Flip on printing `?' at Unicode codepoints in -Tascii, -Tpdf, and -Tps.

[mandoc.git] / term.c

diff --git a/term.c b/term.c
index 1c5e8ff0f0869d8cf86d770aa766bdb80a99a199..bfedefb0ff31e718b8f5f80d5ce3ddd619853414 100644 (file)
--- a/term.c
+++ b/term.c
@@ -1,4 +1,4 @@
-/*     $Id: term.c,v 1.189 2011/05/15 14:50:01 kristaps Exp $ */
+/*     $Id: term.c,v 1.192 2011/05/17 11:55:08 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>
@@ -33,12 +33,9 @@
 #include "term.h"
 #include "main.h"
 
-static void              spec(struct termp *, const char *, size_t);
-static void              res(struct termp *, const char *, size_t);
-static void              bufferc(struct termp *, char);
-static void              adjbuf(struct termp *p, int);
-static void              encode(struct termp *, const char *, size_t);
-
+static void             adjbuf(struct termp *p, int);
+static void             bufferc(struct termp *, char);
+static void             encode(struct termp *, const char *, size_t);
 
 void
 term_free(struct termp *p)
@@ -344,43 +341,6 @@ term_vspace(struct termp *p)
        (*p->endline)(p);
 }
 
-
-static void
-numbered(struct termp *p, const char *word, size_t len)
-{
-       char             c;
-
-       if ('\0' != (c = mchars_num2char(word, len)))
-               encode(p, &c, 1);
-}
-
-
-static void
-spec(struct termp *p, const char *word, size_t len)
-{
-       const char      *rhs;
-       size_t           sz;
-
-       rhs = mchars_spec2str(p->symtab, word, len, &sz);
-       if (rhs) 
-               encode(p, rhs, sz);
-       else if (1 == len)
-               encode(p, word, len);
-}
-
-
-static void
-res(struct termp *p, const char *word, size_t len)
-{
-       const char      *rhs;
-       size_t           sz;
-
-       rhs = mchars_res2str(p->symtab, word, len, &sz);
-       if (rhs)
-               encode(p, rhs, sz);
-}
-
-
 void
 term_fontlast(struct termp *p)
 {
@@ -445,7 +405,6 @@ term_fontpop(struct termp *p)
        p->fonti--;
 }
 
-
 /*
  * Handle pwords, partial words, which may be either a single word or a
  * phrase that cannot be broken down (such as a literal string).  This
@@ -454,7 +413,8 @@ term_fontpop(struct termp *p)
 void
 term_word(struct termp *p, const char *word)
 {
-       const char      *seq;
+       const char      *seq, *cp;
+       char             c;
        int              sz;
        size_t           ssz;
        enum mandoc_esc  esc;
@@ -491,14 +451,24 @@ term_word(struct termp *p, const char *word)
                        break;
 
                switch (esc) {
+               case (ESCAPE_UNICODE):
+                       encode(p, "?", 1);
+                       break;
                case (ESCAPE_NUMBERED):
-                       numbered(p, seq, sz);
+                       if ('\0' != (c = mchars_num2char(seq, sz)))
+                               encode(p, &c, 1);
                        break;
                case (ESCAPE_PREDEF):
-                       res(p, seq, sz);
+                       cp = mchars_res2str(p->symtab, seq, sz, &ssz);
+                       if (NULL != cp)
+                               encode(p, cp, ssz);
                        break;
                case (ESCAPE_SPECIAL):
-                       spec(p, seq, sz);
+                       cp = mchars_spec2str(p->symtab, seq, sz, &ssz);
+                       if (NULL != cp) 
+                               encode(p, cp, ssz);
+                       else if (1 == ssz)
+                               encode(p, seq, sz);
                        break;
                case (ESCAPE_FONTBOLD):
                        term_fontrepl(p, TERMFONT_BOLD);
@@ -522,7 +492,6 @@ term_word(struct termp *p, const char *word)
        }
 }
 
-
 static void
 adjbuf(struct termp *p, int sz)
 {
@@ -536,7 +505,6 @@ adjbuf(struct termp *p, int sz)
                (p->buf, sizeof(int) * (size_t)p->maxcols);
 }
 
-
 static void
 bufferc(struct termp *p, char c)
 {
@@ -547,7 +515,6 @@ bufferc(struct termp *p, char c)
        p->buf[p->col++] = c;
 }
 
-
 static void
 encode(struct termp *p, const char *word, size_t sz)
 {
@@ -592,7 +559,6 @@ encode(struct termp *p, const char *word, size_t sz)
        }
 }
 
-
 size_t
 term_len(const struct termp *p, size_t sz)
 {
@@ -605,7 +571,7 @@ size_t
 term_strlen(const struct termp *p, const char *cp)
 {
        size_t           sz, rsz, i;
-       int              ssz;
+       int              ssz, c;
        const char      *seq, *rhs;
        static const char rej[] = { '\\', ASCII_HYPH, ASCII_NBRSP, '\0' };
 
@@ -621,12 +587,23 @@ term_strlen(const struct termp *p, const char *cp)
                for (i = 0; i < rsz; i++)
                        sz += (*p->width)(p, *cp++);
 
+               c = 0;
                switch (*cp) {
                case ('\\'):
                        cp++;
+                       rhs = NULL;
                        switch (mandoc_escape(&cp, &seq, &ssz)) {
                        case (ESCAPE_ERROR):
                                return(sz);
+                       case (ESCAPE_UNICODE):
+                               c = '?';
+                               /* FALLTHROUGH */
+                       case (ESCAPE_NUMBERED):
+                               if ('\0' != c)
+                                       c = mchars_num2char(seq, ssz);
+                               if ('\0' != c)
+                                       sz += (*p->width)(p, c);
+                               break;
                        case (ESCAPE_PREDEF):
                                rhs = mchars_res2str
                                        (p->symtab, seq, ssz, &rsz);
@@ -642,7 +619,6 @@ term_strlen(const struct termp *p, const char *cp)
                                rsz = ssz;
                                break;
                        default:
-                               rhs = NULL;
                                break;
                        }