-/* $Id: html.c,v 1.80 2009/11/02 06:22:44 kristaps Exp $ */
+/* $Id: html.c,v 1.100 2010/05/25 12:37:20 kristaps Exp $ */
/*
* Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
*
* 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"
+#endif
+
#include <sys/types.h>
#include <assert.h>
#include <string.h>
#include <unistd.h>
+#include "mandoc.h"
#include "out.h"
#include "chars.h"
#include "html.h"
#define UNCONST(a) ((void *)(uintptr_t)(const void *)(a))
-#define DOCTYPE "-//W3C//DTD HTML 4.01//EN"
-#define DTD "http://www.w3.org/TR/html4/strict.dtd"
-
struct htmldata {
const char *name;
int flags;
#define HTML_CLRLINE (1 << 0)
#define HTML_NOSTACK (1 << 1)
+#define HTML_AUTOCLOSE (1 << 2) /* Tag has auto-closure. */
};
static const struct htmldata htmltags[TAG_MAX] = {
{"html", HTML_CLRLINE}, /* TAG_HTML */
{"head", HTML_CLRLINE}, /* TAG_HEAD */
{"body", HTML_CLRLINE}, /* TAG_BODY */
- {"meta", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_META */
+ {"meta", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_META */
{"title", HTML_CLRLINE}, /* TAG_TITLE */
{"div", HTML_CLRLINE}, /* TAG_DIV */
{"h1", 0}, /* TAG_H1 */
{"h2", 0}, /* TAG_H2 */
- {"p", HTML_CLRLINE}, /* TAG_P */
{"span", 0}, /* TAG_SPAN */
- {"link", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
- {"br", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */
+ {"link", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_LINK */
+ {"br", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_BR */
{"a", 0}, /* TAG_A */
{"table", HTML_CLRLINE}, /* TAG_TABLE */
- {"col", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_COL */
+ {"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 */
- {"base", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_BASE */
};
-static const char *const htmlattrs[ATTR_MAX] = {
+static const char *const htmlfonts[HTMLFONT_MAX] = {
+ "roman",
+ "bold",
+ "italic"
+};
+
+static const char *const htmlattrs[ATTR_MAX] = {
"http-equiv",
"content",
"name",
"summary",
};
-#ifdef __linux__
-extern int getsubopt(char **, char * const *, char **);
-#endif
+static void print_spec(struct html *, 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);
-void *
-html_alloc(char *outopts)
+
+static void *
+ml_alloc(char *outopts, enum htmltype type)
{
struct html *h;
const char *toks[4];
exit(EXIT_FAILURE);
}
+ h->type = type;
h->tags.head = NULL;
h->ords.head = NULL;
h->symtab = chars_init(CHARS_HTML);
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)
static void
-print_spec(struct html *h, const char *p, int len)
+print_spec(struct html *h, const char *p, size_t len)
{
const char *rhs;
size_t sz;
- rhs = chars_a2ascii(h->symtab, p, (size_t)len, &sz);
+ rhs = chars_a2ascii(h->symtab, p, len, &sz);
if (NULL == rhs)
return;
static void
-print_res(struct html *h, const char *p, int len)
+print_res(struct html *h, const char *p, size_t len)
{
const char *rhs;
size_t sz;
- rhs = chars_a2res(h->symtab, p, (size_t)len, &sz);
+ rhs = chars_a2res(h->symtab, p, len, &sz);
if (NULL == rhs)
return;
}
-static void
-print_escape(struct html *h, const char **p)
+struct tag *
+print_ofont(struct html *h, enum htmlfont font)
{
- int j, type;
- const char *wp;
-
- wp = *p;
- type = 1;
-
- if (0 == *(++wp)) {
- *p = wp;
- return;
- }
-
- if ('(' == *wp) {
- wp++;
- if (0 == *wp || 0 == *(wp + 1)) {
- *p = 0 == *wp ? wp : wp + 1;
- return;
- }
-
- print_spec(h, wp, 2);
- *p = ++wp;
- return;
+ struct htmlpair tag;
- } else if ('*' == *wp) {
- if (0 == *(++wp)) {
- *p = wp;
- return;
- }
+ h->metal = h->metac;
+ h->metac = font;
- switch (*wp) {
- case ('('):
- wp++;
- if (0 == *wp || 0 == *(wp + 1)) {
- *p = 0 == *wp ? wp : wp + 1;
- return;
- }
-
- print_res(h, wp, 2);
- *p = ++wp;
- return;
- case ('['):
- type = 0;
- break;
- default:
- print_res(h, wp, 1);
- *p = wp;
- return;
- }
-
- } else if ('f' == *wp) {
- if (0 == *(++wp)) {
- *p = wp;
- return;
- }
+ /* FIXME: DECO_ROMAN should just close out preexisting. */
- switch (*wp) {
- case ('B'):
- /* TODO */
- break;
- case ('I'):
- /* TODO */
- break;
- case ('P'):
- /* FALLTHROUGH */
- case ('R'):
- /* TODO */
- break;
- default:
- break;
- }
+ if (h->metaf && h->tags.head == h->metaf)
+ print_tagq(h, h->metaf);
- *p = wp;
- return;
+ PAIR_CLASS_INIT(&tag, htmlfonts[font]);
+ h->metaf = print_otag(h, TAG_SPAN, 1, &tag);
