new: escape sequence handling

[mandoc.git] / roff.c

diff --git a/roff.c b/roff.c
index 0d845c7ca2a22b9bccdbc62bc9510c6ee4ff6e74..7467fe0e11c29305693c5fff6d51a86dbd363647 100644 (file)
--- a/roff.c
+++ b/roff.c
@@ -1,4 +1,4 @@
-/*     $Id: roff.c,v 1.94 2010/07/07 15:04:54 kristaps Exp $ */
+/*     $Id: roff.c,v 1.102 2010/09/04 20:18:53 kristaps Exp $ */

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


@@ -59,7 +59,7 @@ enum  rofft {
        ROFF_rm,
        ROFF_tr,
        ROFF_cblock,
        ROFF_rm,
        ROFF_tr,
        ROFF_cblock,

-       ROFF_ccond,
+       ROFF_ccond, /* FIXME: remove this. */

        ROFF_nr,
        ROFF_MAX
 };
        ROFF_nr,
        ROFF_MAX
 };


@@ -132,10 +132,11 @@ static    const char      *roff_getstrn(const struct roff *,
                                const char *, size_t);
 static enum rofferr     roff_line(ROFF_ARGS);
 static enum rofferr     roff_nr(ROFF_ARGS);
                                const char *, size_t);
 static enum rofferr     roff_line(ROFF_ARGS);
 static enum rofferr     roff_nr(ROFF_ARGS);

-static int              roff_res(struct roff *, int, 
-                               char **, size_t *, int, int *);
+static int              roff_res(struct roff *, 
+                               char **, size_t *, int);

 static void             roff_setstr(struct roff *,
                                const char *, const char *);
 static void             roff_setstr(struct roff *,
                                const char *, const char *);

+static char            *roff_strdup(const char *);

 
 /* See roff_hash_find() */
 
 
 /* See roff_hash_find() */
 


@@ -168,7 +169,7 @@ static      void             roff_free1(struct roff *);
 static enum rofft       roff_hash_find(const char *);
 static void             roff_hash_init(void);
 static void             roffnode_cleanscope(struct roff *);
 static enum rofft       roff_hash_find(const char *);
 static void             roff_hash_init(void);
 static void             roffnode_cleanscope(struct roff *);

-static int              roffnode_push(struct roff *, 
+static void             roffnode_push(struct roff *, 

                                enum rofft, int, int);
 static void             roffnode_pop(struct roff *);
 static enum rofft       roff_parse(const char *, int *);
                                enum rofft, int, int);
 static void             roffnode_pop(struct roff *);
 static enum rofft       roff_parse(const char *, int *);


@@ -247,6 +248,7 @@ roffnode_pop(struct roff *r)
                if (r->rstackpos > -1)
                        r->rstackpos--;
 
                if (r->rstackpos > -1)
                        r->rstackpos--;
 

+       ROFF_DEBUG("roff: popping scope\n");

        r->last = r->last->parent;
        if (p->end)
                free(p->end);
        r->last = r->last->parent;
        if (p->end)
                free(p->end);


@@ -258,16 +260,12 @@ roffnode_pop(struct roff *r)
  * Push a roff node onto the instruction stack.  This must later be
  * removed with roffnode_pop().
  */
  * Push a roff node onto the instruction stack.  This must later be
  * removed with roffnode_pop().
  */

-static int
+static void

 roffnode_push(struct roff *r, enum rofft tok, int line, int col)
 {
        struct roffnode *p;
 
 roffnode_push(struct roff *r, enum rofft tok, int line, int col)
 {
        struct roffnode *p;
 

-       if (NULL == (p = calloc(1, sizeof(struct roffnode)))) {
-               (*r->msg)(MANDOCERR_MEM, r->data, line, col, NULL);
-               return(0);
-       }
-
+       p = mandoc_calloc(1, sizeof(struct roffnode));

        p->tok = tok;
        p->parent = r->last;
        p->line = line;
        p->tok = tok;
        p->parent = r->last;
        p->line = line;


@@ -275,7 +273,6 @@ roffnode_push(struct roff *r, enum rofft tok, int line, int col)
        p->rule = p->parent ? p->parent->rule : ROFFRULE_DENY;
 
        r->last = p;
        p->rule = p->parent ? p->parent->rule : ROFFRULE_DENY;
 
        r->last = p;

-       return(1);

 }
 
 
 }
 
 


@@ -307,15 +304,11 @@ roff_free(struct roff *r)
 
 
 struct roff *
 
 
 struct roff *

-roff_alloc(struct regset *regs, const mandocmsg msg, void *data)
+roff_alloc(struct regset *regs, void *data, const mandocmsg msg)

 {
        struct roff     *r;
 
 {
        struct roff     *r;
 

