line editors: make putstring functions multibyte aware

Symptom on (at least) some Linux systems:

  $ bin/package use
  [blurb skipped]
  «0»37:�/local/src/ksh[dev] $
  «0»37:…/local/src/ksh[dev] $
  «0»37:…/local/src/ksh[dev] $ ▂

If the line editor is emacs, the first prompt is corrupted.
The prompt printed after ^L in emacs is also corrupted.

The cause is the putstring() function in emacs. It calls
ed_putchar() for each byte in the string. This is incorrect for
multibyte characters, as ed_putchar() *is* multibyte aware and
calls mbconv() for each character passed to it.

There is a very similar ed_putstring() function in edit.c that has
the same problem.

There is also a pr_string() function in vi.c that is used for
printing the prompt there; it is not multibyte-aware either, but
the bug is not triggered in vi because pr_string() copies bytes
directly into the output buffer. But it does so without bounds
checking, so a very long prompt could cause a buffer overflow.

src/cmd/ksh93/edit/edit.{c,h}:
- Make ed_putstring() multibyte-aware by using the mbchar macro
  from ast.h. This function is only used for printing the prompt
  and some escape sequences from tput, so also print '?' for
  invalid multibyte characters as this is more useful than just
  copying the invalid byte.
- Make ed_putstring() an extern and declare it in the header.
- While we're here, further cleanup of ed_putchar(). (re: 1eb0b06)

src/cmd/ksh93/edit/{emacs,vi}.c:
- Remove putstring() and pr_string() functions.
- Replace with ed_putstring() calls.

NEWS

@@ -2,6 +2,11 @@ This documents significant changes in the 1.0 branch of ksh 93u+m.
 For full details, see the git log at: https://github.com/ksh93/ksh/tree/1.0
 Uppercase BUG_* IDs are shell bug IDs as used by the Modernish shell library.
 
+2024-03-07:
+
+- Fixed a bug that caused some systems to corrupt the display of multibyte
+  characters in the command prompt.
+
 2024-03-05:
 
 - Fixed a regression, introduced on 2021-02-07, that made the common AST

src/cmd/ksh93/edit/edit.c

@@ -677,11 +677,12 @@ void	ed_setup(Edit_t *ep, int fd, int reedit)
 	}
 }
 
-static void ed_putstring(Edit_t *ep, const char *str)
+void ed_putstring(Edit_t *ep, const char *str)
 {
 	int c;
-	while(c = *str++)
-		ed_putchar(ep,c);
+	mbinit();
+	while (c = mbchar(str))
+		ed_putchar(ep, c < 0 ? '?' : c);
 }
 
 static void ed_nputchar(Edit_t *ep, int n, int c)
@@ -1019,24 +1020,24 @@ void ed_ungetchar(Edit_t *ep,int c)
 
 void	ed_putchar(Edit_t *ep,int c)
 {
-	char buf[8];
 	char *dp = ep->e_outptr;
-	int i,size=1;
+#if SHOPT_MULTIBYTE
+	char buf[8];
+	int size;
+#endif	/* SHOPT_MULTIBYTE */
 	if(!dp)
 		return;
-	buf[0] = c;
 #if SHOPT_MULTIBYTE
 	/* check for place holder */
 	if(c == MARKER)
 		return;
 	if((size = mbconv(buf, (wchar_t)c)) > 1)
 	{
-		for (i = 0; i < (size-1); i++)
+		int i;
+		for (i = 0; i < size; i++)
 			*dp++ = buf[i];
-		c = buf[i];
 	}
 	else
-		buf[0] = c;
 #endif	/* SHOPT_MULTIBYTE */
 	*dp++ = c;
 	*dp = '\0';

