Re: printf.9: document limited conversion specification syntax
Yep, all good Jason McIntyre wrote: > yep, better than mine. ok > jmc > > On 4 April 2021 14:41:13 BST, Klemens Nanni wrote: > >On Sun, Apr 04, 2021 at 07:17:27AM +0100, Jason McIntyre wrote: > >> On Sun, Apr 04, 2021 at 01:22:23AM +0200, Klemens Nanni wrote: > >> > @@ -119,6 +119,17 @@ send formatted strings to the ddb console, and > >are onl > >> > Since each of these kernel functions is a variant of its user > >space > >> > counterpart, this page describes only the differences between the > >user > >> > space and kernel versions. > >> > +.Pp > >> > +The limited syntax of a conversion specification is: > >> > >> hi. i'm fine with your diff, though cannot really judge the actual > >> content. but the "limited syntax" text is a bit ambiguous. you could > >> still expect from it that a full syntax version is supported. > >Right, I do imply that nothing more is supported; this sentence is a > >cheap copy of printf(3)'s sentence > > > > The overall syntax of a conversion specification is: > > > >> maybe spell it out? > >> > >>printf(9) supports a subset of the conversion specification > >>available to printf(3): > >> > >> sth like that? > >Thanks, how about this? > > > >Only a subset of the user space conversion specification is available > >to > > the kernel version: > > > > %[width][size]conversion > > > > Refer to printf(3) for functional details. > > > > > > > >Index: share/man/man9/printf.9 > >=== > >RCS file: /cvs/src/share/man/man9/printf.9,v > >retrieving revision 1.24 > >diff -u -p -r1.24 printf.9 > >--- share/man/man9/printf.9 7 Nov 2015 03:48:25 - 1.24 > >+++ share/man/man9/printf.9 4 Apr 2021 13:39:02 - > >@@ -119,6 +119,18 @@ send formatted strings to the ddb consol > > Since each of these kernel functions is a variant of its user space > > counterpart, this page describes only the differences between the user > > space and kernel versions. > >+.Pp > >+Only a subset of the user space conversion specification is available > >to the > >+kernel version: > >+.Bd -filled -offset indent > >+.Sm off > >+.Cm % > >+.Op Ar width > >+.Op Ar size > >+.Ar conversion > >+.Sm on > >+.Ed > >+.Pp > > Refer to > > .Xr printf 3 > > for functional details. >
Re: printf.9: document limited conversion specification syntax
yep, better than mine. ok jmc On 4 April 2021 14:41:13 BST, Klemens Nanni wrote: >On Sun, Apr 04, 2021 at 07:17:27AM +0100, Jason McIntyre wrote: >> On Sun, Apr 04, 2021 at 01:22:23AM +0200, Klemens Nanni wrote: >> > @@ -119,6 +119,17 @@ send formatted strings to the ddb console, and >are onl >> > Since each of these kernel functions is a variant of its user >space >> > counterpart, this page describes only the differences between the >user >> > space and kernel versions. >> > +.Pp >> > +The limited syntax of a conversion specification is: >> >> hi. i'm fine with your diff, though cannot really judge the actual >> content. but the "limited syntax" text is a bit ambiguous. you could >> still expect from it that a full syntax version is supported. >Right, I do imply that nothing more is supported; this sentence is a >cheap copy of printf(3)'s sentence > > The overall syntax of a conversion specification is: > >> maybe spell it out? >> >> printf(9) supports a subset of the conversion specification >> available to printf(3): >> >> sth like that? >Thanks, how about this? > >Only a subset of the user space conversion specification is available >to > the kernel version: > > %[width][size]conversion > > Refer to printf(3) for functional details. > > > >Index: share/man/man9/printf.9 >=== >RCS file: /cvs/src/share/man/man9/printf.9,v >retrieving revision 1.24 >diff -u -p -r1.24 printf.9 >--- share/man/man9/printf.97 Nov 2015 03:48:25 - 1.24 >+++ share/man/man9/printf.94 Apr 2021 13:39:02 - >@@ -119,6 +119,18 @@ send formatted strings to the ddb consol > Since each of these kernel functions is a variant of its user space > counterpart, this page describes only the differences between the user > space and kernel versions. >+.Pp >+Only a subset of the user space conversion specification is available >to the >+kernel version: >+.Bd -filled -offset indent >+.Sm off >+.Cm % >+.Op Ar width >+.Op Ar size >+.Ar conversion >+.Sm on >+.Ed >+.Pp > Refer to > .Xr printf 3 > for functional details.
