Locale support. I'm checking this in to clean up fall-out in-tree, but

[mandoc.git] / term.c

diff --git a/term.c b/term.c
index 0ea561cb8959775e5d758599a441ae694c3543aa..04d55d6730583bb940c467172906cdb979397be4 100644 (file)
--- a/term.c
+++ b/term.c
@@ -1,4 +1,4 @@
-/*     $Id: term.c,v 1.191 2011/05/15 22:29:50 kristaps Exp $ */
+/*     $Id: term.c,v 1.194 2011/05/17 22:32:45 kristaps Exp $ */

 /*
  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>
 /*
  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>


@@ -36,6 +36,7 @@
 static void             adjbuf(struct termp *p, int);
 static void             bufferc(struct termp *, char);
 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);

+static void             encode1(struct termp *, int);

 
 void
 term_free(struct termp *p)
 
 void
 term_free(struct termp *p)


@@ -69,18 +70,6 @@ term_end(struct termp *p)
        (*p->end)(p);
 }
 
        (*p->end)(p);
 }
 

-
-struct termp *
-term_alloc(enum termenc enc)
-{
-       struct termp    *p;
-
-       p = mandoc_calloc(1, sizeof(struct termp));
-       p->enc = enc;
-       return(p);
-}
-
-

 /*
  * Flush a line of text.  A "line" is loosely defined as being something
  * that should be followed by a newline, regardless of whether it's
 /*
  * Flush a line of text.  A "line" is loosely defined as being something
  * that should be followed by a newline, regardless of whether it's


@@ -415,7 +404,7 @@ term_word(struct termp *p, const char *word)
 {
        const char      *seq, *cp;
        char             c;
 {
        const char      *seq, *cp;
        char             c;

-       int              sz;
+       int              sz, uc;

        size_t           ssz;
        enum mandoc_esc  esc;
 
        size_t           ssz;
        enum mandoc_esc  esc;
 


@@ -451,6 +440,15 @@ term_word(struct termp *p, const char *word)
                        break;
 
                switch (esc) {
                        break;
 
                switch (esc) {

+               case (ESCAPE_UNICODE):
+                       if (TERMENC_ASCII == p->enc) {
+                               encode1(p, '?');
+                               break;
+                       }
+                       uc = mchars_num2uc(seq + 1, sz - 1);
+                       if ('\0' != uc)
+                               encode1(p, uc);
+                       break;

                case (ESCAPE_NUMBERED):
                        if ('\0' != (c = mchars_num2char(seq, sz)))
                                encode(p, &c, 1);
                case (ESCAPE_NUMBERED):
                        if ('\0' != (c = mchars_num2char(seq, sz)))
                                encode(p, &c, 1);


@@ -512,6 +510,33 @@ bufferc(struct termp *p, char c)
        p->buf[p->col++] = c;
 }
 
        p->buf[p->col++] = c;
 }
 

+/*
+ * See encode().
+ * Do this for a single (probably unicode) value.
+ * Does not check for non-decorated glyphs.
+ */
+static void
+encode1(struct termp *p, int c)
+{
+       enum termfont     f;
+
+       if (p->col + 4 >= p->maxcols)
+               adjbuf(p, p->col + 4);
+
+       f = term_fonttop(p);
+
+       if (TERMFONT_NONE == f) {
+               p->buf[p->col++] = c;
+               return;
+       } else if (TERMFONT_UNDER == f) {
+               p->buf[p->col++] = '_';
+       } else
+               p->buf[p->col++] = c;
+
+       p->buf[p->col++] = 8;
+       p->buf[p->col++] = c;
+}
+

 static void
 encode(struct termp *p, const char *word, size_t sz)
 {
 static void
 encode(struct termp *p, const char *word, size_t sz)
 {


@@ -584,6 +609,7 @@ term_strlen(const struct termp *p, const char *cp)
                for (i = 0; i < rsz; i++)
                        sz += (*p->width)(p, *cp++);
 
                for (i = 0; i < rsz; i++)
                        sz += (*p->width)(p, *cp++);
 

+               c = 0;

                switch (*cp) {
                case ('\\'):
                        cp++;
                switch (*cp) {
                case ('\\'):
                        cp++;


@@ -591,6 +617,15 @@ term_strlen(const struct termp *p, const char *cp)
                        switch (mandoc_escape(&cp, &seq, &ssz)) {
                        case (ESCAPE_ERROR):
                                return(sz);
                        switch (mandoc_escape(&cp, &seq, &ssz)) {
                        case (ESCAPE_ERROR):
                                return(sz);

+                       case (ESCAPE_UNICODE):
+                               if (TERMENC_ASCII != p->enc) {
+                                       sz += (*p->width)(p, '?');
+                                       break;
+                               }
+                               c = mchars_num2uc(seq + 1, ssz - 1);
+                               if ('\0' != c)
+                                       sz += (*p->width)(p, c);
+                               break;

                        case (ESCAPE_NUMBERED):
                                c = mchars_num2char(seq, ssz);
                                if ('\0' != c)
                        case (ESCAPE_NUMBERED):
                                c = mchars_num2char(seq, ssz);
                                if ('\0' != c)