=================================================================== RCS file: /cvs/texi2mdoc/main.c,v retrieving revision 1.20 retrieving revision 1.57 diff -u -p -r1.20 -r1.57 --- texi2mdoc/main.c 2015/02/19 16:15:17 1.20 +++ texi2mdoc/main.c 2015/03/01 16:57:39 1.57 @@ -1,4 +1,4 @@ -/* $Id: main.c,v 1.20 2015/02/19 16:15:17 kristaps Exp $ */ +/* $Id: main.c,v 1.57 2015/03/01 16:57:39 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 @@ -27,311 +26,100 @@ #include #include #include -#include #include -/* - * This defines each one of the Texinfo commands that we understand. - * Obviously this only refers to native commands; overriden names are a - * different story. - */ -enum texicmd { - TEXICMD_ACRONYM, - TEXICMD_ACUTE, - TEXICMD_A4PAPER, - TEXICMD_ANCHOR, - TEXICMD_APPENDIX, - TEXICMD_APPENDIXSEC, - TEXICMD_APPENDIXSUBSEC, - TEXICMD_ASTERISK, - TEXICMD_AT, - TEXICMD_AUTHOR, - TEXICMD_B, - TEXICMD_BANG, - TEXICMD_BULLET, - TEXICMD_BYE, - TEXICMD_CENTER, - TEXICMD_CHAPTER, - TEXICMD_CINDEX, - TEXICMD_CIRCUMFLEX, - TEXICMD_CITE, - TEXICMD_CODE, - TEXICMD_COLON, - TEXICMD_COLUMNFRACTIONS, - TEXICMD_COMMAND, - TEXICMD_COMMENT, - TEXICMD_COMMENT_LONG, - TEXICMD_CONTENTS, - TEXICMD_COPYING, - TEXICMD_COPYRIGHT, - TEXICMD_DEFFN, - TEXICMD_DEFFNX, - TEXICMD_DEFMAC, - TEXICMD_DEFMACX, - TEXICMD_DEFTP, - TEXICMD_DEFTPX, - TEXICMD_DEFTYPEFN, - TEXICMD_DEFTYPEFNX, - TEXICMD_DEFTYPEFUN, - TEXICMD_DEFTYPEFUNX, - TEXICMD_DEFTYPEVAR, - TEXICMD_DEFTYPEVARX, - TEXICMD_DEFTYPEVR, - TEXICMD_DEFTYPEVRX, - TEXICMD_DEFUN, - TEXICMD_DEFUNX, - TEXICMD_DEFVAR, - TEXICMD_DEFVARX, - TEXICMD_DEFVR, - TEXICMD_DEFVRX, - TEXICMD_DETAILMENU, - TEXICMD_DFN, - TEXICMD_DIRCATEGORY, - TEXICMD_DIRENTRY, - TEXICMD_DISPLAY, - TEXICMD_DOTS, - TEXICMD_EMAIL, - TEXICMD_EMPH, - TEXICMD_END, - TEXICMD_ENUMERATE, - TEXICMD_ENV, - TEXICMD_ERROR, - TEXICMD_EXAMPLE, - TEXICMD_EXPANSION, - TEXICMD_FILE, - TEXICMD_FINALOUT, - TEXICMD_FINDEX, - TEXICMD_FTABLE, - TEXICMD_FORMAT, - TEXICMD_GRAVE, - TEXICMD_GROUP, - TEXICMD_HEADING, - TEXICMD_HEADINGS, - TEXICMD_HEADITEM, - TEXICMD_HYPHEN, - TEXICMD_I, - TEXICMD_IFCLEAR, - TEXICMD_IFDOCBOOK, - TEXICMD_IFHTML, - TEXICMD_IFINFO, - TEXICMD_IFNOTDOCBOOK, - TEXICMD_IFNOTHTML, - TEXICMD_IFNOTINFO, - TEXICMD_IFNOTPLAINTEXT, - TEXICMD_IFNOTTEX, - TEXICMD_IFNOTXML, - TEXICMD_IFPLAINTEXT, - TEXICMD_IFTEX, - TEXICMD_IFSET, - TEXICMD_IFXML, - TEXICMD_IGNORE, - TEXICMD_IMAGE, - TEXICMD_INCLUDE, - TEXICMD_INDENTBLOCK, - TEXICMD_INSERTCOPYING, - TEXICMD_ITEM, - TEXICMD_ITEMIZE, - TEXICMD_ITEMX, - TEXICMD_KBD, - TEXICMD_KEY, - TEXICMD_KINDEX, - TEXICMD_LATEX, - TEXICMD_MATH, - TEXICMD_MENU, - TEXICMD_MULTITABLE, - TEXICMD_NEED, - TEXICMD_NEWLINE, - TEXICMD_NODE, - TEXICMD_NOINDENT, - TEXICMD_OPTION, - TEXICMD_PXREF, - TEXICMD_QUESTIONMARK, - TEXICMD_QUOTATION, - TEXICMD_PAGE, - TEXICMD_PARINDENT, - TEXICMD_PRINTINDEX, - TEXICMD_R, - TEXICMD_REF, - TEXICMD_RESULT, - TEXICMD_SAMP, - TEXICMD_SANSSERIF, - TEXICMD_SC, - TEXICMD_SECTION, - TEXICMD_SET, - TEXICMD_SETCHAPNEWPAGE, - TEXICMD_SETFILENAME, - TEXICMD_SETTITLE, - TEXICMD_SLANTED, - TEXICMD_SP, - TEXICMD_SPACE, - TEXICMD_SMALLBOOK, - TEXICMD_SMALLDISPLAY, - TEXICMD_SMALLEXAMPLE, - TEXICMD_SMALLFORMAT, - TEXICMD_SMALLINDENTBLOCK, - TEXICMD_SQUIGGLE_LEFT, - TEXICMD_SQUIGGLE_RIGHT, - TEXICMD_STRONG, - TEXICMD_SUBHEADING, - TEXICMD_SUBSECTION, - TEXICMD_SUBTITLE, - TEXICMD_SYNCODEINDEX, - TEXICMD_T, - TEXICMD_TAB, - TEXICMD_TABSYM, - TEXICMD_TABLE, - TEXICMD_TEX, - TEXICMD_TEXSYM, - TEXICMD_TILDE, - TEXICMD_TITLE, - TEXICMD_TITLEFONT, - TEXICMD_TITLEPAGE, - TEXICMD_TOP, - TEXICMD_UMLAUT, - TEXICMD_UNNUMBERED, - TEXICMD_UNNUMBEREDSEC, - TEXICMD_UNNUMBEREDSUBSEC, - TEXICMD_UREF, - TEXICMD_URL, - TEXICMD_VAR, - TEXICMD_VERBATIMINCLUDE, - TEXICMD_VINDEX, - TEXICMD_VSKIP, - TEXICMD_VTABLE, - TEXICMD_W, - TEXICMD_XREF, - TEXICMD__MAX -}; +#include "extern.h" -/* - * The file currently being parsed. - * This keeps track of our location within that file. - */ -struct texifile { - const char *name; /* name of the file */ - size_t line; /* current line (from zero) */ - size_t col; /* current column in line (from zero) */ - char *map; /* mmap'd file */ - size_t mapsz; /* size of mmap */ +#define SECTSZ 4 +static const char *const sects[SECTSZ] = { + "Sh", + "Ss", + "Em", + "No", }; -struct texi; +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 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 doinline(struct texi *, enum texicmd, size_t *); +static void doinclude(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 domultitable(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 dosubsection(struct texi *, enum texicmd, size_t *); +static void dosubsubsection(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 *); -/* - * Callback for functions implementing texi commands. - */ -typedef void (*texicmdfp)(struct texi *, - enum texicmd, const char *, size_t, size_t *); - -/* - * Describes Texinfo commands, whether native or overriden. - */ -struct texitok { - texicmdfp fp; /* callback (or NULL if none) */ - const char *tok; /* name of the token */ - size_t len; /* strlen(tok) */ -}; - -enum texilist { - TEXILIST_NONE = 0, - TEXILIST_ITEM, - TEXILIST_NOITEM, - TEXILIST_TABLE -}; - -/* - * The main parse structure. - * This keeps any necessary information handy. - */ -struct texi { - struct texifile files[64]; /* stack of open files */ - size_t filepos; /* number of open files */ - size_t outcol; /* column in output line */ - char **dirs; /* texi directories */ - size_t dirsz; /* number of texi directories */ - char *title; /* title of document */ - char *subtitle; /* subtitle of document */ - /* - * The following control what we output to the screen. - * The complexity is required to accomodate for mdoc(7). - */ - enum texilist list; /* current list (set recursively) */ - int outmacro; /* if >0, output is in line macro */ - int seenws; /* ws has been seen (and ignored) */ - int