* [PATCH] printf: Add %B convespecifier for binary
@ 2020-04-27 13:35 Alejandro Colomar
2020-04-27 14:46 ` Florian Weimer
0 siblings, 1 reply; 9+ messages in thread
From: Alejandro Colomar @ 2020-04-27 13:35 UTC (permalink / raw)
To: libc-alpha, jrnieder
Hi all,
This patch adds a new feature to the ``printf`` family of functions:
``%B`` conversion specifier for printing unsigned numbers in binary.
Behaviour is exactly as with ``%X``, only changing the base (16 -> 2).
As Jonathan Nieder pointed out in a previous email, ``%b`` is already in
use by ``printf(1)``, so I didn't use it for binary.
I added some tests, and everything looks good.
Alex.
From b2bd279d76f2e729953a43fe8d619ceb927257ba Mon Sep 17 00:00:00 2001
From: Alejandro Colomar <colomar.6.4.3@gmail.com>
Date: Mon, 27 Apr 2020 15:10:25 +0200
Subject: [PATCH] printf: Add %B conversion specifier for binary
---
stdio-common/_itoa.c | 5 ++++
stdio-common/_itowa.c | 4 +++
stdio-common/_itowa.h | 1 +
stdio-common/tst-printf.c | 24 ++++++++++++++--
stdio-common/vfprintf-internal.c | 48 ++++++++++++++++++++------------
sysdeps/generic/_itoa.h | 1 +
6 files changed, 62 insertions(+), 21 deletions(-)
diff --git a/stdio-common/_itoa.c b/stdio-common/_itoa.c
index 5ab6bc982e..c0057eba00 100644
--- a/stdio-common/_itoa.c
+++ b/stdio-common/_itoa.c
@@ -179,6 +179,7 @@ _itoa_word (_ITOA_WORD_TYPE value, char *buflim,
SPECIAL (10);
SPECIAL (16);
SPECIAL (8);
+ SPECIAL (2);
default:
do
*--buflim = digits[value % base];
@@ -245,6 +246,10 @@ _itoa (unsigned long long int value, char *buflim,
unsigned int base,
while (work_hi != 0); \
} \
while (0)
+ case 2:
+ RUN_2N (1);
+ break;
+
case 8:
RUN_2N (3);
break;
diff --git a/stdio-common/_itowa.c b/stdio-common/_itowa.c
index 2b68a85dbe..7543656c3b 100644
--- a/stdio-common/_itowa.c
+++ b/stdio-common/_itowa.c
@@ -141,6 +141,10 @@ _itowa (unsigned long long int value, wchar_t
*buflim, unsigned int base,
while (work_hi != 0); \
} \
while (0)
+ case 2:
+ RUN_2N (1);
+ break;
+
case 8:
RUN_2N (3);
break;
diff --git a/stdio-common/_itowa.h b/stdio-common/_itowa.h
index 91c5fd4f11..b525c761fd 100644
--- a/stdio-common/_itowa.h
+++ b/stdio-common/_itowa.h
@@ -53,6 +53,7 @@ _itowa_word (_ITOA_WORD_TYPE value, wchar_t *buflim,
SPECIAL (10);
SPECIAL (16);
SPECIAL (8);
+ SPECIAL (2);
default:
do
*--bp = digits[value % base];
diff --git a/stdio-common/tst-printf.c b/stdio-common/tst-printf.c
index 52caf04e85..edcf6237cc 100644
--- a/stdio-common/tst-printf.c
+++ b/stdio-common/tst-printf.c
@@ -96,9 +96,19 @@ I am ready for my first lesson today.";
#endif
printf("decimal negative:\t\"%d\"\n", -2345);
+ /* FIXME: GCC 9 ignores `%B' yet */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT(9, "-Wformat-extra-args");
+ printf("binary negative:\t\"%B\"\n", -2345);
+ DIAG_POP_NEEDS_COMMENT;
printf("octal negative:\t\"%o\"\n", -2345);
printf("hex negative:\t\"%x\"\n", -2345);
printf("long decimal number:\t\"%ld\"\n", -123456L);
