=================================================================== RCS file: /cvs/mandoc/html.c,v retrieving revision 1.132 retrieving revision 1.141 diff -u -p -r1.132 -r1.141 --- mandoc/html.c 2011/04/09 15:29:40 1.132 +++ mandoc/html.c 2011/05/17 11:19:45 1.141 @@ -1,4 +1,4 @@ -/* $Id: html.c,v 1.132 2011/04/09 15:29:40 kristaps Exp $ */ +/* $Id: html.c,v 1.141 2011/05/17 11:19:45 kristaps Exp $ */ /* * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons * Copyright (c) 2011 Ingo Schwarze @@ -31,6 +31,7 @@ #include #include "mandoc.h" +#include "libmandoc.h" #include "out.h" #include "html.h" #include "main.h" @@ -93,18 +94,27 @@ static const char *const htmlattrs[ATTR_MAX] = { "colspan", /* ATTR_COLSPAN */ }; -static void print_num(struct html *, const char *, size_t); -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 mandoc_esc); -static void print_attr(struct html *, - const char *, const char *); -static void *ml_alloc(char *, enum htmltype); +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 bufncat(struct html *, const char *, size_t); +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 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 *); +static void *ml_alloc(char *, enum htmltype); static void * ml_alloc(char *outopts, enum htmltype type) @@ -122,7 +132,7 @@ ml_alloc(char *outopts, enum htmltype type) h->type = type; h->tags.head = NULL; - h->symtab = chars_init(CHARS_HTML); + h->symtab = mchars_alloc(); while (outopts && *outopts) switch (getsubopt(&outopts, UNCONST(toks), &v)) { @@ -172,7 +182,7 @@ html_free(void *p) } if (h->symtab) - chars_free(h->symtab); + mchars_free(h->symtab); free(h); } @@ -208,25 +218,14 @@ print_gen_head(struct html *h) } } -/* ARGSUSED */ static void -print_num(struct html *h, const char *p, size_t len) -{ - const char *rhs; - - rhs = chars_num2char(p, len); - if (rhs) - putchar((int)*rhs); -} - -static void print_spec(struct html *h, const char *p, size_t len) { int cp; const char *rhs; size_t sz; - if ((cp = chars_spec2cp(h->symtab, p, len)) > 0) { + if ((cp = mchars_spec2cp(h->symtab, p, len)) > 0) { printf("&#%d;", cp); return; } else if (-1 == cp && 1 == len) { @@ -235,7 +234,7 @@ print_spec(struct html *h, const char *p, size_t len) } else if (-1 == cp) return; - if (NULL != (rhs = chars_spec2str(h->symtab, p, len, &sz))) + if (NULL != (rhs = mchars_spec2str(h->symtab, p, len, &sz))) fwrite(rhs, 1, sz, stdout); } @@ -247,13 +246,13 @@ print_res(struct html *h, const char *p, size_t len) const char *rhs; size_t sz; - if ((cp = chars_res2cp(h->symtab, p, len)) > 0) { + if ((cp = mchars_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))) + if (NULL != (rhs = mchars_res2str(h->symtab, p, len, &sz))) fwrite(rhs, 1, sz, stdout); } @@ -295,12 +294,48 @@ print_metaf(struct html *h, enum mandoc_esc deco) print_otag(h, TAG_I, 0, NULL); } +int +html_strlen(const char *cp) +{ + int ssz, sz; + const char *seq, *p; + /* + * 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. + */ + + sz = 0; + while (NULL != (p = strchr(cp, '\\'))) { + sz += (int)(p - cp); + ++cp; + switch (mandoc_escape(&cp, &seq, &ssz)) { + case (ESCAPE_ERROR): + return(sz); + case (ESCAPE_NUMBERED): + /* FALLTHROUGH */ + case (ESCAPE_PREDEF): + /* FALLTHROUGH */ + case (ESCAPE_SPECIAL): + sz++; + break; + default: + break; + } + } + + assert(sz >= 0); + return(sz + strlen(cp)); +} + 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; enum mandoc_esc esc; static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' }; @@ -339,7 +374,9 @@ print_encode(struct html *h, const char *p, int norecu switch (esc) { case (ESCAPE_NUMBERED): - print_num(h, seq, len); + c = mchars_num2char(seq, len); + if ('\0' != c) + putchar(c); break; case (ESCAPE_PREDEF): print_res(h, seq, len); @@ -427,7 +464,7 @@ print_otag(struct html *h, enum htmltag tag, print_attr(h, "lang", "en"); } - /* Accomodate for XML "well-formed" singleton escaping. */ + /* Accommodate for XML "well-formed" singleton escaping. */ if (HTML_AUTOCLOSE & htmltags[tag].flags) switch (h->type) { @@ -460,28 +497,9 @@ print_ctag(struct html *h, enum htmltag tag) } } - 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) - puts(""); -} - - -static void -print_doctype(struct html *h) -{ const char *doctype; const char *dtd; const char *name; @@ -493,6 +511,7 @@ print_doctype(struct html *h) dtd = "http://www.w3.org/TR/html4/strict.dtd"; break; default: + puts(""); name = "html"; doctype = "-//W3C//DTD XHTML 1.0 Strict//EN"; dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"; @@ -582,7 +601,6 @@ print_stagq(struct html *h, const struct tag *suntil) } } - void bufinit(struct html *h) { @@ -591,28 +609,27 @@ bufinit(struct html *h) h->buflen = 0; } - void bufcat_style(struct html *h, const char *key, const char *val) { bufcat(h, key); - bufncat(h, ":", 1); + bufcat(h, ":"); bufcat(h, val); - bufncat(h, ";", 1); + bufcat(h, ";"); } - void bufcat(struct html *h, const char *p) { - bufncat(h, p, strlen(p)); + h->buflen = strlcat(h->buf, p, BUFSIZ); + assert(h->buflen < BUFSIZ); + h->buflen--; } - void -buffmt(struct html *h, const char *fmt, ...) +bufcat_fmt(struct html *h, const char *fmt, ...) { va_list ap; @@ -623,19 +640,15 @@ buffmt(struct html *h, const char *fmt, ...) h->buflen = strlen(h->buf); } - -void +static 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); + assert(h->buflen + sz + 1 < BUFSIZ); + strncat(h->buf, p, sz); h->buflen += sz; } - void buffmt_includes(struct html *h, const char *name) { @@ -659,7 +672,6 @@ buffmt_includes(struct html *h, const char *name) bufcat(h, pp); } - void buffmt_man(struct html *h, const char *name, const char *sec) @@ -676,7 +688,7 @@ buffmt_man(struct html *h, bufcat(h, sec ? sec : "1"); break; case('N'): - buffmt(h, name); + bufcat_fmt(h, name); break; default: bufncat(h, p, 2); @@ -688,57 +700,17 @@ buffmt_man(struct html *h, bufcat(h, pp); } - void bufcat_su(struct html *h, const char *p, const struct roffsu *su) { double v; - const char *u; v = su->scale; + if (SCALE_MM == su->unit && 0.0 == (v /= 100.0)) + v = 1.0; - 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 html_idcat(char *dst, const char *src, int sz)