-       if (NULL == (r = calloc(1, sizeof(struct roff)))) {
-               (*msg)(MANDOCERR_MEM, data, 0, 0, NULL);
-               return(0);
-       }
-
+       r = mandoc_calloc(1, sizeof(struct roff));

        r->regs = regs;
        r->msg = msg;
        r->data = data;
        r->regs = regs;
        r->msg = msg;
        r->data = data;


@@ -332,14 +325,14 @@ roff_alloc(struct regset *regs, const mandocmsg msg, void *data)
  * is processed. 
  */
 static int
  * is processed. 
  */
 static int

-roff_res(struct roff *r, int ln, char **bufp,
-               size_t *szp, int pos, int *offs)
+roff_res(struct roff *r, char **bufp, size_t *szp, int pos)

 {
        const char      *cp, *cpp, *st, *res;
        int              i, maxl;
        size_t           nsz;
        char            *n;
 
 {
        const char      *cp, *cpp, *st, *res;
        int              i, maxl;
        size_t           nsz;
        char            *n;
 

+       /* LINTED */

        for (cp = &(*bufp)[pos]; (cpp = strstr(cp, "\\*")); cp++) {
                cp = cpp + 2;
                switch (*cp) {
        for (cp = &(*bufp)[pos]; (cpp = strstr(cp, "\\*")); cp++) {
                cp = cpp + 2;
                switch (*cp) {


@@ -406,7 +399,7 @@ roff_parseln(struct roff *r, int ln, char **bufp,
         * words to fill in.
         */
 
         * words to fill in.
         */
 

-       if (r->first_string && ! roff_res(r, ln, bufp, szp, pos, offs))
+       if (r->first_string && ! roff_res(r, bufp, szp, pos))

                return(ROFF_RERUN);
 
        /*
                return(ROFF_RERUN);
 
        /*


@@ -434,8 +427,8 @@ roff_parseln(struct roff *r, int ln, char **bufp,
        if (r->last) {
                t = r->last->tok;
                assert(roffs[t].sub);
        if (r->last) {
                t = r->last->tok;
                assert(roffs[t].sub);

-               ROFF_DEBUG("roff: intercept scoped context: %s\n", 
-                               roffs[t].name);
+               ROFF_DEBUG("roff: intercept scoped context: %s, [%s]\n", 
+                               roffs[t].name, &(*bufp)[pos]);

                return((*roffs[t].sub)
                                (r, t, bufp, szp, 
                                 ln, pos, pos, offs));
                return((*roffs[t].sub)
                                (r, t, bufp, szp, 
                                 ln, pos, pos, offs));


@@ -650,8 +643,7 @@ roff_block(ROFF_ARGS)
                        pos++;
        }
 
                        pos++;
        }
 

-       if ( ! roffnode_push(r, tok, ln, ppos))
-               return(ROFF_ERR);
+       roffnode_push(r, tok, ln, ppos);

 
        if ('\0' == (*bufp)[pos])
                return(ROFF_IGN);
 
        if ('\0' == (*bufp)[pos])
                return(ROFF_IGN);


@@ -673,12 +665,7 @@ roff_block(ROFF_ARGS)
        if (1 == sz && '.' == (*bufp)[sv])
                return(ROFF_IGN);
 
        if (1 == sz && '.' == (*bufp)[sv])
                return(ROFF_IGN);
 

-       r->last->end = malloc(sz + 1);
-
-       if (NULL == r->last->end) {
-               (*r->msg)(MANDOCERR_MEM, r->data, ln, pos, NULL);
-               return(ROFF_ERR);
-       }
+       r->last->end = mandoc_malloc(sz + 1);

 
        memcpy(r->last->end, *bufp + sv, sz);
        r->last->end[(int)sz] = '\0';
 
        memcpy(r->last->end, *bufp + sv, sz);
        r->last->end[(int)sz] = '\0';


@@ -762,7 +749,6 @@ roff_cond_sub(ROFF_ARGS)
 {
        enum rofft       t;
        enum roffrule    rr;
 {
        enum rofft       t;
        enum roffrule    rr;

-       struct roffnode *l;

 
        ppos = pos;
        rr = r->last->rule;
 
        ppos = pos;
        rr = r->last->rule;


@@ -772,14 +758,15 @@ roff_cond_sub(ROFF_ARGS)
         * continue. 
         */
 
         * continue. 
         */
 

-       l = r->last;

        roffnode_cleanscope(r);
 
        roffnode_cleanscope(r);
 

-       if (l != r->last)
-               return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT);
-
-       if (ROFF_MAX == (t = roff_parse(*bufp, &pos)))
+       if (ROFF_MAX == (t = roff_parse(*bufp, &pos))) {
+               if ('\\' == (*bufp)[pos] && '}' == (*bufp)[pos + 1])
+                       return(roff_ccond
+                               (r, ROFF_ccond, bufp, szp, 
+                                ln, pos, pos + 2, offs));

                return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT);
                return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT);

