bash-5.0 distribution sources and documentation

lib/sh/Makefile.in

@@ -67,8 +67,8 @@ LOCAL_DEFS = @LOCAL_DEFS@
 
 INCLUDES = -I. -I../.. -I$(topdir) -I$(topdir)/lib -I$(BASHINCDIR) -I$(srcdir) $(INTL_INC)
 
-CCFLAGS = ${PROFILE_FLAGS} ${INCLUDES} $(DEFS) $(LOCAL_DEFS) $(LOCAL_CFLAGS) \
-	  $(CFLAGS) $(CPPFLAGS) 
+CCFLAGS = ${ADDON_CFLAGS} ${PROFILE_FLAGS} ${INCLUDES} $(DEFS) $(LOCAL_DEFS) \
+	  $(LOCAL_CFLAGS) $(CFLAGS) $(CPPFLAGS) 
 
 GCC_LINT_FLAGS = -Wall -Wshadow -Wpointer-arith -Wcast-qual \
 		 -Wcast-align -Wstrict-prototypes -Wconversion \
@@ -92,7 +92,8 @@ CSOURCES = clktck.c clock.c getcwd.c getenv.c oslib.c setlinebuf.c \
 	   mktime.c strftime.c mbschr.c zcatfd.c zmapfd.c winsize.c eaccess.c \
 	   wcsdup.c fpurge.c zgetline.c mbscmp.c uconvert.c ufuncs.c \
 	   casemod.c dprintf.c input_avail.c mbscasecmp.c fnxform.c \
-	   strchrnul.c unicode.c wcswidth.c wcsnwidth.c shmbchar.c strdup.c
+	   strchrnul.c unicode.c wcswidth.c wcsnwidth.c shmbchar.c strdup.c \
+	   utf8.c
 
 # The header files for this library.
 HSOURCES = 
@@ -107,7 +108,7 @@ OBJECTS = clktck.o clock.o getenv.o oslib.o setlinebuf.o strnlen.o \
 	  fmtullong.o fmtumax.o zcatfd.o zmapfd.o winsize.o wcsdup.o \
 	  fpurge.o zgetline.o mbscmp.o uconvert.o ufuncs.o casemod.o \
 	  input_avail.o mbscasecmp.o fnxform.o unicode.o shmbchar.o \
-	  wcsnwidth.o ${LIBOBJS}
+	  utf8.o wcsnwidth.o ${LIBOBJS}
 
 SUPPORT = Makefile
 
@@ -200,6 +201,7 @@ tmpfile.o: tmpfile.c
 uconvert.o: uconvert.c
 ufuncs.o: ufuncs.c
 unicode.o: unicode.c
+utf8.o: utf8.c
 vprint.o: vprint.c
 wcsdup.o: wcsdup.c
 wcsnwidth.o: wcsnwidth.c
@@ -277,6 +279,7 @@ tmpfile.o: ${BUILD_DIR}/config.h
 uconvert.o: ${BUILD_DIR}/config.h
 ufuncs.o: ${BUILD_DIR}/config.h
 unicode.o: ${BUILD_DIR}/config.h
+utf8.o: ${BUILD_DIR}/config.h
 vprint.o: ${BUILD_DIR}/config.h
 wcsdup.o: ${BUILD_DIR}/config.h
 wcsnwidth.o: ${BUILD_DIR}/config.h
@@ -612,6 +615,10 @@ unicode.o: ${topdir}/bashansi.h ${BASHINCDIR}/ansi_stdlib.h
 unicode.o: ${BASHINCDIR}/stdc.h
 unicode.o: ${topdir}/xmalloc.h
 
+utf8.o: ${topdir}/bashansi.h
+utf8.o: ${BASHINCDIR}/ansi_stdlib.h
+utf8.o: ${BASHINCDIR}/shmbutil.h ${BASHINCDIR}/shmbchar.h
+
 winsize.o: ${BASHINCDIR}/stdc.h
 winsize.o: ${topdir}/xmalloc.h
 winsize.o: ${topdir}/bashtypes.h

lib/sh/casemod.c

@@ -229,7 +229,10 @@ singlebyte:
       else
 	{
 	  m = mbrtowc (&wc, string + start, end - start, &state);
-	  if (MB_INVALIDCH (m) || m == 1)
+	  /* Have to go through wide case conversion even for single-byte
+	     chars, to accommodate single-byte characters where the
+	     corresponding upper or lower case equivalent is multibyte. */
+	  if (MB_INVALIDCH (m))
 	    {
 	      wc = (unsigned char)string[start];
 	      goto singlebyte;

lib/sh/clock.c

@@ -32,7 +32,13 @@
 #include <stdio.h>
 #include <stdc.h>
 
-extern long get_clk_tck __P((void));
+#include <bashintl.h>
+
+#ifndef locale_decpoint
+extern int locale_decpoint PARAMS((void));
+#endif
+
+extern long get_clk_tck PARAMS((void));
 
 void
 clock_t_to_secs (t, sp, sfp)
@@ -76,6 +82,6 @@ print_clock_t (fp, t)
   minutes = timestamp / 60;
   seconds = timestamp % 60;
 
-  fprintf (fp, "%ldm%d.%03ds",  minutes, seconds, seconds_fraction);
+  fprintf (fp, "%ldm%d%c%03ds",  minutes, seconds, locale_decpoint(), seconds_fraction);
 }
 #endif /* HAVE_TIMES */

lib/sh/eaccess.c

@@ -91,7 +91,9 @@ sh_stat (path, finfo)
     }
   if (path[0] == '/' && path[1] == 'd' && strncmp (path, "/dev/fd/", 8) == 0)
     {
-#if !defined (HAVE_DEV_FD)
+      /* If stating /dev/fd/n doesn't produce the same results as fstat of
+	 FD N, then define DEV_FD_STAT_BROKEN */
+#if !defined (HAVE_DEV_FD) || defined (DEV_FD_STAT_BROKEN)
       intmax_t fd;
       int r;
 

lib/sh/getenv.c

@@ -69,7 +69,7 @@ getenv (name)
       if (var && exported_p (var))
 	return (value_cell (var));
     }
-  else
+  else if (environ)
     {
       register int i, len;
 

lib/sh/mbschr.c

@@ -28,9 +28,12 @@
 #include "shmbutil.h"
 
 extern int locale_mb_cur_max;
+extern int locale_utf8locale;
 
 #undef mbschr
 
+extern char *utf8_mbschr (const char *, int);	/* XXX */
+
 /* In some locales, the non-first byte of some multibyte characters have
    the same value as some ascii character.  Faced with these strings, a
    legacy strchr() might return the wrong value. */
@@ -49,6 +52,9 @@ mbschr (s, c)
   mbstate_t state;
   size_t strlength, mblength;
 
+  if (locale_utf8locale && c < 0x80)
+    return (utf8_mbschr (s, c));		/* XXX */
+
   /* The locale encodings with said weird property are BIG5, BIG5-HKSCS,
      GBK, GB18030, SHIFT_JIS, and JOHAB.  They exhibit the problem only
      when c >= 0x30.  We can therefore use the faster bytewise search if

lib/sh/mbscmp.c

@@ -1,6 +1,6 @@
 /* mbscmp - multibyte string comparison. */
 
-/* Copyright (C) 1995-2015 Free Software Foundation, Inc.
+/* Copyright (C) 1995-2018 Free Software Foundation, Inc.
 
    This file is part of GNU Bash, the Bourne Again SHell.
 
@@ -26,6 +26,10 @@
 #include <stddef.h>
 #include <string.h>
 
+extern int locale_utf8locale;
+
+extern int utf8_mbscmp (const char *, const char *);
+
 /* Compare MBS1 and MBS2.  */
 int
 mbscmp (mbs1, mbs2)

lib/sh/pathcanon.c

@@ -227,7 +227,7 @@ sh_canonpath (path, flags)
       if (result[2] == '\0')	/* short-circuit for bare `//' */
 	result[1] = '\0';
       else
-	strcpy (result, result + 1);
+	memmove (result, result + 1, strlen (result + 1) + 1);
     }
 
   return (result);

lib/sh/pathphys.c

@@ -245,7 +245,7 @@ error:
       if (result[2] == '\0')	/* short-circuit for bare `//' */
 	result[1] = '\0';
       else
-	strcpy (result, result + 1);
+	memmove (result, result + 1, strlen (result + 1) + 1);
     }
 
   return (result);

lib/sh/shmbchar.c

@@ -1,4 +1,4 @@
-/* Copyright (C) 2001, 2006, 2009, 2010, 2012, 2015 Free Software Foundation, Inc.
+/* Copyright (C) 2001, 2006, 2009, 2010, 2012, 2015-2018 Free Software Foundation, Inc.
 
    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
@@ -20,9 +20,15 @@
 #include <stdlib.h>
 #include <limits.h>
 
+#include <errno.h>
+
 #include <shmbutil.h>
 #include <shmbchar.h>
 
+#ifndef errno
+extern int errno;
+#endif
+
 #if IS_BASIC_ASCII
 
 /* Bit table of characters in the ISO C "basic character set".  */
@@ -37,6 +43,13 @@ const unsigned int is_basic_table [UCHAR_MAX / 32 + 1] =
 
 #endif /* IS_BASIC_ASCII */
 
+extern int locale_utf8locale;
+
+extern char *utf8_mbsmbchar (const char *);
+extern int utf8_mblen (const char *, size_t);
+
+/* Count the number of characters in S, counting multi-byte characters as a
+   single character. */
 size_t
 mbstrlen (s)
      const char *s;
@@ -65,6 +78,8 @@ mbstrlen (s)
 }
 
 /* Return pointer to first multibyte char in S, or NULL if none. */
+/* XXX - if we know that the locale is UTF-8, we can just check whether or
+   not any byte has the eighth bit turned on */
 char *
 mbsmbchar (s)
      const char *s;
@@ -74,13 +89,19 @@ mbsmbchar (s)
   mbstate_t mbs = { 0 };
   int mb_cur_max;
 
+  if (locale_utf8locale)
+    return (utf8_mbsmbchar (s));	/* XXX */
+
   mb_cur_max = MB_CUR_MAX;
   for (t = (char *)s; *t; t++)
     {
       if (is_basic (*t))
 	continue;
 
-      clen = mbrlen (t, mb_cur_max, &mbs);
+      if (locale_utf8locale)		/* not used if above code active */
+	clen = utf8_mblen (t, mb_cur_max);
+      else
+	clen = mbrlen (t, mb_cur_max, &mbs);
 
       if (clen == 0)
         return 0;

lib/sh/shquote.c

@@ -228,7 +228,8 @@ sh_un_double_quote (string)
    going through the shell parser, which will protect the internal
    quoting characters.  TABLE, if set, points to a map of the ascii code
    set with char needing to be backslash-quoted if table[char]==1.  FLAGS,
-   if 1, causes tildes to be quoted as well. */
+   if 1, causes tildes to be quoted as well.  If FLAGS&2, backslash-quote
+   other shell blank characters. */
    
 char *
 sh_backslash_quote (string, table, flags)
@@ -273,6 +274,8 @@ sh_backslash_quote (string, table, flags)
         /* Tildes are special at the start of a word or after a `:' or `='
 	   (technically unquoted, but it doesn't make a difference in practice) */
 	*r++ = '\\';
+      else if ((flags&2) && shellblank((unsigned char)c))
+	*r++ = '\\';
       *r++ = c;
     }
 

lib/sh/snprintf.c

@@ -142,9 +142,11 @@ extern char *fmtullong __P((unsigned long long int, int, char *, size_t, int));
    302 / 1000 is log10 (2) rounded up;
    add one for integer division truncation;
    add one more for a minus sign if t is signed.  */
+#ifndef INT_STRLEN_BOUND
 #define INT_STRLEN_BOUND(t) \
   ((sizeof (t) * CHAR_BIT - TYPE_SIGNED (t)) * 302 / 1000 \
      + 1 + TYPE_SIGNED (t))
+#endif
 
 /* conversion flags */
 #define PF_ALTFORM	0x00001		/* # */

lib/sh/strcasecmp.c

@@ -32,7 +32,7 @@ int
 strncasecmp (string1, string2, count)
      const char *string1;
      const char *string2;
-     int count;
+     size_t count;
 {
   register const char *s1;
   register const char *s2;

lib/sh/strtrans.c

@@ -230,8 +230,6 @@ ansic_quote (str, flags, rlen)
   *r++ = '$';
   *r++ = '\'';
 
-  s = str;
-
   for (s = str; c = *s; s++)
     {
       b = l = 1;		/* 1 == add backslash; 0 == no backslash */
@@ -305,11 +303,9 @@ ansic_wshouldquote (string)
 {
   const wchar_t *wcs;
   wchar_t wcc;
-
   wchar_t *wcstr = NULL;
   size_t slen;
 
-
   slen = mbstowcs (wcstr, string, 0);
 
   if (slen == (size_t)-1)

lib/sh/timeval.c

@@ -25,6 +25,13 @@
 #include <sys/types.h>
 #include <posixtime.h>
 
+#include <bashintl.h>
+#include <stdc.h>
+
+#ifndef locale_decpoint
+extern int locale_decpoint PARAMS((void));
+#endif
+
 #include <stdio.h>
 
 struct timeval *
@@ -140,6 +147,6 @@ print_timeval (fp, tvp)
   minutes = timestamp / 60;
   seconds = timestamp % 60;
 
-  fprintf (fp, "%ldm%d.%03ds",  minutes, seconds, seconds_fraction);
+  fprintf (fp, "%ldm%d%c%03ds",  minutes, seconds, locale_decpoint (), seconds_fraction);
 }
 #endif /* HAVE_TIMEVAL */

lib/sh/ufuncs.c

@@ -37,8 +37,16 @@
 #include <unistd.h>
 #endif
 
+#include <errno.h>
+#if !defined (errno)
+extern int errno;
+#endif /* !errno */
+
 #if defined (HAVE_SELECT)
 #  include "posixselect.h"
+#  include "quit.h"
+#  include "trap.h"
+#  include "stat-time.h"
 #endif
 
 /* A version of `alarm' using setitimer if it's available. */
@@ -84,17 +92,50 @@ falarm (secs, usecs)
 /* A version of sleep using fractional seconds and select.  I'd like to use
    `usleep', but it's already taken */
 
-#if defined (HAVE_TIMEVAL) && defined (HAVE_SELECT)
+#if defined (HAVE_TIMEVAL) && (defined (HAVE_SELECT) || defined (HAVE_PSELECT))
 int
 fsleep(sec, usec)
      unsigned int sec, usec;
 {
+  int e, r;
+  sigset_t blocked_sigs, prevmask;
+#if defined (HAVE_PSELECT)
+  struct timespec ts;
+#else
   struct timeval tv;
+#endif
 
+  sigemptyset (&blocked_sigs);
+#  if defined (SIGCHLD)
+  sigaddset (&blocked_sigs, SIGCHLD);
+#  endif
+
+#if defined (HAVE_PSELECT)
+  ts.tv_sec = sec;
+  ts.tv_nsec = usec * 1000;
+#else
+  sigemptyset (&prevmask);
   tv.tv_sec = sec;
   tv.tv_usec = usec;
+#endif /* !HAVE_PSELECT */
 
-  return select(0, (fd_set *)0, (fd_set *)0, (fd_set *)0, &tv);
+  do
+    {
+#if defined (HAVE_PSELECT)
+      r = pselect(0, (fd_set *)0, (fd_set *)0, (fd_set *)0, &ts, &blocked_sigs);
+#else
+      sigprocmask (SIG_SETMASK, &blocked_sigs, &prevmask);
+      r = select(0, (fd_set *)0, (fd_set *)0, (fd_set *)0, &tv);
+      sigprocmask (SIG_SETMASK, &prevmask, NULL);
+#endif
+      e = errno;
+      if (r < 0 && errno == EINTR)
+	QUIT;		/* just signals, no traps */
+      errno = e;
+    }
+  while (r < 0 && errno == EINTR);
+
+  return r;
 }
 #else /* !HAVE_TIMEVAL || !HAVE_SELECT */
 int

lib/sh/unicode.c

@@ -1,6 +1,6 @@
 /* unicode.c - functions to convert unicode characters */
 
-/* Copyright (C) 2010-2015 Free Software Foundation, Inc.
+/* Copyright (C) 2010-2016 Free Software Foundation, Inc.
 
