[PATCH 05/11] Fix width computation

[Robbie Harwood]

From: Vladimir 'phcoder' Serbinenko <phcoder@gmail.com>

Signed-off-by: Robbie Harwood <rharwood@redhat.com>
---
 lib/argp-fmtstream.c | 80 +++++++++++++++++++++++++++++++++++++-------
 lib/argp-help.c      |  3 +-
 lib/mbswidth.c       | 15 +++++++++
 lib/mbswidth.h       |  4 +++
 4 files changed, 88 insertions(+), 14 deletions(-)

diff --git a/lib/argp-fmtstream.c b/lib/argp-fmtstream.c
index 197eebab8..f679751b9 100644
--- a/lib/argp-fmtstream.c
+++ b/lib/argp-fmtstream.c
@@ -28,9 +28,11 @@
 #include <errno.h>
 #include <stdarg.h>
 #include <ctype.h>
+#include <wchar.h>
 
 #include "argp-fmtstream.h"
 #include "argp-namefrob.h"
+#include "mbswidth.h"
 
 #ifndef ARGP_FMTSTREAM_USE_LINEWRAP
 
@@ -115,6 +117,51 @@ weak_alias (__argp_fmtstream_free, argp_fmtstream_free)
 #endif
 #endif
 
+
+/* Return the pointer to the first character that doesn't fit in l columns.  */
+static inline const ptrdiff_t
+add_width (const char *ptr, const char *end, size_t l)
+{
+  mbstate_t ps;
+  const char *ptr0 = ptr;
+
+  memset (&ps, 0, sizeof (ps));
+
+  while (ptr < end)
+    {
+      wchar_t wc;
+      size_t s, k;
+
+      s = mbrtowc (&wc, ptr, end - ptr, &ps);
+      if (s == (size_t) -1)
+       break;
+      if (s == (size_t) -2)
+       {
+         if (1 >= l)
+           break;
+         l--;
+         ptr++;
+         continue;
+       }
+
+      if (wc == '\e' && ptr + 3 < end
+         && ptr[1] == '[' && (ptr[2] == '0' || ptr[2] == '1')
+         && ptr[3] == 'm')
+       {
+         ptr += 4;
+         continue;
+       }
+
+      k = wcwidth (wc);
+
+      if (k >= l)
+       break;
+      l -= k;
+      ptr += s;
+    }
+  return ptr - ptr0;
+}
+
 /* Process FS's buffer so that line wrapping is done from POINT_OFFS to the
    end of its buffer.  This code is mostly from glibc stdio/linewrap.c.  */
 void
@@ -168,13 +215,15 @@ __argp_fmtstream_update (argp_fmtstream_t fs)
       if (!nl)
         {
           /* The buffer ends in a partial line.  */
+          size_t display_width = mbsnwidth (buf, fs->p - buf,
+                                            MBSW_STOP_AT_NUL);
 
-          if (fs->point_col + len < fs->rmargin)
+          if (fs->point_col + display_width < fs->rmargin)
             {
               /* The remaining buffer text is a partial line and fits
                  within the maximum line width.  Advance point for the
                  characters to be written and stop scanning.  */
-              fs->point_col += len;
+              fs->point_col += display_width;
               break;
             }
           else
@@ -182,14 +231,18 @@ __argp_fmtstream_update (argp_fmtstream_t fs)
                the end of the buffer.  */
             nl = fs->p;
         }
-      else if (fs->point_col + (nl - buf) < (ssize_t) fs->rmargin)
-        {
-          /* The buffer contains a full line that fits within the maximum
-             line width.  Reset point and scan the next line.  */
-          fs->point_col = 0;
-          buf = nl + 1;
-          continue;
-        }
+      else
+       {
+         size_t display_width = mbsnwidth (buf, nl - buf, MBSW_STOP_AT_NUL);
+         if (display_width < (ssize_t) fs->rmargin)
+           {
+             /* The buffer contains a full line that fits within the maximum
+                line width.  Reset point and scan the next line.  */
+             fs->point_col = 0;
+             buf = nl + 1;
+             continue;
+           }
+       }
 
       /* This line is too long.  */
       r = fs->rmargin - 1;
@@ -225,7 +278,7 @@ __argp_fmtstream_update (argp_fmtstream_t fs)
           char *p, *nextline;
           int i;
 