+ /* FIXME: GCC 9 ignores `%B' yet */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT(9, "-Wformat-extra-args");
+ printf("long binary negative:\t\"%lB\"\n", -2345L);
+ DIAG_POP_NEEDS_COMMENT;
printf("long octal negative:\t\"%lo\"\n", -2345L);
printf("long unsigned decimal number:\t\"%lu\"\n", -123456L);
printf("zero-padded LDN:\t\"%010ld\"\n", -123456L);
@@ -311,6 +321,14 @@ rfg2 (void)
sprintf (buf, "%09.*u", prec, 33);
if (strcmp (buf, " 0000033") != 0)
printf ("got: '%s', expected: '%s'\n", buf, " 0000033");
+ prec = 7;
+ /* FIXME: GCC 9 ignores `%B' yet */
+ DIAG_PUSH_NEEDS_COMMENT;
+ DIAG_IGNORE_NEEDS_COMMENT(9, "-Wformat-extra-args");
+ sprintf (buf, "%08.*B", prec, 33);
+ DIAG_POP_NEEDS_COMMENT;
+ if (strcmp (buf, " 0100001") != 0)
+ printf ("got: '%s', expected: '%s'\n", buf, " 0100001");
prec = 3;
sprintf (buf, "%04.*x", prec, 33);
if (strcmp (buf, " 021") != 0)
@@ -333,12 +351,12 @@ rfg3 (void)
int h = 1234;
sprintf (buf,
- "%1$*5$d %2$*6$hi %3$*7$lo %4$*8$f %9$*12$e %10$*13$g %11$*14$s",
+ "%1$*5$d %2$*6$hi %2$*7$hB %3$*7$lo %4$*8$f %9$*12$e %10$*13$g
%11$*14$s",
i, h, l, d, 8, 5, 14, 14, d, g, s, 14, 3, 14);
if (strcmp (buf,
- " 12345 1234 11145401322 321.765432 3.217654e+02 5
test-string") != 0)
+ " 12345 1234 10011010010 11145401322 321.765432
3.217654e+02 5 test-string") != 0)
printf ("got: '%s', expected: '%s'\n", buf,
- " 12345 1234 11145401322 321.765432 3.217654e+02 5
test-string");
+ " 12345 1234 10011010010 11145401322 321.765432
3.217654e+02 5 test-string");
}
#define TEST_FUNCTION do_test ()
diff --git a/stdio-common/vfprintf-internal.c
b/stdio-common/vfprintf-internal.c
index 3be92d4b6e..e31edbc75c 100644
--- a/stdio-common/vfprintf-internal.c
+++ b/stdio-common/vfprintf-internal.c
@@ -270,7 +270,7 @@ static const uint8_t jump_table[] =
/* '4' */ 8, /* '5' */ 8, /* '6' */ 8, /* '7' */ 8,
/* '8' */ 8, /* '9' */ 8, 0, 0,
0, 0, 0, 0,
- 0, /* 'A' */ 26, 0, /* 'C' */ 25,
+ 0, /* 'A' */ 26, /* 'B' */ 30, /* 'C' */ 25,
0, /* 'E' */ 19, /* F */ 19, /* 'G' */ 19,
0, /* 'I' */ 29, 0, 0,
/* 'L' */ 12, 0, 0, 0,
@@ -324,7 +324,7 @@ static const uint8_t jump_table[] =
#define STEP0_3_TABLE \
/* Step 0: at the beginning. */ \
- static JUMP_TABLE_TYPE step0_jumps[30] = \
+ static JUMP_TABLE_TYPE step0_jumps[31] = \
{ \
REF (form_unknown), \
REF (flag_space), /* for ' ' */ \
@@ -356,9 +356,10 @@ static const uint8_t jump_table[] =
REF (mod_ptrdiff_t), /* for 't' */ \
REF (mod_intmax_t), /* for 'j' */ \
REF (flag_i18n), /* for 'I' */ \
+ REF (form_bin) /* for 'B' */ \
}; \
/* Step 1: after processing width. */ \
- static JUMP_TABLE_TYPE step1_jumps[30] = \
+ static JUMP_TABLE_TYPE step1_jumps[31] = \
{ \
REF (form_unknown), \
REF (form_unknown), /* for ' ' */ \
@@ -389,10 +390,11 @@ static const uint8_t jump_table[] =
REF (form_floathex), /* for 'A', 'a' */ \
REF (mod_ptrdiff_t), /* for 't' */ \
REF (mod_intmax_t), /* for 'j' */ \
- REF (form_unknown) /* for 'I' */ \
+ REF (form_unknown), /* for 'I' */ \
