ajosey@rdg.opengroup.org> writes:
> Enclosed is a markup for these in SUSV2. I'd suggest all the
> ones marked SUSV2 be considered with the exception of
> those marked removed in SUSV3.
I agree, we should remove them.
> On Feb 28, 5:15pm in "Remaining Interfaces", Stuart Anderson wrote:
> >
> > All,
> > We scrubbed a lot of lingering interfaces at the meeting
> > this week. Attached is a short list of the remaining interfaces that
> > need to be cleaned up. I see a fgew that I think I know what to do about,
> > but I need everyone else to looks over this list, and let me know what
> > the disposition of the interface should be. (Included or Excluded), and
> > what the reference Specification should be.
> >
> > asprintf
GNU extension, see <stdio.h>
> > closelog SUSV2
> > dprintf
GNU extension, see <stdio.h>
> > drem
<math.h>: comes from BSD 4.3. Since drem and remainder are the same
AFAIK, we should not include drem.
> > encrypt SUSV2
> > _errno
Internal glibc symbol, don't add it.
> > fstat SUSV2
> > fts_children
> > fts_close
> > fts_open
> > fts_read
> > fts_set
man fts_children tells me:
BSD 4.4. The fts utility is expected to be included in a future IEEE
Std1003.1-1988 (``POSIX'') revision.
We can add all of these.
> > ftw SUSV2
> > getchar_unlocked SUSV2 (POSIX.1c)
> > getc_unlocked SUSV2 (POSIX.1c)
> > getlogin SUSV2
> > gettimeofday SUSV2
> > glob SUSV2
> > glob64
Should be added.
> > globfree SUSV2
> > globfree64
Should be added
> > glob_pattern_p
GNU extension.
> > htonl SUSV2
> > htons SUSV2
> > ilogbf SUSV3 /c99
> > ilogbl SUSV3 /c99
> > inet_addr SUSV2
> > inet_aton
>From BSD 4.3
> > inet_lnaof SUSV2 (removed in SUSV3)
> > inet_makeaddr SUSV2 (removed in SUSv3)
> > inet_netof SUSV2 (removed in SUSV3)
> > inet_network SUSV2 (removed in SUSV3)
I agree, let's remove those.
> > inet_ntoa SUSV2
> > ioperm
> > iopl
ioperm and iopl are hardware specific <sys/io.h> has this:
/* If TURN_ON is TRUE, request for permission to do direct i/o on the
port numbers in the range [FROM,FROM+NUM-1]. Otherwise, turn I/O
permission off for that range. This call requires root privileges.
Portability note: not all Linux platforms support this call. Most
platforms based on the PC I/O architecture probably will, however.
E.g., Linux/Alpha for Alpha PCs supports this. */
Should we really add them?
> > j0f
> > j0l
> > j1f
> > j1l
> > jnf
> > jnl
These are extensions of j0,j1,jn - let's add them.
> > lockf SUSV2
> > _longjmp SUSV2
> > madvise (probably a relation to posix_madvise based
> on the draft before the posix_ prefix was added)
> > mknod SUSV2
> > ntohl SUSV2
> > ntohs SUSV2
> > scandir
BSD 4.3, <dirent.h>
> > scandir64
If we add scandir, we should add also scandir64.
> > _setjmp SUSV2
> > ssignal
The comment says:
/* SVID names for the same things. */
Let's remove it - we don't need duplicates
> > sysconf SUSV2
> > _sys_errlist
Mangle for sys_errlist, please add it.
> > tmpfile SUSV2
> > tmpnam SUSV2
> > tmpnam_r
Reentrant version of tmpnam - GNU extension.
> > ttyname SUSV2
> > ttyname_r SUSV2 (POSIX.1c)
> > uncompress
zlib?
> > vasprintf
> > vdprintf
both are GNU extension
> > zlibVersion
zlib?
Stuart, I can write some short manual entries as I did already. Send
me a list of functions that you can't find yourself documentation to
and I'll write something up.
Andreas
--
Andreas Jaeger
SuSE Labs aj@suse.de
private aj@arthur.inka.de
http://www.suse.de/~aj
Reply to: