* gettext-tools/src/message.h (enum format_type): Add format_sh_printf.
(NFORMATS): Increment.
* gettext-tools/src/message.c (format_language, format_language_pretty): Add an
entry for format_sh_printf.
* gettext-tools/src/format.h (formatstring_sh_printf): New declaration.
* gettext-tools/src/format.c (formatstring_parsers): Add an entry for
format_sh_printf.
* gettext-tools/src/format-sh-printf.c: New file, based on
gettext-tools/src/format-awk.c.
* gettext-tools/src/FILES: Mention it.
* gettext-tools/src/x-sh.h (SCANNERS_SH): Use formatstring_sh_printf as
secondary format string type.
* gettext-tools/src/xgettext.c (xgettext_record_flag): Update accordingly.
* gettext-tools/src/x-sh.c (init_flag_table_sh): Register gettext, ngettext with
flag 'pass-sh-printf-format'. Register 'printf' with flag 'sh-printf-format'.
* gettext-tools/src/Makefile.am (FORMAT_SOURCE): Add format-sh-printf.c.
* gettext-tools/libgettextpo/Makefile.am (libgettextpo_la_AUXSOURCES): Likewise.
* gettext-tools/doc/gettext.texi (PO Files): Mention sh-printf-format.
(sh-format): Document also the sh-printf-format strings.
* gettext-tools/doc/lang-sh.texi (sh): Mention the coreutils 'printf' command.
* gettext-tools/tests/xgettext-sh-1: Add a test case with a printf invocation.
* gettext-tools/tests/format-sh-printf-1: New file, based on
gettext-tools/tests/format-awk-1.
* gettext-tools/tests/format-sh-printf-2: New file, based on
gettext-tools/tests/format-awk-2.
* gettext-tools/tests/Makefile.am (TESTS): Add them.
* NEWS: Mention the change.
in a context that requires a format string. You can override this
heuristic by using a comment of the form /* xgettext: c-format */.
* Shell:
+ - xgettext now recognizes format strings in the 'printf' command syntax.
+ They are marked as 'sh-printf-format' in POT and PO files.
- xgettext now recognizes the \c, \u, and \U escape sequences in dollar-
single-quoted strings $'...'.
@kwindex sh-format@r{ flag}
@itemx no-sh-format
@kwindex no-sh-format@r{ flag}
-Likewise for Shell, see @ref{sh-format}.
+Likewise for Shell format strings, see @ref{sh-format}.
+
+@item sh-printf-format
+@kwindex sh-printf-format@r{ flag}
+@itemx no-sh-printf-format
+@kwindex no-sh-printf-format@r{ flag}
+Likewise for Shell @code{printf} format strings, see @ref{sh-format}.
@item awk-format
@kwindex awk-format@r{ flag}
@node sh-format
@subsection Shell Format Strings
+There are two kinds of format strings in shell scripts:
+those with dollar notation for placeholders,
+called @emph{Shell format strings}
+and labelled as @samp{sh-format},
+and those acceptable to the @samp{printf} command (or shell built-in command),
+called @emph{Shell @code{printf} format strings}
+and labelled as @samp{sh-printf-format}.
+
Shell format strings, as supported by GNU gettext and the @samp{envsubst}
program, are strings with references to shell variables in the form
@code{$@var{variable}} or @code{$@{@var{variable}@}}. References of the form
ASCII characters, not start with a digit and be nonempty; otherwise such
a variable reference is ignored.
+Shell @code{printf} format strings are the format strings supported
+by the POSIX @samp{printf} command
+(@url{https://pubs.opengroup.org/onlinepubs/9799919799/utilities/printf.html}),
+including the floating-point conversion specifiers
+@code{a}, @code{A}, @code{e}, @code{E}, @code{f}, @code{F}, @code{g}, @code{G},
+but without the obsolescent @code{b} conversion specifier.
+Extensions by the GNU coreutils @samp{printf} command
+(@url{https://www.gnu.org/software/coreutils/manual/html_node/printf-invocation.html})
+are not supported:
+use of @samp{*} or @samp{*@var{m}$} as width or precision;
+use of size specifiers @code{h}, @code{l}, @code{j}, @code{z}, @code{t} (ignored);
+and the escape sequences @code{\c},
+@code{\x@var{nn}}, @code{\u@var{nnnn}}, @code{\U@var{nnnnnnnn}}.
+Extensions by the GNU bash @samp{printf} built-in
+(@url{https://www.gnu.org/software/bash/manual/html_node/Bash-Builtins.html#index-printf})
+are not supported either:
+use of @samp{*} as width or precision;
+use of size specifiers @code{h}, @code{l}, @code{j}, @code{z}, @code{t} (ignored);
+the @code{%b}, @code{%q}, @code{%Q}, @code{%T}, @code{%n} directives;
+and the escape sequences
+@code{\x@var{nn}}, @code{\u@var{nnnn}}, @code{\U@var{nnnnnnnn}}.
+
@node awk-format
@subsection awk Format Strings
@c This file is part of the GNU gettext manual.
-@c Copyright (C) 1995-2024 Free Software Foundation, Inc.
+@c Copyright (C) 1995-2025 Free Software Foundation, Inc.
@c See the file gettext.texi for copying conditions.
@node sh
@code{xgettext}
@item Formatting with positions
----
-@c Not yet: It requires support in GNU coreutils, GNU bash, dash, etc.
-@c @url{https://pubs.opengroup.org/onlinepubs/9799919799/utilities/printf.html,
-@c @code{printf}}
+A POSIX compliant
+@url{https://pubs.opengroup.org/onlinepubs/9799919799/utilities/printf.html,
+ @code{printf}}
+command, such as the one from GNU coreutils 9.6 or newer.
+@c GNU Bash built-in?
@item Portability
fully portable
../src/format-go.c \
../src/format-ruby.c \
../src/format-sh.c \
+ ../src/format-sh-printf.c \
../src/format-awk.c \
../src/format-lua.c \
../src/format-pascal.c \
format-go.c Format string handling for Go.
format-ruby.c Format string handling for Ruby.
format-sh.c Format string handling for Shell.
+format-sh-printf.c Format string handling for Shell, printf syntax.
format-awk.c Format string handling for awk.
format-lua.c Format string handling for Lua.
format-pascal.c Format string handling for Object Pascal.
format-go.c \
format-ruby.c \
format-sh.c \
+ format-sh-printf.c \
format-awk.c \
format-lua.c \
format-pascal.c \
--- /dev/null
+/* Shell printf format strings.
+ Copyright (C) 2001-2025 Free Software Foundation, Inc.
+ Written by Bruno Haible <bruno@clisp.org>, 2025.
+
+ This program is free software: you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program 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 General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <https://www.gnu.org/licenses/>. */
+
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include <stdbool.h>
+#include <stdlib.h>
+
+#include "format.h"
+#include "c-ctype.h"
+#include "xalloc.h"
+#include "xvasprintf.h"
+#include "format-invalid.h"
+#include "gettext.h"
+
+#define _(str) gettext (str)
+
+/* Shell printf format strings are described in
+ * POSIX:
+ <https://pubs.opengroup.org/onlinepubs/9799919799/utilities/printf.html>
+ <https://pubs.opengroup.org/onlinepubs/9799919799/basedefs/V1_chap05.html#tag_05>
+ * The GNU coreutils documentation:
+ <https://www.gnu.org/software/coreutils/manual/html_node/printf-invocation.html>
+ * The GNU bash documentation:
+ <https://www.gnu.org/software/bash/manual/html_node/Bash-Builtins.html#index-printf>
+
+ The format string consists of
+ - plain text,
+ - directives, that start with '%',
+ - escape sequences, that start with a backslash and don't contain '%'.
+
+ The set of supported directives and escape sequences is documented in gettext.texi.
+
+ A directive
+ - starts with '%' or '%m$' where m is a positive integer,
+ - is optionally followed by any of the characters '#', '0', '-', ' ', '+',
+ each of which acts as a flag,
+ - is optionally followed by a width specification: a nonempty digit sequence,
+ [not in POSIX: '*' (reads an argument) or '*m$']
+ - is optionally followed by '.' and a precision specification: an optional
+ nonempty digit sequence,
+ [not in POSIX: '*' (reads an argument) or '*m$']
+ - [not in POSIX: is optionally followed by a size specifier, one of
+ 'hh' 'h' 'l' 'll' 'L' 'q' 'j' 'z' 't']
+ - is finished by a specifier
+ - 'c', that needs a character argument,
+ - 's', that needs a string argument,
+ - 'i', 'd', that need a signed integer argument,
+ - 'u', 'o', 'x', 'X', that need an unsigned integer argument,
+ - [optional in POSIX, but supported here:] 'e', 'E', 'f', 'F', 'g', 'G',
+ 'a', 'A', that need a floating-point argument.
+ Additionally there is the directive '%%', which takes no argument.
+ Numbered ('%m$' or '*m$') and unnumbered argument specifications cannot
+ be used in the same string.
+
+ The valid escape sequences are:
+ \\ \a \b \f \n \r \t \v
+ \nnn with 1 to 3 octal digits n
+ [not in POSIX: \c \xnn \unnnn \Unnnnnnnn]
+ */
+
+enum format_arg_type
+{
+ FAT_NONE,
+ FAT_CHARACTER,
+ FAT_STRING,
+ FAT_INTEGER,
+ FAT_UNSIGNED_INTEGER,
+ FAT_FLOAT
+};
+
+struct numbered_arg
+{
+ unsigned int number;
+ enum format_arg_type type;
+};
+
+struct spec
+{
+ unsigned int directives;
+ /* We consider a directive as "likely intentional" if it does not contain a
+ space. This prevents xgettext from flagging strings like "100% complete"
+ as 'sh-printf-format' if they don't occur in a context that requires a
+ format string. */
+ unsigned int likely_intentional_directives;
+ unsigned int numbered_arg_count;
+ struct numbered_arg *numbered;
+};
+
+
+static int
+numbered_arg_compare (const void *p1, const void *p2)
+{
+ unsigned int n1 = ((const struct numbered_arg *) p1)->number;
+ unsigned int n2 = ((const struct numbered_arg *) p2)->number;
+
+ return (n1 > n2 ? 1 : n1 < n2 ? -1 : 0);
+}
+
+static void *
+format_parse (const char *format, bool translated, char *fdi,
+ char **invalid_reason)
+{
+ const char *const format_start = format;
+ struct spec spec;
+ unsigned int numbered_allocated;
+ unsigned int unnumbered_arg_count;
+ struct spec *result;
+
+ spec.directives = 0;
+ spec.likely_intentional_directives = 0;
+ spec.numbered_arg_count = 0;
+ spec.numbered = NULL;
+ numbered_allocated = 0;
+ unnumbered_arg_count = 0;
+
+ for (; *format != '\0';)
+ /* Invariant: spec.numbered_arg_count == 0 || unnumbered_arg_count == 0. */
+ if (*format == '%')
+ {
+ /* A directive. */
+ bool likely_intentional = true;
+
+ FDI_SET (format, FMTDIR_START);
+ format++;
+ spec.directives++;
+
+ if (*format != '%')
+ {
+ unsigned int number = 0;
+ enum format_arg_type type;
+
+ if (c_isdigit (*format))
+ {
+ const char *f = format;
+ unsigned int m = 0;
+
+ do
+ {
+ m = 10 * m + (*f - '0');
+ f++;
+ }
+ while (c_isdigit (*f));
+
+ if (*f == '$')
+ {
+ if (m == 0)
+ {
+ *invalid_reason = INVALID_ARGNO_0 (spec.directives);
+ FDI_SET (f, FMTDIR_ERROR);
+ goto bad_format;
+ }
+ number = m;
+ format = ++f;
+ }
+ }
+
+ /* Parse flags. */
+ while (*format == ' ' || *format == '+' || *format == '-'
+ || *format == '#' || *format == '0')
+ {
+ if (*format == ' ')
+ likely_intentional = false;
+ format++;
+ }
+
+ /* Parse width. */
+ if (c_isdigit (*format))
+ {
+ do format++; while (c_isdigit (*format));
+ }
+
+ /* Parse precision. */
+ if (*format == '.')
+ {
+ format++;
+
+ while (c_isdigit (*format))
+ format++;
+ }
+
+ switch (*format)
+ {
+ case 'c':
+ type = FAT_CHARACTER;
+ break;
+ case 's':
+ type = FAT_STRING;
+ break;
+ case 'i': case 'd':
+ type = FAT_INTEGER;
+ break;
+ case 'u': case 'o': case 'x': case 'X':
+ type = FAT_UNSIGNED_INTEGER;
+ break;
+ case 'e': case 'E': case 'f': case 'F': case 'g': case 'G':
+ case 'a': case 'A':
+ type = FAT_FLOAT;
+ break;
+ default:
+ if (*format == '\0')
+ {
+ *invalid_reason = INVALID_UNTERMINATED_DIRECTIVE ();
+ FDI_SET (format - 1, FMTDIR_ERROR);
+ }
+ else
+ {
+ *invalid_reason =
+ INVALID_CONVERSION_SPECIFIER (spec.directives, *format);
+ FDI_SET (format, FMTDIR_ERROR);
+ }
+ goto bad_format;
+ }
+
+ if (number)
+ {
+ /* Numbered argument. */
+
+ /* Numbered and unnumbered specifications are exclusive. */
+ if (unnumbered_arg_count > 0)
+ {
+ *invalid_reason = INVALID_MIXES_NUMBERED_UNNUMBERED ();
+ FDI_SET (format, FMTDIR_ERROR);
+ goto bad_format;
+ }
+
+ if (numbered_allocated == spec.numbered_arg_count)
+ {
+ numbered_allocated = 2 * numbered_allocated + 1;
+ spec.numbered = (struct numbered_arg *) xrealloc (spec.numbered, numbered_allocated * sizeof (struct numbered_arg));
+ }
+ spec.numbered[spec.numbered_arg_count].number = number;
+ spec.numbered[spec.numbered_arg_count].type = type;
+ spec.numbered_arg_count++;
+ }
+ else
+ {
+ /* Unnumbered argument. */
+
+ /* Numbered and unnumbered specifications are exclusive. */
+ if (spec.numbered_arg_count > 0)
+ {
+ *invalid_reason = INVALID_MIXES_NUMBERED_UNNUMBERED ();
+ FDI_SET (format, FMTDIR_ERROR);
+ goto bad_format;
+ }
+
+ if (numbered_allocated == unnumbered_arg_count)
+ {
+ numbered_allocated = 2 * numbered_allocated + 1;
+ spec.numbered = (struct numbered_arg *) xrealloc (spec.numbered, numbered_allocated * sizeof (struct numbered_arg));
+ }
+ spec.numbered[unnumbered_arg_count].number = unnumbered_arg_count + 1;
+ spec.numbered[unnumbered_arg_count].type = type;
+ unnumbered_arg_count++;
+ }
+ }
+
+ if (likely_intentional)
+ spec.likely_intentional_directives++;
+ FDI_SET (format, FMTDIR_END);
+
+ format++;
+ }
+ else if (*format == '\\')
+ {
+ /* An escape sequence. */
+ FDI_SET (format, FMTDIR_START);
+ format++;
+
+ switch (*format)
+ {
+ case '\\':
+ case 'a':
+ case 'b':
+ case 'f':
+ case 'n':
+ case 'r':
+ case 't':
+ case 'v':
+ format++;
+ break;
+
+ case '0': case '1': case '2': case '3': case '4': case '5': case '6':
+ case '7':
+ format++;
+ if (*format >= '0' && *format <= '7')
+ {
+ format++;
+ if (*format >= '0' && *format <= '7')
+ format++;
+ }
+ break;
+
+ default:
+ if (*format == '\0')
+ {
+ *invalid_reason =
+ xstrdup (_("The string ends in the middle of an escape sequence."));
+ FDI_SET (format - 1, FMTDIR_ERROR);
+ }
+ else
+ {
+ *invalid_reason =
+ (c_isprint (*format)
+ ? ((*format == 'c'
+ || *format == 'x'
+ || *format == 'u' || *format == 'U')
+ ? xasprintf (_("The escape sequence '%c%c' is unsupported (not in POSIX)."), '\\', *format)
+ : xasprintf (_("The escape sequence '%c%c' is invalid."), '\\', *format))
+ : xstrdup (_("This escape sequence is invalid.")));
+ FDI_SET (format, FMTDIR_ERROR);
+ }
+ goto bad_format;
+ }
+ FDI_SET (format - 1, FMTDIR_END);
+ }
+ else
+ format++;
+
+ /* Convert the unnumbered argument array to numbered arguments. */
+ if (unnumbered_arg_count > 0)
+ spec.numbered_arg_count = unnumbered_arg_count;
+ /* Sort the numbered argument array, and eliminate duplicates. */
+ else if (spec.numbered_arg_count > 1)
+ {
+ unsigned int i, j;
+ bool err;
+
+ qsort (spec.numbered, spec.numbered_arg_count,
+ sizeof (struct numbered_arg), numbered_arg_compare);
+
+ /* Remove duplicates: Copy from i to j, keeping 0 <= j <= i. */
+ err = false;
+ for (i = j = 0; i < spec.numbered_arg_count; i++)
+ if (j > 0 && spec.numbered[i].number == spec.numbered[j-1].number)
+ {
+ enum format_arg_type type1 = spec.numbered[i].type;
+ enum format_arg_type type2 = spec.numbered[j-1].type;
+ enum format_arg_type type_both;
+
+ if (type1 == type2)
+ type_both = type1;
+ else
+ {
+ /* Incompatible types. */
+ type_both = FAT_NONE;
+ if (!err)
+ *invalid_reason =
+ INVALID_INCOMPATIBLE_ARG_TYPES (spec.numbered[i].number);
+ err = true;
+ }
+
+ spec.numbered[j-1].type = type_both;
+ }
+ else
+ {
+ if (j < i)
+ {
+ spec.numbered[j].number = spec.numbered[i].number;
+ spec.numbered[j].type = spec.numbered[i].type;
+ }
+ j++;
+ }
+ spec.numbered_arg_count = j;
+ if (err)
+ /* *invalid_reason has already been set above. */
+ goto bad_format;
+ }
+
+ result = XMALLOC (struct spec);
+ *result = spec;
+ return result;
+
+ bad_format:
+ if (spec.numbered != NULL)
+ free (spec.numbered);
+ return NULL;
+}
+
+static void
+format_free (void *descr)
+{
+ struct spec *spec = (struct spec *) descr;
+
+ if (spec->numbered != NULL)
+ free (spec->numbered);
+ free (spec);
+}
+
+static int
+format_get_number_of_directives (void *descr)
+{
+ struct spec *spec = (struct spec *) descr;
+
+ return spec->directives;
+}
+
+static bool
+format_is_unlikely_intentional (void *descr)
+{
+ struct spec *spec = (struct spec *) descr;
+
+ return spec->likely_intentional_directives == 0;
+}
+
+static bool
+format_check (void *msgid_descr, void *msgstr_descr, bool equality,
+ formatstring_error_logger_t error_logger, void *error_logger_data,
+ const char *pretty_msgid, const char *pretty_msgstr)
+{
+ struct spec *spec1 = (struct spec *) msgid_descr;
+ struct spec *spec2 = (struct spec *) msgstr_descr;
+ bool err = false;
+
+ if (spec1->numbered_arg_count + spec2->numbered_arg_count > 0)
+ {
+ unsigned int i, j;
+ unsigned int n1 = spec1->numbered_arg_count;
+ unsigned int n2 = spec2->numbered_arg_count;
+
+ /* Check that the argument numbers are the same.