+ REF (form_bin) /* for 'B' */ \
}; \
/* Step 2: after processing precision. */ \
- static JUMP_TABLE_TYPE step2_jumps[30] = \
+ static JUMP_TABLE_TYPE step2_jumps[31] = \
{ \
REF (form_unknown), \
REF (form_unknown), /* for ' ' */ \
@@ -423,10 +425,11 @@ static const uint8_t jump_table[] =
REF (form_floathex), /* for 'A', 'a' */ \
REF (mod_ptrdiff_t), /* for 't' */ \
REF (mod_intmax_t), /* for 'j' */ \
- REF (form_unknown) /* for 'I' */ \
+ REF (form_unknown), /* for 'I' */ \
+ REF (form_bin) /* for 'B' */ \
}; \
/* Step 3a: after processing first 'h' modifier. */ \
- static JUMP_TABLE_TYPE step3a_jumps[30] = \
+ static JUMP_TABLE_TYPE step3a_jumps[31] = \
{ \
REF (form_unknown), \
REF (form_unknown), /* for ' ' */ \
@@ -457,10 +460,11 @@ static const uint8_t jump_table[] =
REF (form_unknown), /* for 'A', 'a' */ \
REF (form_unknown), /* for 't' */ \
REF (form_unknown), /* for 'j' */ \
- REF (form_unknown) /* for 'I' */ \
+ REF (form_unknown), /* for 'I' */ \
+ REF (form_bin) /* for 'B' */ \
}; \
/* Step 3b: after processing first 'l' modifier. */ \
- static JUMP_TABLE_TYPE step3b_jumps[30] = \
+ static JUMP_TABLE_TYPE step3b_jumps[31] = \
{ \
REF (form_unknown), \
REF (form_unknown), /* for ' ' */ \
@@ -491,12 +495,13 @@ static const uint8_t jump_table[] =
REF (form_floathex), /* for 'A', 'a' */ \
REF (form_unknown), /* for 't' */ \
REF (form_unknown), /* for 'j' */ \
- REF (form_unknown) /* for 'I' */ \
+ REF (form_unknown), /* for 'I' */ \
+ REF (form_bin) /* for 'B' */ \
}
#define STEP4_TABLE \
/* Step 4: processing format specifier. */ \
- static JUMP_TABLE_TYPE step4_jumps[30] = \
+ static JUMP_TABLE_TYPE step4_jumps[31] = \
{ \
REF (form_unknown), \
REF (form_unknown), /* for ' ' */ \
@@ -527,7 +532,8 @@ static const uint8_t jump_table[] =
REF (form_floathex), /* for 'A', 'a' */ \
REF (form_unknown), /* for 't' */ \
REF (form_unknown), /* for 'j' */ \
- REF (form_unknown) /* for 'I' */ \
+ REF (form_unknown), /* for 'I' */ \
+ REF (form_bin) /* for 'B' */ \
}
@@ -597,6 +603,12 @@ static const uint8_t jump_table[] =
goto LABEL (unsigned_number); \
/* NOTREACHED */ \
\
+ LABEL (form_bin): \
+ /* Unsigned binary integer. */ \
+ base = 2; \
+ goto LABEL (unsigned_number); \
+ /* NOTREACHED */ \
+ \
LABEL (form_octal): \
/* Unsigned octal integer. */ \
base = 8; \
@@ -644,7 +656,7 @@ static const uint8_t jump_table[] =
{ \
/* Put the number in WORK. */ \
string = _itoa (number.longlong, workend, base, \
- spec == L_('X')); \
+ spec == L_('X') || spec == L_('B')); \
if (group && grouping) \
string = group_number (work_buffer, string, workend, \
grouping, thousands_sep); \
@@ -701,7 +713,7 @@ static const uint8_t jump_table[] =
{ \
/* Put the number in WORK. */ \
string = _itoa_word (number.word, workend, base, \
- spec == L_('X')); \
+ spec == L_('X') || spec == L_('B')); \
if (group && grouping) \
string = group_number (work_buffer, string, workend, \
grouping, thousands_sep); \