Re: printf.9: document limited conversion specification syntax
On Sun, Apr 04, 2021 at 07:17:27AM +0100, Jason McIntyre wrote: > On Sun, Apr 04, 2021 at 01:22:23AM +0200, Klemens Nanni wrote: > > @@ -119,6 +119,17 @@ send formatted strings to the ddb console, and are onl > > Since each of these kernel functions is a variant of its user space > > counterpart, this page describes only the differences between the user > > space and kernel versions. > > +.Pp > > +The limited syntax of a conversion specification is: > > hi. i'm fine with your diff, though cannot really judge the actual > content. but the "limited syntax" text is a bit ambiguous. you could > still expect from it that a full syntax version is supported. Right, I do imply that nothing more is supported; this sentence is a cheap copy of printf(3)'s sentence The overall syntax of a conversion specification is: > maybe spell it out? > > printf(9) supports a subset of the conversion specification > available to printf(3): > > sth like that? Thanks, how about this? Only a subset of the user space conversion specification is available to the kernel version: %[width][size]conversion Refer to printf(3) for functional details. Index: share/man/man9/printf.9 === RCS file: /cvs/src/share/man/man9/printf.9,v retrieving revision 1.24 diff -u -p -r1.24 printf.9 --- share/man/man9/printf.9 7 Nov 2015 03:48:25 - 1.24 +++ share/man/man9/printf.9 4 Apr 2021 13:39:02 - @@ -119,6 +119,18 @@ send formatted strings to the ddb consol Since each of these kernel functions is a variant of its user space counterpart, this page describes only the differences between the user space and kernel versions. +.Pp +Only a subset of the user space conversion specification is available to the +kernel version: +.Bd -filled -offset indent +.Sm off +.Cm % +.Op Ar width +.Op Ar size +.Ar conversion +.Sm on +.Ed +.Pp Refer to .Xr printf 3 for functional details.
Re: printf.9: document limited conversion specification syntax
On Sun, Apr 04, 2021 at 01:22:23AM +0200, Klemens Nanni wrote:
> print(9) says it only documents differences and points at printf(3) for
> identical/shared bits:
>
>DESCRIPTION
> [...]
>
> Since each of these kernel functions is a variant of its user space
> counterpart, this page describes only the differences between the user
> space and kernel versions. Refer to printf(3) for functional details.
>
>FORMAT OPTIONS
> The kernel functions don't support as many formatting specifiers as their
> user space counterparts. In addition to the floating point formatting
> specifiers, the following integer type specifiers are not supported:
>
> %hhArgument of char type. This format specifier is accepted by the
> kernel but will be handled as %h.
>
> %j Argument of intmax_t or uintmax_t type.
>
> %t Argument of ptrdiff_t type.
>
> The kernel functions support some additional formatting specifiers:
>
> %b Bit field expansion. [...]
>
>
> So reading this, I get the impression that print(3)'s entire conversion
> specifications syntax is supported (merely with a few less `conversion'
> specifiers):
>
> %[argno$][flags][width][.precision][size]conversion
>
> but printf(9) supports much fewer specifiers, afaict amounting to
>
> %[width][size]conversion
>
> That's fine, but code using printf(3) syntax might actually compile
> without warnings (like this example did for me) and be useless afterall:
>
> + printf("%1$s: node %2$x:\n"
> + " mclk_rate is %3$u (0x%3$x)\n",
> + __func__, sc->sc_node, mclk_rate);
>
> Apr 4 00:32:29 pine64 /bsd: mclk_rate is $u (0x $x)
>
>
> I don't fancy expanding syntax or adding checks to the kernel's printf.c
> but documenting it seems needed.
>
> Feedback? Objections? OK?
>
> diff 5b147494c5d594d5ec97ca57686c9e3527b87e29 /usr/src
> blob - c31fb0cb42e9af426c42306a54339ec7f7f375a7
> file + share/man/man9/printf.9
> --- share/man/man9/printf.9
> +++ share/man/man9/printf.9
> @@ -119,6 +119,17 @@ send formatted strings to the ddb console, and are onl
> Since each of these kernel functions is a variant of its user space
> counterpart, this page describes only the differences between the user
> space and kernel versions.
> +.Pp
> +The limited syntax of a conversion specification is:
hi. i'm fine with your diff, though cannot really judge the actual
content. but the "limited syntax" text is a bit ambiguous. you could
still expect from it that a full syntax version is supported.
maybe spell it out?
printf(9) supports a subset of the conversion specification
available to printf(3):
sth like that?
jmc
> +.Bd -filled -offset indent
> +.Sm off
> +.Cm %
> +.Op Ar width
> +.Op Ar size
> +.Ar conversion
> +.Sm on
> +.Ed
> +.Pp
> Refer to
> .Xr printf 3
> for functional details.
>
printf.9: document limited conversion specification syntax
print(9) says it only documents differences and points at printf(3) for
identical/shared bits:
DESCRIPTION
[...]
Since each of these kernel functions is a variant of its user space
counterpart, this page describes only the differences between the user
space and kernel versions. Refer to printf(3) for functional details.
FORMAT OPTIONS
The kernel functions don't support as many formatting specifiers as their
user space counterparts. In addition to the floating point formatting
specifiers, the following integer type specifiers are not supported:
%hhArgument of char type. This format specifier is accepted by the
kernel but will be handled as %h.
%j Argument of intmax_t or uintmax_t type.
%t Argument of ptrdiff_t type.
The kernel functions support some additional formatting specifiers:
%b Bit field expansion. [...]
So reading this, I get the impression that print(3)'s entire conversion
specifications syntax is supported (merely with a few less `conversion'
specifiers):
%[argno$][flags][width][.precision][size]conversion
but printf(9) supports much fewer specifiers, afaict amounting to
%[width][size]conversion
That's fine, but code using printf(3) syntax might actually compile
without warnings (like this example did for me) and be useless afterall:
+ printf("%1$s: node %2$x:\n"
+ " mclk_rate is %3$u (0x%3$x)\n",
+ __func__, sc->sc_node, mclk_rate);
Apr 4 00:32:29 pine64 /bsd: mclk_rate is $u (0x $x)
I don't fancy expanding syntax or adding checks to the kernel's printf.c
but documenting it seems needed.
Feedback? Objections? OK?
diff 5b147494c5d594d5ec97ca57686c9e3527b87e29 /usr/src
blob - c31fb0cb42e9af426c42306a54339ec7f7f375a7
file + share/man/man9/printf.9
--- share/man/man9/printf.9
+++ share/man/man9/printf.9
@@ -119,6 +119,17 @@ send formatted strings to the ddb console, and are onl
Since each of these kernel functions is a variant of its user space
counterpart, this page describes only the differences between the user
space and kernel versions.
+.Pp
+The limited syntax of a conversion specification is:
+.Bd -filled -offset indent
+.Sm off
+.Cm %
+.Op Ar width
+.Op Ar size
+.Ar conversion
+.Sm on
+.Ed
+.Pp
Refer to
.Xr printf 3
for functional details.
