=================================================================== RCS file: /cvs/texi2mdoc/main.c,v retrieving revision 1.30 retrieving revision 1.66 diff -u -p -r1.30 -r1.66 --- texi2mdoc/main.c 2015/02/23 11:44:30 1.30 +++ texi2mdoc/main.c 2015/03/11 12:51:41 1.66 @@ -1,4 +1,4 @@ -/* $Id: main.c,v 1.30 2015/02/23 11:44:30 kristaps Exp $ */ +/* $Id: main.c,v 1.66 2015/03/11 12:51:41 kristaps Exp $ */ /* * Copyright (c) 2015 Kristaps Dzonsons * @@ -14,12 +14,11 @@ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ -#include -#include - +#if defined(__linux__) || defined(__MINT__) +# define _GNU_SOURCE /* memmem */ +#endif #include #include -#include #include #include #include @@ -27,87 +26,108 @@ #include #include #include -#include #include #include "extern.h" +#define HAVE_INDEX 1 + +/* + * Texinfo can change the "meaning" of its section headings: chapter, + * section, subsection, etc., can be promoted and/or demoted to other + * levels of heading. + * Thus, we use an offset and just jump into this array. + */ #define SECTSZ 4 static const char *const sects[SECTSZ] = { - "Sh", - "Ss", - "Em", - "No", + "Sh", /* Chapters (sections) */ + "Ss", /* Sections (subsections) */ + "Em", /* Subsections (subsubsection) */ + "Sy", /* Subsubsections (...). */ }; -static void doaccent(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doblock(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dobracket(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dobye(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dodefindex(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dodefn(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dodisplay(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doenumerate(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doexample(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doignargn(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doignblock(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doignbracket(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doignline(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doinline(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doinclude(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doitem(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doitemize(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dolink(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void domacro(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void domath(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void domultitable(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doquotation(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dotable(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dotop(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dosecoffs(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dosection(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dosp(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dosubsection(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dosubsubsection(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dosymbol(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dotab(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dotitle(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void dovalue(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doverb(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doverbinclude(struct texi *, enum texicmd, const char *, size_t, size_t *); +static void doaccent(struct texi *, enum texicmd, size_t *); +static void doblock(struct texi *, enum texicmd, size_t *); +static void dobracket(struct texi *, enum texicmd, size_t *); +static void dobye(struct texi *, enum texicmd, size_t *); +static void docopying(struct texi *, enum texicmd, size_t *); +static void dodefindex(struct texi *, enum texicmd, size_t *); +static void dodefn(struct texi *, enum texicmd, size_t *); +static void dodisplay(struct texi *, enum texicmd, size_t *); +static void doend(struct texi *, enum texicmd, size_t *); +static void doenumerate(struct texi *, enum texicmd, size_t *); +static void doexample(struct texi *, enum texicmd, size_t *); +static void doignargn(struct texi *, enum texicmd, size_t *); +static void doignblock(struct texi *, enum texicmd, size_t *); +static void doignbracket(struct texi *, enum texicmd, size_t *); +static void doignline(struct texi *, enum texicmd, size_t *); +static void doindex(struct texi *, enum texicmd, size_t *); +static void doinline(struct texi *, enum texicmd, size_t *); +static void doinclude(struct texi *, enum texicmd, size_t *); +static void doinsertcopying(struct texi *, enum texicmd, size_t *); +static void doitem(struct texi *, enum texicmd, size_t *); +static void doitemize(struct texi *, enum texicmd, size_t *); +static void dolink(struct texi *, enum texicmd, size_t *); +static void domacro(struct texi *, enum texicmd, size_t *); +static void domath(struct texi *, enum texicmd, size_t *); +static void domenu(struct texi *, enum texicmd, size_t *); +static void domultitable(struct texi *, enum texicmd, size_t *); +static void donode(struct texi *, enum texicmd, size_t *); +static void doprintindex(struct texi *, enum texicmd, size_t *); +static void doquotation(struct texi *, enum texicmd, size_t *); +static void dotable(struct texi *, enum texicmd, size_t *); +static void dotop(struct texi *, enum texicmd, size_t *); +static void dosecoffs(struct texi *, enum texicmd, size_t *); +static void dosection(struct texi *, enum texicmd, size_t *); +static void dosp(struct texi *, enum texicmd, size_t *); +static void dosymbol(struct texi *, enum texicmd, size_t *); +static void dotab(struct texi *, enum texicmd, size_t *); +static void dotitle(struct texi *, enum texicmd, size_t *); +static void dovalue(struct texi *, enum texicmd, size_t *); +static void doverb(struct texi *, enum texicmd, size_t *); +static void doverbatim(struct texi *, enum texicmd, size_t *); +static void doverbinclude(struct texi *, enum texicmd, size_t *); static const struct texitok __texitoks[TEXICMD__MAX] = { /* TEXICMD__BEGIN */ + { doignline, "afourpaper", 10 }, /* TEXICMD_A4PAPER */ + { dosymbol, "AA", 2 }, /* TEXICMD_AA */ + { dosymbol, "aa", 2 }, /* TEXICMD_AASMALL */ { doignargn, "acronym", 7 }, /* TEXICMD_ACRONYM */ { doaccent, "'", 1 }, /* TEXICMD_ACUTE */ - { doignline, "afourpaper", 10 }, /* TEXICMD_A4PAPER */ + { dosymbol, "AE", 2 }, /* TEXICMD_AE */ + { dosymbol, "ae", 2 }, /* TEXICMD_AESMALL */ { doignbracket, "anchor", 6 }, /* TEXICMD_ANCHOR */ { dosection, "appendix", 8 }, /* TEXICMD_APPENDIX */ { dosection, "appendixsec", 11 }, /* TEXICMD_APPENDIXSEC */ - { dosubsection, "appendixsubsec", 14 }, /* TEXICMD_APPENDIXSUBSEC */ - { dosubsubsection, "appendixsubsubsec", 17 }, /* TEXICMD_APPENDIXSUBSUBSEC */ + { dosection, "appendixsubsec", 14 }, /* TEXICMD_APPENDIXSUBSEC */ + { dosection, "appendixsubsubsec", 17 }, /* TEXICMD_APPENDIXSUBSUBSEC */ { doinline, "asis", 4 }, /* TEXICMD_ASIS */ { dosymbol, "*", 1 }, /* TEXICMD_ASTERISK */ { dosymbol, "@", 1 }, /* TEXICMD_AT */ { doignline, "author", 6 }, /* TEXICMD_AUTHOR */ - { doinline, "b", 1 }, /* TEXICMD_BOLD */ + { doinline, "b", 1 }, /* TEXICMD_B */ + { dosymbol, "\\", 1 }, /* TEXICMD_BACKSLASH */ { dosymbol, "!", 1 }, /* TEXICMD_BANG */ { dosymbol, "bullet", 6 }, /* TEXICMD_BULLET */ { dobye, "bye", 3 }, /* TEXICMD_BYE */ + { doblock, "cartouche", 9 }, /* TEXICMD_CARTOUCHE */ + { doaccent, ",", 1 }, /* TEXICMD_CEDILLA */ { doignline, "center", 6 }, /* TEXICMD_CENTER */ { dosection, "chapter", 7 }, /* TEXICMD_CHAPTER */ - { doignline, "cindex", 6 }, /* TEXICMD_CINDEX */ + { doindex, "cindex", 6 }, /* TEXICMD_CINDEX */ { doaccent, "^", 1 }, /* TEXICMD_CIRCUMFLEX */ { doinline, "cite", 4 }, /* TEXICMD_CITE */ { dovalue, "clear", 5 }, /* TEXICMD_CLEAR */ { doinline, "code", 4 }, /* TEXICMD_CODE */ { dosymbol, ":", 1 }, /* TEXICMD_COLON */ { NULL, "columnfractions", 15 }, /* TEXICMD_COLUMNFRACTIONS */ + { dosymbol, "comma", 5 }, /* TEXICMD_COMMA */ { doinline, "command", 7 }, /* TEXICMD_COMMAND */ { doignline, "c", 1 }, /* TEXICMD_COMMENT */ { doignline, "comment", 7 }, /* TEXICMD_COMMENT_LONG */ { doignline, "contents", 8 }, /* TEXICMD_CONTENTS */ - { doignblock, "copying", 7 }, /* TEXICMD_COPYING */ + { docopying, "copying", 7 }, /* TEXICMD_COPYING */ { dosymbol, "copyright", 9 }, /* TEXICMD_COPYRIGHT */ { dodefindex, "defcodeindex", 12 }, /* TEXICMD_DEFCODEINDEX */ { dodefn, "deffn", 5 }, /* TEXICMD_DEFFN */ @@ -121,6 +141,8 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { dodefn, "deftypefnx", 10 }, /* TEXICMD_DEFTYPEFNX */ { dodefn, "deftypefun", 10 }, /* TEXICMD_DEFTYPEFUN */ { dodefn, "deftypefunx", 11 }, /* TEXICMD_DEFTYPEFUNX */ + { dodefn, "deftypemethod", 13 }, /* TEXICMD_DEFTYPEMETHOD */ + { dodefn, "deftypemethodx", 14 }, /* TEXICMD_DEFTYPEMETHODX */ { dodefn, "deftypevar", 10 }, /* TEXICMD_DEFTYPEVAR */ { dodefn, "deftypevarx", 11 }, /* TEXICMD_DEFTYPEVARX */ { dodefn, "deftypevr", 9 }, /* TEXICMD_DEFTYPEVR */ @@ -131,30 +153,53 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { dodefn, "defvarx", 7 }, /* TEXICMD_DEFVARX */ { dodefn, "defvr", 5 }, /* TEXICMD_DEFVR */ { dodefn, "defvrx", 6 }, /* TEXICMD_DEFVRX */ - { doignblock, "detailmenu", 10 }, /* TEXICMD_DETAILMENU */ + { domenu, "detailmenu", 10 }, /* TEXICMD_DETAILMENU */ { doinline, "dfn", 3 }, /* TEXICMD_DFN */ + { dosymbol, "DH", 2 }, /* TEXICMD_DH */ + { dosymbol, "dh", 2 }, /* TEXICMD_DHSMALL */ { doignline, "dircategory", 11 }, /* TEXICMD_DIRCATEGORY */ { doignblock, "direntry", 8 }, /* TEXICMD_DIRENTRY */ { dodisplay, "display", 7 }, /* TEXICMD_DISPLAY */ { doignbracket, "dmn", 3 }, /* TEXICMD_DMN */ + { doignblock, "documentdescription", 19 }, /* TEXICMD_DOCUMENTDESCRIPTION */ + { doignline, "documentencoding", 16 }, /* TEXICMD_DOCUMENTENCODING */ + { doignline, "documentlanguage", 16 }, /* TEXICMD_DOCUMENTLANGUAGE */ + { doaccent, "dotaccent", 9 }, /* TEXICMD_DOTACCENT */ + { doaccent, "dotless", 7 }, /* TEXICMD_DOTLESS */ { dosymbol, "dots", 4 }, /* TEXICMD_DOTS */ { dolink, "email", 5 }, /* TEXICMD_EMAIL */ { doinline, "emph", 4 }, /* TEXICMD_EMPH */ - { NULL, "end", 3 }, /* TEXICMD_END */ + { doend, "end", 3 }, /* TEXICMD_END */ + { dosymbol, "enddots", 7 }, /* TEXICMD_ENDDOTS */ { doenumerate, "enumerate", 9 }, /* TEXICMD_ENUMERATE */ { doinline, "env", 3 }, /* TEXICMD_ENV */ { dosymbol, "equiv", 5 }, /* TEXICMD_EQUIV */ { dosymbol, "error", 5 }, /* TEXICMD_ERROR */ + { dosymbol, "euro", 4 }, /* TEXICMD_EURO */ { doexample, "example", 7 }, /* TEXICMD_EXAMPLE */ + { dosymbol, "exclamdown", 10 }, /* TEXICMD_EXCLAMDOWN */ { doignline, "exdent", 6 }, /* TEXICMD_EXDENT */ { dosymbol, "expansion", 9 }, /* TEXICMD_EXPANSION */ { doinline, "file", 4 }, /* TEXICMD_FILE */ { doignline, "finalout", 8 }, /* TEXICMD_FINALOUT */ - { doignline, "findex", 6 }, /* TEXICMD_FINDEX */ + { doindex, "findex", 6 }, /* TEXICMD_FINDEX */ + { doblock, "flushleft", 9 }, /* TEXICMD_FLUSHLEFT */ + { doblock, "flushright", 10 }, /* TEXICMD_FLUSHRIGHT */ + { doignline, "firstparagraphindent", 20 }, /* TEXICMD_FIRSTPARAGRAPHINDENT */ + { doignbracket, "footnote", 8 }, /* TEXICMD_FOOTNOTE */ + { doignline, "footnotestyle", 13 }, /* TEXICMD_FOOTNOTESTYLE */ { dotable, "ftable", 6 }, /* TEXICMD_FTABLE */ { dodisplay, "format", 6 }, /* TEXICMD_FORMAT */ + { dosymbol, "geq", 3 }, /* TEXICMD_GEQ */ { doaccent, "`", 1 }, /* TEXICMD_GRAVE */ { doblock, "group", 5 }, /* TEXICMD_GROUP */ + { dosymbol, "guillemetleft", 13 }, /* TEXICMD_GUILLEMETLEFT */ + { dosymbol, "guillemetright", 14 }, /* TEXICMD_GUILLEMETRIGHT */ + { dosymbol, "guillemotleft", 13 }, /* TEXICMD_GUILLEMOTLEFT */ + { dosymbol, "guillemotright", 14 }, /* TEXICMD_GUILLEMOTRIGHT */ + { dosymbol, "guilsinglleft", 13 }, /* TEXICMD_GUILSINGLLEFT */ + { dosymbol, "guilsinglright", 14 }, /* TEXICMD_GUILSINGLRIGHT */ + { doaccent, "H", 1 }, /* TEXICMD_H */ { dosection, "heading", 7 }, /* TEXICMD_HEADING */ { doignline, "headings", 8 }, /* TEXICMD_HEADINGS */ { doitem, "headitem", 8 }, /* TEXICMD_HEADITEM */ @@ -179,49 +224,73 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { doignbracket, "image", 5 }, /* TEXICMD_IMAGE */ { doinclude, "include", 7 }, /* TEXICMD_INCLUDE */ { dodisplay, "indentblock", 11 }, /* TEXICMD_INDENTBLOCK */ - { doignline, "", 0 }, /* TEXICMD_USER_INDEX */ - { doignline, "insertcopying", 13 }, /* TEXICMD_INSERTCOPYING */ + { dolink, "indicateurl", 11 }, /* TEXICMD_INDICATEURL */ + { dolink, "inforef", 7 }, /* TEXICMD_INFOREF */ + { doinsertcopying, "insertcopying", 13 }, /* TEXICMD_INSERTCOPYING */ { doitem, "item", 4 }, /* TEXICMD_ITEM */ { doitemize, "itemize", 7 }, /* TEXICMD_ITEMIZE */ { doitem, "itemx", 5 }, /* TEXICMD_ITEMX */ { doinline, "kbd", 3 }, /* TEXICMD_KBD */ { dobracket, "key", 3 }, /* TEXICMD_KEY */ { doignline, "kindex", 6 }, /* TEXICMD_KINDEX */ + { dosymbol, "L", 1 }, /* TEXICMD_L */ { dosymbol, "LaTeX", 5 }, /* TEXICMD_LATEX */ + { dosymbol, "leq", 3 }, /* TEXICMD_LEQ */ { dosecoffs, "lowersections", 13 }, /* TEXICMD_LOWERSECTIONS */ + { dosymbol, "l", 1 }, /* TEXICMD_LSMALL */ { domacro, "macro", 5 }, /* TEXICMD_MACRO */ + { doaccent, "=", 1 }, /* TEXICMD_MACRON */ { domath, "math", 4 }, /* TEXICMD_MATH */ - { doignblock, "menu", 4 }, /* TEXICMD_MENU */ + { domenu, "menu", 4 }, /* TEXICMD_MENU */ { dosymbol, "minus", 5 }, /* TEXICMD_MINUS */ { domultitable, "multitable", 10 }, /* TEXICMD_MULTITABLE */ { doignline, "need", 4 }, /* TEXICMD_NEED */ { dosymbol, "\n", 1 }, /* TEXICMD_NEWLINE */ - { doignline, "node", 4 }, /* TEXICMD_NODE */ + { donode, "node", 4 }, /* TEXICMD_NODE */ { doignline, "noindent", 8 }, /* TEXICMD_NOINDENT */ + { dosymbol, "O", 1 }, /* TEXICMD_O */ + { dosymbol, "OE", 2 }, /* TEXICMD_OE */ + { dosymbol, "oe", 2 }, /* TEXICMD_OESMALL */ + { doaccent, "ogonek", 6 }, /* TEXICMD_OGONEK */ { doinline, "option", 6 }, /* TEXICMD_OPTION */ - { dolink, "pxref", 5 }, /* TEXICMD_PXREF */ - { dosymbol, "?", 1 }, /* TEXICMD_QUESTIONMARK */ - { doquotation, "quotation", 9 }, /* TEXICMD_QUOTATION */ + { dosymbol, "ordf", 4 }, /* TEXICMD_ORDF */ + { dosymbol, "ordm", 4 }, /* TEXICMD_ORDM */ + { dosymbol, "o", 1 }, /* TEXICMD_OSMALL */ { doignline, "page", 4 }, /* TEXICMD_PAGE */ { doignline, "paragraphindent", 15 }, /* TEXICMD_PARINDENT */ { dosymbol, ".", 1 }, /* TEXICMD_PERIOD */ { doignline, "pindex", 6 }, /* TEXICMD_PINDEX */ - { doignline, "printindex", 10 }, /* TEXICMD_PRINTINDEX */ + { dosymbol, "pounds", 6 }, /* TEXICMD_POUNDS */ + { doprintindex, "printindex", 10 }, /* TEXICMD_PRINTINDEX */ + { dolink, "pxref", 5 }, /* TEXICMD_PXREF */ + { dosymbol, "questiondown", 12 }, /* TEXICMD_QUESTIONDOWN */ + { dosymbol, "?", 1 }, /* TEXICMD_QUESTIONMARK */ + { doquotation, "quotation", 9 }, /* TEXICMD_QUOTATION */ + { dosymbol, "quotedblbase", 12 }, /* TEXICMD_QUOTEDBLBASE */ + { dosymbol, "quotedblleft", 12 }, /* TEXICMD_QUOTEDBLLEFT */ + { dosymbol, "quotedblright", 13 }, /* TEXICMD_QUOTEDBLRIGHT */ + { dosymbol, "quotesinglbase", 14 }, /* TEXICMD_QUOTESINGLBASE */ + { dosymbol, "quoteleft", 9 }, /* TEXICMD_QUOTELEFT */ + { dosymbol, "quoteright", 10 }, /* TEXICMD_QUOTERIGHT */ { doinline, "r", 1 }, /* TEXICMD_R */ { dosecoffs, "raisesections", 13 }, /* TEXICMD_RAISESECTIONS */ { dobracket, "ref", 3 }, /* TEXICMD_REF */ { doignline, "refill", 6 }, /* TEXICMD_REFILL */ + { dosymbol, "registeredsymbol", 16 }, /* TEXICMD_REGISTEREDSYMBOL */ { dosymbol, "result", 6 }, /* TEXICMD_RESULT */ + { doaccent, "ringaccent", 10 }, /* TEXICMD_RINGACCENT */ { doinline, "samp", 4 }, /* TEXICMD_SAMP */ { doinline, "sansserif", 9 }, /* TEXICMD_SANSSERIF */ - { dobracket, "sc", 2 }, /* TEXICMD_SC */ + { doinline, "sc", 2 }, /* TEXICMD_SC */ { dosection, "section", 7 }, /* TEXICMD_SECTION */ { dovalue, "set", 3 }, /* TEXICMD_SET */ { doignline, "setchapternewpage", 17 }, /* TEXICMD_SETCHAPNEWPAGE */ + { doignline, "setcontentsaftertitlepage", 25 }, /* TEXICMD_SETCONTENTSAFTER */ { doignline, "setfilename", 11 }, /* TEXICMD_SETFILENAME */ { dotitle, "settitle", 8 }, /* TEXICMD_SETTITLE */ { doignline, "shortcontents", 13 }, /* TEXICMD_SHORTCONTENTS */ { doinline, "slanted", 7 }, /* TEXICMD_SLANTED */ + { dosymbol, "/", 1 }, /* TEXICMD_SLASH */ { dosp, "sp", 2 }, /* TEXICMD_SP */ { dosymbol, " ", 1 }, /* TEXICMD_SPACE */ { doignline, "smallbook", 9 }, /* TEXICMD_SMALLBOOK */ @@ -231,39 +300,53 @@ static const struct texitok __texitoks[TEXICMD__MAX] = { dodisplay, "smallindentblock", 16 }, /* TEXICMD_SMALLINDENTBLOCK */ { dosymbol, "{", 1 }, /* TEXICMD_SQUIGGLE_LEFT */ { dosymbol, "}", 1 }, /* TEXICMD_SQUIGGLE_RIGHT */ + { dosymbol, "ss", 2 }, /* TEXICMD_SS */ { doinline, "strong", 6 }, /* TEXICMD_STRONG */ - { dosubsection, "subheading", 10 }, /* TEXICMD_SUBHEADING */ - { dosubsection, "subsection", 10 }, /* TEXICMD_SUBSECTION */ - { dosubsubsection, "subsubheading", 13 }, /* TEXICMD_SUBSUBHEADING */ - { dosubsubsection, "subsubsection", 13 }, /* TEXICMD_SUBSUBSECTION */ + { dosection, "subheading", 10 }, /* TEXICMD_SUBHEADING */ + { dosection, "subsection", 10 }, /* TEXICMD_SUBSECTION */ + { dosection, "subsubheading", 13 }, /* TEXICMD_SUBSUBHEADING */ + { dosection, "subsubsection", 13 }, /* TEXICMD_SUBSUBSECTION */ { doignline, "subtitle", 8 }, /* TEXICMD_SUBTITLE */ { doignline, "summarycontents", 15 }, /* TEXICMD_SUMMARYCONTENTS */ - { doignline, "synindex", 8 }, /* TEXICMD_SYNINDEX */ - { doignline, "syncodeindex", 12 }, /* TEXICMD_SYNCODEINDEX */ + { dodefindex, "synindex", 8 }, /* TEXICMD_SYNINDEX */ + { dodefindex, "syncodeindex", 12 }, /* TEXICMD_SYNCODEINDEX */ { doinline, "t", 1 }, /* TEXICMD_T */ { dotab, "tab", 3 }, /* TEXICMD_TAB */ { dosymbol, "\t", 1 }, /* TEXICMD_TABSYM */ { dotable, "table", 5 }, /* TEXICMD_TABLE */ { doignblock, "tex", 3 }, /* TEXICMD_TEX */ { dosymbol, "TeX", 3 }, /* TEXICMD_TEXSYM */ + { dosymbol, "textdegree", 10 }, /* TEXICMD_TEXTDEGREE */ + { dosymbol, "TH", 2 }, /* TEXICMD_TH */ + { dosymbol, "th", 2 }, /* TEXICMD_THSMALL */ + { dosymbol, "tie", 3 }, /* TEXICMD_TIE */ + { doaccent, "tieaccent", 9 }, /* TEXICMD_TIEACCENT */ { doaccent, "~", 1 }, /* TEXICMD_TILDE */ - { doignline, "tindex", 6 }, /* TEXICMD_TINDEX */ + { doindex, "tindex", 6 }, /* TEXICMD_TINDEX */ { doignline, "title", 5 }, /* TEXICMD_TITLE */ { dobracket, "titlefont", 9 }, /* TEXICMD_TITLEFONT */ { doignblock, "titlepage", 9 }, /* TEXICMD_TITLEPAGE */ { dotop, "top", 3 }, /* TEXICMD_TOP */ + { doindex, "tpindex", 7 }, /* TEXICMD_TPINDEX */ + { doaccent, "u", 1 }, /* TEXICMD_U */ + { doaccent, "ubaraccent", 10 }, /* TEXICMD_UBARACCENT */ + { doaccent, "udotaccent", 10 }, /* TEXICMD_UDOTACCENT */ { doaccent, "\"", 1 }, /* TEXICMD_UMLAUT */ { dosection, "unnumbered", 10 }, /* TEXICMD_UNNUMBERED */ { dosection, "unnumberedsec", 13 }, /* TEXICMD_UNNUMBEREDSEC */ - { dosubsection, "unnumberedsubsec", 16 }, /* TEXICMD_UNNUMBEREDSUBSEC */ - { dosubsubsection, "unnumberedsubsubsec", 19 }, /* TEXICMD_UNNUMBEREDSUBSUBSEC */ + { dosection, "unnumberedsubsec", 16 }, /* TEXICMD_UNNUMBEREDSUBSEC */ + { dosection, "unnumberedsubsubsec", 19 }, /* TEXICMD_UNNUMBEREDSUBSUBSEC */ { dolink, "uref", 4 }, /* TEXICMD_UREF */ { dolink, "url", 3 }, /* TEXICMD_URL */ + { doignline, "", 0 }, /* TEXICMD_USER_INDEX */ + { doaccent, "v", 1 }, /* TEXICMD_V */ { dovalue, "value", 5 }, /* TEXICMD_VALUE */ { doinline, "var", 3 }, /* TEXICMD_VAR */ - { doverb, "verbatim", 8 }, /* TEXICMD_VERBATIM */ + { doverb, "verb", 4 }, /* TEXICMD_VERB */ + { doverbatim, "verbatim", 8 }, /* TEXICMD_VERBATIM */ { doverbinclude, "verbatiminclude", 15 }, /* TEXICMD_VERBATIMINCLUDE */ - { doignline, "vindex", 6 }, /* TEXICMD_VINDEX */ + { doindex, "vindex", 6 }, /* TEXICMD_VINDEX */ + { doindex, "vrindex", 7 }, /* TEXICMD_VRINDEX */ { dosp, "vskip", 5 }, /* TEXICMD_VSKIP */ { dotable, "vtable", 6 }, /* TEXICMD_VTABLE */ { dobracket, "w", 1 }, /* TEXICMD_W */ @@ -273,49 +356,48 @@ static const struct texitok __texitoks[TEXICMD__MAX] = const struct texitok *const texitoks = __texitoks; +/* + * Define new indexes either by assignment or aliasing (both of these + * accept the first argument as the new index). + */ static void -dodefindex(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dodefindex(struct texi *p, enum texicmd cmd, size_t *pos) { - size_t start, end; - char *cp; + size_t start; - while (*pos < sz && isws(buf[*pos])) - advance(p, buf, pos); - - start = end = *pos; - while (end < sz && ! ismspace(buf[end])) - end++; - - if (start == end) { - advanceeoln(p, buf, sz, pos, 1); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + start = *pos; + while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) + advance(p, pos); + if (*pos == BUFSZ(p)) { + texiwarn(p, "unexpected EOF"); return; - } else if (NULL == (cp = malloc(end - start + 1))) - texiabort(p, NULL); - - memcpy(cp, &buf[start], end - start); - cp[end - start] = '\0'; - - p->indexs = realloc(p->indexs, - sizeof(char *) * (p->indexsz + 1)); - - if (NULL == p->indexs) - texiabort(p, NULL); - p->indexs[p->indexsz++] = cp; + } + if (0 == *pos - start) + texiwarn(p, "zero-length index definition"); + else + texindex_add(p, &BUF(p)[start], *pos - start); + advance(p, pos); } +/* + * Handle both possible "define function" (type, etc.) classes: where + * we'll have a body and without one (suffixed with "x"). + */ static void -dodefn(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dodefn(struct texi *p, enum texicmd cmd, size_t *pos) { const char *blk; blk = NULL; switch (cmd) { case (TEXICMD_DEFFN): + case (TEXICMD_DEFMAC): case (TEXICMD_DEFTP): case (TEXICMD_DEFTYPEFN): case (TEXICMD_DEFTYPEFUN): + case (TEXICMD_DEFTYPEMETHOD): case (TEXICMD_DEFTYPEVAR): case (TEXICMD_DEFTYPEVR): case (TEXICMD_DEFUN): @@ -329,15 +411,21 @@ dodefn(struct texi *p, enum texicmd cmd, if (p->ign) { NULL != blk ? - parseto(p, buf, sz, pos, blk) : - parseeoln(p, buf, sz, pos); + parseto(p, pos, blk) : + parseeoln(p, pos); return; } - if (NULL != blk) - texivspace(p); + if (p->seenvs >= 0) { + teximacro(p, "Pp"); + p->seenvs = -1; + } switch (cmd) { + case (TEXICMD_DEFTYPEMETHOD): + case (TEXICMD_DEFTYPEMETHODX): + texiputchars(p, "Method"); + break; case (TEXICMD_DEFMAC): case (TEXICMD_DEFMACX): texiputchars(p, "Macro"); @@ -355,17 +443,19 @@ dodefn(struct texi *p, enum texicmd cmd, texiputchars(p, "Function"); break; default: - parselinearg(p, buf, sz, pos); + parselinearg(p, pos); break; } - texiputchars(p, ":\n"); + p->seenvs = 0; + texiputchar(p, ':'); + texiputchar(p, '\n'); switch (cmd) { case (TEXICMD_DEFMAC): case (TEXICMD_DEFMACX): teximacroopen(p, "Dv"); - while (parselinearg(p, buf, sz, pos)) + while (parselinearg(p, pos)) /* Spin. */ ; teximacroclose(p); break; @@ -374,10 +464,10 @@ dodefn(struct texi *p, enum texicmd cmd, case (TEXICMD_DEFUN): case (TEXICMD_DEFUNX): teximacroopen(p, "Fo"); - parselinearg(p, buf, sz, pos); + parselinearg(p, pos); teximacroclose(p); teximacroopen(p, "Fa"); - while (parselinearg(p, buf, sz, pos)) + while (parselinearg(p, pos)) /* Spin. */ ; teximacroclose(p); teximacro(p, "Fc"); @@ -386,14 +476,16 @@ dodefn(struct texi *p, enum texicmd cmd, case (TEXICMD_DEFTYPEFUNX): case (TEXICMD_DEFTYPEFN): case (TEXICMD_DEFTYPEFNX): + case (TEXICMD_DEFTYPEMETHOD): + case (TEXICMD_DEFTYPEMETHODX): teximacroopen(p, "Ft"); - parselinearg(p, buf, sz, pos); + parselinearg(p, pos); teximacroclose(p); teximacroopen(p, "Fo"); - parselinearg(p, buf, sz, pos); + parselinearg(p, pos); teximacroclose(p); teximacroopen(p, "Fa"); - while (parselinearg(p, buf, sz, pos)) + while (parselinearg(p, pos)) /* Spin. */ ; teximacroclose(p); teximacro(p, "Fc"); @@ -405,7 +497,7 @@ dodefn(struct texi *p, enum texicmd cmd, case (TEXICMD_DEFTYPEVR): case (TEXICMD_DEFTYPEVRX): teximacroopen(p, "Vt"); - while (parselinearg(p, buf, sz, pos)) + while (parselinearg(p, pos)) /* Spin. */ ; teximacroclose(p); break; @@ -414,7 +506,7 @@ dodefn(struct texi *p, enum texicmd cmd, case (TEXICMD_DEFVR): case (TEXICMD_DEFVRX): teximacroopen(p, "Va"); - while (parselinearg(p, buf, sz, pos)) + while (parselinearg(p, pos)) /* Spin. */ ; teximacroclose(p); break; @@ -422,14 +514,58 @@ dodefn(struct texi *p, enum texicmd cmd, abort(); } + if (NULL == blk) + return; + + /* + * All "block" definitions have their block bodies indented + * unless they have the "x" form of the command following. + * E.g., + * @deffn some function + * @deffnx another + * An explanation. + * @end deffn + * With this loop, we delay opening the indented block until we + * skipped past conformant macros. + */ + for (;;) { + switch (peekcmd(p, *pos)) { + case (TEXICMD_DEFFNX): + case (TEXICMD_DEFMACX): + case (TEXICMD_DEFTPX): + case (TEXICMD_DEFTYPEFNX): + case (TEXICMD_DEFTYPEFUNX): + case (TEXICMD_DEFTYPEMETHODX): + case (TEXICMD_DEFTYPEVARX): + case (TEXICMD_DEFTYPEVRX): + case (TEXICMD_DEFUNX): + case (TEXICMD_DEFVARX): + case (TEXICMD_DEFVRX): + texivspace(p); + parseeoln(p, pos); + continue; + default: + break; + } + break; + } + + if (TEXICMD_END == peekcmd(p, *pos)) { + parseto(p, pos, blk); + return; + } + texivspace(p); - if (NULL != blk) - parseto(p, buf, sz, pos, blk); + teximacro(p, "Bd -filled -offset indent -compact"); + p->seenvs = -1; + parseto(p, pos, blk); + p->seenvs = 0; + teximacro(p, "Ed"); + texivspace(p); } static void -domacro(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +domacro(struct texi *p, enum texicmd cmd, size_t *pos) { size_t start, end, endtoksz, len; struct teximacro m; @@ -437,42 +573,58 @@ domacro(struct texi *p, enum texicmd cmd, memset(&m, 0, sizeof(struct teximacro)); - while (*pos < sz && isws(buf[*pos])) - advance(p, buf, pos); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); - for (start = end = *pos; end < sz; end++) - if (ismspace(buf[end]) || '{' == buf[end]) + for (start = end = *pos; end < BUFSZ(p); end++) + if (ismspace(BUF(p)[end]) || '{' == BUF(p)[end]) break; if (start == end) texierr(p, "zero-length macro name"); - advanceto(p, buf, pos, end); + advanceto(p, pos, end); m.key = malloc(end - start + 1); if (NULL == m.key) texiabort(p, NULL); - memcpy(m.key, &buf[start], end - start); + memcpy(m.key, &BUF(p)[start], end - start); m.key[end - start] = '\0'; - m.args = argparse(p, buf, sz, pos, &m.argsz); - advanceeoln(p, buf, sz, pos, 0); + m.args = argparse(p, pos, &m.argsz, 0); + /* Note: we advance to the beginning of the macro. */ + advanceeoln(p, pos, 1); + + /* + * According to the Texinfo manual, the macro ends on the + * newline subsequent the @end macro. + * That's COMPLETELY FUCKING WRONG. + * It ends inclusive the newline, which is why so many macros + * say things like @r{hello}@c, where the subsequent @c swallows + * the newline. + * However, it does swallow the leading newline, so look for the + * @end macro without the leading newline else we might look + * past empty macros. + */ start = *pos; - endtok = "\n@end macro\n"; + endtok = "@end macro\n"; endtoksz = strlen(endtok); - blk = memmem(&buf[start], sz, endtok, endtoksz); + blk = memmem(&BUF(p)[start], BUFSZ(p) - start, endtok, endtoksz); if (NULL == blk) texierr(p, "unterminated macro body"); - while (&buf[*pos] != blk) - advance(p, buf, pos); - assert('\n' == buf[*pos]); - advance(p, buf, pos); - len = blk - &buf[start]; + /* Roll us back one character. */ + while (&BUF(p)[*pos] != blk) + advance(p, pos); + assert('@' == BUF(p)[*pos]); + if ('\n' != BUF(p)[*pos - 1]) + texierr(p, "cannot handle @end macro in-line"); + + len = blk - &BUF(p)[start]; m.value = malloc(len + 1); if (NULL == m.value) texiabort(p, NULL); - memcpy(m.value, &buf[start], len); + memcpy(m.value, &BUF(p)[start], len); m.value[len] = '\0'; p->macros = realloc @@ -483,60 +635,88 @@ domacro(struct texi *p, enum texicmd cmd, texiabort(p, NULL); p->macros[p->macrosz++] = m; - advanceeoln(p, buf, sz, pos, 1); + advanceeoln(p, pos, 1); } static void -doignblock(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doignblock(struct texi *p, enum texicmd cmd, size_t *pos) { - char end[32]; - const char *term; - size_t endsz, endpos; + char end[32], start[32]; + const char *endt, *startt; + size_t esz, ssz, newpos, stack; /* - * We want to completely ignore everything in these blocks, so - * simply jump to the @end block. + * FIXME: this is cheating. + * These tokens are supposed to begin on a newline. + * However, if we do that, then we would need to check within + * the loop for trailer (or leading, as the case may be) + * newline, and that's just a bit too complicated right now. + * This is becasue + * @ifset BAR + * @ifset FOO + * @end ifset + * @end ifset + * won't work right now: we'd read after the first "@end ifset" + * to the next line, then look for the next line after that. */ - endsz = snprintf(end, sizeof(end), - "\n@end %s\n", texitoks[cmd].tok); - assert(endsz < sizeof(end)); + ssz = snprintf(start, sizeof(start), + "@%s", texitoks[cmd].tok); + assert(ssz < sizeof(start)); + esz = snprintf(end, sizeof(end), + "@end %s\n", texitoks[cmd].tok); + assert(esz < sizeof(end)); + stack = 1; - /* - * Look up where our end token occurs. - * Set our end position based on the relative offset of that - * from our current position, or the EOF if we don't have a - * proper ending point. + /* + * Here we look for the end token "end" somewhere in the file in + * front of us. + * It's not that easy, of course: if we have a nested block, + * then there'll be an "end" token of the same kind between us. + * Thus, we keep track of scopes for matching "end" blocks. */ - term = memmem(&buf[*pos], sz, end, endsz); - endpos = NULL == term ? sz : - *pos + term - &buf[*pos]; - assert(endpos <= sz); - while (*pos < endpos) - advance(p, buf, pos); + while (stack > 0 && *pos < BUFSZ(p)) { + if (stack > 64) + texierr(p, "run-away nested stack?"); + endt = memmem(&BUF(p)[*pos], BUFSZ(p) - *pos, end, esz); + startt = memmem(&BUF(p)[*pos], BUFSZ(p) - *pos, start, ssz); + if (NULL == endt) { + texiwarn(p, "unterminated \"%s\" " + "block", texitoks[cmd].tok); + *pos = BUFSZ(p); + break; + } - /* Only do this if we're not already at the end. */ - if (endpos < sz) - advanceto(p, buf, pos, endpos + endsz); + newpos = *pos; + if (NULL == startt || startt > endt) { + newpos += esz + (size_t)(endt - &BUF(p)[*pos]); + stack--; + } else { + newpos += ssz + (size_t)(startt - &BUF(p)[*pos]); + stack++; + } + + assert(newpos <= BUFSZ(p)); + while (*pos < newpos) + advance(p, pos); + } } static void -doblock(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doblock(struct texi *p, enum texicmd cmd, size_t *pos) { - parseto(p, buf, sz, pos, texitoks[cmd].tok); + parseto(p, pos, texitoks[cmd].tok); } static void -doinline(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doinline(struct texi *p, enum texicmd cmd, size_t *pos) { const char *macro = NULL; switch (cmd) { case (TEXICMD_CODE): case (TEXICMD_KBD): + /* FIXME: quote around @samp{} */ case (TEXICMD_SAMP): case (TEXICMD_T): macro = "Li"; @@ -572,61 +752,163 @@ doinline(struct texi *p, enum texicmd cmd, } if (NULL == macro || p->literal || TEXILIST_TABLE == p->list) { - parsebracket(p, buf, sz, pos); + if (TEXICMD_SC == cmd) + p->uppercase++; + parsebracket(p, pos, 0); + if (TEXICMD_SC == cmd) + p->uppercase--; return; } + /* + * If we haven't seen any whitespace, then we don't want the + * subsequent macro to insert any whitespace. + */ + if (p->outmacro && 0 == p->seenws) { + teximacroopen(p, "Ns"); + teximacroclose(p); + } + teximacroopen(p, macro); p->seenws = 0; - parsebracket(p, buf, sz, pos); - texipunctuate(p, buf, sz, pos); + if (TEXICMD_CODE == cmd) + p->literal++; + if (TEXICMD_SC == cmd) + p->uppercase++; + parsebracket(p, pos, 0); + if (TEXICMD_SC == cmd) + p->uppercase--; + if (TEXICMD_CODE == cmd) + p->literal--; + texipunctuate(p, pos); teximacroclose(p); } static void -doverb(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doverb(struct texi *p, enum texicmd cmd, size_t *pos) { + char delim; + size_t start; + + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + if (*pos == BUFSZ(p) || '{' != BUF(p)[*pos]) + return; + advance(p, pos); + if (*pos == BUFSZ(p)) + return; + + delim = BUF(p)[*pos]; + advance(p, pos); + /* Make sure we flush out our initial whitespace... */ + if (p->seenws && p->outcol && 0 == p->literal) + texiputchar(p, ' '); + p->seenws = 0; + start = *pos; + /* Read until we see the delimiter then end-brace. */ + while (*pos < BUFSZ(p) - 1) { + if (BUF(p)[*pos] == delim && BUF(p)[*pos + 1] == '}') + break; + advance(p, pos); + } + if (*pos >= BUFSZ(p) - 1) + return; + texiputbuf(p, start, *pos); + + /* Make sure we read after the end-brace. */ + assert(delim == BUF(p)[*pos]); + advance(p, pos); + assert('}' == BUF(p)[*pos]); + advance(p, pos); +} + +static void +doinsertcopying(struct texi *p, enum texicmd cmd, size_t *pos) +{ + + advanceeoln(p, pos, 0); + if (NULL != p->copying) + texisplice(p, p->copying, p->copyingsz, *pos); +} + +static void +docopying(struct texi *p, enum texicmd cmd, size_t *pos) +{ const char *end, *term; size_t endsz, endpos; - advanceeoln(p, buf, sz, pos, 1); + /* We retain our starting (but not ending) newlines. */ + end = "\n@end copying\n"; + endsz = strlen(end); + advanceeoln(p, pos, 0); + if (*pos == BUFSZ(p)) { + texiwarn(p, "unterminated \"%s\"", texitoks[cmd].tok); + return; + } - /* We end at exactly this token. */ + term = memmem(&BUF(p)[*pos], BUFSZ(p) - *pos, end, endsz); + if (NULL == term) { + texiwarn(p, "unterminated \"%s\"", texitoks[cmd].tok); + endpos = BUFSZ(p); + } else + endpos = *pos + (size_t)(term - &BUF(p)[*pos]); + + assert(endpos <= BUFSZ(p)); + assert('\n' == BUF(p)[*pos]); + advance(p, pos); + + p->copying = malloc(endpos - *pos + 1); + p->copyingsz = endpos - *pos; + memcpy(p->copying, &BUF(p)[*pos], p->copyingsz); + p->copying[endpos - *pos] = '\0'; + + while (*pos < endpos) + advance(p, pos); + if (*pos < BUFSZ(p)) + advanceto(p, pos, endpos + endsz); +} + +static void +doverbatim(struct texi *p, enum texicmd cmd, size_t *pos) +{ + const char *end, *term; + size_t endsz, endpos; + + /* We read from the @verbatim\n newline inclusive! */ end = "\n@end verbatim\n"; endsz = strlen(end); + advanceeoln(p, pos, 0); + if (*pos == BUFSZ(p)) { + texiwarn(p, "unterminated \"%s\"", texitoks[cmd].tok); + return; + } - /* - * Look up where our end token occurs. - * Set our end position based on the relative offset of that - * from our current position. - */ - term = memmem(&buf[*pos], sz, end, endsz); - endpos = NULL == term ? sz : - *pos + term - &buf[*pos]; + term = memmem(&BUF(p)[*pos], BUFSZ(p) - *pos, end, endsz); + if (NULL == term) { + texiwarn(p, "unterminated \"%s\"", texitoks[cmd].tok); + endpos = BUFSZ(p); + } else + endpos = *pos + (size_t)(term - &BUF(p)[*pos]); - teximacro(p, "Bd -literal -offset indent"); - assert(endpos <= sz); + assert(endpos <= BUFSZ(p)); + assert('\n' == BUF(p)[*pos]); + advance(p, pos); + texivspace(p); + teximacro(p, "Bd -literal -offset indent -compact"); + p->seenvs = -1; while (*pos < endpos) { - if (buf[*pos] == '\n') - p->outcol = 0; - else - p->outcol++; - if (*pos > 0 && '.' == buf[*pos]) - if ('\n' == buf[*pos - 1]) - fputs("\\&", stdout); - putchar(buf[*pos]); - if ('\\' == buf[*pos]) - putchar('e'); - advance(p, buf, pos); + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); } + p->seenvs = 0; teximacro(p, "Ed"); - advanceto(p, buf, pos, endpos + endsz); + texivspace(p); + if (*pos < BUFSZ(p)) + advanceto(p, pos, endpos + endsz); } static void -doverbinclude(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doverbinclude(struct texi *p, enum texicmd cmd, size_t *pos) { char fname[PATH_MAX], path[PATH_MAX]; int rc; @@ -634,22 +916,22 @@ doverbinclude(struct texi *p, enum texicmd cmd, const char *v; enum texicmd type; - while (*pos < sz && ' ' == buf[*pos]) - advance(p, buf, pos); + while (*pos < BUFSZ(p) && ' ' == BUF(p)[*pos]) + advance(p, pos); - for (i = 0; *pos < sz && '\n' != buf[*pos]; ) { + for (i = 0; *pos < BUFSZ(p) && '\n' != BUF(p)[*pos]; ) { if (i == sizeof(fname) - 1) break; - if ('@' != buf[*pos]) { - fname[i++] = buf[*pos]; - advance(p, buf, pos); + if ('@' != BUF(p)[*pos]) { + fname[i++] = BUF(p)[*pos]; + advance(p, pos); continue; } - type = texicmd(p, buf, *pos, sz, &end, NULL); - advanceto(p, buf, pos, end); + type = texicmd(p, *pos, &end, NULL); + advanceto(p, pos, end); if (TEXICMD_VALUE != type) texierr(p, "unknown verbatiminclude command"); - v = valueblookup(p, buf, sz, pos); + v = valueblookup(p, pos); if (NULL == v) continue; while ('\0' != *v) { @@ -663,7 +945,7 @@ doverbinclude(struct texi *p, enum texicmd cmd, if (i == 0) texierr(p, "path too short"); - else if ('\n' != buf[*pos]) + else if ('\n' != BUF(p)[*pos]) texierr(p, "path too long"); else if ('/' == fname[0]) texierr(p, "no absolute paths"); @@ -683,8 +965,7 @@ doverbinclude(struct texi *p, enum texicmd cmd, } static void -doinclude(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doinclude(struct texi *p, enum texicmd cmd, size_t *pos) { char fname[PATH_MAX], path[PATH_MAX]; size_t i, end; @@ -692,23 +973,23 @@ doinclude(struct texi *p, enum texicmd cmd, const char *v; enum texicmd type; - while (*pos < sz && ' ' == buf[*pos]) - advance(p, buf, pos); + while (*pos < BUFSZ(p) && ' ' == BUF(p)[*pos]) + advance(p, pos); /* Read in the filename. */ - for (i = 0; *pos < sz && '\n' != buf[*pos]; ) { + for (i = 0; *pos < BUFSZ(p) && '\n' != BUF(p)[*pos]; ) { if (i == sizeof(fname) - 1) break; - if ('@' != buf[*pos]) { - fname[i++] = buf[*pos]; - advance(p, buf, pos); + if ('@' != BUF(p)[*pos]) { + fname[i++] = BUF(p)[*pos]; + advance(p, pos); continue; } - type = texicmd(p, buf, *pos, sz, &end, NULL); - advanceto(p, buf, pos, end); + type = texicmd(p, *pos, &end, NULL); + advanceto(p, pos, end); if (TEXICMD_VALUE != type) texierr(p, "unknown include command"); - v = valueblookup(p, buf, sz, pos); + v = valueblookup(p, pos); if (NULL == v) continue; while ('\0' != *v) { @@ -722,7 +1003,7 @@ doinclude(struct texi *p, enum texicmd cmd, if (i == 0) texierr(p, "path too short"); - else if ('\n' != buf[*pos]) + else if ('\n' != BUF(p)[*pos]) texierr(p, "path too long"); else if ('/' == fname[0]) texierr(p, "no absolute paths"); @@ -749,52 +1030,54 @@ doinclude(struct texi *p, enum texicmd cmd, } static void -dobracket(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dobracket(struct texi *p, enum texicmd cmd, size_t *pos) { - parsebracket(p, buf, sz, pos); + parsebracket(p, pos, 0); } static void -dodisplay(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dodisplay(struct texi *p, enum texicmd cmd, size_t *pos) { + advanceeoln(p, pos, 1); + texivspace(p); + switch (cmd) { case (TEXICMD_FORMAT): case (TEXICMD_SMALLFORMAT): - teximacro(p, "Bd -filled"); + teximacro(p, "Bd -filled -compact"); break; default: - teximacro(p, "Bd -filled -offset indent"); + teximacro(p, "Bd -filled -offset indent -compact"); break; } - p->seenvs = 1; - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, buf, sz, pos, 1); - parseto(p, buf, sz, pos, texitoks[cmd].tok); + p->seenvs = -1; + parseto(p, pos, texitoks[cmd].tok); + p->seenvs = 0; teximacro(p, "Ed"); + texivspace(p); } static void -doexample(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doexample(struct texi *p, enum texicmd cmd, size_t *pos) { - teximacro(p, "Bd -literal -offset indent"); - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, buf, sz, pos, 1); + advanceeoln(p, pos, 1); + + texivspace(p); + teximacro(p, "Bd -literal -offset indent -compact"); p->literal++; - parseto(p, buf, sz, pos, texitoks[cmd].tok); + parseto(p, pos, texitoks[cmd].tok); p->literal--; + p->seenvs = 0; teximacro(p, "Ed"); + texivspace(p); } static void -dobye(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dobye(struct texi *p, enum texicmd cmd, size_t *pos) { texiexit(p); @@ -802,90 +1085,147 @@ dobye(struct texi *p, enum texicmd cmd, } static void -dotitle(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dotitle(struct texi *p, enum texicmd cmd, size_t *pos) { - size_t start, end; + size_t start; - while (*pos < sz && isws(buf[*pos])) - advance(p, buf, pos); - start = end = *pos; - while (end < sz && '\n' != buf[end]) - end++; - advanceeoln(p, buf, sz, pos, 1); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + + /* We want to suck down the entire line, inclusive \n. */ + start = *pos; + while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) { + if ('@' == BUF(p)[*pos]) + advance(p, pos); + advance(p, pos); + } + if (*pos < BUFSZ(p)) + advance(p, pos); + + /* Copy this into a buffer. */ free(p->subtitle); - p->subtitle = malloc(end - start + 1); - if (NULL == p->subtitle) + if (NULL == (p->subtitle = malloc(*pos - start + 1))) texiabort(p, NULL); - memcpy(p->subtitle, &buf[start], end - start); - p->subtitle[end - start] = '\0'; + memcpy(p->subtitle, &BUF(p)[start], *pos - start); + p->subtitle[*pos - start] = '\0'; } static void -doaccent(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doaccent(struct texi *p, enum texicmd cmd, size_t *pos) { + int brace = 0; - if (*pos == sz) + if (*pos == BUFSZ(p)) { + texiwarn(p, "truncated: @%s", texitoks[cmd].tok); return; - advance(p, buf, pos); + } + + /* Pad us with space, if necessary. */ + if (p->seenws && p->outcol && 0 == p->literal) { + texiputchar(p, ' '); + p->seenws = 0; + } + + /* + * If we're braced, then that's that. + * Otherwise, in a special Texinfo case: if we're a non + * alphabetic command of one letter, then the next character is + * the critical one. + * Otherwise, space can sit between us and our argument. + */ + if ('{' == BUF(p)[*pos]) { + brace = 1; + advance(p, pos); + } else if (isalpha((unsigned char)texitoks[cmd].tok[0])) + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + + if (*pos == BUFSZ(p)) { + texiwarn(p, "truncated: @%s", texitoks[cmd].tok); + return; + } + switch (cmd) { case (TEXICMD_ACUTE): - switch (buf[*pos]) { + switch (BUF(p)[*pos]) { case ('a'): case ('A'): case ('e'): case ('E'): case ('i'): case ('I'): case ('o'): case ('O'): case ('u'): case ('U'): texiputchars(p, "\\(\'"); - texiputchar(p, buf[*pos]); break; default: - texiputchar(p, buf[*pos]); + texiwarn(p, "ignoring accent"); + break; } + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); break; + case (TEXICMD_CEDILLA): + if ('c' == BUF(p)[*pos] || 'C' == BUF(p)[*pos]) + texiputchars(p, "\\(,"); + else + texiwarn(p, "ignoring accent"); + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); + break; case (TEXICMD_CIRCUMFLEX): - switch (buf[*pos]) { + switch (BUF(p)[*pos]) { case ('a'): case ('A'): case ('e'): case ('E'): case ('i'): case ('I'): case ('o'): case ('O'): case ('u'): case ('U'): texiputchars(p, "\\(^"); - texiputchar(p, buf[*pos]); break; default: - texiputchar(p, buf[*pos]); + texiwarn(p, "ignoring accent"); + break; } + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); break; + case (TEXICMD_DOTLESS): + if ('i' == BUF(p)[*pos] || 'j' == BUF(p)[*pos]) + texiputchars(p, "\\(."); + else + texiwarn(p, "ignoring accent"); + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); + break; case (TEXICMD_GRAVE): - switch (buf[*pos]) { + switch (BUF(p)[*pos]) { case ('a'): case ('A'): case ('e'): case ('E'): case ('i'): case ('I'): case ('o'): case ('O'): case ('u'): case ('U'): texiputchars(p, "\\(`"); - texiputchar(p, buf[*pos]); break; default: - texiputchar(p, buf[*pos]); + texiwarn(p, "ignoring accent"); + break; } + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); break; case (TEXICMD_TILDE): - switch (buf[*pos]) { + switch (BUF(p)[*pos]) { case ('a'): case ('A'): case ('n'): case ('N'): case ('o'): case ('O'): texiputchars(p, "\\(~"); - texiputchar(p, buf[*pos]); break; default: - texiputchar(p, buf[*pos]); + texiwarn(p, "ignoring accent"); + break; } + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); break; case (TEXICMD_UMLAUT): - switch (buf[*pos]) { + switch (BUF(p)[*pos]) { case ('a'): case ('A'): case ('e'): case ('E'): case ('i'): case ('I'): @@ -893,28 +1233,64 @@ doaccent(struct texi *p, enum texicmd cmd, case ('u'): case ('U'): case ('y'): texiputchars(p, "\\(:"); - texiputchar(p, buf[*pos]); break; default: - texiputchar(p, buf[*pos]); + texiwarn(p, "ignoring accent"); + break; } + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); break; default: - abort(); + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); + break; } + + if (brace) { + while (*pos < BUFSZ(p) && '}' != BUF(p)[*pos]) { + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); + } + if (*pos < BUFSZ(p)) + advance(p, pos); + } + + switch (cmd) { + case (TEXICMD_TIEACCENT): + texiputchar(p, ']'); + break; + case (TEXICMD_DOTACCENT): + texiputchar(p, '*'); + break; + default: + break; + } } static void -dosymbol(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dosymbol(struct texi *p, enum texicmd cmd, size_t *pos) { - if (p->seenws && p->outcol && 0 == p->literal) { + /* Remember to pad us. */ + if (p->seenws && p->outcol && 0 == p->literal) texiputchar(p, ' '); - p->seenws = 0; - } + p->seenws = 0; + switch (cmd) { + case (TEXICMD_AA): + texiputchars(p, "\\(oA"); + break; + case (TEXICMD_AASMALL): + texiputchars(p, "\\(oa"); + break; + case (TEXICMD_AE): + texiputchars(p, "\\(AE"); + break; + case (TEXICMD_AESMALL): + texiputchars(p, "\\(ae"); + break; case (TEXICMD_ASTERISK): case (TEXICMD_NEWLINE): case (TEXICMD_SPACE): @@ -924,16 +1300,29 @@ dosymbol(struct texi *p, enum texicmd cmd, case (TEXICMD_AT): texiputchar(p, '@'); break; + case (TEXICMD_BACKSLASH): + texiputchar(p, '\\'); + break; case (TEXICMD_BANG): texiputchar(p, '!'); break; case (TEXICMD_BULLET): texiputchars(p, "\\(bu"); break; + case (TEXICMD_COMMA): + texiputchar(p, ','); + break; case (TEXICMD_COPYRIGHT): texiputchars(p, "\\(co"); break; + case (TEXICMD_DH): + texiputchars(p, "\\(-D"); + break; + case (TEXICMD_DHSMALL): + texiputchars(p, "\\(Sd"); + break; case (TEXICMD_DOTS): + case (TEXICMD_ENDDOTS): texiputchars(p, "..."); break; case (TEXICMD_EQUIV): @@ -942,24 +1331,107 @@ dosymbol(struct texi *p, enum texicmd cmd, case (TEXICMD_ERROR): texiputchars(p, "error\\(->"); break; + case (TEXICMD_EURO): + texiputchars(p, "\\(Eu"); + break; + case (TEXICMD_EXCLAMDOWN): + texiputchars(p, "\\(r!"); + break; case (TEXICMD_EXPANSION): texiputchars(p, "\\(->"); break; + case (TEXICMD_GEQ): + texiputchars(p, "\\(>="); + break; + case (TEXICMD_GUILLEMETLEFT): + case (TEXICMD_GUILLEMOTLEFT): + texiputchars(p, "\\(Fo"); + break; + case (TEXICMD_GUILLEMETRIGHT): + case (TEXICMD_GUILLEMOTRIGHT): + texiputchars(p, "\\(Fc"); + break; + case (TEXICMD_GUILSINGLLEFT): + texiputchars(p, "\\(fo"); + break; + case (TEXICMD_GUILSINGLRIGHT): + texiputchars(p, "\\(fc"); + break; + case (TEXICMD_L): + texiputchars(p, "\\(/L"); + break; case (TEXICMD_LATEX): texiputchars(p, "LaTeX"); break; + case (TEXICMD_LEQ): + texiputchars(p, "\\(<="); + break; + case (TEXICMD_LSMALL): + texiputchars(p, "\\(/l"); + break; case (TEXICMD_MINUS): texiputchars(p, "\\-"); break; + case (TEXICMD_O): + texiputchars(p, "\\(/O"); + break; + case (TEXICMD_OE): + texiputchars(p, "\\(OE"); + break; + case (TEXICMD_OESMALL): + texiputchars(p, "\\(oe"); + break; + case (TEXICMD_ORDF): + texiputchars(p, "a"); + break; + case (TEXICMD_ORDM): + texiputchars(p, "o"); + break; + case (TEXICMD_OSMALL): + texiputchars(p, "\\(/o"); + break; case (TEXICMD_PERIOD): texiputchar(p, '.'); break; + case (TEXICMD_POUNDS): + texiputchars(p, "\\(Po"); + break; + case (TEXICMD_QUESTIONDOWN): + texiputchars(p, "\\(r?"); + break; case (TEXICMD_QUESTIONMARK): texiputchar(p, '?'); break; + case (TEXICMD_QUOTEDBLBASE): + texiputchars(p, "\\(Bq"); + break; + case (TEXICMD_QUOTEDBLLEFT): + texiputchars(p, "\\(lq"); + break; + case (TEXICMD_QUOTEDBLRIGHT): + texiputchars(p, "\\(rq"); + break; + case (TEXICMD_QUOTESINGLBASE): + texiputchars(p, "\\(bq"); + break; + case (TEXICMD_QUOTELEFT): + texiputchars(p, "\\(oq"); + break; + case (TEXICMD_QUOTERIGHT): + texiputchars(p, "\\(cq"); + break; + case (TEXICMD_REGISTEREDSYMBOL): + texiputchars(p, "\\(rg"); + break; case (TEXICMD_RESULT): texiputchars(p, "\\(rA"); break; + case (TEXICMD_SLASH): + texiputchar(p, '/'); + break; + case (TEXICMD_SS): + texiputchars(p, "\\(ss"); + break; case (TEXICMD_SQUIGGLE_LEFT): texiputchars(p, "{"); break; @@ -969,6 +1441,18 @@ dosymbol(struct texi *p, enum texicmd cmd, case (TEXICMD_TEXSYM): texiputchars(p, "TeX"); break; + case (TEXICMD_TEXTDEGREE): + texiputchars(p, "\\(de"); + break; + case (TEXICMD_TH): + texiputchars(p, "\\(TP"); + break; + case (TEXICMD_THSMALL): + texiputchars(p, "\\(Tp"); + break; + case (TEXICMD_TIE): + texiputchars(p, "\\ "); + break; case (TEXICMD_COLON): case (TEXICMD_HYPHEN): break; @@ -976,128 +1460,334 @@ dosymbol(struct texi *p, enum texicmd cmd, abort(); } - if (texitoks[cmd].len > 1) - doignbracket(p, cmd, buf, sz, pos); + /* Alphabetic commands have braces we ignore. */ + if (isalpha((unsigned char)texitoks[cmd].tok[0])) + doignbracket(p, cmd, pos); } static void -doquotation(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doquotation(struct texi *p, enum texicmd cmd, size_t *pos) { teximacro(p, "Qo"); - parseto(p, buf, sz, pos, "quotation"); + parseto(p, pos, "quotation"); teximacro(p, "Qc"); } +static int +indexcmp(const void *p1, const void *p2) +{ + + return(strcasecmp(*(const char **)p1, *(const char **)p2)); +} + static void -domath(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doprintindex(struct texi *p, enum texicmd cmd, size_t *pos) { - size_t nest; + size_t i, j, start, end, len; +#if HAVE_INDEX + char *cp; +#endif - /* - * Math handling is different from everything else. - * We don't allow any subcomponents, and we ignore the rules in - * terms of @-commands. - * This departs from GNU's rules, but whatever. - */ - while (*pos < sz && isws(buf[*pos])) - advance(p, buf, pos); - if (*pos == sz || '{' != buf[*pos]) + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + start = *pos; + while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) + advance(p, pos); + if ((end = *pos) == BUFSZ(p)) { + texiwarn(p, "unexpected EOF"); return; - advance(p, buf, pos); - if (p->seenws && p->outcol && 0 == p->literal) - texiputchar(p, ' '); - p->seenws = 0; - for (nest = 1; *pos < sz && nest > 0; ) { - if ('{' == buf[*pos]) - nest++; - else if ('}' == buf[*pos]) - if (0 == --nest) - continue; - texiputchar(p, buf[*pos]); - advance(p, buf, pos); } - if (*pos == sz) + + advance(p, pos); + if (0 == (len = end - start)) { + texiwarn(p, "zero-length index"); return; - assert('}' == buf[*pos]); - advance(p, buf, pos); + } + + /* Look for the index in our table. */ + for (i = 0; i < p->indexsz; i++) { + if (strlen(p->indexs[i].name) != len) + continue; + if (strncmp(p->indexs[i].name, &BUF(p)[start], len)) + continue; + break; + } + + if (i == p->indexsz) { + texiwarn(p, "cannot find index"); + return; + } else if (0 == p->indexs[i].indexsz) + return; + + /* Alphabetically sort our indices. */ + qsort(p->indexs[i].index, + p->indexs[i].indexsz, sizeof(char *), indexcmp); + + texivspace(p); + teximacro(p, "Bl -tag -width Ds -compact"); + for (j = 0; j < p->indexs[i].indexsz; j++) { + teximacroopen(p, "It"); +#if HAVE_INDEX + teximacroopen(p, "Lkx"); + texiputchars(p, "\"idx"); + texiputchars(p, p->indexs[i].name); + cp = p->indexs[i].index[j]; + while ('\n' != *cp) { + assert('\0' != *cp); + texiputchar(p, *cp++); + } + texiputchars(p, "\" \""); + p->literal++; +#endif + texisplice(p, p->indexs[i].index[j], + strlen(p->indexs[i].index[j]), *pos); + parseeoln(p, pos); +#if HAVE_INDEX + p->literal--; + texiputchars(p, "\""); + teximacroclose(p); +#endif + teximacroclose(p); + } + p->seenvs = 0; + teximacro(p, "El"); + texivspace(p); } static void -dovalue(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +donode(struct texi *p, enum texicmd cmd, size_t *pos) { - size_t start, end, i; + int sv = p->seenvs; + + if (NULL != p->chapters) { + advanceeoln(p, pos, 1); + return; + } + +#if HAVE_INDEX + p->seenvs = -1; + teximacroopen(p, "Ix"); + texiputchars(p, "node"); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + while (*pos < BUFSZ(p)) { + if (BUF(p)[*pos] == ',') + break; + else if (BUF(p)[*pos] == '\n') + break; + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); + } + teximacroclose(p); +#endif + advanceeoln(p, pos, 1); + p->seenvs = sv; +} + +/* + * This handles both menu and detailedmenu. + * The syntax of these is fairly... unspecific, but what we do here + * seems to work with most manuals. + */ +static void +domenu(struct texi *p, enum texicmd cmd, size_t *pos) +{ + size_t start, end, sv; + + if (NULL != p->chapters) { + doignblock(p, cmd, pos); + return; + } + + advanceeoln(p, pos, 1); + + /* + * Parse past initial stuff. + * TODO: the manual says we're supposed to make this in bold or + * something. + */ + while (*pos < BUFSZ(p)) { + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + if ('*' != BUF(p)[*pos]) { + if (TEXICMD_END == peeklinecmd(p, *pos)) + break; + parseeoln(p, pos); + } else + break; + } + + texivspace(p); + teximacro(p, "Bl -tag -width Ds -compact"); + while (*pos < BUFSZ(p)) { + /* + * Read to next menu item. + * We simply parse every line until we get a magic '*'. + * These lines might occur interspersed OR as the + * description of an entry. + * Either way it's in the `It' block. + */ + if (0 == p->seenws) + p->seenws = *pos < BUFSZ(p) && isws(BUF(p)[*pos]); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + if ('*' != BUF(p)[*pos]) { + if (TEXICMD_END == peeklinecmd(p, *pos)) + break; + parseeoln(p, pos); + continue; + } + + /* Now we're parsing a menu item. */ + advance(p, pos); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + start = sv = *pos; + while (*pos < BUFSZ(p) && ':' != BUF(p)[*pos]) + advance(p, pos); + end = *pos; + advance(p, pos); + + if (*pos == BUFSZ(p)) { + texiwarn(p, "bad menu syntax"); + break; + } else if (':' != BUF(p)[*pos]) { + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + teximacroopen(p, "It"); +#ifdef HAVE_INDEX + teximacroopen(p, "Lkx"); + texiputchars(p, "\"node"); +#endif + while (*pos < BUFSZ(p)) { + switch (BUF(p)[*pos]) { + case ('\t'): + case ('\n'): + case (','): + break; + case ('.'): + if (*pos + 1 == BUFSZ(p)) { + advance(p, pos); + continue; + } + if (' ' == BUF(p)[*pos + 1]) { + advance(p, pos); + break; + } + /* FALLTHROUGH */ + default: + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); + continue; + } + advance(p, pos); + break; + } +#ifdef HAVE_INDEX + texiputchars(p, "\" \""); + for (start = sv; start < end; start++) + texiputchar(p, BUF(p)[start]); + texiputchar(p, '"'); +#endif + } else { + advance(p, pos); + p->seenvs = 0; + teximacroopen(p, "It"); +#ifdef HAVE_INDEX + teximacroopen(p, "Lkx"); + texiputchars(p, "\"node"); + for (start = sv; start < end; start++) + texiputchar(p, BUF(p)[start]); + texiputchars(p, "\" \""); +#endif + for (start = sv; start < end; start++) + texiputchar(p, BUF(p)[start]); +#ifdef HAVE_INDEX + texiputchar(p, '"'); +#endif + } +#ifdef HAVE_INDEX + teximacroclose(p); +#endif + teximacroclose(p); + } + + advanceeoln(p, pos, 0); + p->seenvs = 0; + teximacro(p, "El"); + texivspace(p); +} + +static void +domath(struct texi *p, enum texicmd cmd, size_t *pos) +{ + + parsebracket(p, pos, 1); +} + +static void +dovalue(struct texi *p, enum texicmd cmd, size_t *pos) +{ + size_t start, end; char *key, *val; const char *cp; if (TEXICMD_SET == cmd) { - while (*pos < sz && isws(buf[*pos])) - advance(p, buf, pos); - for (start = end = *pos; end < sz; end++) - if (ismspace(buf[end])) + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + for (start = end = *pos; end < BUFSZ(p); end++) + if (ismspace(BUF(p)[end])) break; /* We don't allow empty keys. */ if (start == end) return; - advanceto(p, buf, pos, end); + advanceto(p, pos, end); key = malloc(end - start + 1); if (NULL == key) texiabort(p, NULL); - memcpy(key, &buf[start], end - start); + memcpy(key, &BUF(p)[start], end - start); key[end - start] = '\0'; - while (*pos < sz && isws(buf[*pos])) - advance(p, buf, pos); - for (start = end = *pos; end < sz; end++) - if ('\n' == buf[end]) + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + for (start = end = *pos; end < BUFSZ(p); end++) + if ('\n' == BUF(p)[end]) break; /* We do allow empty values. */ - advanceeoln(p, buf, sz, pos, 1); + advanceeoln(p, pos, 1); val = malloc(end - start + 1); if (NULL == val) texiabort(p, NULL); - memcpy(val, &buf[start], end - start); + memcpy(val, &BUF(p)[start], end - start); val[end - start] = '\0'; valueadd(p, key, val); } else if (TEXICMD_VALUE == cmd) { if (p->seenws) texiputchar(p, ' '); p->seenws = 0; - if (NULL != (cp = valueblookup(p, buf, sz, pos))) { - for (i = 0; i < p->valstackpos; i++) - if (cp == p->valstack[i]) - break; - if (i < p->valstackpos) - texierr(p, "recursive value"); - if (64 == p->valstackpos) - texierr(p, "too many nested values"); - p->valstack[p->valstackpos++] = cp; - parseeof(p, cp, strlen(cp)); - p->valstackpos--; - } else + if (NULL != (cp = valueblookup(p, pos))) + texisplice(p, cp, strlen(cp), *pos); + else texiputchars(p, "{No value}"); } else if (TEXICMD_IFCLEAR == cmd) { - if (NULL != valuellookup(p, buf, sz, pos)) - doignblock(p, cmd, buf, sz, pos); + if (NULL != valuellookup(p, pos)) + doignblock(p, cmd, pos); else - parseto(p, buf, sz, pos, texitoks[cmd].tok); + parseto(p, pos, texitoks[cmd].tok); } else if (TEXICMD_IFSET == cmd) { - if (NULL == valuellookup(p, buf, sz, pos)) - doignblock(p, cmd, buf, sz, pos); + if (NULL == valuellookup(p, pos)) + doignblock(p, cmd, pos); else - parseto(p, buf, sz, pos, texitoks[cmd].tok); + parseto(p, pos, texitoks[cmd].tok); } else if (TEXICMD_CLEAR == cmd) - valuelclear(p, buf, sz, pos); + valuelclear(p, pos); } static void -dolink(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dolink(struct texi *p, enum texicmd cmd, size_t *pos) { int c; @@ -1107,40 +1797,44 @@ dolink(struct texi *p, enum texicmd cmd, break; case (TEXICMD_UREF): case (TEXICMD_URL): + case (TEXICMD_INDICATEURL): teximacroopen(p, "Lk"); break; case (TEXICMD_XREF): texiputchars(p, "See Section"); - teximacroopen(p, "Qq"); + teximacroopen(p, "Dq"); break; case (TEXICMD_PXREF): texiputchars(p, "see Section"); - teximacroopen(p, "Qq"); + teximacroopen(p, "Dq"); break; + case (TEXICMD_INFOREF): + texiputchars(p, "See Info file node"); + teximacroopen(p, "Dq"); + break; default: abort(); } - c = parsearg(p, buf, sz, pos, 0); + c = parsearg(p, pos, 0); p->ign++; while (c > 0) - c = parsearg(p, buf, sz, pos, 1); + c = parsearg(p, pos, 1); p->ign--; - texipunctuate(p, buf, sz, pos); + texipunctuate(p, pos); teximacroclose(p); } static void -doignargn(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doignargn(struct texi *p, enum texicmd cmd, size_t *pos) { int c; - c = parsearg(p, buf, sz, pos, 0); + c = parsearg(p, pos, 0); p->ign++; while (c > 0) - c = parsearg(p, buf, sz, pos, 1); + c = parsearg(p, pos, 1); p->ign--; } @@ -1165,47 +1859,9 @@ sectioner(struct texi *p, int sec) } static void -dosubsubsection(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dosecoffs(struct texi *p, enum texicmd cmd, size_t *pos) { - int sec; - sec = sectioner(p, 3); - - /* We don't have a subsubsubsection, so make one up. */ - texivspace(p); - teximacroopen(p, sects[sec]); - parseeoln(p, buf, sz, pos); - teximacroclose(p); - texivspace(p); -} - -static void -dosubsection(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) -{ - int sec; - - sec = sectioner(p, 2); - - if (p->outmacro) - texierr(p, "\"%s\" in open line scope!?", sects[sec]); - else if (p->literal) - texierr(p, "\"%s\" in a literal scope!?", sects[sec]); - - /* We don't have a subsubsection, so make one up. */ - texivspace(p); - teximacroopen(p, sects[sec]); - parseeoln(p, buf, sz, pos); - teximacroclose(p); - texivspace(p); -} - -static void -dosecoffs(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) -{ - if (TEXICMD_RAISESECTIONS == cmd) p->secoffs++; else @@ -1213,15 +1869,16 @@ dosecoffs(struct texi *p, enum texicmd cmd, } static void -dosection(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dosection(struct texi *p, enum texicmd cmd, size_t *pos) { int sec; switch (cmd) { + case (TEXICMD_TOP): + sec = 0; + break; case (TEXICMD_APPENDIX): case (TEXICMD_CHAPTER): - case (TEXICMD_TOP): case (TEXICMD_UNNUMBERED): sec = sectioner(p, 0); break; @@ -1231,6 +1888,18 @@ dosection(struct texi *p, enum texicmd cmd, case (TEXICMD_UNNUMBEREDSEC): sec = sectioner(p, 1); break; + case (TEXICMD_APPENDIXSUBSEC): + case (TEXICMD_SUBHEADING): + case (TEXICMD_SUBSECTION): + case (TEXICMD_UNNUMBEREDSUBSEC): + sec = sectioner(p, 2); + break; + case (TEXICMD_APPENDIXSUBSUBSEC): + case (TEXICMD_SUBSUBHEADING): + case (TEXICMD_SUBSUBSECTION): + case (TEXICMD_UNNUMBEREDSUBSUBSEC): + sec = sectioner(p, 3); + break; default: abort(); } @@ -1240,72 +1909,57 @@ dosection(struct texi *p, enum texicmd cmd, else if (p->literal) texierr(p, "\"%s\" in a literal scope!?", sects[sec]); + if (sec < 2) + p->seenvs = -1; + else + texivspace(p); + + if (0 == sec && NULL != p->chapters) { + teximdocclose(p, 0); + teximdocopen(p, pos); + } + teximacroopen(p, sects[sec]); - parseeoln(p, buf, sz, pos); + parseeoln(p, pos); teximacroclose(p); - p->seenvs = 1; + + if (sec < 2) + p->seenvs = -1; + else + texivspace(p); } static void -dosp(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dotop(struct texi *p, enum texicmd cmd, size_t *pos) { - texivspace(p); - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, buf, sz, pos, 1); + if (--p->ign) + texierr(p, "@top command while ignoring"); + + if (NULL == p->chapters) + teximdocopen(p, pos); + dosection(p, cmd, pos); } static void -dotop(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dosp(struct texi *p, enum texicmd cmd, size_t *pos) { - const char *cp; - time_t t; - char date[32]; - if (--p->ign) - texierr(p, "@top command while ignoring (%d)", p->ign); - - /* - * Here we print our standard mdoc(7) prologue. - * We use the title set with @settitle for the `Nd' description - * and the source document filename (the first one as invoked on - * the command line) for the title. - * The date is set to the current date. - */ - t = time(NULL); - strftime(date, sizeof(date), "%F", localtime(&t)); - - teximacroopen(p, "Dd"); - texiputchars(p, date); - teximacroclose(p); - teximacroopen(p, "Dt"); - for (cp = p->title; '\0' != *cp; cp++) - texiputchar(p, toupper(*cp)); - texiputchars(p, " 7"); - teximacroclose(p); - teximacro(p, "Os"); - teximacro(p, "Sh NAME"); - teximacroopen(p, "Nm"); - texiputchars(p, p->title); - teximacroclose(p); - teximacroopen(p, "Nd"); - texiputchars(p, NULL != p->subtitle ? - p->subtitle : "Unknown description"); - teximacroclose(p); - p->seenvs = 1; - dosection(p, cmd, buf, sz, pos); + advanceeoln(p, pos, 1); + if (p->literal) + texiputchar(p, '\n'); + else + texivspace(p); } static void -doitem(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doitem(struct texi *p, enum texicmd cmd, size_t *pos) { /* Multitable is using raw tbl(7). */ if (TEXILIST_TABLE == p->list) { - texiputchar(p, '\n'); + if (p->outcol > 0) + texiputchar(p, '\n'); return; } @@ -1316,9 +1970,11 @@ doitem(struct texi *p, enum texicmd cmd, switch (p->list) { case (TEXILIST_ITEM): + p->seenvs = -1; teximacroopen(p, "It"); break; case (TEXILIST_NOITEM): + p->seenvs = -1; teximacro(p, "It"); break; default: @@ -1327,8 +1983,7 @@ doitem(struct texi *p, enum texicmd cmd, } /* Trick so we don't start with Pp. */ - p->seenvs = 1; - parseeoln(p, buf, sz, pos); + parseeoln(p, pos); if (TEXILIST_ITEM == p->list) teximacroclose(p); @@ -1337,51 +1992,56 @@ doitem(struct texi *p, enum texicmd cmd, } static void -dotab(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dotab(struct texi *p, enum texicmd cmd, size_t *pos) { /* This command is only useful in @multitable. */ - if (TEXILIST_TABLE == p->list) + if (TEXILIST_TABLE == p->list && p->outcol) texiputchar(p, '\t'); } static void -domultitable(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +domultitable(struct texi *p, enum texicmd cmd, size_t *pos) { enum texilist sv = p->list; + int svliteral = p->literal; enum texicmd type; size_t i, end, columns; + texivspace(p); p->list = TEXILIST_TABLE; + /* + * TS/TE blocks aren't "in mdoc(7)", so we can disregard the + * fact that we're in literal mode right now. + */ + p->literal = 0; teximacro(p, "TS"); columns = 0; /* Advance to the first argument... */ - while (*pos < sz && isws(buf[*pos])) - advance(p, buf, pos); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); /* Make sure we don't print anything when scanning. */ p->ign++; - if ('@' == buf[*pos]) { + if (*pos < BUFSZ(p) && '@' == BUF(p)[*pos]) { /* * Look for @columnfractions. * We ignore these, but we do use the number of * arguments to set the number of columns that we'll * have. */ - type = texicmd(p, buf, *pos, sz, &end, NULL); - advanceto(p, buf, pos, end); + type = texicmd(p, *pos, &end, NULL); + advanceto(p, pos, end); if (TEXICMD_COLUMNFRACTIONS != type) texierr(p, "unknown multitable command"); - while (*pos < sz && '\n' != buf[*pos]) { - while (*pos < sz && isws(buf[*pos])) - advance(p, buf, pos); - while (*pos < sz && ! isws(buf[*pos])) { - if ('\n' == buf[*pos]) + while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) { + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + while (*pos < BUFSZ(p) && ! isws(BUF(p)[*pos])) { + if ('\n' == BUF(p)[*pos]) break; - advance(p, buf, pos); + advance(p, pos); } columns++; } @@ -1392,7 +2052,7 @@ domultitable(struct texi *p, enum texicmd cmd, * tbl(7) figure it out. * So use this only to count arguments. */ - while (parselinearg(p, buf, sz, pos) > 0) + while (parselinearg(p, pos) > 0) columns++; p->ign--; @@ -1402,85 +2062,155 @@ domultitable(struct texi *p, enum texicmd cmd, texiputchar(p, ' '); texiputchar(p, 'l'); } - texiputchars(p, ".\n"); + + texiputchar(p, '.'); + texiputchar(p, '\n'); p->outmacro++; - parseto(p, buf, sz, pos, texitoks[cmd].tok); + parseto(p, pos, texitoks[cmd].tok); p->outmacro--; teximacro(p, "TE"); + p->literal = svliteral; p->list = sv; + texivspace(p); } static void -dotable(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dotable(struct texi *p, enum texicmd cmd, size_t *pos) { enum texilist sv = p->list; + advanceeoln(p, pos, 1); + p->list = TEXILIST_ITEM; - teximacro(p, "Bl -tag -width Ds"); - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, buf, sz, pos, 1); - p->seenvs = 1; - parseto(p, buf, sz, pos, texitoks[cmd].tok); + texivspace(p); + teximacro(p, "Bl -tag -width Ds -compact"); + parseto(p, pos, texitoks[cmd].tok); + p->seenvs = 0; teximacro(p, "El"); + texivspace(p); p->list = sv; } static void -doenumerate(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doend(struct texi *p, enum texicmd cmd, size_t *pos) { + size_t start; + + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + start = *pos; + while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) + advance(p, pos); + + texiwarn(p, "unexpected \"end\": %.*s", + (int)(*pos - start), &BUF(p)[start]); + advanceeoln(p, pos, 1); +} + +static void +doenumerate(struct texi *p, enum texicmd cmd, size_t *pos) +{ enum texilist sv = p->list; + advanceeoln(p, pos, 1); + p->list = TEXILIST_NOITEM; - teximacro(p, "Bl -enum"); - p->seenvs = 1; - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, buf, sz, pos, 1); - parseto(p, buf, sz, pos, "enumerate"); + texivspace(p); + teximacro(p, "Bl -enum -compact"); + parseto(p, pos, texitoks[cmd].tok); + p->seenvs = 0; teximacro(p, "El"); + texivspace(p); p->list = sv; } static void -doitemize(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doitemize(struct texi *p, enum texicmd cmd, size_t *pos) { enum texilist sv = p->list; + advanceeoln(p, pos, 1); + p->list = TEXILIST_NOITEM; - teximacro(p, "Bl -bullet"); - p->seenvs = 1; - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, buf, sz, pos, 1); - parseto(p, buf, sz, pos, "itemize"); + texivspace(p); + teximacro(p, "Bl -bullet -compact"); + parseto(p, pos, texitoks[cmd].tok); + p->seenvs = 0; teximacro(p, "El"); + texivspace(p); p->list = sv; } static void -doignbracket(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doignbracket(struct texi *p, enum texicmd cmd, size_t *pos) { p->ign++; - parsebracket(p, buf, sz, pos); + parsebracket(p, pos, 0); p->ign--; } static void -doignline(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +doindex(struct texi *p, enum texicmd cmd, size_t *pos) { + size_t start, end, len; - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, buf, sz, pos, 1); + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + + start = *pos; + while (*pos < BUFSZ(p) && '\n' != BUF(p)[*pos]) + advance(p, pos); + + if (*pos == BUFSZ(p)) { + texiwarn(p, "unexpected EOF"); + return; + } + + advance(p, pos); + end = *pos; + if (0 == (len = end - start)) { + texiwarn(p, "zero-length index"); + return; + } + + /* Two-letter combos we can look up verbatim. */ + if (7 == texitoks[cmd].len) { + texindex(p, texitoks[cmd].tok, 2, &BUF(p)[start], len); + return; + } + + assert(6 == texitoks[cmd].len); + /* Newer one-letter combos need to be mapped. */ + switch (texitoks[cmd].tok[0]) { + case ('c'): + texindex(p, "cp", 2, &BUF(p)[start], len); + break; + case ('v'): + texindex(p, "vr", 2, &BUF(p)[start], len); + break; + case ('f'): + texindex(p, "fn", 2, &BUF(p)[start], len); + break; + case ('t'): + texindex(p, "tp", 2, &BUF(p)[start], len); + break; + default: + abort(); + } } +static void +doignline(struct texi *p, enum texicmd cmd, size_t *pos) +{ + + advanceeoln(p, pos, 1); +} + /* * Parse colon-separated directories from "cp" (if not NULL) and returns * the array of pointers. - * Prepends "base" to the array. + * Prepends "base" to the array, if found. * This does NOT sanitise the directories! */ static char ** @@ -1488,26 +2218,27 @@ parsedirs(struct texi *p, const char *base, const char { char *tok, *str, *tofree; const char *cpp; - size_t i; + size_t i = 0; char **dirs; - *sz = NULL != (cpp = cp) ? 2 : 1; - if (*sz > 1) - for ( ; NULL != (cpp = strchr(cpp, ':')); (*sz)++) + /* Count up our expected arguments. */ + *sz = NULL != base; + if (NULL != (cpp = cp)) + for ((*sz)++; NULL != (cpp = strchr(cpp, ':')); (*sz)++) cpp++; + if (0 == *sz) + return(NULL); if (NULL == (dirs = calloc(*sz, sizeof(char *)))) texiabort(p, NULL); - else if (NULL == (dirs[0] = strdup(base))) + if (NULL != base && NULL == (dirs[i++] = strdup(base))) texiabort(p, NULL); - if (NULL == cp) return(dirs); - if (NULL == (tofree = tok = str = strdup(cp))) texiabort(p, NULL); - for (i = 1; NULL != (tok = strsep(&str, ":")); i++) + for ( ; NULL != (tok = strsep(&str, ":")); i++) if (NULL == (dirs[i] = strdup(tok))) texiabort(p, NULL); @@ -1520,7 +2251,7 @@ main(int argc, char *argv[]) { struct texi texi; int c; - char *path, *dir; + char *dirpath, *dir, *ccp; const char *progname, *Idir, *cp; progname = strrchr(argv[0], '/'); @@ -1530,10 +2261,16 @@ main(int argc, char *argv[]) ++progname; memset(&texi, 0, sizeof(struct texi)); + texi.ign = 1; + texi.outfile = stdout; + texi.seenvs = -1; Idir = NULL; - while (-1 != (c = getopt(argc, argv, "I:"))) + while (-1 != (c = getopt(argc, argv, "C:I:"))) switch (c) { + case ('C'): + texi.chapters = optarg; + break; case ('I'): Idir = optarg; break; @@ -1542,33 +2279,39 @@ main(int argc, char *argv[]) } argv += optind; - if (0 == (argc -= optind)) - goto usage; + argc -= optind; - if (NULL == (path = strdup(argv[0]))) - texiabort(&texi, NULL); - else if (NULL == (dir = dirname(path))) - texiabort(&texi, NULL); + /* Add the default Texinfo indices. */ + texindex_add(&texi, "cp", 2); + texindex_add(&texi, "vr", 2); + texindex_add(&texi, "tp", 2); + texindex_add(&texi, "fn", 2); - free(path); + if (argc > 0) { + if (NULL == (dirpath = strdup(argv[0]))) + texiabort(&texi, NULL); + if (NULL == (dir = dirname(dirpath))) + texiabort(&texi, NULL); + if (NULL != (cp = strrchr(argv[0], '/'))) + texi.title = strdup(cp + 1); + else + texi.title = strdup(argv[0]); + if (NULL == texi.title) + texiabort(&texi, NULL); + else if (NULL != (ccp = strchr(texi.title, '.'))) + *ccp = '\0'; + texi.dirs = parsedirs(&texi, dir, Idir, &texi.dirsz); + free(dirpath); + parsefile(&texi, argv[0], 1); + } else { + texi.title = strdup("Unknown Manual"); + texi.dirs = parsedirs(&texi, NULL, Idir, &texi.dirsz); + parsestdin(&texi); + } - if (NULL != (cp = strrchr(argv[0], '/'))) - texi.title = strdup(cp + 1); - else - texi.title = strdup(argv[0]); - - if (NULL == texi.title) - texiabort(&texi, NULL); - else if (NULL != (path = strchr(texi.title, '.'))) - *path = '\0'; - - texi.ign = 1; - texi.dirs = parsedirs(&texi, dir, Idir, &texi.dirsz); - parsefile(&texi, argv[0], 1); - /* We shouldn't get here. */ texiexit(&texi); - return(EXIT_FAILURE); + exit(EXIT_SUCCESS); usage: - fprintf(stderr, "usage: %s [-Idirs] file\n", progname); + fprintf(stderr, "usage: %s [-Cdir] [-Idirs] [file]\n", progname); return(EXIT_FAILURE); }