+ Both arrays are sorted. We search for the first difference. */
+ for (i = 0, j = 0; i < n1 || j < n2; )
+ {
+ int cmp = (i >= n1 ? 1 :
+ j >= n2 ? -1 :
+ spec1->numbered[i].number > spec2->numbered[j].number ? 1 :
+ spec1->numbered[i].number < spec2->numbered[j].number ? -1 :
+ 0);
+
+ if (cmp > 0)
+ {
+ if (error_logger)
+ error_logger (error_logger_data,
+ _("a format specification for argument %u, as in '%s', doesn't exist in '%s'"),
+ spec2->numbered[j].number, pretty_msgstr,
+ pretty_msgid);
+ err = true;
+ break;
+ }
+ else if (cmp < 0)
+ {
+ if (equality)
+ {
+ if (error_logger)
+ error_logger (error_logger_data,
+ _("a format specification for argument %u doesn't exist in '%s'"),
+ spec1->numbered[i].number, pretty_msgstr);
+ err = true;
+ break;
+ }
+ else
+ i++;
+ }
+ else
+ j++, i++;
+ }
+ /* Check the argument types are the same. */
+ if (!err)
+ for (i = 0, j = 0; j < n2; )
+ {
+ if (spec1->numbered[i].number == spec2->numbered[j].number)
+ {
+ if (spec1->numbered[i].type != spec2->numbered[j].type)
+ {
+ if (error_logger)
+ error_logger (error_logger_data,
+ _("format specifications in '%s' and '%s' for argument %u are not the same"),
+ pretty_msgid, pretty_msgstr,
+ spec2->numbered[j].number);
+ err = true;
+ break;
+ }
+ j++, i++;
+ }
+ else
+ i++;
+ }
+ }
+
+ return err;
+}
+
+
+struct formatstring_parser formatstring_sh_printf =
+{
+ format_parse,
+ format_free,
+ format_get_number_of_directives,
+ format_is_unlikely_intentional,
+ format_check
+};
+
+
+#ifdef TEST
+
+/* Test program: Print the argument list specification returned by
+ format_parse for strings read from standard input. */
+
+#include <stdio.h>
+
+static void
+format_print (void *descr)
+{
+ struct spec *spec = (struct spec *) descr;
+ unsigned int last;
+ unsigned int i;
+
+ if (spec == NULL)
+ {
+ printf ("INVALID");
+ return;
+ }
+
+ printf ("(");
+ last = 1;
+ for (i = 0; i < spec->numbered_arg_count; i++)
+ {
+ unsigned int number = spec->numbered[i].number;
+
+ if (i > 0)
+ printf (" ");
+ if (number < last)
+ abort ();
+ for (; last < number; last++)
+ printf ("_ ");
+ switch (spec->numbered[i].type)
+ {
+ case FAT_CHARACTER:
+ printf ("c");
+ break;
+ case FAT_STRING:
+ printf ("s");
+ break;
+ case FAT_INTEGER:
+ printf ("i");
+ break;
+ case FAT_UNSIGNED_INTEGER:
+ printf ("[unsigned]i");
+ break;
+ case FAT_FLOAT:
+ printf ("f");
+ break;
+ default:
+ abort ();
+ }
+ last = number + 1;
+ }
+ printf (")");
+}
+
+int
+main ()
+{
+ for (;;)
+ {
+ char *line = NULL;
+ size_t line_size = 0;
+ int line_len;
+ char *invalid_reason;
+ void *descr;
+
+ line_len = getline (&line, &line_size, stdin);
+ if (line_len < 0)
+ break;
+ if (line_len > 0 && line[line_len - 1] == '\n')
+ line[--line_len] = '\0';
+
+ invalid_reason = NULL;
+ descr = format_parse (line, false, NULL, &invalid_reason);
+
+ format_print (descr);
+ printf ("\n");
+ if (descr == NULL)
+ printf ("%s\n", invalid_reason);
+
+ free (invalid_reason);
+ free (line);
+ }
+
+ return 0;
+}
+
+/*
+ * For Emacs M-x compile
+ * Local Variables:
+ * compile-command: "/bin/sh ../libtool --tag=CC --mode=link gcc -o a.out -static -O -g -Wall -I.. -I../gnulib-lib -I../../gettext-runtime/intl -DHAVE_CONFIG_H -DTEST format-sh-printf.c ../gnulib-lib/libgettextlib.la"
+ * End:
+ */
+
+#endif /* TEST */
/* format_go */ &formatstring_go,
/* format_ruby */ &formatstring_ruby,
/* format_sh */ &formatstring_sh,
+ /* format_sh_printf */ &formatstring_sh_printf,
/* format_awk */ &formatstring_awk,
/* format_lua */ &formatstring_lua,
/* format_pascal */ &formatstring_pascal,
extern LIBGETTEXTSRC_DLL_VARIABLE struct formatstring_parser formatstring_go;
extern LIBGETTEXTSRC_DLL_VARIABLE struct formatstring_parser formatstring_ruby;
extern LIBGETTEXTSRC_DLL_VARIABLE struct formatstring_parser formatstring_sh;
+extern LIBGETTEXTSRC_DLL_VARIABLE struct formatstring_parser formatstring_sh_printf;
extern LIBGETTEXTSRC_DLL_VARIABLE struct formatstring_parser formatstring_awk;
extern LIBGETTEXTSRC_DLL_VARIABLE struct formatstring_parser formatstring_lua;
extern LIBGETTEXTSRC_DLL_VARIABLE struct formatstring_parser formatstring_pascal;
/* format_go */ "go",
/* format_ruby */ "ruby",
/* format_sh */ "sh",
+ /* format_sh_printf */ "sh-printf",
/* format_awk */ "awk",
/* format_lua */ "lua",
/* format_pascal */ "object-pascal",
/* format_go */ "Go",
/* format_ruby */ "Ruby",
/* format_sh */ "Shell",
+ /* format_sh_printf */ "Shell printf",
/* format_awk */ "awk",
/* format_lua */ "Lua",
/* format_pascal */ "Object Pascal",
format_go,
format_ruby,
format_sh,
+ format_sh_printf,
format_awk,
format_lua,
format_pascal,
format_gfc_internal,
format_ycp
};
-#define NFORMATS 35 /* Number of format_type enum values. */
+#define NFORMATS 36 /* Number of format_type enum values. */
extern LIBGETTEXTSRC_DLL_VARIABLE const char *const format_language[NFORMATS];
extern LIBGETTEXTSRC_DLL_VARIABLE const char *const format_language_pretty[NFORMATS];
init_flag_table_sh ()
{
xgettext_record_flag ("gettext:1:pass-sh-format");
+ xgettext_record_flag ("gettext:1:pass-sh-printf-format");
xgettext_record_flag ("ngettext:1:pass-sh-format");
+ xgettext_record_flag ("ngettext:1:pass-sh-printf-format");
xgettext_record_flag ("ngettext:2:pass-sh-format");
+ xgettext_record_flag ("ngettext:2:pass-sh-printf-format");
xgettext_record_flag ("eval_gettext:1:sh-format");
xgettext_record_flag ("eval_ngettext:1:sh-format");
xgettext_record_flag ("eval_ngettext:2:sh-format");
xgettext_record_flag ("eval_pgettext:2:sh-format");
xgettext_record_flag ("eval_npgettext:2:sh-format");
xgettext_record_flag ("eval_npgettext:3:sh-format");
+ xgettext_record_flag ("printf:1:sh-printf-format");
}
/* xgettext sh backend.
- Copyright (C) 2003, 2006, 2014, 2018, 2020 Free Software Foundation, Inc.
+ Copyright (C) 2003-2025 Free Software Foundation, Inc.
Written by Bruno Haible <bruno@clisp.org>, 2003.
This program is free software: you can redistribute it and/or modify
#define SCANNERS_SH \
{ "Shell", extract_sh, NULL, \
- &flag_table_sh, &formatstring_sh, NULL }, \
+ &flag_table_sh, \
+ &formatstring_sh, &formatstring_sh_printf }, \
/* Scan a shell script file and add its translatable strings to mdlp. */
extern void extract_sh (FILE *fp, const char *real_filename,
name_start, name_end,
argnum, value, pass);
break;
+ case format_sh_printf:
+ flag_context_list_table_insert (&flag_table_sh, XFORMAT_SECONDARY,
+ name_start, name_end,
+ argnum, value, pass);
+ break;
case format_awk:
flag_context_list_table_insert (&flag_table_awk, XFORMAT_PRIMARY,
name_start, name_end,
format-rust-1 format-rust-2 \
format-scheme-1 format-scheme-2 \
format-sh-1 format-sh-2 \
+ format-sh-printf-1 format-sh-printf-2 \
format-tcl-1 format-tcl-2 format-tcl-3 \
format-ycp-1 format-ycp-2 \
plural-1 plural-2 plural-3 plural-4 \
--- /dev/null
+#! /bin/sh
+. "${srcdir=.}/init.sh"; path_prepend_ . ../src
+
+# Test recognition of Shell printf format strings.
