=================================================================== RCS file: /cvs/mandoc/html.c,v retrieving revision 1.32 retrieving revision 1.54 diff -u -p -r1.32 -r1.54 --- mandoc/html.c 2009/09/17 08:21:42 1.32 +++ mandoc/html.c 2009/10/03 15:26:26 1.54 @@ -1,4 +1,4 @@ -/* $Id: html.c,v 1.32 2009/09/17 08:21:42 kristaps Exp $ */ +/* $Id: html.c,v 1.54 2009/10/03 15:26:26 kristaps Exp $ */ /* * Copyright (c) 2008, 2009 Kristaps Dzonsons * @@ -14,49 +14,22 @@ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ +#include #include #include #include #include #include +#include +#include #include "chars.h" -#include "mdoc.h" -#include "man.h" +#include "html.h" #define DOCTYPE "-//W3C//DTD HTML 4.01//EN" #define DTD "http://www.w3.org/TR/html4/strict.dtd" -enum htmltag { - TAG_HTML, - TAG_HEAD, - TAG_BODY, - TAG_META, - TAG_TITLE, - TAG_DIV, - TAG_H1, - TAG_H2, - TAG_P, - TAG_SPAN, - TAG_LINK, - TAG_BR, - TAG_A, - TAG_MAX -}; - -enum htmlattr { - ATTR_HTTPEQUIV, - ATTR_CONTENT, - ATTR_NAME, - ATTR_REL, - ATTR_HREF, - ATTR_TYPE, - ATTR_MEDIA, - ATTR_CLASS, - ATTR_MAX -}; - struct htmldata { char *name; int flags; @@ -69,7 +42,7 @@ static const struct htmldata htmltags[TAG_MAX] = { {"head", HTML_CLRLINE}, /* TAG_HEAD */ {"body", HTML_CLRLINE}, /* TAG_BODY */ {"meta", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_META */ - {"title", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_TITLE */ + {"title", HTML_CLRLINE}, /* TAG_TITLE */ {"div", HTML_CLRLINE}, /* TAG_DIV */ {"h1", 0}, /* TAG_H1 */ {"h2", 0}, /* TAG_H2 */ @@ -78,6 +51,14 @@ static const struct htmldata htmltags[TAG_MAX] = { {"link", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */ {"br", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */ {"a", 0}, /* TAG_A */ + {"table", HTML_CLRLINE}, /* TAG_TABLE */ + {"col", HTML_CLRLINE | HTML_NOSTACK}, /* 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] = { @@ -88,399 +69,120 @@ static const char *const htmlattrs[ATTR_MAX] = { "href", "type", "media", - "class" + "class", + "style", + "width", + "valign", + "target", }; -struct htmlpair { - enum htmlattr key; - char *val; -}; +#ifdef __linux__ +extern int getsubopt(char **, char * const *, char **); +#endif -struct tag { - enum htmltag tag; - SLIST_ENTRY(tag) entry; -}; - -SLIST_HEAD(tagq, tag); - -struct html { - int flags; -#define HTML_NOSPACE (1 << 0) -#define HTML_NEWLINE (1 << 1) - struct tagq stack; - void *symtab; -}; - -#define MDOC_ARGS const struct mdoc_meta *m, \ - const struct mdoc_node *n, \ - struct html *h -#define MAN_ARGS const struct man_meta *m, \ - const struct man_node *n, \ - struct html *h -struct htmlmdoc { - int (*pre)(MDOC_ARGS); - void (*post)(MDOC_ARGS); -}; - -static void print_gen_doctype(struct html *); -static void print_gen_head(struct html *); -static void print_mdoc(MDOC_ARGS); -static void print_mdoc_head(MDOC_ARGS); -static void print_mdoc_title(MDOC_ARGS); -static void print_mdoc_node(MDOC_ARGS); -static void print_man(MAN_ARGS); -static void print_man_head(MAN_ARGS); -static void print_man_body(MAN_ARGS); -static struct tag *print_otag(struct html *, enum htmltag, - int, const struct htmlpair *); -static void print_tagq(struct html *, const struct tag *); -static void print_stagq(struct html *, const struct tag *); -static void print_ctag(struct