[v6,04/11] stdio-common: Add __translated_number_width
Checks
| Context |
Check |
Description |
| dj/TryBot-apply_patch |
success
|
Patch applied to master at the time it was sent
|
Commit Message
This function will be used to compute the width of a number
after i18n digit translation.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
---
include/printf.h | 13 +++++++-
stdio-common/Makefile | 1 +
stdio-common/translated_number_width.c | 42 ++++++++++++++++++++++++++
3 files changed, 55 insertions(+), 1 deletion(-)
create mode 100644 stdio-common/translated_number_width.c
Comments
Hi,
On Fri, Dec 16, 2022 at 09:15:22PM +0100, Florian Weimer via Libc-alpha wrote:
> This function will be used to compute the width of a number
> after i18n digit translation.
I haven't bisected, but I suppose it was this changeset that introduced
a regression reported by the strace test suite.
glibc-2.36$ cat <<'EOF' |gcc -xc - && ./a.out
#include <stdio.h>
int main() { printf("%03d\n", 1); return 0; }
EOF
001
The same test on master prints 0001 instead of 001.
* Dmitry V. Levin:
> Hi,
>
> On Fri, Dec 16, 2022 at 09:15:22PM +0100, Florian Weimer via Libc-alpha wrote:
>> This function will be used to compute the width of a number
>> after i18n digit translation.
>
> I haven't bisected, but I suppose it was this changeset that introduced
> a regression reported by the strace test suite.
>
> glibc-2.36$ cat <<'EOF' |gcc -xc - && ./a.out
> #include <stdio.h>
> int main() { printf("%03d\n", 1); return 0; }
> EOF
> 001
>
> The same test on master prints 0001 instead of 001.
Sorry, I can't reproduce it. Did you build glibc yourself from upstream
sources, or did you get the build from somewhere else?
Thanks,
Florian
On Mon, Jan 30, 2023 at 11:42:21AM +0100, Florian Weimer via Libc-alpha wrote:
> * Dmitry V. Levin:
>
> > Hi,
> >
> > On Fri, Dec 16, 2022 at 09:15:22PM +0100, Florian Weimer via Libc-alpha wrote:
> >> This function will be used to compute the width of a number
> >> after i18n digit translation.
> >
> > I haven't bisected, but I suppose it was this changeset that introduced
> > a regression reported by the strace test suite.
> >
> > glibc-2.36$ cat <<'EOF' |gcc -xc - && ./a.out
> > #include <stdio.h>
> > int main() { printf("%03d\n", 1); return 0; }
> > EOF
> > 001
> >
> > The same test on master prints 0001 instead of 001.
>
> Sorry, I can't reproduce it. Did you build glibc yourself from upstream
> sources, or did you get the build from somewhere else?
This was initially reported by the strace test suite running on
rawhide-test.fedorainfracloud.org, and I suppose the patch submitted
by Andreas today fixes it.
* Dmitry V. Levin:
> On Mon, Jan 30, 2023 at 11:42:21AM +0100, Florian Weimer via Libc-alpha wrote:
>> * Dmitry V. Levin:
>>
>> > Hi,
>> >
>> > On Fri, Dec 16, 2022 at 09:15:22PM +0100, Florian Weimer via Libc-alpha wrote:
>> >> This function will be used to compute the width of a number
>> >> after i18n digit translation.
>> >
>> > I haven't bisected, but I suppose it was this changeset that introduced
>> > a regression reported by the strace test suite.
>> >
>> > glibc-2.36$ cat <<'EOF' |gcc -xc - && ./a.out
>> > #include <stdio.h>
>> > int main() { printf("%03d\n", 1); return 0; }
>> > EOF
>> > 001
>> >
>> > The same test on master prints 0001 instead of 001.
>>
>> Sorry, I can't reproduce it. Did you build glibc yourself from upstream
>> sources, or did you get the build from somewhere else?
>
> This was initially reported by the strace test suite running on
> rawhide-test.fedorainfracloud.org, and I suppose the patch submitted
> by Andreas today fixes it.
Andreas' fix is for %#03o, though. Perhaps the reduction above is
incorrect?
Thanks,
Florian
On Mon, Jan 30, 2023 at 12:03:24PM +0100, Florian Weimer wrote:
> * Dmitry V. Levin:
>
> > On Mon, Jan 30, 2023 at 11:42:21AM +0100, Florian Weimer via Libc-alpha wrote:
> >> * Dmitry V. Levin:
> >>
> >> > Hi,
> >> >
> >> > On Fri, Dec 16, 2022 at 09:15:22PM +0100, Florian Weimer via Libc-alpha wrote:
> >> >> This function will be used to compute the width of a number
> >> >> after i18n digit translation.
> >> >
> >> > I haven't bisected, but I suppose it was this changeset that introduced
> >> > a regression reported by the strace test suite.
> >> >
> >> > glibc-2.36$ cat <<'EOF' |gcc -xc - && ./a.out
> >> > #include <stdio.h>
> >> > int main() { printf("%03d\n", 1); return 0; }
> >> > EOF
> >> > 001
> >> >
> >> > The same test on master prints 0001 instead of 001.
> >>
> >> Sorry, I can't reproduce it. Did you build glibc yourself from upstream
> >> sources, or did you get the build from somewhere else?
> >
> > This was initially reported by the strace test suite running on
> > rawhide-test.fedorainfracloud.org, and I suppose the patch submitted
> > by Andreas today fixes it.
>
> Andreas' fix is for %#03o, though. Perhaps the reduction above is
> incorrect?
Oops, I posted the wrong reduction, it was originally %#03o indeed.
@@ -53,7 +53,18 @@ int __wprintf_function_invoke (void *, printf_function callback,
#include <bits/types/locale_t.h>
-/* Now define the internal interfaces. */
+/* Returns the width (as for printf, in bytes) of the converted ASCII
+ number in the characters in the range [FIRST, LAST). The range
+ must only contain ASCII digits. The caller is responsible for
+ avoiding overflow.
+
+ This function is used during non-wide digit translation. Wide
+ digit translate produces one wide character per ASCII digit,
+ so the width is simply LAST - FIRST. */
+int __translated_number_width (locale_t loc,
+ const char *first, const char *last)
+ attribute_hidden;
+
extern int __printf_fphex (FILE *, const struct printf_info *,
const void *const *) attribute_hidden;
extern int __printf_fp (FILE *, const struct printf_info *,
@@ -85,6 +85,7 @@ routines := \
tmpfile64 \
tmpnam \
tmpnam_r \
+ translated_number_width \
vfprintf \
vfprintf-internal \
vfscanf \
new file mode 100644
@@ -0,0 +1,42 @@
+/* Compute the printf width of a sequence of ASCII digits.
+ Copyright (C) 2022 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <https://www.gnu.org/licenses/>. */
+
+#include <assert.h>
+#include <limits.h>
+#include <locale/localeinfo.h>
+#include <printf.h>
+
+int
+__translated_number_width (locale_t loc, const char *first, const char *last)
+{
+ struct lc_ctype_data *ctype = loc->__locales[LC_CTYPE]->private;
+
+ if (ctype->outdigit_bytes_all_equal > 0)
+ return (last - first) * ctype->outdigit_bytes_all_equal;
+ else
+ {
+ /* Digits have varying length, so the fast path cannot be used. */
+ int digits = 0;
+ for (const char *p = first; p < last; ++p)
+ {
+ assert ('0' <= *p && *p <= '9');
+ digits += ctype->outdigit_bytes[*p - '0'];
+ }
+ return digits;
+ }
+}