]> git.cameronkatri.com Git - mandoc.git/blobdiff - html.c
git.ckatri.com / mandoc.git / blobdiff
? search:
summary | shortlog | log | commit | commitdiff | tree
raw | inline | side by side
Reduce memory and time consumption on certain malformed input files
[mandoc.git] / html.c
diff --git a/html.c b/html.c
index 138d3e4c6802140ef31a205096975edfce1f9cee..da9808579a36347313540777410ac6949d5bc221 100644 (file)
--- a/html.c
+++ b/html.c
@@ -1,6 +1,7 @@
-/*     $Id: html.c,v 1.112 2010/09/04 20:18:53 kristaps Exp $ */
+/*     $Id: html.c,v 1.184 2014/12/20 00:20:11 schwarze Exp $ */
 /*
 /*
- * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
+ * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
+ * Copyright (c) 2011, 2012, 2013, 2014 Ingo Schwarze <schwarze@openbsd.org>
  *
  * Permission to use, copy, modify, and distribute this software for any
  * purpose with or without fee is hereby granted, provided that the above
  *
  * Permission to use, copy, modify, and distribute this software for any
  * purpose with or without fee is hereby granted, provided that the above
@@ -14,9 +15,7 @@
  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  */
  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  */
-#ifdef HAVE_CONFIG_H
 #include "config.h"
 #include "config.h"
-#endif
 
 #include <sys/types.h>
 
 
 #include <sys/types.h>
 
@@ -30,8 +29,8 @@
 #include <unistd.h>
 
 #include "mandoc.h"
 #include <unistd.h>
 
 #include "mandoc.h"
+#include "mandoc_aux.h"
 #include "out.h"
 #include "out.h"
-#include "chars.h"
 #include "html.h"
 #include "main.h"
 
 #include "html.h"
 #include "main.h"
 
@@ -57,84 +56,110 @@ static     const struct htmldata htmltags[TAG_MAX] = {
        {"br",          HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_BR */
        {"a",           0}, /* TAG_A */
        {"table",       HTML_CLRLINE}, /* TAG_TABLE */
        {"br",          HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_BR */
        {"a",           0}, /* TAG_A */
        {"table",       HTML_CLRLINE}, /* TAG_TABLE */
+       {"tbody",       HTML_CLRLINE}, /* TAG_TBODY */
        {"col",         HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_COL */
        {"tr",          HTML_CLRLINE}, /* TAG_TR */
        {"td",          HTML_CLRLINE}, /* TAG_TD */
        {"li",          HTML_CLRLINE}, /* TAG_LI */
        {"ul",          HTML_CLRLINE}, /* TAG_UL */
        {"ol",          HTML_CLRLINE}, /* TAG_OL */
        {"col",         HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_COL */
        {"tr",          HTML_CLRLINE}, /* TAG_TR */
        {"td",          HTML_CLRLINE}, /* TAG_TD */
        {"li",          HTML_CLRLINE}, /* TAG_LI */
        {"ul",          HTML_CLRLINE}, /* TAG_UL */
        {"ol",          HTML_CLRLINE}, /* TAG_OL */
+       {"dl",          HTML_CLRLINE}, /* TAG_DL */
+       {"dt",          HTML_CLRLINE}, /* TAG_DT */
+       {"dd",          HTML_CLRLINE}, /* TAG_DD */
+       {"blockquote",  HTML_CLRLINE}, /* TAG_BLOCKQUOTE */
+       {"pre",         HTML_CLRLINE }, /* TAG_PRE */
+       {"b",           0 }, /* TAG_B */
+       {"i",           0 }, /* TAG_I */
+       {"code",        0 }, /* TAG_CODE */
+       {"small",       0 }, /* TAG_SMALL */
+       {"style",       HTML_CLRLINE}, /* TAG_STYLE */
+       {"math",        HTML_CLRLINE}, /* TAG_MATH */
+       {"mrow",        0}, /* TAG_MROW */
+       {"mi",          0}, /* TAG_MI */
+       {"mo",          0}, /* TAG_MO */
+       {"msup",        0}, /* TAG_MSUP */
+       {"msub",        0}, /* TAG_MSUB */
+       {"msubsup",     0}, /* TAG_MSUBSUP */
+       {"mfrac",       0}, /* TAG_MFRAC */
+       {"msqrt",       0}, /* TAG_MSQRT */
+       {"mfenced",     0}, /* TAG_MFENCED */
+       {"mtable",      0}, /* TAG_MTABLE */
+       {"mtr",         0}, /* TAG_MTR */
+       {"mtd",         0}, /* TAG_MTD */
+       {"munderover",  0}, /* TAG_MUNDEROVER */
+       {"munder",      0}, /* TAG_MUNDER*/
+       {"mover",       0}, /* TAG_MOVER*/
 };
 
 };
 
-static const char      *const htmlfonts[HTMLFONT_MAX] = {
-       "roman",
-       "bold",
-       "italic"
+static const char      *const htmlattrs[ATTR_MAX] = {
+       "name", /* ATTR_NAME */
+       "rel", /* ATTR_REL */
+       "href", /* ATTR_HREF */
+       "type", /* ATTR_TYPE */
+       "media", /* ATTR_MEDIA */
+       "class", /* ATTR_CLASS */
+       "style", /* ATTR_STYLE */
+       "id", /* ATTR_ID */
+       "colspan", /* ATTR_COLSPAN */
+       "charset", /* ATTR_CHARSET */
+       "open", /* ATTR_OPEN */
+       "close", /* ATTR_CLOSE */
+       "mathvariant", /* ATTR_MATHVARIANT */
 };
 
 };
 
-static const char      *const htmlattrs[ATTR_MAX] = {
-       "http-equiv",
-       "content",
-       "name",
-       "rel",
-       "href",
-       "type",
-       "media",
-       "class",
-       "style",
-       "width",
-       "valign",
-       "target",
-       "id",
-       "summary",
+static const char      *const roffscales[SCALE_MAX] = {
+       "cm", /* SCALE_CM */
+       "in", /* SCALE_IN */
+       "pc", /* SCALE_PC */
+       "pt", /* SCALE_PT */
+       "em", /* SCALE_EM */
+       "em", /* SCALE_MM */
+       "ex", /* SCALE_EN */
+       "ex", /* SCALE_BU */
+       "em", /* SCALE_VS */
+       "ex", /* SCALE_FS */
 };
 
 };
 
-static void              print_spec(struct html *, enum roffdeco,
-                               const char *, size_t);
-static void              print_res(struct html *, const char *, size_t);
-static void              print_ctag(struct html *, enum htmltag);
-static void              print_doctype(struct html *);
-static void              print_xmltype(struct html *);
-static int               print_encode(struct html *, const char *, int);
-static void              print_metaf(struct html *, enum roffdeco);
-static void              print_attr(struct html *, 
-                               const char *, const char *);
-static void             *ml_alloc(char *, enum htmltype);
-
-
-static void *
-ml_alloc(char *outopts, enum htmltype type)
+static void     bufncat(struct html *, const char *, size_t);
+static void     print_ctag(struct html *, struct tag *);
+static int      print_escape(char);
+static int      print_encode(struct html *, const char *, int);
+static void     print_metaf(struct html *, enum mandoc_esc);
+static void     print_attr(struct html *, const char *, const char *);
+
+
+void *
+html_alloc(const struct mchars *mchars, char *outopts)
 {
        struct html     *h;
 {
        struct html     *h;
-       const char      *toks[4];
+       const char      *toks[5];
        char            *v;
 
        toks[0] = "style";
        toks[1] = "man";
        toks[2] = "includes";
        char            *v;
 
        toks[0] = "style";
        toks[1] = "man";
        toks[2] = "includes";
-       toks[3] = NULL;
+       toks[3] = "fragment";
+       toks[4] = NULL;
 
 
-       h = calloc(1, sizeof(struct html));
-       if (NULL == h) {
-               perror(NULL);
-               exit((int)MANDOCLEVEL_SYSERR);
-       }
+       h = mandoc_calloc(1, sizeof(struct html));
 
 
-       h->type = type;
        h->tags.head = NULL;
        h->tags.head = NULL;
-       h->ords.head = NULL;
-       h->symtab = chars_init(CHARS_HTML);
+       h->symtab = mchars;
 
        while (outopts && *outopts)
                switch (getsubopt(&outopts, UNCONST(toks), &v)) {
 
        while (outopts && *outopts)
                switch (getsubopt(&outopts, UNCONST(toks), &v)) {
-               case (0):
+               case 0:
                        h->style = v;
                        break;
                        h->style = v;
                        break;
-               case (1):
+               case 1:
                        h->base_man = v;
                        break;
                        h->base_man = v;
                        break;
-               case (2):
+               case 2:
                        h->base_includes = v;
                        break;
                        h->base_includes = v;
                        break;
+               case 3:
+                       h->oflags |= HTML_FRAGMENT;
+                       break;
                default:
                        break;
                }
                default:
                        break;
                }
@@ -142,64 +167,43 @@ ml_alloc(char *outopts, enum htmltype type)
        return(h);
 }
 
        return(h);
 }
 
-void *
-html_alloc(char *outopts)
-{
-
-       return(ml_alloc(outopts, HTML_HTML_4_01_STRICT));
-}
-
-
-void *
-xhtml_alloc(char *outopts)
-{
-
-       return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT));
-}
-
-
 void
 html_free(void *p)
 {
        struct tag      *tag;
 void
 html_free(void *p)
 {
        struct tag      *tag;
-       struct ord      *ord;
        struct html     *h;
 
        h = (struct html *)p;
 
        struct html     *h;
 
        h = (struct html *)p;
 
-       while ((ord = h->ords.head) != NULL) { 
-               h->ords.head = ord->next;
-               free(ord);
-       }
-
        while ((tag = h->tags.head) != NULL) {
        while ((tag = h->tags.head) != NULL) {
-               h->tags.head = tag->next;       
+               h->tags.head = tag->next;
                free(tag);
        }
                free(tag);
        }
-       
-       if (h->symtab)
-               chars_free(h->symtab);
 
        free(h);
 }
 
 
        free(h);
 }
 
-
 void
 print_gen_head(struct html *h)
 {
        struct htmlpair  tag[4];
 void
 print_gen_head(struct html *h)
 {
        struct htmlpair  tag[4];
+       struct tag      *t;
 
 
-       tag[0].key = ATTR_HTTPEQUIV;
-       tag[0].val = "Content-Type";
-       tag[1].key = ATTR_CONTENT;
-       tag[1].val = "text/html; charset=utf-8";
-       print_otag(h, TAG_META, 2, tag);
+       tag[0].key = ATTR_CHARSET;
+       tag[0].val = "utf-8";
+       print_otag(h, TAG_META, 1, tag);
 
 
-       tag[0].key = ATTR_NAME;
-       tag[0].val = "resource-type";
-       tag[1].key = ATTR_CONTENT;
-       tag[1].val = "document";
-       print_otag(h, TAG_META, 2, tag);
+       /*
+        * Print a default style-sheet.
+        */
+       t = print_otag(h, TAG_STYLE, 0, NULL);
+       print_text(h, "table.head, table.foot { width: 100%; }\n"
+             "td.head-rtitle, td.foot-os { text-align: right; }\n"
+             "td.head-vol { text-align: center; }\n"
+             "table.foot td { width: 50%; }\n"
+             "table.head td { width: 33%; }\n"
+             "div.spacer { margin: 1em 0; }\n");
+       print_tagq(h, t);
 
        if (h->style) {
                tag[0].key = ATTR_REL;
 
        if (h->style) {
                tag[0].key = ATTR_REL;
@@ -214,168 +218,236 @@ print_gen_head(struct html *h)
        }
 }
 
        }
 }
 
-
-static void
-print_spec(struct html *h, enum roffdeco d, const char *p, size_t len)
-{
-       int              cp;
-       const char      *rhs;
-       size_t           sz;
-
-       if ((cp = chars_spec2cp(h->symtab, p, len)) > 0) {
-               printf("&#%d;", cp);
-               return;
-       } else if (-1 == cp && DECO_SSPECIAL == d) {
-               fwrite(p, 1, len, stdout);
-               return;
-       } else if (-1 == cp)
-               return;
-
-       if (NULL != (rhs = chars_spec2str(h->symtab, p, len, &sz)))
-               fwrite(rhs, 1, sz, stdout);
-}
-
-
 static void
 static void
-print_res(struct html *h, const char *p, size_t len)
+print_metaf(struct html *h, enum mandoc_esc deco)
 {
 {
-       int              cp;
-       const char      *rhs;
-       size_t           sz;
-
-       if ((cp = chars_res2cp(h->symtab, p, len)) > 0) {
-               printf("&#%d;", cp);
-               return;
-       } else if (-1 == cp)
-               return;
-
-       if (NULL != (rhs = chars_res2str(h->symtab, p, len, &sz)))
-               fwrite(rhs, 1, sz, stdout);
-}
+       enum htmlfont    font;
 
 
+       switch (deco) {
+       case ESCAPE_FONTPREV:
+               font = h->metal;
+               break;
+       case ESCAPE_FONTITALIC:
+               font = HTMLFONT_ITALIC;
+               break;
+       case ESCAPE_FONTBOLD:
+               font = HTMLFONT_BOLD;
+               break;
+       case ESCAPE_FONTBI:
+               font = HTMLFONT_BI;
+               break;
+       case ESCAPE_FONT:
+               /* FALLTHROUGH */
+       case ESCAPE_FONTROMAN:
+               font = HTMLFONT_NONE;
+               break;
+       default:
+               abort();
+               /* NOTREACHED */
+       }
 
 
-struct tag *
-print_ofont(struct html *h, enum htmlfont font)
-{
-       struct htmlpair  tag;
+       if (h->metaf) {
+               print_tagq(h, h->metaf);
+               h->metaf = NULL;
+       }
 
        h->metal = h->metac;
        h->metac = font;
 
 
        h->metal = h->metac;
        h->metac = font;
 
-       /* FIXME: DECO_ROMAN should just close out preexisting. */
+       switch (font) {
+       case HTMLFONT_ITALIC:
+               h->metaf = print_otag(h, TAG_I, 0, NULL);
+               break;
+       case HTMLFONT_BOLD:
+               h->metaf = print_otag(h, TAG_B, 0, NULL);
+               break;
+       case HTMLFONT_BI:
+               h->metaf = print_otag(h, TAG_B, 0, NULL);
+               print_otag(h, TAG_I, 0, NULL);
+               break;
+       default:
+               break;
+       }
+}
 