    This file is part of GNU Bash, the Bourne Again SHell.
 
@@ -55,6 +55,8 @@ extern const char *locale_charset __P((void));
 extern char *get_locale_var __P((char *));
 #endif
 
+extern int locale_utf8locale;
+
 static int u32init = 0;
 static int utf8locale = 0;
 #if defined (HAVE_ICONV)
@@ -219,12 +221,12 @@ u32toutf16 (c, s)
   int l;
 
   l = 0;
-  if (c < 0x0d800)
+  if (c < 0x0d800 || (c >= 0x0e000 && c <= 0x0ffff))
     {
       s[0] = (unsigned short) (c & 0xFFFF);
       l = 1;
     }
-  else if (c >= 0x0e000 && c <= 0x010ffff)
+  else if (c >= 0x10000 && c <= 0x010ffff)
     {
       c -= 0x010000;
       s[0] = (unsigned short)((c >> 10) + 0xd800);
@@ -265,28 +267,21 @@ u32cconv (c, s)
     return n;
 #endif
 
-#if HAVE_NL_LANGINFO
-  codeset = nl_langinfo (CODESET);
-  if (STREQ (codeset, "UTF-8"))
-    {
-      n = u32toutf8 (c, s);
-      return n;
-    }
-#endif
-
 #if HAVE_ICONV
   /* this is mostly from coreutils-8.5/lib/unicodeio.c */
   if (u32init == 0)
     {
-#  if HAVE_LOCALE_CHARSET
-      charset = locale_charset ();	/* XXX - fix later */
-#  else
-      charset = stub_charset ();
-#  endif
-      if (STREQ (charset, "UTF-8"))
-	utf8locale = 1;
-      else
+      utf8locale = locale_utf8locale;
+      localconv = (iconv_t)-1;
+      if (utf8locale == 0)
 	{
+#if HAVE_LOCALE_CHARSET
+	  charset = locale_charset ();
+#elif HAVE_NL_LANGINFO
+	  charset = nl_langinfo (CODESET);
+#else
+	  charset = stub_charset ();
+#endif
 	  localconv = iconv_open (charset, "UTF-8");
 	  if (localconv == (iconv_t)-1)
 	    /* We assume ASCII when presented with an unknown encoding. */
@@ -295,6 +290,8 @@ u32cconv (c, s)
       u32init = 1;
     }
 
+  /* NL_LANGINFO and locale_charset used when setting locale_utf8locale */
+  
   /* If we have a UTF-8 locale, convert to UTF-8 and return converted value. */
   n = u32toutf8 (c, s);
   if (utf8locale)
@@ -315,12 +312,8 @@ u32cconv (c, s)
 
   if (iconv (localconv, (ICONV_CONST char **)&iptr, &sn, &optr, &obytesleft) == (size_t)-1)
     {
-#if 1
       /* You get ISO C99 escape sequences if iconv fails */      
       n = u32tocesc (c, s);
-#else
-      /* You get UTF-8 if iconv fails */
-#endif
       return n;
     }
 
@@ -332,7 +325,10 @@ u32cconv (c, s)
   return (optr - obuf);
 #endif	/* HAVE_ICONV */
 
-  n = u32tocesc (c, s);	/* fallback is ISO C99 escape sequences */
+  if (locale_utf8locale)
+    n = u32toutf8 (c, s);
+  else
+    n = u32tocesc (c, s);	/* fallback is ISO C99 escape sequences */
   return n;
 }
 #else

