=================================================================== RCS file: /cvs/mandoc/html.c,v retrieving revision 1.172 retrieving revision 1.178 diff -u -p -r1.172 -r1.178 --- mandoc/html.c 2014/09/27 11:17:19 1.172 +++ mandoc/html.c 2014/10/27 13:31:04 1.178 @@ -1,6 +1,6 @@ -/* $Id: html.c,v 1.172 2014/09/27 11:17:19 kristaps Exp $ */ +/* $Id: html.c,v 1.178 2014/10/27 13:31:04 schwarze Exp $ */ /* - * Copyright (c) 2008, 2009, 2010, 2011, 2014 Kristaps Dzonsons + * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons * Copyright (c) 2011, 2012, 2013, 2014 Ingo Schwarze * * Permission to use, copy, modify, and distribute this software for any @@ -74,6 +74,22 @@ static const struct htmldata htmltags[TAG_MAX] = { {"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 htmlattrs[ATTR_MAX] = { @@ -87,6 +103,9 @@ static const char *const htmlattrs[ATTR_MAX] = { "id", /* ATTR_ID */ "colspan", /* ATTR_COLSPAN */ "charset", /* ATTR_CHARSET */ + "open", /* ATTR_OPEN */ + "close", /* ATTR_CLOSE */ + "mathvariant", /* ATTR_MATHVARIANT */ }; static const char *const roffscales[SCALE_MAX] = { @@ -418,8 +437,18 @@ print_encode(struct html *h, const char *p, int norecu case ESCAPE_UNICODE: /* Skip past "u" header. */ c = mchars_num2uc(seq + 1, len - 1); - if ('\0' != c) - printf("&#x%x;", c); + + /* + * XXX Security warning: + * For now, forbid Unicode obfuscation of ASCII + * characters. An audit of the callers is + * required before this can be removed. + */ + + if (c < 0x80) + c = 0xFFFD; + + printf("&#x%x;", c); break; case ESCAPE_NUMBERED: c = mchars_num2char(seq, len); @@ -428,11 +457,12 @@ print_encode(struct html *h, const char *p, int norecu break; case ESCAPE_SPECIAL: c = mchars_spec2cp(h->symtab, seq, len); - if (c > 0) + if (c <= 0) + break; + if (c < 0x20 || c > 0x7e) printf("&#%d;", c); - else if (-1 == c && 1 == len && - !print_escape(*seq)) - putchar((int)*seq); + else if ( ! print_escape(c)) + putchar(c); break; case ESCAPE_NOSPACE: if ('\0' == *p)