=================================================================== RCS file: /cvs/mandoc/html.c,v retrieving revision 1.54 retrieving revision 1.100 diff -u -p -r1.54 -r1.100 --- mandoc/html.c 2009/10/03 15:26:26 1.54 +++ mandoc/html.c 2010/05/25 12:37:20 1.100 @@ -1,4 +1,4 @@ -/* $Id: html.c,v 1.54 2009/10/03 15:26:26 kristaps Exp $ */ +/* $Id: html.c,v 1.100 2010/05/25 12:37:20 kristaps Exp $ */ /* * Copyright (c) 2008, 2009 Kristaps Dzonsons * @@ -14,54 +14,66 @@ * 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 -#include #include -#include +#include +#include #include +#include #include #include #include +#include "mandoc.h" +#include "out.h" #include "chars.h" #include "html.h" +#include "main.h" -#define DOCTYPE "-//W3C//DTD HTML 4.01//EN" -#define DTD "http://www.w3.org/TR/html4/strict.dtd" +#define UNCONST(a) ((void *)(uintptr_t)(const void *)(a)) struct htmldata { - char *name; + 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", @@ -74,36 +86,47 @@ static const char *const htmlattrs[ATTR_MAX] = { "width", "valign", "target", + "id", + "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; - char *toks[4], *v; + const char *toks[4]; + char *v; toks[0] = "style"; toks[1] = "man"; toks[2] = "includes"; toks[3] = NULL; - if (NULL == (h = calloc(1, sizeof(struct html)))) - return(NULL); - - SLIST_INIT(&h->tags); - SLIST_INIT(&h->ords); - - if (NULL == (h->symtab = chars_init(CHARS_HTML))) { - free(h); - return(NULL); + h = calloc(1, sizeof(struct html)); + if (NULL == h) { + perror(NULL); + exit(EXIT_FAILURE); } + h->type = type; + h->tags.head = NULL; + h->ords.head = NULL; + h->symtab = chars_init(CHARS_HTML); + while (outopts && *outopts) - switch (getsubopt(&outopts, toks, &v)) { + switch (getsubopt(&outopts, UNCONST(toks), &v)) { case (0): h->style = v; break; @@ -120,7 +143,22 @@ html_alloc(char *outopts) 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) { @@ -130,20 +168,16 @@ html_free(void *p) h = (struct html *)p; - while ( ! SLIST_EMPTY(&h->ords)) { - ord = SLIST_FIRST(&h->ords); - SLIST_REMOVE_HEAD(&h->ords, entry); + while ((ord = h->ords.head) != NULL) { + h->ords.head = ord->next; free(ord); } - while ( ! SLIST_EMPTY(&h->tags)) { - tag = SLIST_FIRST(&h->tags); - SLIST_REMOVE_HEAD(&h->tags, entry); + while ((tag = h->tags.head) != NULL) { + h->tags.head = tag->next; free(tag); } - if (h->buf) - free(h->buf); if (h->symtab) chars_free(h->symtab); @@ -183,161 +217,159 @@ print_gen_head(struct html *h) 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; - int i; 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; - for (i = 0; i < (int)sz; i++) - putchar(rhs[i]); + fwrite(rhs, 1, sz, stdout); } 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; - int i; 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; - for (i = 0; i < (int)sz; i++) - putchar(rhs[i]); + fwrite(rhs, 1, sz, stdout); } -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; + struct htmlpair tag; - wp = *p; - type = 1; + h->metal = h->metac; + h->metac = font; - if (0 == *(++wp)) { - *p = wp; - return; + /* FIXME: DECO_ROMAN should just close out preexisting. */ + + if (h->metaf && h->tags.head == h->metaf) + print_tagq(h, h->metaf); + + PAIR_CLASS_INIT(&tag, htmlfonts[font]); + h->metaf = print_otag(h, TAG_SPAN, 1, &tag); + return(h->metaf); +} + + +static void +print_metaf(struct html *h, enum roffdeco deco) +{ + enum htmlfont font; + + 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 ('(' == *wp) { - wp++; - if (0 == *wp || 0 == *(wp + 