+       }

 
        /*
         * A denied conditional must evaluate its children if and only
 
        /*
         * A denied conditional must evaluate its children if and only


@@ -811,6 +798,8 @@ roff_cond_text(ROFF_ARGS)
         * scope permits us to do so.
         */
 
         * scope permits us to do so.
         */
 

+       /* FIXME: use roff_ccond? */
+

        st = &(*bufp)[pos];
        if (NULL == (ep = strstr(st, "\\}"))) {
                roffnode_cleanscope(r);
        st = &(*bufp)[pos];
        if (NULL == (ep = strstr(st, "\\}"))) {
                roffnode_cleanscope(r);


@@ -905,8 +894,7 @@ roff_cond(ROFF_ARGS)
                return(ROFF_ERR);
        }
 
                return(ROFF_ERR);
        }
 

-       if ( ! roffnode_push(r, tok, ln, ppos))
-               return(ROFF_ERR);
+       roffnode_push(r, tok, ln, ppos);

 
        r->last->rule = rule;
 
 
        r->last->rule = rule;
 


@@ -969,30 +957,40 @@ roff_cond(ROFF_ARGS)
 static enum rofferr
 roff_ds(ROFF_ARGS)
 {
 static enum rofferr
 roff_ds(ROFF_ARGS)
 {

-       char *name, *string, *end;
+       char            *name, *string;
+
+       /*
+        * A symbol is named by the first word following the macro
+        * invocation up to a space.  Its value is anything after the
+        * name's trailing whitespace and optional double-quote.  Thus,
+        *
+        *  [.ds foo "bar  "     ]
+        *
+        * will have `bar  "     ' as its value.
+        */

 
        name = *bufp + pos;
        if ('\0' == *name)
                return(ROFF_IGN);
 
        string = name;
 
        name = *bufp + pos;
        if ('\0' == *name)
                return(ROFF_IGN);
 
        string = name;

+       /* Read until end of name. */

        while (*string && ' ' != *string)
                string++;
        while (*string && ' ' != *string)
                string++;

+
+       /* Nil-terminate name. */

        if (*string)
        if (*string)

-               *(string++) = NULL;
-       if (*string && '"' == *string)
-               string++;
+               *(string++) = '\0';
+       
+       /* Read past spaces. */

        while (*string && ' ' == *string)
                string++;
        while (*string && ' ' == *string)
                string++;

-       end = string;
-       while (*end)
-               end++;
-       if (string < end) {
-               end--;
-               if (*end == '"')
-                       *end = '\0';
-       }

 
 

+       /* Read passed initial double-quote. */
+       if (*string && '"' == *string)
+               string++;
+
+       /* The rest is the value. */

        roff_setstr(r, name, string);
        return(ROFF_IGN);
 }
        roff_setstr(r, name, string);
        return(ROFF_IGN);
 }


@@ -1041,6 +1039,27 @@ roff_nr(ROFF_ARGS)
 }
 
 
 }
 
 

+static char *
+roff_strdup(const char *name)
+{
+       char            *namecopy, *sv;
+
+       /* 
+        * This isn't a nice simple mandoc_strdup() because we must
+        * handle roff's stupid double-escape rule. 
+        */
+       sv = namecopy = mandoc_malloc(strlen(name) + 1);
+       while (*name) {
+               if ('\\' == *name && '\\' == *(name + 1))
+                       name++;
+               *namecopy++ = *name++;
+       }
+
+       *namecopy = '\0';
+       return(sv);
+}
+
+

 static void
 roff_setstr(struct roff *r, const char *name, const char *string)
 {
 static void
 roff_setstr(struct roff *r, const char *name, const char *string)
 {


@@ -1060,7 +1079,9 @@ roff_setstr(struct roff *r, const char *name, const char *string)
        } else
                free(n->string);
 
        } else
                free(n->string);
 

-       n->string = string ? strdup(string) : NULL;
+       /* Don't use mandoc_strdup: clean out double-escapes. */
+       n->string = string ? roff_strdup(string) : NULL;
+       ROFF_DEBUG("roff: new symbol: [%s] = [%s]\n", name, n->string);

 }
 
 
 }
 
 


@@ -1070,7 +1091,7 @@ roff_getstrn(const struct roff *r, const char *name, size_t len)
        const struct roffstr *n;
 
        n = r->first_string;
        const struct roffstr *n;
 
        n = r->first_string;

-       while (n && (strncmp(name, n->name, len) || '\0' != n->name[len]))
+       while (n && (strncmp(name, n->name, len) || '\0' != n->name[(int)len]))

                n = n->next;
 
        return(n ? n->string : NULL);
                n = n->next;
 
        return(n ? n->string : NULL);