lib/sh/utf8.c

@@ -0,0 +1,147 @@
+/* utf8.c - UTF-8 character handling functions */
+
+/* Copyright (C) 2018 Free Software Foundation, Inc.
+
+   This file is part of GNU Bash, the Bourne Again SHell.
+
+   Bash 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.
+
+   Bash 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 Bash.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+#include <config.h>
+
+#ifdef HAVE_STDLIB_H
+#  include <stdlib.h>
+#endif
+
+#include "bashansi.h"
+#include "shmbutil.h"
+
+extern int locale_mb_cur_max;
+extern int locale_utf8locale;
+
+#if defined (HANDLE_MULTIBYTE)
+
+char *
+utf8_mbschr (s, c)
+     const char *s;
+     int c;
+{
+  return strchr (s, c);		/* for now */
+}
+
+int
+utf8_mbscmp (s1, s2)
+     const char *s1, *s2;
+{
+  /* Use the fact that the UTF-8 encoding preserves lexicographic order.  */
+  return strcmp (s1, s2);
+}
+
+char *
+utf8_mbsmbchar (str)
+     const char *str;
+{
+  register char *s;
+
+  for (s = (char *)str; *s; s++)
+    if ((*s & 0xc0) == 0x80)
+      return s;
+  return (0);
+}
+
+int
+utf8_mbsnlen(src, srclen, maxlen)
+     const char *src;
+     size_t srclen;
+     int maxlen;
+{
+  register int sind, count;
+
+  for (sind = count = 0; src[sind] && sind <= maxlen; sind++)
+    {
+      if ((src[sind] & 0xc0) != 0x80)
+	count++;
+    }
+  return (count);
+}
+
+/* Adapted from GNU gnulib */
+int
+utf8_mblen (s, n)
+     const char *s;
+     size_t n;
+{
+  unsigned char c, c1;
+
+  if (s == 0)
+    return (0);	/* no shift states */
+  if (n <= 0)
+    return (-1);
+
+  c = (unsigned char)*s;
+  if (c < 0x80)
+    return (c != 0);
+  if (c >= 0xc2)
+    {
+      c1 = (unsigned char)s[1];
+      if (c < 0xe0)
+	{
+	  if (n >= 2 && (s[1] ^ 0x80) < 0x40)
+	    return 2;
+	}
+      else if (c < 0xf0)
+	{
+	  if (n >= 3
+		&& (s[1] ^ 0x80) < 0x40 && (s[2] ^ 0x80) < 0x40
+		&& (c >= 0xe1 || c1 >= 0xa0)
+		&& (c != 0xed || c1 < 0xa0))
+	    return 3;
+	}
+      else if (c < 0xf8)
+	{
+	  if (n >= 4
+		&& (s[1] ^ 0x80) < 0x40 && (s[2] ^ 0x80) < 0x40
+	 	&& (s[3] ^ 0x80) < 0x40
+		&& (c >= 0xf1 || c1 >= 0x90)
+		&& (c < 0xf4 || (c == 0xf4 && c1 < 0x90)))
+	    return 4;
+	}
+    }
+  /* invalid or incomplete multibyte character */
+  return -1;
+}
+
+/* We can optimize this if we know the locale is UTF-8, but needs to handle
+   malformed byte sequences. */
+size_t
+utf8_mbstrlen(s)
+     const char *s;
+{
+  size_t clen, nc;
+  int mb_cur_max;
+
+  nc = 0;
+  mb_cur_max = MB_CUR_MAX;
+  while (*s && (clen = (size_t)utf8_mblen(s, mb_cur_max)) != 0)
+    {
+      if (MB_INVALIDCH(clen))
+	clen = 1;	/* assume single byte */
+
+      s += clen;
+      nc++;
+    }
+  return nc;
+}
+
+#endif

lib/sh/zread.c

@@ -1,6 +1,6 @@
 /* zread - read data from file descriptor into buffer with retries */
 
-/* Copyright (C) 1999-2002 Free Software Foundation, Inc.
+/* Copyright (C) 1999-2017 Free Software Foundation, Inc.
 
    This file is part of GNU Bash, the Bourne Again SHell.
 
@@ -53,6 +53,7 @@ zread (fd, buf, len)
 {
   ssize_t r;
 
+  check_signals ();	/* check for signals before a blocking read */
   while ((r = read (fd, buf, len)) < 0 && errno == EINTR)
     /* XXX - bash-5.0 */
     /* We check executing_builtin and run traps here for backwards compatibility */
@@ -103,6 +104,7 @@ zreadintr (fd, buf, len)
      char *buf;
      size_t len;
 {
+  check_signals ();
   return (read (fd, buf, len));
 }