 
-       if (h->metaf && h->tags.head == h->metaf)
-               print_tagq(h, h->metaf);
+int
+html_strlen(const char *cp)
+{
+       size_t           rsz;
+       int              skip, sz;
+
+       /*
+        * Account for escaped sequences within string length
+        * calculations.  This follows the logic in term_strlen() as we
+        * must calculate the width of produced strings.
+        * Assume that characters are always width of "1".  This is
+        * hacky, but it gets the job done for approximation of widths.
+        */
 
 
-       PAIR_CLASS_INIT(&tag, htmlfonts[font]);
-       h->metaf = print_otag(h, TAG_SPAN, 1, &tag);
-       return(h->metaf);
+       sz = 0;
+       skip = 0;
+       while (1) {
+               rsz = strcspn(cp, "\\");
+               if (rsz) {
+                       cp += rsz;
+                       if (skip) {
+                               skip = 0;
+                               rsz--;
+                       }
+                       sz += rsz;
+               }
+               if ('\0' == *cp)
+                       break;
+               cp++;
+               switch (mandoc_escape(&cp, NULL, NULL)) {
+               case ESCAPE_ERROR:
+                       return(sz);
+               case ESCAPE_UNICODE:
+                       /* FALLTHROUGH */
+               case ESCAPE_NUMBERED:
+                       /* FALLTHROUGH */
+               case ESCAPE_SPECIAL:
+                       if (skip)
+                               skip = 0;
+                       else
+                               sz++;
+                       break;
+               case ESCAPE_SKIPCHAR:
+                       skip = 1;
+                       break;
+               default:
+                       break;
+               }
+       }
+       return(sz);
 }
 
 }
 
-
-static void
-print_metaf(struct html *h, enum roffdeco deco)
+static int
+print_escape(char c)
 {
 {
-       enum htmlfont    font;
 
 
-       switch (deco) {
-       case (DECO_PREVIOUS):
-               font = h->metal;
+       switch (c) {
+       case '<':
+               printf("&lt;");
                break;
                break;
-       case (DECO_ITALIC):
-               font = HTMLFONT_ITALIC;
+       case '>':
+               printf("&gt;");
                break;
                break;
-       case (DECO_BOLD):
-               font = HTMLFONT_BOLD;
+       case '&':
+               printf("&amp;");
                break;
                break;
-       case (DECO_ROMAN):
-               font = HTMLFONT_NONE;
+       case '"':
+               printf("&quot;");
+               break;
+       case ASCII_NBRSP:
+               putchar('-');
+               break;
+       case ASCII_HYPH:
+               putchar('-');
+               /* FALLTHROUGH */
+       case ASCII_BREAK:
                break;
        default:
                break;
        default:
-               abort();
-               /* NOTREACHED */
+               return(0);
        }
        }
-
-       (void)print_ofont(h, font);
+       return(1);
 }
 
 }
 
-
 static int
 print_encode(struct html *h, const char *p, int norecurse)
 {
        size_t           sz;
 static int
 print_encode(struct html *h, const char *p, int norecurse)
 {
        size_t           sz;
-       int              len, nospace;
+       int              c, len, nospace;
        const char      *seq;
        const char      *seq;
-       enum roffdeco    deco;
-       static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' };
+       enum mandoc_esc  esc;
+       static const char rejs[9] = { '\\', '<', '>', '&', '"',
+               ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' };
 
        nospace = 0;
 
 
        nospace = 0;
 
-       for (; *p; p++) {
+       while ('\0' != *p) {
+               if (HTML_SKIPCHAR & h->flags && '\\' != *p) {
+                       h->flags &= ~HTML_SKIPCHAR;
+                       p++;
+                       continue;
+               }
+
                sz = strcspn(p, rejs);
 
                fwrite(p, 1, sz, stdout);
                sz = strcspn(p, rejs);
 
                fwrite(p, 1, sz, stdout);
-               p += /* LINTED */
-                       sz;
+               p += (int)sz;
 
 
-               if ('<' == *p) {
-                       printf("&lt;");
-                       continue;
-               } else if ('>' == *p) {
-                       printf("&gt;");
-                       continue;
-               } else if ('&' == *p) {
-                       printf("&amp;");
-                       continue;
-               } else if (ASCII_HYPH == *p) {
-                       /*
-                        * Note: "soft hyphens" aren't graphically
-                        * displayed when not breaking the text; we want
-                        * them to be displayed.
-                        */
-                       /*printf("&#173;");*/
-                       putchar('-');
-                       continue;
-               } else if ('\0' == *p)
+               if ('\0' == *p)
                        break;
 
                        break;
 
-               seq = ++p;
-               len = a2roffdeco(&deco, &seq, &sz);
+               if (print_escape(*p++))
+                       continue;
 
 
-               switch (deco) {
-               case (DECO_RESERVED):
-                       print_res(h, seq, sz);
+               esc = mandoc_escape(&p, &seq, &len);
+               if (ESCAPE_ERROR == esc)
                        break;
                        break;
-               case (DECO_SSPECIAL):
+
+               switch (esc) {
+               case ESCAPE_FONT:
                        /* FALLTHROUGH */
                        /* FALLTHROUGH */
-               case (DECO_SPECIAL):
-                       print_spec(h, deco, seq, sz);
-                       break;
-               case (DECO_PREVIOUS):
+               case ESCAPE_FONTPREV:
                        /* FALLTHROUGH */
                        /* FALLTHROUGH */
-               case (DECO_BOLD):
+               case ESCAPE_FONTBOLD:
                        /* FALLTHROUGH */
                        /* FALLTHROUGH */
-               case (DECO_ITALIC):
+               case ESCAPE_FONTITALIC:
                        /* FALLTHROUGH */
                        /* FALLTHROUGH */
-               case (DECO_ROMAN):
-                       if (norecurse)
-                               break;
-                       print_metaf(h, deco);
-                       break;
+               case ESCAPE_FONTBI:
+                       /* FALLTHROUGH */
+               case ESCAPE_FONTROMAN:
+                       if (0 == norecurse)
+                               print_metaf(h, esc);
+                       continue;
+               case ESCAPE_SKIPCHAR:
+                       h->flags |= HTML_SKIPCHAR;
+                       continue;
                default:
                        break;
                }
 
                default:
                        break;
                }
 
-               p += len - 1;
+               if (h->flags & HTML_SKIPCHAR) {
+                       h->flags &= ~HTML_SKIPCHAR;
+                       continue;
+               }
 
 
-               if (DECO_NOSPACE == deco && '\0' == *(p + 1))
-                       nospace = 1;
+               switch (esc) {
+               case ESCAPE_UNICODE:
+                       /* Skip past "u" header. */
+                       c = mchars_num2uc(seq + 1, len - 1);
+                       break;
+               case ESCAPE_NUMBERED:
+                       c = mchars_num2char(seq, len);
+                       if (c < 0)
+                               continue;
+                       break;
+               case ESCAPE_SPECIAL:
+                       c = mchars_spec2cp(h->symtab, seq, len);
+                       if (c <= 0)
+                               continue;
+                       break;
+               case ESCAPE_NOSPACE:
+                       if ('\0' == *p)
+                               nospace = 1;
+                       continue;
+               default:
+                       continue;
+               }
+               if ((c < 0x20 && c != 0x09) ||
+                   (c > 0x7E && c < 0xA0))
+                       c = 0xFFFD;
+               if (c > 0x7E)
+                       printf("&#%d;", c);
+               else if ( ! print_escape(c))
+                       putchar(c);
        }
 
        return(nospace);
 }
 
        }
 
        return(nospace);
 }
 
-
 static void
 print_attr(struct html *h, const char *key, const char *val)
 {
 static void
 print_attr(struct html *h, const char *key, const char *val)
 {
@@ -384,9 +456,8 @@ print_attr(struct html *h, const char *key, const char *val)
        putchar('\"');
 }
 
        putchar('\"');
 }
 
-
 struct tag *
 struct tag *
-print_otag(struct html *h, enum htmltag tag, 
+print_otag(struct html *h, enum htmltag tag,
                int sz, const struct htmlpair *p)
 {
        int              i;
                int sz, const struct htmlpair *p)
 {
        int              i;
@@ -395,11 +466,7 @@ print_otag(struct html *h, enum htmltag tag,
        /* Push this tags onto the stack of open scopes. */
 
        if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
        /* Push this tags onto the stack of open scopes. */
 
        if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
-               t = malloc(sizeof(struct tag));
-               if (NULL == t) {
-                       perror(NULL);
-                       exit((int)MANDOCLEVEL_SYSERR);
-               }
+               t = mandoc_malloc(sizeof(struct tag));
                t->tag = tag;
                t->next = h->tags.head;
                h->tags.head = t;
                t->tag = tag;
                t->next = h->tags.head;
                h->tags.head = t;
@@ -428,115 +495,55 @@ print_otag(struct html *h, enum htmltag tag,
        for (i = 0; i < sz; i++)
                print_attr(h, htmlattrs[p[i].key], p[i].val);
 
        for (i = 0; i < sz; i++)
                print_attr(h, htmlattrs[p[i].key], p[i].val);
 
-       /* Add non-overridable attributes. */
-
-       if (TAG_HTML == tag && HTML_XHTML_1_0_STRICT == h->type) {
-               print_attr(h, "xmlns", "http://www.w3.org/1999/xhtml");
-               print_attr(h, "xml:lang", "en");
-               print_attr(h, "lang", "en");
-       }
-
-       /* Accomodate for XML "well-formed" singleton escaping. */
+       /* Accommodate for "well-formed" singleton escaping. */
 
        if (HTML_AUTOCLOSE & htmltags[tag].flags)
 
        if (HTML_AUTOCLOSE & htmltags[tag].flags)
-               switch (h->type) {
-               case (HTML_XHTML_1_0_STRICT):
-                       putchar('/');
-                       break;
-               default:
-                       break;
-               }
+               putchar('/');
 
        putchar('>');
 
        h->flags |= HTML_NOSPACE;
 
        putchar('>');
 
        h->flags |= HTML_NOSPACE;
+
+       if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags)
+               putchar('\n');
+
        return(t);
 }
 
        return(t);
 }
 
-
 static void
 static void
-print_ctag(struct html *h, enum htmltag tag)
+print_ctag(struct html *h, struct tag *tag)
 {
 {
-       
-       printf("</%s>", htmltags[tag].name);
-       if (HTML_CLRLINE & htmltags[tag].flags) {
+
+       /*
+        * Remember to close out and nullify the current
+        * meta-font and table, if applicable.
+        */
+       if (tag == h->metaf)
+               h->metaf = NULL;
+       if (tag == h->tblt)
+               h->tblt = NULL;
+
+       printf("</%s>", htmltags[tag->tag].name);
+       if (HTML_CLRLINE & htmltags[tag->tag].flags) {
                h->flags |= HTML_NOSPACE;
                putchar('\n');
                h->flags |= HTML_NOSPACE;
                putchar('\n');
-       } 
-}
+       }
 
 
+       h->tags.head = tag->next;
+       free(tag);
+}
 
 void
 print_gen_decls(struct html *h)
 {
 
 
 void
 print_gen_decls(struct html *h)
 {
 
-       print_xmltype(h);
-       print_doctype(h);
-}
-
-
-static void
-print_xmltype(struct html *h)
-{
-
-       if (HTML_XHTML_1_0_STRICT == h->type)
-               printf("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
+       puts("<!DOCTYPE html>");
 }
 
 }
 
-
-static void
-print_doctype(struct html *h)
-{
-       const char      *doctype;
-       const char      *dtd;
-       const char      *name;
-
-       switch (h->type) {
-       case (HTML_HTML_4_01_STRICT):
-               name = "HTML";
-               doctype = "-//W3C//DTD HTML 4.01//EN";
-               dtd = "http://www.w3.org/TR/html4/strict.dtd";
-               break;
-       default:
-               name = "html";
-               doctype = "-//W3C//DTD XHTML 1.0 Strict//EN";
-               dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";
-               break;
-       }
-
-       printf("<!DOCTYPE %s PUBLIC \"%s\" \"%s\">\n", 
-                       name, doctype, dtd);
-}
-
-
 void
 print_text(struct html *h, const char *word)
 {
 
 void
 print_text(struct html *h, const char *word)
 {
 
-       if (word[0] && '\0' == word[1])
-               switch (word[0]) {
-               case('.'):
-                       /* FALLTHROUGH */
-               case(','):
-                       /* FALLTHROUGH */
-               case(';'):
-                       /* FALLTHROUGH */
-               case(':'):
-                       /* FALLTHROUGH */
-               case('?'):
-                       /* FALLTHROUGH */
-               case('!'):
-                       /* FALLTHROUGH */
-               case(')'):
-                       /* FALLTHROUGH */
-               case(']'):
-                       if ( ! (HTML_IGNDELIM & h->flags))
-                               h->flags |= HTML_NOSPACE;
-                       break;
-               default:
-                       break;
-               }
-
        if ( ! (HTML_NOSPACE & h->flags)) {
                /* Manage keeps! */
                if ( ! (HTML_KEEP & h->flags)) {
        if ( ! (HTML_NOSPACE & h->flags)) {
                /* Manage keeps! */
                if ( ! (HTML_KEEP & h->flags)) {
@@ -547,27 +554,37 @@ print_text(struct html *h, const char *word)
                        printf("&#160;");
        }
 
                        printf("&#160;");
        }
 
+       assert(NULL == h->metaf);
+       switch (h->metac) {
+       case HTMLFONT_ITALIC:
+               h->metaf = print_otag(h, TAG_I, 0, NULL);
+               break;
+       case HTMLFONT_BOLD:
+               h->metaf = print_otag(h, TAG_B, 0, NULL);
+               break;
+       case HTMLFONT_BI:
+               h->metaf = print_otag(h, TAG_B, 0, NULL);
+               print_otag(h, TAG_I, 0, NULL);
+               break;
+       default:
+               break;
+       }
+
        assert(word);
        assert(word);
-       if ( ! print_encode(h, word, 0))
+       if ( ! print_encode(h, word, 0)) {
                if ( ! (h->flags & HTML_NONOSPACE))
                        h->flags &= ~HTML_NOSPACE;
                if ( ! (h->flags & HTML_NONOSPACE))
                        h->flags &= ~HTML_NOSPACE;
+               h->flags &= ~HTML_NONEWLINE;
+       } else
+               h->flags |= HTML_NOSPACE | HTML_NONEWLINE;
 