+ return(h->metaf);
+}
- } else if ('[' != *wp) {
- print_spec(h, wp, 1);
- *p = wp;
- return;
- }
- wp++;
- for (j = 0; *wp && ']' != *wp; wp++, j++)
- /* Loop... */ ;
+static void
+print_metaf(struct html *h, enum roffdeco deco)
+{
+ enum htmlfont font;
- if (0 == *wp) {
- *p = wp;
- return;
+ switch (deco) {
+ case (DECO_PREVIOUS):
+ font = h->metal;
+ break;
+ case (DECO_ITALIC):
+ font = HTMLFONT_ITALIC;
+ break;
+ case (DECO_BOLD):
+ font = HTMLFONT_BOLD;
+ break;
+ case (DECO_ROMAN):
+ font = HTMLFONT_NONE;
+ break;
+ default:
+ abort();
+ /* NOTREACHED */
}
- if (type)
- print_spec(h, wp - j, j);
- else
- print_res(h, wp - j, j);
-
- *p = wp;
+ (void)print_ofont(h, font);
}
-static void
-print_encode(struct html *h, const char *p)
+static int
+print_encode(struct html *h, const char *p, int norecurse)
{
size_t sz;
+ int len, nospace;
+ const char *seq;
+ enum roffdeco deco;
+ static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' };
+
+ nospace = 0;
for (; *p; p++) {
- sz = strcspn(p, "\\<>&");
+ sz = strcspn(p, rejs);
fwrite(p, 1, sz, stdout);
p += /* LINTED */
sz;
- if ('\\' == *p) {
- print_escape(h, &p);
+ if ('<' == *p) {
+ printf("<");
+ continue;
+ } else if ('>' == *p) {
+ printf(">");
+ continue;
+ } else if ('&' == *p) {
+ printf("&");
+ 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("­");*/
+ putchar('-');
continue;
} else if ('\0' == *p)
break;
- if ('<' == *p)
- printf("<");
- else if ('>' == *p)
- printf(">");
- else if ('&' == *p)
- printf("&");
+ seq = ++p;
+ len = a2roffdeco(&deco, &seq, &sz);
+
+ switch (deco) {
+ case (DECO_RESERVED):
+ print_res(h, seq, sz);
+ break;
+ case (DECO_SPECIAL):
+ print_spec(h, seq, sz);
+ break;
+ case (DECO_PREVIOUS):
+ /* FALLTHROUGH */
+ case (DECO_BOLD):
+ /* FALLTHROUGH */
+ case (DECO_ITALIC):
+ /* FALLTHROUGH */
+ case (DECO_ROMAN):
+ if (norecurse)
+ break;
+ print_metaf(h, deco);
+ break;
+ default:
+ break;
+ }
+
+ p += len - 1;
+
+ if (DECO_NOSPACE == deco && '\0' == *(p + 1))
+ nospace = 1;
}
+
+ return(nospace);
+}
+
+
+static void
+print_attr(struct html *h, const char *key, const char *val)
+{
+ printf(" %s=\"", key);
+ (void)print_encode(h, val, 1);
+ putchar('\"');
}
int i;
struct tag *t;
+ /* Push this tags onto the stack of open scopes. */
+
if ( ! (HTML_NOSTACK & htmltags[tag].flags)) {
t = malloc(sizeof(struct tag));
if (NULL == t) {
if ( ! (HTML_CLRLINE & htmltags[tag].flags))
putchar(' ');
+ /* Print out the tag name and attributes. */
+
printf("<%s", htmltags[tag].name);
- for (i = 0; i < sz; i++) {
- printf(" %s=\"", htmlattrs[p[i].key]);
- assert(p->val);
- print_encode(h, p[i].val);
- putchar('\"');
+ 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. */
+
+ if (HTML_AUTOCLOSE & htmltags[tag].flags)
+ switch (h->type) {
+ case (HTML_XHTML_1_0_STRICT):
+ putchar('/');
+ break;
+ default:
+ break;
+ }
+
putchar('>');
h->flags |= HTML_NOSPACE;
- if (HTML_CLRLINE & htmltags[tag].flags)
- h->flags |= HTML_NEWLINE;
- else
- h->flags &= ~HTML_NEWLINE;
-
return(t);
}
-/* ARGSUSED */
static void
print_ctag(struct html *h, enum htmltag tag)
{
printf("</%s>", htmltags[tag].name);
if (HTML_CLRLINE & htmltags[tag].flags) {
h->flags |= HTML_NOSPACE;
- h->flags |= HTML_NEWLINE;
putchar('\n');
- } else
- h->flags &= ~HTML_NEWLINE;
+ }
}
-/* ARGSUSED */
void
-print_gen_doctype(struct html *h)
+print_gen_decls(struct html *h)
{
-
- printf("<!DOCTYPE HTML PUBLIC \"%s\" \"%s\">", DOCTYPE, DTD);
+
+ 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\"?>");
+}
+
+
+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);
}
case(')'):
/* FALLTHROUGH */
case(']'):
- /* FALLTHROUGH */
- case('}'):
if ( ! (HTML_IGNDELIM & h->flags))
h->flags |= HTML_NOSPACE;
break;
if ( ! (h->flags & HTML_NOSPACE))
putchar(' ');
- h->flags &= ~HTML_NOSPACE;
- h->flags &= ~HTML_NEWLINE;
-
- if (p)
- print_encode(h, p);
+ assert(p);
+ if ( ! print_encode(h, p, 0))
+ h->flags &= ~HTML_NOSPACE;
+ /*
+ * Note that we don't process the pipe: the parser sees it as
+ * punctuation, but we don't in terms of typography.
+ */
if (*p && 0 == *(p + 1))
switch (*p) {
case('('):
/* FALLTHROUGH */
case('['):
- /* FALLTHROUGH */
- case('{'):
h->flags |= HTML_NOSPACE;
break;
default:
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);
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);
git.ckatri.com / mandoc.git / blobdiff