+
+escape_backslashes='s/\\/\\\\/g'
+LC_ALL=C sed -e "$escape_backslashes" <<\EOF > f-sp-1.data
+# Valid: no argument
+"abc%%"
+# Valid: one character argument
+"abc%c"
+# Valid: one string argument
+"abc%s"
+# Valid: one integer argument
+"abc%i"
+# Valid: one integer argument
+"abc%d"
+# Valid: one integer argument
+"abc%o"
+# Valid: one integer argument
+"abc%u"
+# Valid: one integer argument
+"abc%x"
+# Valid: one integer argument
+"abc%X"
+# Valid: one floating-point argument
+"abc%e"
+# Valid: one floating-point argument
+"abc%E"
+# Valid: one floating-point argument
+"abc%f"
+# Valid: one floating-point argument
+"abc%F"
+# Valid: one floating-point argument
+"abc%g"
+# Valid: one floating-point argument
+"abc%G"
+# Valid: one floating-point argument
+"abc%a"
+# Valid: one floating-point argument
+"abc%A"
+# Valid: one argument with flags
+"abc%0#g"
+# Valid: one argument with width
+"abc%2g"
+# Invalid: one argument with width
+"abc%*g"
+# Valid: one argument with precision
+"abc%.4g"
+# Invalid: one argument with precision
+"abc%.*g"
+# Valid: one argument with width and precision
+"abc%14.4g"
+# Invalid: one argument with width and precision
+"abc%14.*g"
+# Invalid: one argument with width and precision
+"abc%*.4g"
+# Invalid: one argument with width and precision
+"abc%*.*g"
+# Invalid: unterminated
+"abc%"
+# Invalid: unknown format specifier
+"abc%y"
+# Invalid: flags after width
+"abc%*0g"
+# Valid: null precision
+"abc%.f"
+# Invalid: twice precision
+"abc%.4.2g"
+# Valid: three arguments
+"abc%d%u%u"
+# Valid: a numbered argument
+"abc%1$d"
+# Invalid: zero
+"abc%0$d"
+# Valid: two-digit numbered arguments
+"abc%11$def%10$dgh%9$dij%8$dkl%7$dmn%6$dop%5$dqr%4$dst%3$duv%2$dwx%1$dyz"
+# Invalid: unterminated number
+"abc%1"
+# Invalid: flags before number
+"abc%+1$d"
+# Valid: three arguments, two with same number
+"abc%1$4x,%2$c,%1$u"
+# Invalid: argument with conflicting types
+"abc%1$4x,%2$c,%1$s"
+# Valid: no conflict
+"abc%1$4x,%2$c,%1$u"
+# Invalid: mixing of numbered and unnumbered arguments
+"abc%d%2$x"
+# Valid: numbered argument with constant precision
+"abc%1$.9x"
+# Invalid: mixing of numbered and unnumbered arguments
+"abc%1$.*x"
+# Valid: missing non-final argument
+"abc%2$x%3$s"
+# Valid: permutation
+"abc%2$ddef%1$d"
+# Valid: multiple uses of same argument
+"abc%2$xdef%1$sghi%2$x"
+# Invalid: one argument with width
+"abc%2$#*1$g"
+# Invalid: one argument with width and precision
+"abc%3$*2$.*1$g"
+# Invalid: zero
+"abc%2$*0$.*1$g"
+# Valid: escape sequence
+"abc%%def\\"
+# Valid: escape sequence
+"abc%%def\a"
+# Valid: escape sequence
+"abc%%def\b"
+# Valid: escape sequence
+"abc%%def\f"
+# Valid: escape sequence
+"abc%%def\n"
+# Valid: escape sequence
+"abc%%def\r"
+# Valid: escape sequence
+"abc%%def\t"
+# Valid: escape sequence
+"abc%%def\v"
+# Valid: escape sequence
+"abc%%def\066"
+# Invalid: escape sequence
+"abc%%def\"
+# Invalid: escape sequence
+"abc%%def\""
+# Invalid: escape sequence
+"abc%%def\c"
+# Invalid: escape sequence
+"abc%%def\x32"
+# Invalid: escape sequence
+"abc%%def\u20ac"
+# Invalid: escape sequence
+"abc%%def\U0001F41C"
+# Invalid: escape sequence
+"abc%%def\%d"
+EOF
+
+: ${XGETTEXT=xgettext}
+n=0
+while read comment; do
+ # Note: The 'read' command processes backslashes. ('read -r' is not portable.)
+ read string
+ n=`expr $n + 1`
+ escape_backslashes='s/\\/\\\\/g'
+ escape_dollars='s/\$/\\\$/g'
+ string=`echo "$string" | LC_ALL=C sed -e "$escape_backslashes" -e "$escape_dollars"`
+ cat <<EOF > f-sp-1-$n.in
+gettext ${string};
+EOF
+ ${XGETTEXT} -L Shell -o f-sp-1-$n.po f-sp-1-$n.in || Exit 1
+ test -f f-sp-1-$n.po || Exit 1
+ fail=
+ if echo "$comment" | grep 'Valid:' > /dev/null; then
+ if grep sh-printf-format f-sp-1-$n.po > /dev/null; then
+ :
+ else
+ fail=yes
+ fi
+ else
+ if grep sh-printf-format f-sp-1-$n.po > /dev/null; then
+ fail=yes
+ else
+ :
+ fi
+ fi
+ if test -n "$fail"; then
+ echo "Format string recognition error:" 1>&2
+ cat f-sp-1-$n.in 1>&2
+ echo "Got:" 1>&2
+ cat f-sp-1-$n.po 1>&2
+ Exit 1
+ fi
+ rm -f f-sp-1-$n.in f-sp-1-$n.po
+done < f-sp-1.data
+
+Exit 0
--- /dev/null
+#! /bin/sh
+. "${srcdir=.}/init.sh"; path_prepend_ . ../src
+
+# Test checking of Shell printf format strings.