 
-       /* 
-        * Note that we don't process the pipe: the parser sees it as
-        * punctuation, but we don't in terms of typography.
-        */
-       if (word[0] && '\0' == word[1])
-               switch (word[0]) {
-               case('('):
-                       /* FALLTHROUGH */
-               case('['):
-                       h->flags |= HTML_NOSPACE;
-                       break;
-               default:
-                       break;
-               }
-}
+       if (h->metaf) {
+               print_tagq(h, h->metaf);
+               h->metaf = NULL;
+       }
 
 
+       h->flags &= ~HTML_IGNDELIM;
+}
 
 void
 print_tagq(struct html *h, const struct tag *until)
 
 void
 print_tagq(struct html *h, const struct tag *until)
@@ -575,17 +592,12 @@ print_tagq(struct html *h, const struct tag *until)
        struct tag      *tag;
 
        while ((tag = h->tags.head) != NULL) {
        struct tag      *tag;
 
        while ((tag = h->tags.head) != NULL) {
-               if (tag == h->metaf)
-                       h->metaf = NULL;
-               print_ctag(h, tag->tag);
-               h->tags.head = tag->next;
-               free(tag);
+               print_ctag(h, tag);
                if (until && tag == until)
                        return;
        }
 }
 
                if (until && tag == until)
                        return;
        }
 }
 
-
 void
 print_stagq(struct html *h, const struct tag *suntil)
 {
 void
 print_stagq(struct html *h, const struct tag *suntil)
 {
@@ -594,14 +606,21 @@ print_stagq(struct html *h, const struct tag *suntil)
        while ((tag = h->tags.head) != NULL) {
                if (suntil && tag == suntil)
                        return;
        while ((tag = h->tags.head) != NULL) {
                if (suntil && tag == suntil)
                        return;
-               if (tag == h->metaf)
-                       h->metaf = NULL;
-               print_ctag(h, tag->tag);
-               h->tags.head = tag->next;
-               free(tag);
+               print_ctag(h, tag);
        }
 }
 
        }
 }
 
+void
+print_paragraph(struct html *h)
+{
+       struct tag      *t;
+       struct htmlpair  tag;
+
+       PAIR_CLASS_INIT(&tag, "spacer");
+       t = print_otag(h, TAG_DIV, 1, &tag);
+       print_tagq(h, t);
+}
+
 
 void
 bufinit(struct html *h)
 
 void
 bufinit(struct html *h)
@@ -611,62 +630,63 @@ bufinit(struct html *h)
        h->buflen = 0;
 }
 
        h->buflen = 0;
 }
 
-
 void
 bufcat_style(struct html *h, const char *key, const char *val)
 {
 
        bufcat(h, key);
 void
 bufcat_style(struct html *h, const char *key, const char *val)
 {
 
        bufcat(h, key);
-       bufncat(h, ":", 1);
+       bufcat(h, ":");
        bufcat(h, val);
        bufcat(h, val);
-       bufncat(h, ";", 1);
+       bufcat(h, ";");
 }
 
 }
 
-
 void
 bufcat(struct html *h, const char *p)
 {
 
 void
 bufcat(struct html *h, const char *p)
 {
 
-       bufncat(h, p, strlen(p));
-}
+       /*
+        * XXX This is broken and not easy to fix.
+        * When using the -Oincludes option, buffmt_includes()
+        * may pass in strings overrunning BUFSIZ, causing a crash.
+        */
 
 
+       h->buflen = strlcat(h->buf, p, BUFSIZ);
+       assert(h->buflen < BUFSIZ);
+}
 
 void
 
 void
-buffmt(struct html *h, const char *fmt, ...)
+bufcat_fmt(struct html *h, const char *fmt, ...)
 {
        va_list          ap;
 
        va_start(ap, fmt);
 {
        va_list          ap;
 
        va_start(ap, fmt);
-       (void)vsnprintf(h->buf + (int)h->buflen, 
-                       BUFSIZ - h->buflen - 1, fmt, ap);
+       (void)vsnprintf(h->buf + (int)h->buflen,
+           BUFSIZ - h->buflen - 1, fmt, ap);
        va_end(ap);
        h->buflen = strlen(h->buf);
 }
 
        va_end(ap);
        h->buflen = strlen(h->buf);
 }
 
-
-void
+static void
 bufncat(struct html *h, const char *p, size_t sz)
 {
 
 bufncat(struct html *h, const char *p, size_t sz)
 {
 
-       if (h->buflen + sz > BUFSIZ - 1)
-               sz = BUFSIZ - 1 - h->buflen;
-
-       (void)strncat(h->buf, p, sz);
+       assert(h->buflen + sz + 1 < BUFSIZ);
+       strncat(h->buf, p, sz);
        h->buflen += sz;
 }
 