1)) { - *p = 0 == *wp ? wp : wp + 1; - return; - } + (void)print_ofont(h, font); +} - print_spec(h, wp, 2); - *p = ++wp; - return; - } else if ('*' == *wp) { - if (0 == *(++wp)) { - *p = wp; - return; - } +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' }; - switch (*wp) { - case ('('): - wp++; - if (0 == *wp || 0 == *(wp + 1)) { - *p = 0 == *wp ? wp : wp + 1; - return; - } + nospace = 0; - print_res(h, wp, 2); - *p = ++wp; - return; - case ('['): - type = 0; + for (; *p; p++) { + sz = strcspn(p, rejs); + + fwrite(p, 1, sz, stdout); + p += /* LINTED */ + sz; + + 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; - default: - print_res(h, wp, 1); - *p = wp; - return; - } - - } else if ('f' == *wp) { - if (0 == *(++wp)) { - *p = wp; - return; - } - switch (*wp) { - case ('B'): - /* TODO */ + seq = ++p; + len = a2roffdeco(&deco, &seq, &sz); + + switch (deco) { + case (DECO_RESERVED): + print_res(h, seq, sz); break; - case ('I'): - /* TODO */ + case (DECO_SPECIAL): + print_spec(h, seq, sz); break; - case ('P'): + case (DECO_PREVIOUS): /* FALLTHROUGH */ - case ('R'): - /* TODO */ + case (DECO_BOLD): + /* FALLTHROUGH */ + case (DECO_ITALIC): + /* FALLTHROUGH */ + case (DECO_ROMAN): + if (norecurse) + break; + print_metaf(h, deco); break; default: break; } - *p = wp; - return; + p += len - 1; - } else if ('[' != *wp) { - print_spec(h, wp, 1); - *p = wp; - return; + if (DECO_NOSPACE == deco && '\0' == *(p + 1)) + nospace = 1; } - wp++; - for (j = 0; *wp && ']' != *wp; wp++, j++) - /* Loop... */ ; - - if (0 == *wp) { - *p = wp; - return; - } - - if (type) - print_spec(h, wp - j, j); - else - print_res(h, wp - j, j); - - *p = wp; + return(nospace); } static void -print_encode(struct html *h, const char *p) +print_attr(struct html *h, const char *key, const char *val) { - - for (; *p; p++) { - if ('\\' == *p) { - print_escape(h, &p); - continue; - } - switch (*p) { - case ('<'): - printf("<"); - break; - case ('>'): - printf(">"); - break; - case ('&'): - printf("&"); - break; - default: - putchar(*p); - break; - } - } + printf(" %s=\"", key); + (void)print_encode(h, val, 1); + putchar('\"'); } @@ -348,61 +380,111 @@ print_otag(struct html *h, enum htmltag tag, int i; struct tag *t; + /* Push this tags onto the stack of open scopes. */ + if ( ! (HTML_NOSTACK & htmltags[tag].flags)) { - if (NULL == (t = malloc(sizeof(struct tag)))) - err(EXIT_FAILURE, "malloc"); + t = malloc(sizeof(struct tag)); + if (NULL == t) { + perror(NULL); + exit(EXIT_FAILURE); + } t->tag = tag; - SLIST_INSERT_HEAD(&h->tags, t, entry); + t->next = h->tags.head; + h->tags.head = t; } else t = NULL; if ( ! (HTML_NOSPACE & h->flags)) if ( ! (HTML_CLRLINE & htmltags[tag].flags)) - printf(" "); + 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); - printf("\""); + 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"); } - printf(">"); - h->flags |= HTML_NOSPACE; - if (HTML_CLRLINE & htmltags[tag].flags) - h->flags |= HTML_NEWLINE; - else - h->flags &= ~HTML_NEWLINE; + /* 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; return(t); } -/* ARGSUSED */ static void print_ctag(struct html *h, enum htmltag tag) { printf("", htmltags[tag].name); - if (HTML_CLRLINE & htmltags[tag].flags) + if (HTML_CLRLINE & htmltags[tag].flags) { h->flags |= HTML_NOSPACE; - if (HTML_CLRLINE & htmltags[tag].flags) - h->flags |= HTML_NEWLINE; - else - h->flags &= ~HTML_NEWLINE; + putchar('\n'); + } } -/* ARGSUSED */ void -print_gen_doctype(struct html *h) +print_gen_decls(struct html *h) { - - printf("", DOCTYPE, DTD); + + print_xmltype(h); + print_doctype(h); } +static void +print_xmltype(struct html *h) +{ + + if (HTML_XHTML_1_0_STRICT == h->type) + printf(""); +} + + +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("\n", + name, doctype, dtd); +} + + void print_text(struct html *h, const char *p) { @@ -424,8 +506,6 @@ print_text(struct html *h, const char *p) case(')'): /* FALLTHROUGH */ case(']'): - /* FALLTHROUGH */ - case('}'): if ( ! (HTML_IGNDELIM & h->flags)) h->flags |= HTML_NOSPACE; break; @@ -434,21 +514,21 @@ print_text(struct html *h, const char *p) } if ( ! (h->flags & HTML_NOSPACE)) - printf(" "); + putchar(' '); - h->flags &= ~HTML_NOSPACE; - h->flags &= ~HTML_NEWLINE; + assert(p); + if ( ! print_encode(h, p, 0)) + h->flags &= ~HTML_NOSPACE; - if (p) - print_encode(h, p); - + /* + * 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: @@ -462,10 +542,11 @@ print_tagq(struct html *h, const struct tag *until) { struct tag *tag; - while ( ! SLIST_EMPTY(&h->tags)) { - tag = SLIST_FIRST(&h->tags); + while ((tag = h->tags.head) != NULL) { + if (tag == h->metaf) + h->metaf = NULL; print_ctag(h, tag->tag); - SLIST_REMOVE_HEAD(&h->tags, entry); + h->tags.head = tag->next; free(tag); if (until && tag == until) return; @@ -478,12 +559,198 @@ print_stagq(struct html *h, const struct tag *suntil) { struct tag *tag; - while ( ! SLIST_EMPTY(&h->tags)) { - tag = SLIST_FIRST(&h->tags); + while ((tag = h->tags.head) != NULL) { if (suntil && tag == suntil) return; + if (tag == h->metaf) + h->metaf = NULL; print_ctag(h, tag->tag); - SLIST_REMOVE_HEAD(&h->tags, entry); + h->tags.head = tag->next; free(tag); + } +} + + +void +bufinit(struct html *h) +{ + + h->buf[0] = '\0'; + h->buflen = 0; +} + + +void +bufcat_style(struct html *h, const char *key, const char *val) +{ + + bufcat(h, key); + bufncat(h, ":", 1); + bufcat(h, val); + bufncat(h, ";", 1); +} + + +void +bufcat(struct html *h, const char *p) +{ + + bufncat(h, p, strlen(p)); +} + + +void +buffmt(struct html *h, const char *fmt, ...) +{ + va_list ap; + + va_start(ap, fmt); + (void)vsnprintf(h->buf + (int)h->buflen, + BUFSIZ - h->buflen - 1, fmt, ap); + va_end(ap); + h->buflen = strlen(h->buf); +} + + +void +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); + h->buflen += sz; +} + + +void +buffmt_includes(struct html *h, const char *name) +{ + const char *p, *pp; + + pp = h->base_includes; + + while (NULL != (p = strchr(pp, '%'))) { + bufncat(h, pp, (size_t)(p - pp)); + switch (*(p + 1)) { + case('I'): + bufcat(h, name); + break; + default: + bufncat(h, p, 2); + break; + } + pp = p + 2; + } + if (pp) + bufcat(h, pp); +} + + +void +buffmt_man(struct html *h, + const char *name, const char *sec) +{ + const char *p, *pp; + + pp = h->base_man; + + /* LINTED */ + while (NULL != (p = strchr(pp, '%'))) { + bufncat(h, pp, (size_t)(p - pp)); + switch (*(p + 1)) { + case('S'): + bufcat(h, sec ? sec : "1"); + break; + case('N'): + buffmt(h, name); + break; + default: + bufncat(h, p, 2); + break; + } + pp = p + 2; + } + if (pp) + bufcat(h, pp); +} + + +void +bufcat_su(struct html *h, const char *p, const struct roffsu *su) +{ + double v; + const char *u; + + v = su->scale; + + 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; + } + + if (su->pt) + buffmt(h, "%s: %f%s;", p, v, u); + else + /* LINTED */ + buffmt(h, "%s: %d%s;", p, (int)v, u); +} + + +void +html_idcat(char *dst, const char *src, int sz) +{ + int ssz; + + assert(sz); + + /* Cf. . */ + + 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; } }