@@ -720,8 +732,8 @@ static const uint8_t jump_table[] =
{ \
width -= workend - string + prec; \
\
- if (number.word != 0 && alt && base == 16) \
- /* Account for 0X hex marker. */ \
+ if (number.word != 0 && alt && (base == 16 || base == 2)) \
+ /* Account for 0X hex marker or 0B bin marker. */ \
width -= 2; \
\
if (is_negative || showsign || space) \
@@ -740,7 +752,7 @@ static const uint8_t jump_table[] =
else if (space) \
outchar (L_(' ')); \
\
- if (number.word != 0 && alt && base == 16) \
+ if (number.word != 0 && alt && (base == 16 || base == 2)) \
{ \
outchar (L_('0')); \
outchar (spec); \
@@ -771,7 +783,7 @@ static const uint8_t jump_table[] =
--width; \
} \
\
- if (number.word != 0 && alt && base == 16) \
+ if (number.word != 0 && alt && (base == 16 || base == 2)) \
{ \
outchar (L_('0')); \
outchar (spec); \
diff --git a/sysdeps/generic/_itoa.h b/sysdeps/generic/_itoa.h
index 790f88a6d5..5d2a0720f6 100644
--- a/sysdeps/generic/_itoa.h
+++ b/sysdeps/generic/_itoa.h
@@ -76,6 +76,7 @@ _itoa_word (_ITOA_WORD_TYPE value, char *buflim,
SPECIAL (10);
SPECIAL (16);
SPECIAL (8);
+ SPECIAL (2);
default:
do
*--buflim = digits[value % base];
--
2.26.2
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH] printf: Add %B convespecifier for binary
2020-04-27 13:35 [PATCH] printf: Add %B convespecifier for binary Alejandro Colomar
@ 2020-04-27 14:46 ` Florian Weimer
2020-04-27 15:27 ` Andreas Schwab
` (2 more replies)
0 siblings, 3 replies; 9+ messages in thread
From: Florian Weimer @ 2020-04-27 14:46 UTC (permalink / raw)
To: Alejandro Colomar via Libc-alpha; +Cc: jrnieder, Alejandro Colomar
* Alejandro Colomar via Libc-alpha:
> This patch adds a new feature to the ``printf`` family of functions:
>
> ``%B`` conversion specifier for printing unsigned numbers in binary.
>
> Behaviour is exactly as with ``%X``, only changing the base (16 -> 2).
>
> As Jonathan Nieder pointed out in a previous email, ``%b`` is already in
> use by ``printf(1)``, so I didn't use it for binary.
Do any other libcs already have a such a feature?
In the past, we ran into problems when ISO C standardized specifies we
already used for something else. I do not know if there are any plans
to add binary literals to C (to match C++). If they were added, maybe
it would make sense to add the corresponding printf format specifier
at the same time.
To be honest, due to the present level of C standards activity, I
think it is very hard to extend ISO C functions at this time, I'm afraid.
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH] printf: Add %B convespecifier for binary
2020-04-27 14:46 ` Florian Weimer
@ 2020-04-27 15:27 ` Andreas Schwab
2020-04-27 15:40 ` Florian Weimer
2020-04-27 20:42 ` Joseph Myers
2021-10-20 16:57 ` Alejandro Colomar
2 siblings, 1 reply; 9+ messages in thread
From: Andreas Schwab @ 2020-04-27 15:27 UTC (permalink / raw)
To: Florian Weimer; +Cc: Alejandro Colomar via Libc-alpha, Alejandro Colomar
In any case, this needs documentation.
