lseek(2) BSD buglet

lseek(2) BSD buglet

[der Mouse]

lseek(2) (noticed on 3.23, 3.25 has similar wording) says, in part,

       SVr1-3 returns long instead of off_t, BSD returns int.

While the BSDs are not a single monolithic entity, NetBSD has returned
off_t from lseek(2) since 1.5 at the very latest, meaning late 2000.
I'd be surprised if there's a non-orphaned BSD left that still uses
int.

I'm not sure what wording would be best.  Depends on whether you think
it's worth mentioning the historical practice from the days when BSD
deserved the B in its name or not, I'd say.  Perhaps something like

	..., BSD from Berkeley returned int but modern BSD variants
	have switched to off_t.

or

	..., most (all?) BSDs use off_t.

or some such.

Note, if you want to write back to me for any reason, you will not get
through from gmail (nor hotmail nor yahoo webmail; there seems to be
something about running a big webmailer that's incompatible with
running a civilized shop).  Depending on where you're sending from, you
may get better results from with mouse-S783fYmB3Ccdnm+yROfE0A@public.gmane.org and/or
mouse-N7zJMUF80UksA/PxXw9srA@public.gmane.org - actually, sending to all three might not be a bad
idea.

Of course, if you don't need to write back, that's no issue. :)

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse-NcIlfYlhY41Ekk/fuQG8RGD2FQJk+8+b@public.gmane.org
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B
--
To unsubscribe from this list: send the line "unsubscribe linux-man" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: lseek(2) BSD buglet

[Christoph Hellwig]

On Wed, Aug 11, 2010 at 02:13:35PM -0400, der Mouse wrote:
> lseek(2) (noticed on 3.23, 3.25 has similar wording) says, in part,
> 
>        SVr1-3 returns long instead of off_t, BSD returns int.
> 
> While the BSDs are not a single monolithic entity, NetBSD has returned
> off_t from lseek(2) since 1.5 at the very latest, meaning late 2000.
> I'd be surprised if there's a non-orphaned BSD left that still uses
> int.
> 
> I'm not sure what wording would be best.  Depends on whether you think
> it's worth mentioning the historical practice from the days when BSD
> deserved the B in its name or not, I'd say.  Perhaps something like

I think removing it is fine.  Historical SysV and BSD behaviour might have
been interesting in the early 90s, but these days it's something for a
history book and not a manual page.
--
To unsubscribe from this list: send the line "unsubscribe linux-man" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: lseek(2) BSD buglet

[der Mouse]

>> lseek(2) (noticed on 3.23, 3.25 has similar wording) says, in part,

>>        SVr1-3 returns long instead of off_t, BSD returns int.

> I think removing it is fine.  Historical SysV and BSD behaviour might
> have been interesting in the early 90s, but these days it's something
> for a history book and not a manual page.

Sounds like a reasonable point of view to me.

However, since the manpages are also one of the places people go when
trying to figure out old code, I'd probably not delete it entirely but
instead replace it with something alojng the lines of

	Hisorical practice has varied, but modern systems uniformly use
	off_t.

But, really, even that may be more than is worth bothering with; it's
been a long time since lseek returning non-off_t mattered to anyone but
hardcore retrocomputing geeks.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse-NcIlfYlhY41Ekk/fuQG8RGD2FQJk+8+b@public.gmane.org
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B
--
To unsubscribe from this list: send the line "unsubscribe linux-man" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: lseek(2) BSD buglet

[Michael Kerrisk]

Hi,

On Sat, Aug 14, 2010 at 4:09 PM, der Mouse <mouse-NcIlfYlhY41Ekk/fuQG8RGD2FQJk+8+b@public.gmane.org> wrote:
>>> lseek(2) (noticed on 3.23, 3.25 has similar wording) says, in part,
>
>>>        SVr1-3 returns long instead of off_t, BSD returns int.
>
>> I think removing it is fine.  Historical SysV and BSD behaviour might
>> have been interesting in the early 90s, but these days it's something
>> for a history book and not a manual page.

I agree. I removed this entire line.

> Sounds like a reasonable point of view to me.
>
> However, since the manpages are also one of the places people go when
> trying to figure out old code, I'd probably not delete it entirely but
> instead replace it with something alojng the lines of
>
>        Hisorical practice has varied, but modern systems uniformly use
>        off_t.
>
> But, really, even that may be more than is worth bothering with; it's
> been a long time since lseek returning non-off_t mattered to anyone but
> hardcore retrocomputing geeks.

Yes.

Thanks for the note. The change will be in Linux man-pages-3.27.

Cheers,

Michael

-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Author of "The Linux Programming Interface"; http://man7.org/tlpi/
--
To unsubscribe from this list: send the line "unsubscribe linux-man" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html