-          p = buf + (r + 1 - fs->point_col);
+          p = buf + add_width (buf, fs->p, (r + 1 - fs->point_col));
           while (p >= buf && !isblank ((unsigned char) *p))
             --p;
           nextline = p + 1;     /* This will begin the next line.  */
@@ -243,7 +296,7 @@ __argp_fmtstream_update (argp_fmtstream_t fs)
             {
               /* A single word that is greater than the maximum line width.
                  Oh well.  Put it on an overlong line by itself.  */
-              p = buf + (r + 1 - fs->point_col);
+              p = buf + add_width (buf, fs->p, (r + 1 - fs->point_col));
               /* Find the end of the long word.  */
               if (p < nl)
                 do
@@ -277,7 +330,8 @@ __argp_fmtstream_update (argp_fmtstream_t fs)
               && fs->p > nextline)
             {
               /* The margin needs more blanks than we removed.  */
-              if (fs->end - fs->p > fs->wmargin + 1)
+              if (mbsnwidth (fs->p, fs->end - fs->p, MBSW_STOP_AT_NUL)
+                  > fs->wmargin + 1)
                 /* Make some space for them.  */
                 {
                   size_t mv = fs->p - nextline;
diff --git a/lib/argp-help.c b/lib/argp-help.c
index 80cdb4493..84013b002 100644
--- a/lib/argp-help.c
+++ b/lib/argp-help.c
@@ -52,6 +52,7 @@
 #include "argp.h"
 #include "argp-fmtstream.h"
 #include "argp-namefrob.h"
+#include "mbswidth.h"
 
 #ifndef SIZE_MAX
 # define SIZE_MAX ((size_t) -1)
@@ -1547,7 +1548,7 @@ argp_args_usage (const struct argp *argp, const struct 
argp_state *state,
 
       /* Manually do line wrapping so that it (probably) won't get wrapped at
          any embedded spaces.  */
-      space (stream, 1 + nl - cp);
+      space (stream, 1 + mbsnwidth (cp, nl - cp, MBSW_STOP_AT_NUL));
 
       __argp_fmtstream_write (stream, cp, nl - cp);
     }
diff --git a/lib/mbswidth.c b/lib/mbswidth.c
index 7f599359c..1820f51c1 100644
--- a/lib/mbswidth.c
+++ b/lib/mbswidth.c
@@ -38,6 +38,14 @@
 /* Get INT_MAX.  */
 #include <limits.h>
 
+#ifndef FALLTHROUGH
+# if __GNUC__ < 7
+#  define FALLTHROUGH ((void) 0)
+# else
+#  define FALLTHROUGH __attribute__ ((__fallthrough__))
+# endif
+#endif
+
 /* Returns the number of columns needed to represent the multibyte
    character string pointed to by STRING.  If a non-printable character
    occurs, and MBSW_REJECT_UNPRINTABLE is specified, -1 is returned.
@@ -90,6 +98,10 @@ mbsnwidth (const char *string, size_t nbytes, int flags)
               p++;
               width++;
               break;
+            case '\0':
+              if (flags & MBSW_STOP_AT_NUL)
+                return width;
+              FALLTHROUGH;
             default:
               /* If we have a multibyte sequence, scan it up to its end.  */
               {
@@ -168,6 +180,9 @@ mbsnwidth (const char *string, size_t nbytes, int flags)
     {
       unsigned char c = (unsigned char) *p++;
 
+      if (c == 0 && (flags & MBSW_STOP_AT_NUL))
+        return width;
+
       if (isprint (c))
         {
           if (width == INT_MAX)
diff --git a/lib/mbswidth.h b/lib/mbswidth.h
index ec39d7339..744e48eea 100644
--- a/lib/mbswidth.h
+++ b/lib/mbswidth.h
@@ -40,6 +40,10 @@ extern "C" {
    control characters and 1 otherwise.  */
 #define MBSW_REJECT_UNPRINTABLE 2
 
+/* If this bit is set \0 is treated as the end of string.
+   Otherwise it's treated as a normal one column width character.  */
+#define MBSW_STOP_AT_NUL 4
+
 
 /* Returns the number of screen columns needed for STRING.  */
 #define mbswidth gnu_mbswidth  /* avoid clash with UnixWare 7.1.1 function */
-- 
2.33.0