Andreas.
--
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510 2552 DF73 E780 A9DA AEC1
"And now for something completely different."
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH] printf: Add %B convespecifier for binary
2020-04-27 15:27 ` Andreas Schwab
@ 2020-04-27 15:40 ` Florian Weimer
2020-04-27 20:42 ` Jonathan Nieder
0 siblings, 1 reply; 9+ messages in thread
From: Florian Weimer @ 2020-04-27 15:40 UTC (permalink / raw)
To: Andreas Schwab; +Cc: Alejandro Colomar via Libc-alpha, Alejandro Colomar
* Andreas Schwab:
> In any case, this needs documentation.
True, but do you think we would accept such a change, without a
perspective towards C standardization?
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH] printf: Add %B convespecifier for binary
2020-04-27 15:40 ` Florian Weimer
@ 2020-04-27 20:42 ` Jonathan Nieder
0 siblings, 0 replies; 9+ messages in thread
From: Jonathan Nieder @ 2020-04-27 20:42 UTC (permalink / raw)
To: Florian Weimer
Cc: Andreas Schwab, Alejandro Colomar, Alejandro Colomar via Libc-alpha
Florian Weimer wrote:
> * Andreas Schwab:
>> In any case, this needs documentation.
>
> True, but do you think we would accept such a change, without a
> perspective towards C standardization?
In fact, even without the standardization question I don't think we
would accept it: the benefit doesn't justify the cost in complexity
(interop, added code, and so on).
Thanks,
Jonathan
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH] printf: Add %B convespecifier for binary
2020-04-27 14:46 ` Florian Weimer
2020-04-27 15:27 ` Andreas Schwab
@ 2020-04-27 20:42 ` Joseph Myers
2021-10-20 16:57 ` Alejandro Colomar
2 siblings, 0 replies; 9+ messages in thread
From: Joseph Myers @ 2020-04-27 20:42 UTC (permalink / raw)
To: Florian Weimer; +Cc: Alejandro Colomar via Libc-alpha, Alejandro Colomar
On Mon, 27 Apr 2020, Florian Weimer wrote:
> In the past, we ran into problems when ISO C standardized specifies we
> already used for something else. I do not know if there are any plans
And while the rule in the standard is that only lowercase letters are
reserved for future standardization as printf and scanf formats, in
practice uppercase letters have sometimes been added as well (C99 added
'F' and 'A'). When you allow for the fixed-point formats in TR 18037, the
only lowercase letters remaining available for standardization for both
printf and scanf are bqvwy (of which glibc is using 'q' in an obsolescent
extension, so if the standard added its own 'q' formats we'd need to add
new function variants or symbol versions).
--
Joseph S. Myers
joseph@codesourcery.com
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH] printf: Add %B convespecifier for binary
2020-04-27 14:46 ` Florian Weimer
2020-04-27 15:27 ` Andreas Schwab
2020-04-27 20:42 ` Joseph Myers
@ 2021-10-20 16:57 ` Alejandro Colomar
2021-10-20 17:04 ` Florian Weimer
2 siblings, 1 reply; 9+ messages in thread
From: Alejandro Colomar @ 2021-10-20 16:57 UTC (permalink / raw)
To: Florian Weimer, Libc-alpha
Hi Florian (and other glibc programmers),
On 4/27/20 4:46 PM, Florian Weimer wrote:
> * Alejandro Colomar via Libc-alpha:
>
>> This patch adds a new feature to the ``printf`` family of functions:
>>
>> ``%B`` conversion specifier for printing unsigned numbers in binary.
>>
>> Behaviour is exactly as with ``%X``, only changing the base (16 -> 2).
>>
>> As Jonathan Nieder pointed out in a previous email, ``%b`` is already in
>> use by ``printf(1)``, so I didn't use it for binary.
>
> Do any other libcs already have a such a feature?
No, AFAIK.