        h->buflen += sz;
 }
 
-
 void
 buffmt_includes(struct html *h, const char *name)
 {
        const char      *p, *pp;
 
        pp = h->base_includes;
 void
 buffmt_includes(struct html *h, const char *name)
 {
        const char      *p, *pp;
 
        pp = h->base_includes;
-       
+
+       bufinit(h);
        while (NULL != (p = strchr(pp, '%'))) {
                bufncat(h, pp, (size_t)(p - pp));
                switch (*(p + 1)) {
        while (NULL != (p = strchr(pp, '%'))) {
                bufncat(h, pp, (size_t)(p - pp));
                switch (*(p + 1)) {
-               case('I'):
+               case'I':
                        bufcat(h, name);
                        break;
                default:
                        bufcat(h, name);
                        break;
                default:
@@ -679,24 +699,22 @@ buffmt_includes(struct html *h, const char *name)
                bufcat(h, pp);
 }
 
                bufcat(h, pp);
 }
 
-
 void
 void
-buffmt_man(struct html *h, 
-               const char *name, const char *sec)
+buffmt_man(struct html *h, const char *name, const char *sec)
 {
        const char      *p, *pp;
 
        pp = h->base_man;
 {
        const char      *p, *pp;
 
        pp = h->base_man;
-       
-       /* LINTED */
+
+       bufinit(h);
        while (NULL != (p = strchr(pp, '%'))) {
                bufncat(h, pp, (size_t)(p - pp));
                switch (*(p + 1)) {
        while (NULL != (p = strchr(pp, '%'))) {
                bufncat(h, pp, (size_t)(p - pp));
                switch (*(p + 1)) {
-               case('S'):
+               case 'S':
                        bufcat(h, sec ? sec : "1");
                        break;
                        bufcat(h, sec ? sec : "1");
                        break;
-               case('N'):
-                       buffmt(h, name);
+               case 'N':
+                       bufcat_fmt(h, "%s", name);
                        break;
                default:
                        bufncat(h, p, 2);
                        break;
                default:
                        bufncat(h, p, 2);
@@ -708,81 +726,26 @@ buffmt_man(struct html *h,
                bufcat(h, pp);
 }
 
                bufcat(h, pp);
 }
 
-
 void
 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
 {
        double           v;
 void
 bufcat_su(struct html *h, const char *p, const struct roffsu *su)
 {
        double           v;
-       const char      *u;
 
        v = su->scale;
 
        v = su->scale;
+       if (SCALE_MM == su->unit && 0.0 == (v /= 100.0))
+               v = 1.0;
+       else if (SCALE_BU == su->unit)
+               v /= 24.0;
 
 
-       switch (su->unit) {
-       case (SCALE_CM):
-               u = "cm";
-               break;
-       case (SCALE_IN):
-               u = "in";
-               break;
-       case (SCALE_PC):
-               u = "pc";
-               break;
-       case (SCALE_PT):
-               u = "pt";
-               break;
-       case (SCALE_EM):
-               u = "em";
-               break;
-       case (SCALE_MM):
-               if (0 == (v /= 100))
-                       v = 1;
-               u = "em";
-               break;
-       case (SCALE_EN):
-               u = "ex";
-               break;
-       case (SCALE_BU):
-               u = "ex";
-               break;
-       case (SCALE_VS):
-               u = "em";
-               break;
-       default:
-               u = "ex";
-               break;
-       }
-
-       /* 
-        * XXX: the CSS spec isn't clear as to which types accept
-        * integer or real numbers, so we just make them all decimals.
-        */
-       buffmt(h, "%s: %.2f%s;", p, v, u);
+       bufcat_fmt(h, "%s: %.2f%s;", p, v, roffscales[su->unit]);
 }
 
 }
 
-
 void
 void
-html_idcat(char *dst, const char *src, int sz)
+bufcat_id(struct html *h, const char *src)
 {
 {
-       int              ssz;
-
-       assert(sz);
 
        /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
 
 
        /* Cf. <http://www.w3.org/TR/html4/types.html#h-6.2>. */
 
-       for ( ; *dst != '\0' && sz; dst++, sz--)
-               /* Jump to end. */ ;
-
-       assert(sz > 2);
-
-       /* We can't start with a number (bah). */
-
-       *dst++ = 'x';
-       *dst = '\0';
-       sz--;
-
-       for ( ; *src != '\0' && sz > 1; src++) {
-               ssz = snprintf(dst, (size_t)sz, "%.2x", *src);
-               sz -= ssz;
-               dst += ssz;
-       }
+       while ('\0' != *src)
+               bufcat_fmt(h, "%.2x", *src++);
 }
 }
A git conversion mirror of mandoc.bsd.lv