+
+cat <<\EOF > f-sp-2.data
+# Valid: %% doesn't count
+msgid "abc%%def"
+msgstr "xyz"
+# Invalid: invalid msgstr
+msgid "abc%%def"
+msgstr "xyz%"
+# Valid: same arguments
+msgid "abc%s%gdef"
+msgstr "xyz%s%g"
+# Valid: same arguments, with different widths
+msgid "abc%2sdef"
+msgstr "xyz%3s"
+# Valid: same arguments but in numbered syntax
+msgid "abc%s%gdef"
+msgstr "xyz%1$s%2$g"
+# Valid: permutation
+msgid "abc%s%g%cdef"
+msgstr "xyz%3$c%2$g%1$s"
+# Invalid: too few arguments
+msgid "abc%2$udef%1$s"
+msgstr "xyz%1$s"
+# Invalid: too few arguments
+msgid "abc%sdef%u"
+msgstr "xyz%s"
+# Invalid: too many arguments
+msgid "abc%udef"
+msgstr "xyz%uvw%c"
+# Valid: same numbered arguments, with different widths
+msgid "abc%2$5s%1$4s"
+msgstr "xyz%2$4s%1$5s"
+# Invalid: missing argument
+msgid "abc%2$sdef%1$u"
+msgstr "xyz%1$u"
+# Invalid: missing argument
+msgid "abc%1$sdef%2$u"
+msgstr "xyz%2$u"
+# Invalid: added argument
+msgid "abc%1$udef"
+msgstr "xyz%1$uvw%2$c"
+# Valid: type compatibility
+msgid "abc%i"
+msgstr "xyz%d"
+# Valid: type compatibility
+msgid "abc%o"
+msgstr "xyz%u"
+# Valid: type compatibility
+msgid "abc%u"
+msgstr "xyz%x"
+# Valid: type compatibility
+msgid "abc%u"
+msgstr "xyz%X"
+# Valid: type compatibility
+msgid "abc%e"
+msgstr "xyz%E"
+# Valid: type compatibility
+msgid "abc%e"
+msgstr "xyz%f"
+# Valid: type compatibility
+msgid "abc%e"
+msgstr "xyz%F"
+# Valid: type compatibility
+msgid "abc%e"
+msgstr "xyz%g"
+# Valid: type compatibility
+msgid "abc%e"
+msgstr "xyz%G"
+# Valid: type compatibility
+msgid "abc%e"
+msgstr "xyz%a"
+# Valid: type compatibility
+msgid "abc%e"
+msgstr "xyz%A"
+# Invalid: type incompatibility
+msgid "abc%c"
+msgstr "xyz%s"
+# Invalid: type incompatibility
+msgid "abc%c"
+msgstr "xyz%i"
+# Invalid: type incompatibility
+msgid "abc%c"
+msgstr "xyz%o"
+# Invalid: type incompatibility
+msgid "abc%c"
+msgstr "xyz%e"
+# Invalid: type incompatibility
+msgid "abc%s"
+msgstr "xyz%i"
+# Invalid: type incompatibility
+msgid "abc%s"
+msgstr "xyz%o"
+# Invalid: type incompatibility
+msgid "abc%s"
+msgstr "xyz%e"
+# Invalid: type incompatibility
+msgid "abc%i"
+msgstr "xyz%o"
+# Invalid: type incompatibility
+msgid "abc%i"
+msgstr "xyz%e"
+# Invalid: type incompatibility
+msgid "abc%u"
+msgstr "xyz%e"
+EOF
+
+: ${MSGFMT=msgfmt}
+n=0
+while read comment; do
+ read msgid_line
+ read msgstr_line
+ n=`expr $n + 1`
+ cat <<EOF > f-sp-2-$n.po
+#, sh-printf-format
+${msgid_line}
+${msgstr_line}
+EOF
+ fail=
+ if echo "$comment" | grep 'Valid:' > /dev/null; then
+ if ${MSGFMT} --check-format -o f-sp-2-$n.mo f-sp-2-$n.po; then
+ :
+ else
+ fail=yes
+ fi
+ else
+ ${MSGFMT} --check-format -o f-sp-2-$n.mo f-sp-2-$n.po 2> /dev/null
+ if test $? = 1; then
+ :
+ else
+ fail=yes
+ fi
+ fi
+ if test -n "$fail"; then
+ echo "Format string checking error:" 1>&2
+ cat f-sp-2-$n.po 1>&2
+ Exit 1
+ fi
+ rm -f f-sp-2-$n.po f-sp-2-$n.mo
+done < f-sp-2.data
+
+Exit 0
#!/bin/sh
. "${srcdir=.}/init.sh"; path_prepend_ . ../src
-# Test of Shell support: escape sequences, string concatenation,
+# Test of Shell support: escape sequences, format strings, string concatenation,
# strings with embedded expressions.
# Note! This file contains unescaped ASCII control characters. Edit carefully!
echo `echo \`gettext $'depth_2_dollar_posix_1_\cvab\cVcd\c[ef\c\\gh\c]ij\c?kl'\``
echo `echo \`gettext $'depth_2_dollar_bash_0_\Eab'\``
+# Test format strings.
+
+printf "`gettext 'User name: %s\nUser ID: %u'`"'\n' "$USER" `id -u`
+
# Test string concatenation.
gettext "concat_0_""part2"
msgid "depth_2_dollar_bash_0_\eab"
msgstr ""
+#, sh-printf-format
+msgid "User name: %s\\nUser ID: %u"
+msgstr ""
+
msgid "concat_0_part2"
msgstr ""
projects / thirdparty / gettext.git / commitdiff