src/cmd/ksh93/edit/emacs.c

@@ -171,7 +171,7 @@ typedef enum
 
 static void draw(Emacs_t*,Draw_t);
 static int escape(Emacs_t*,genchar*, int);
-static void putstring(Emacs_t*,char*);
+static int dosearch(Emacs_t*,genchar*,int);
 static void search(Emacs_t*,genchar*,int);
 static void setcursor(Emacs_t*,int, int);
 static void show_info(Emacs_t*,const char*);
@@ -599,7 +599,7 @@ int ed_emacsread(void *context, int fd,char *buff,int scend, int reedit)
 				if (ep->terminal == PAPER)
 				{
 					putchar(ep->ed,'\n');
-					putstring(ep,Prompt);
+					ed_putstring(ep->ed,Prompt);
 				}
 				c = ed_getchar(ep->ed,0);
 				if (c != usrkill)
@@ -613,7 +613,7 @@ int ed_emacsread(void *context, int fd,char *buff,int scend, int reedit)
 				{
 					ep->terminal = PAPER;
 					putchar(ep->ed,'\n');
-					putstring(ep,Prompt);
+					ed_putstring(ep->ed,Prompt);
 				}
 			}
 			continue;
@@ -782,14 +782,6 @@ static void show_info(Emacs_t *ep,const char *str)
 	draw(ep,UPDATE);
 }
 
-static void putstring(Emacs_t* ep,char *sp)
-{
-	int c;
-	while (c= *sp++)
-		 putchar(ep->ed,c);
-}
-
-
 static int escape(Emacs_t* ep,genchar *out,int count)
 {
 	int i,value;
@@ -1572,7 +1564,7 @@ static void draw(Emacs_t *ep,Draw_t option)
 			return;
 		}
 		*ep->cursor = '\0';
-		putstring(ep,Prompt);	/* start with prompt */
+		ed_putstring(ep->ed,Prompt);	/* start with prompt */
 	}
 
 	/*********************

src/cmd/ksh93/edit/vi.c

@@ -181,7 +181,6 @@ static int	getcount(Vi_t*,int);
 static void	getline(Vi_t*,int);
 static int	getrchar(Vi_t*);
 static int	mvcursor(Vi_t*,int);
-static void	pr_string(Vi_t*,const char*);
 static void	refresh(Vi_t*,int);
 static void	replace(Vi_t*,int, int);
 static void	restore_v(Vi_t*);
@@ -1688,20 +1687,6 @@ static int mvcursor(Vi_t* vp,int motion)
 	return 1;
 }
 
-/*
- * print a string
- */
-
-static void pr_string(Vi_t *vp, const char *sp)
-{
-	/*** copy string sp ***/
-	char *ptr = editb.e_outptr;
-	while(*sp)
-		*ptr++ = *sp++;
-	editb.e_outptr = ptr;
-	return;
-}
-
 /*{	VI_REDRAW( )
  *
  *	Print the prompt and force a total refresh.
@@ -1714,7 +1699,7 @@ static void pr_string(Vi_t *vp, const char *sp)
 void vi_redraw(void *ep)
 {
 	Vi_t	*vp = (Vi_t*)ep;
-	pr_string(vp,Prompt);
+	ed_putstring(vp->ed,Prompt);
 	window[0] = '\0';
 	cur_phys = vp->first_wind;
 	vp->ofirst_wind = INVALID;

src/cmd/ksh93/include/edit.h

@@ -179,6 +179,7 @@ typedef struct edit
 #define TCAP_ERASE_EOS	"cd"
 
 extern void	ed_putchar(Edit_t*, int);
+extern void	ed_putstring(Edit_t*, const char*);
 extern void	ed_ringbell(void);
 extern void	ed_setup(Edit_t*,int, int);
 extern void	ed_flush(Edit_t*);

src/cmd/ksh93/include/version.h

@@ -18,7 +18,7 @@
 
 #define SH_RELEASE_FORK	"93u+m"		/* only change if you develop a new ksh93 fork */
 #define SH_RELEASE_SVER	"1.0.9-beta"	/* semantic version number: https://semver.org */
-#define SH_RELEASE_DATE	"2024-03-05"	/* must be in this format for $((.sh.version)) */
+#define SH_RELEASE_DATE	"2024-03-07"	/* must be in this format for $((.sh.version)) */
 #define SH_RELEASE_CPYR	"(c) 2020-2024 Contributors to ksh " SH_RELEASE_FORK
 
 /* Scripts sometimes field-split ${.sh.version}, so don't change amount of whitespace. */