html *, enum htmltag); -static void print_encode(struct html *, const char *); -static void print_escape(struct html *, const char **); -static void print_text(struct html *, const char *); -static void print_res(struct html *, const char *, int); -static void print_spec(struct html *, const char *, int); -static int mdoc_root_pre(MDOC_ARGS); - -static int mdoc_ar_pre(MDOC_ARGS); -static int mdoc_fl_pre(MDOC_ARGS); -static int mdoc_nd_pre(MDOC_ARGS); -static int mdoc_nm_pre(MDOC_ARGS); -static int mdoc_ns_pre(MDOC_ARGS); -static int mdoc_op_pre(MDOC_ARGS); -static void mdoc_op_post(MDOC_ARGS); -static int mdoc_pp_pre(MDOC_ARGS); -static int mdoc_sh_pre(MDOC_ARGS); -static int mdoc_ss_pre(MDOC_ARGS); -static int mdoc_xr_pre(MDOC_ARGS); - -static const struct htmlmdoc mdocs[MDOC_MAX] = { - {NULL, NULL}, /* Ap */ - {NULL, NULL}, /* Dd */ - {NULL, NULL}, /* Dt */ - {NULL, NULL}, /* Os */ - {mdoc_sh_pre, NULL }, /* Sh */ - {mdoc_ss_pre, NULL }, /* Ss */ - {mdoc_pp_pre, NULL}, /* Pp */ - {NULL, NULL}, /* D1 */ - {NULL, NULL}, /* Dl */ - {NULL, NULL}, /* Bd */ - {NULL, NULL}, /* Ed */ - {NULL, NULL}, /* Bl */ - {NULL, NULL}, /* El */ - {NULL, NULL}, /* It */ - {NULL, NULL}, /* Ad */ - {NULL, NULL}, /* An */ - {mdoc_ar_pre, NULL}, /* Ar */ - {NULL, NULL}, /* Cd */ - {NULL, NULL}, /* Cm */ - {NULL, NULL}, /* Dv */ - {NULL, NULL}, /* Er */ - {NULL, NULL}, /* Ev */ - {NULL, NULL}, /* Ex */ - {NULL, NULL}, /* Fa */ - {NULL, NULL}, /* Fd */ - {mdoc_fl_pre, NULL}, /* Fl */ - {NULL, NULL}, /* Fn */ - {NULL, NULL}, /* Ft */ - {NULL, NULL}, /* Ic */ - {NULL, NULL}, /* In */ - {NULL, NULL}, /* Li */ - {mdoc_nd_pre, NULL}, /* Nd */ - {mdoc_nm_pre, NULL}, /* Nm */ - {mdoc_op_pre, mdoc_op_post}, /* Op */ - {NULL, NULL}, /* Ot */ - {NULL, NULL}, /* Pa */ - {NULL, NULL}, /* Rv */ - {NULL, NULL}, /* St */ - {NULL, NULL}, /* Va */ - {NULL, NULL}, /* Vt */ - {mdoc_xr_pre, NULL}, /* Xr */ - {NULL, NULL}, /* %A */ - {NULL, NULL}, /* %B */ - {NULL, NULL}, /* %D */ - {NULL, NULL}, /* %I */ - {NULL, NULL}, /* %J */ - {NULL, NULL}, /* %N */ - {NULL, NULL}, /* %O */ - {NULL, NULL}, /* %P */ - {NULL, NULL}, /* %R */ - {NULL, NULL}, /* %T */ - {NULL, NULL}, /* %V */ - {NULL, NULL}, /* Ac */ - {NULL, NULL}, /* Ao */ - {NULL, NULL}, /* Aq */ - {NULL, NULL}, /* At */ - {NULL, NULL}, /* Bc */ - {NULL, NULL}, /* Bf */ - {NULL, NULL}, /* Bo */ - {NULL, NULL}, /* Bq */ - {NULL, NULL}, /* Bsx */ - {NULL, NULL}, /* Bx */ - {NULL, NULL}, /* Db */ - {NULL, NULL}, /* Dc */ - {NULL, NULL}, /* Do */ - {NULL, NULL}, /* Dq */ - {NULL, NULL}, /* Ec */ - {NULL, NULL}, /* Ef */ - {NULL, NULL}, /* Em */ - {NULL, NULL}, /* Eo */ - {NULL, NULL}, /* Fx */ - {NULL, NULL}, /* Ms */ - {NULL, NULL}, /* No */ - {mdoc_ns_pre, NULL}, /* Ns */ - {NULL, NULL}, /* Nx */ - {NULL, NULL}, /* Ox */ - {NULL, NULL}, /* Pc */ - {NULL, NULL}, /* Pf */ - {NULL, NULL}, /* Po */ - {NULL, NULL}, /* Pq */ - {NULL, NULL}, /* Qc */ - {NULL, NULL}, /* Ql */ - {NULL, NULL}, /* Qo */ - {NULL, NULL}, /* Qq */ - {NULL, NULL}, /* Re */ - {NULL, NULL}, /* Rs */ - {NULL, NULL}, /* Sc */ - {NULL, NULL}, /* So */ - {NULL, NULL}, /* Sq */ - {NULL, NULL}, /* Sm */ - {NULL, NULL}, /* Sx */ - {NULL, NULL}, /* Sy */ - {NULL, NULL}, /* Tn */ - {NULL, NULL}, /* Ux */ - {NULL, NULL}, /* Xc */ - {NULL, NULL}, /* Xo */ - {NULL, NULL}, /* Fo */ - {NULL, NULL}, /* Fc */ - {NULL, NULL}, /* Oo */ - {NULL, NULL}, /* Oc */ - {NULL, NULL}, /* Bk */ - {NULL, NULL}, /* Ek */ - {NULL, NULL}, /* Bt */ - {NULL, NULL}, /* Hf */ - {NULL, NULL}, /* Fr */ - {NULL, NULL}, /* Ud */ - {NULL, NULL}, /* Lb */ - {NULL, NULL}, /* Lp */ - {NULL, NULL}, /* Lk */ - {NULL, NULL}, /* Mt */ - {NULL, NULL}, /* Brq */ - {NULL, NULL}, /* Bro */ - {NULL, NULL}, /* Brc */ - {NULL, NULL}, /* %C */ - {NULL, NULL}, /* Es */ - {NULL, NULL}, /* En */ - {NULL, NULL}, /* Dx */ - {NULL, NULL}, /* %Q */ - {NULL, NULL}, /* br */ - {NULL, NULL}, /* sp */ -}; - -void -html_mdoc(void *arg, const struct mdoc *m) -{ - struct html *h; - struct tag *t; - - h = (struct html *)arg; - - print_gen_doctype(h); - t = print_otag(h, TAG_HTML, 0, NULL); - print_mdoc(mdoc_meta(m), mdoc_node(m), h); - print_tagq(h, t); - - printf("\n"); -} - -void -html_man(void *arg, const struct man *m) -{ - struct html *h; - struct tag *t; - - h = (struct html *)arg; - - print_gen_doctype(h); - t = print_otag(h, TAG_HTML, 0, NULL); - print_man(man_meta(m), man_node(m), h); - print_tagq(h, t); - - printf("\n"); -} - void * -html_alloc(void) +html_alloc(char *outopts) { struct html *h; + char *toks[4], *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->stack); + SLIST_INIT(&h->tags); + SLIST_INIT(&h->ords); + if (NULL == (h->symtab = chars_init(CHARS_HTML))) { free(h); return(NULL); } + + while (outopts && *outopts) + switch (getsubopt(&outopts, toks, &v)) { + case (0): + h->style = v; + break; + case (1): + h->base_man = v; + break; + case (2): + h->base_includes = v; + break; + default: + break; + } + return(h); } + void html_free(void *p) { struct tag *tag; + struct ord *ord; struct html *h; h = (struct html *)p; - while ( ! SLIST_EMPTY(&h->stack)) { - tag = SLIST_FIRST(&h->stack); - SLIST_REMOVE_HEAD(&h->stack, entry); + while ( ! SLIST_EMPTY(&h->ords)) { + ord = SLIST_FIRST(&h->ords); + SLIST_REMOVE_HEAD(&h->ords, entry); + free(ord); + } + + while ( ! SLIST_EMPTY(&h->tags)) { + tag = SLIST_FIRST(&h->tags); + SLIST_REMOVE_HEAD(&h->tags, entry); free(tag); } + + if (h->buf) + free(h->buf); + if (h->symtab) + chars_free(h->symtab); + free(h); } -static void -print_mdoc(MDOC_ARGS) -{ - struct tag *t; - t = print_otag(h, TAG_HEAD, 0, NULL); - print_mdoc_head(m, n, h); - print_tagq(h, t); - - t = print_otag(h, TAG_BODY, 0, NULL); - print_mdoc_title(m, n, h); - print_mdoc_node(m, n, h); - print_tagq(h, t); -} - -static void +void print_gen_head(struct html *h) { - struct htmlpair meta0[2]; - struct htmlpair meta1[2]; - struct htmlpair link[4]; + struct htmlpair tag[4]; - meta0[0].key = ATTR_HTTPEQUIV; - meta0[0].val = "Content-Type"; - meta0[1].key = ATTR_CONTENT; - meta0[1].val = "text/html; charest-utf-8"; + 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); - meta1[0].key = ATTR_NAME; - meta1[0].val = "resource-type"; - meta1[1].key = ATTR_CONTENT; - meta1[1].val = "document"; + 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); - link[0].key = ATTR_REL; - link[0].val = "stylesheet"; - link[1].key = ATTR_HREF; - link[1].val = "style.css"; /* XXX */ - link[2].key = ATTR_TYPE; - link[2].val = "text/css"; - link[3].key = ATTR_MEDIA; - link[3].val = "all"; - - print_otag(h, TAG_META, 2, meta0); - print_otag(h, TAG_META, 2, meta1); - print_otag(h, TAG_LINK, 