>
> In the past, we ran into problems when ISO C standardized specifies we
> already used for something else. I do not know if there are any plans
> to add binary literals to C (to match C++). If they were added, maybe
> it would make sense to add the corresponding printf format specifier
> at the same time.
C2x has added binary literals.
<http://open-std.org/jtc1/sc22/wg14/www/docs/n2596.pdf>
6.4.4.1 Integer constants
Syntax
1 integer-constant:
decimal-constant integer-suffixopt
octal-constant integer-suffixopt
hexadecimal-constant integer-suffixopt
binary-constant integer-suffixopt
decimal-constant:
nonzero-digit
decimal-constant digit
octal-constant:
0
octal-constant octal-digit
hexadecimal-constant:
hexadecimal-prefix hexadecimal-digit
hexadecimal-constant hexadecimal-digit
hexadecimal-prefix: one of
0x 0X
binary-prefix: one of
0b 0B
nonzero-digit: one of
1 2 3 4 5 6 7 8 9
octal-digit: one of
0 1 2 3 4 5 6 7
hexadecimal-digit: one of
0 1 2 3 4 5 6 7 8 9
a b c d e f
A B C D E F
binary-digit: one of
0 1
>
> To be honest, due to the present level of C standards activity, I
> think it is very hard to extend ISO C functions at this time, I'm afraid.
>
Any changes in the acceptance (or actually non-acceptance) after C2x
having added binary literals?
Thanks,
Alex
--
Alejandro Colomar
Linux man-pages comaintainer; https://www.kernel.org/doc/man-pages/
http://www.alejandro-colomar.es/
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH] printf: Add %B convespecifier for binary
2021-10-20 16:57 ` Alejandro Colomar
@ 2021-10-20 17:04 ` Florian Weimer
2021-10-20 17:56 ` Joseph Myers
0 siblings, 1 reply; 9+ messages in thread
From: Florian Weimer @ 2021-10-20 17:04 UTC (permalink / raw)
To: Alejandro Colomar; +Cc: Libc-alpha, jrnieder
* Alejandro Colomar:
> Any changes in the acceptance (or actually non-acceptance) after C2x
> having added binary literals?
Joseph posted a patch for the similar C2X feature, I think. It is
still awaiting review.
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH] printf: Add %B convespecifier for binary
2021-10-20 17:04 ` Florian Weimer
@ 2021-10-20 17:56 ` Joseph Myers
0 siblings, 0 replies; 9+ messages in thread
From: Joseph Myers @ 2021-10-20 17:56 UTC (permalink / raw)
To: Florian Weimer; +Cc: Alejandro Colomar, Libc-alpha
On Wed, 20 Oct 2021, Florian Weimer wrote:
> * Alejandro Colomar:
>
> > Any changes in the acceptance (or actually non-acceptance) after C2x
> > having added binary literals?
>
> Joseph posted a patch for the similar C2X feature, I think. It is
> still awaiting review.
Yes:
<https://sourceware.org/pipermail/libc-alpha/2021-October/131905.html>.
I haven't yet implemented the corresponding strtol / scanf support, which
as discussed will be more complicated, needing 32 new __isoc23_* entry
points (44 on platforms with two long double variants, 56 on powerpc64le)
to keep full compatibility with pre-C23 code not expecting 0b to start an
integer for strtol bases 0 or 2 or scanf %i (but I did the scanf format
checking support in GCC along with the printf format checking support).
--
Joseph S. Myers
joseph@codesourcery.com
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2021-10-20 17:56 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-04-27 13:35 [PATCH] printf: Add %B convespecifier for binary Alejandro Colomar
2020-04-27 14:46 ` Florian Weimer
2020-04-27 15:27 ` Andreas Schwab
2020-04-27 15:40 ` Florian Weimer
2020-04-27 20:42 ` Jonathan Nieder
2020-04-27 20:42 ` Joseph Myers
2021-10-20 16:57 ` Alejandro Colomar
2021-10-20 17:04 ` Florian Weimer
2021-10-20 17:56 ` Joseph Myers
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).