4, link); -} - -/* ARGSUSED */ -static void -print_mdoc_head(MDOC_ARGS) -{ - - print_gen_head(h); - print_otag(h, TAG_TITLE, 0, NULL); - print_encode(h, m->title); -} - -/* ARGSUSED */ -static void -print_mdoc_title(MDOC_ARGS) -{ - - /* TODO */ -} - -static void -print_mdoc_node(MDOC_ARGS) -{ - int child; - struct tag *t; - - child = 1; - t = SLIST_FIRST(&h->stack); - - switch (n->type) { - case (MDOC_ROOT): - child = mdoc_root_pre(m, n, h); - break; - case (MDOC_TEXT): - print_text(h, n->string); - break; - default: - if (mdocs[n->tok].pre) - child = (*mdocs[n->tok].pre)(m, n, h); - break; + if (h->style) { + tag[0].key = ATTR_REL; + tag[0].val = "stylesheet"; + tag[1].key = ATTR_HREF; + tag[1].val = h->style; + tag[2].key = ATTR_TYPE; + tag[2].val = "text/css"; + tag[3].key = ATTR_MEDIA; + tag[3].val = "all"; + print_otag(h, TAG_LINK, 4, tag); } - - if (child && n->child) - print_mdoc_node(m, n->child, h); - - print_stagq(h, t); - - switch (n->type) { - case (MDOC_ROOT): - break; - case (MDOC_TEXT): - break; - default: - if (mdocs[n->tok].post) - (*mdocs[n->tok].post)(m, n, h); - break; - } - - if (n->next) - print_mdoc_node(m, n->next, h); } -static void -print_man(MAN_ARGS) -{ - struct tag *t; - t = print_otag(h, TAG_HEAD, 0, NULL); - print_man_head(m, n, h); - print_tagq(h, t); - - t = print_otag(h, TAG_BODY, 0, NULL); - print_man_body(m, n, h); - print_tagq(h, t); -} - -/* ARGSUSED */ static void -print_man_head(MAN_ARGS) -{ - - print_gen_head(h); - print_otag(h, TAG_TITLE, 0, NULL); - print_encode(h, m->title); -} - -/* ARGSUSED */ -static void -print_man_body(MAN_ARGS) -{ - - /* TODO */ -} - -static void print_spec(struct html *h, const char *p, int len) { const char *rhs; @@ -495,6 +197,7 @@ print_spec(struct html *h, const char *p, int len) putchar(rhs[i]); } + static void print_res(struct html *h, const char *p, int len) { @@ -510,6 +213,7 @@ print_res(struct html *h, const char *p, int len) putchar(rhs[i]); } + static void print_escape(struct html *h, const char **p) { @@ -615,16 +319,29 @@ print_encode(struct html *h, const char *p) { for (; *p; p++) { - if ('\\' != *p) { - putchar(*p); + if ('\\' == *p) { + print_escape(h, &p); continue; } - print_escape(h, &p); + switch (*p) { + case ('<'): + printf("<"); + break; + case ('>'): + printf(">"); + break; + case ('&'): + printf("&"); + break; + default: + putchar(*p); + break; + } } } -static struct tag * +struct tag * print_otag(struct html *h, enum htmltag tag, int sz, const struct htmlpair *p) { @@ -635,7 +352,7 @@ print_otag(struct html *h, enum htmltag tag, if (NULL == (t = malloc(sizeof(struct tag)))) err(EXIT_FAILURE, "malloc"); t->tag = tag; - SLIST_INSERT_HEAD(&h->stack, t, entry); + SLIST_INSERT_HEAD(&h->tags, t, entry); } else t = NULL; @@ -678,15 +395,15 @@ print_ctag(struct html *h, enum htmltag tag) /* ARGSUSED */ -static void +void print_gen_doctype(struct html *h) { - printf("\n", DOCTYPE, DTD); + printf("", DOCTYPE, DTD); } -static void +void print_text(struct html *h, const char *p) { @@ -709,7 +426,8 @@ print_text(struct html *h, const char *p) case(']'): /* FALLTHROUGH */ case('}'): - h->flags |= HTML_NOSPACE; + if ( ! (HTML_IGNDELIM & h->flags)) + h->flags |= HTML_NOSPACE; break; default: break; @@ -739,15 +457,15 @@ print_text(struct html *h, const char *p) } -static void +void print_tagq(struct html *h, const struct tag *until) { struct tag *tag; - while ( ! SLIST_EMPTY(&h->stack)) { - tag = SLIST_FIRST(&h->stack); + while ( ! SLIST_EMPTY(&h->tags)) { + tag = SLIST_FIRST(&h->tags); print_ctag(h, tag->tag); - SLIST_REMOVE_HEAD(&h->stack, entry); + SLIST_REMOVE_HEAD(&h->tags, entry); free(tag); if (until && tag == until) return; @@ -755,191 +473,17 @@ print_tagq(struct html *h, const struct tag *until) } -static void +void print_stagq(struct html *h, const struct tag *suntil) { struct tag *tag; - while ( ! SLIST_EMPTY(&h->stack)) { - tag = SLIST_FIRST(&h->stack); + while ( ! SLIST_EMPTY(&h->tags)) { + tag = SLIST_FIRST(&h->tags); if (suntil && tag == suntil) return; print_ctag(h, tag->tag); - SLIST_REMOVE_HEAD(&h->stack, entry); + SLIST_REMOVE_HEAD(&h->tags, entry); free(tag); } -} - - -/* ARGSUSED */ -static int -mdoc_root_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - tag.key = ATTR_CLASS; - tag.val = "body"; - - print_otag(h, TAG_DIV, 1, &tag); - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_ss_pre(MDOC_ARGS) -{ - - if (MDOC_BODY == n->type) - print_otag(h, TAG_P, 0, NULL); - if (MDOC_HEAD == n->type) - print_otag(h, TAG_H2, 0, NULL); - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_fl_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - tag.key = ATTR_CLASS; - tag.val = "flag"; - - print_otag(h, TAG_SPAN, 1, &tag); - print_text(h, "\\-"); - h->flags |= HTML_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_pp_pre(MDOC_ARGS) -{ - - print_otag(h, TAG_BR, 0, NULL); - print_otag(h, TAG_BR, 0, NULL); - return(0); -} - - -/* ARGSUSED */ -static int -mdoc_nd_pre(MDOC_ARGS) -{ - - if (MDOC_BODY == n->type) - print_text(h, "\\(en"); - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_op_pre(MDOC_ARGS) -{ - - if (MDOC_BODY == n->type) { - print_text(h, "\\(lB"); - h->flags |= HTML_NOSPACE; - } - return(1); -} - - -/* ARGSUSED */ -static void -mdoc_op_post(MDOC_ARGS) -{ - - if (MDOC_BODY != n->type) - return; - h->flags |= HTML_NOSPACE; - print_text(h, "\\(rB"); -} - - -static int -mdoc_nm_pre(MDOC_ARGS) -{ - struct htmlpair class; - - if ( ! (HTML_NEWLINE & h->flags)) - if (SEC_SYNOPSIS == n->sec) - print_otag(h, TAG_BR, 0, NULL); - - class.key = ATTR_CLASS; - class.val = "name"; - - print_otag(h, TAG_SPAN, 1, &class); - if (NULL == n->child) - print_text(h, m->name); - - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_sh_pre(MDOC_ARGS) -{ - - if (MDOC_BODY == n->type) - print_otag(h, TAG_P, 0, NULL); - if (MDOC_HEAD == n->type) - print_otag(h, TAG_H1, 0, NULL); - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_xr_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - tag.key = ATTR_HREF; - tag.val = "#"; /* TODO */ - - print_otag(h, TAG_A, 1, &tag); - - n = n->child; - print_text(h, n->string); - if (NULL == (n = n->next)) - return(0); - - h->flags |= HTML_NOSPACE; - print_text(h, "("); - h->flags |= HTML_NOSPACE; - print_text(h, n->string); - h->flags |= HTML_NOSPACE; - print_text(h, ")"); - - return(0); -} - - -/* ARGSUSED */ -static int -mdoc_ns_pre(MDOC_ARGS) -{ - - h->flags |= HTML_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_ar_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - tag.key = ATTR_CLASS; - tag.val = "arg"; - - print_otag(h, TAG_SPAN, 1, &tag); - return(1); }