seenvs; /* newline has been Pp'd */ - int ign; /* if >0, don't print anything */ - int literal; /* if >0, literal context */ -}; - -/* Texi disregards spaces and tabs. */ -#define isws(_x) \ - (' ' == (_x) || '\t' == (_x)) -#define ismspace(_x) \ - (isws((_x)) || '\n' == (_x)) - -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 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 dofont(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 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 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 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 doverbinclude(struct texi *, enum texicmd, const char *, size_t, size_t *); - -static const struct texitok texitoks[TEXICMD__MAX] = { +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 */ + { doinline, "asis", 4 }, /* TEXICMD_ASIS */ { dosymbol, "*", 1 }, /* TEXICMD_ASTERISK */ { dosymbol, "@", 1 }, /* TEXICMD_AT */ { doignline, "author", 6 }, /* TEXICMD_AUTHOR */ - { dofont, "b", 1 }, /* TEXICMD_BOLD */ + { doinline, "b", 1 }, /* TEXICMD_BOLD */ { 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 */ { doaccent, "^", 1 }, /* TEXICMD_CIRCUMFLEX */ - { dofont, "code", 4 }, /* TEXICMD_CODE */ - { dofont, "cite", 4 }, /* TEXICMD_CITE */ + { 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 */ { dosymbol, "copyright", 9 }, /* TEXICMD_COPYRIGHT */ + { dodefindex, "defcodeindex", 12 }, /* TEXICMD_DEFCODEINDEX */ { dodefn, "deffn", 5 }, /* TEXICMD_DEFFN */ { dodefn, "deffnx", 6 }, /* TEXICMD_DEFFNX */ + { dodefindex, "defindex", 8 }, /* TEXICMD_DEFINDEX */ { dodefn, "defmac", 6 }, /* TEXICMD_DEFMAC */ { dodefn, "defmacx", 7 }, /* TEXICMD_DEFMACX */ { dodefn, "deftp", 5 }, /* TEXICMD_DEFTP */ @@ -340,6 +128,8 @@ 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 */ @@ -351,35 +141,62 @@ static const struct texitok texitoks[TEXICMD__MAX] = { { dodefn, "defvr", 5 }, /* TEXICMD_DEFVR */ { dodefn, "defvrx", 6 }, /* TEXICMD_DEFVRX */ { doignblock, "detailmenu", 10 }, /* TEXICMD_DETAILMENU */ - { dofont, "dfn", 3 }, /* TEXICMD_DFN */ + { 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 */ - { dofont, "emph", 4 }, /* TEXICMD_EMPH */ - { NULL, "end", 3 }, /* TEXICMD_END */ + { doinline, "emph", 4 }, /* TEXICMD_EMPH */ + { 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 */ + { 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 */ + { doignblock, "html", 4 }, /* TEXICMD_HTML */ { dosymbol, "-", 1 }, /* TEXICMD_HYPHEN */ - { dofont, "i", 1 }, /* TEXICMD_I */ - { doignblock, "ifclear", 7 }, /* TEXICMD_IFCLEAR */ + { doinline, "i", 1 }, /* TEXICMD_I */ + { dovalue, "ifclear", 7 }, /* TEXICMD_IFCLEAR */ { doignblock, "ifdocbook", 9 }, /* TEXICMD_IFDOCBOOK */ { doignblock, "ifhtml", 6 }, /* TEXICMD_IFHTML */ - { doignblock, "ifinfo", 6 }, /* TEXICMD_IFINFO */ + { doblock, "ifinfo", 6 }, /* TEXICMD_IFINFO */ { doblock, "ifnotdocbook", 12 }, /* TEXICMD_IFNOTDOCBOOK */ { doblock, "ifnothtml", 9 }, /* TEXICMD_IFNOTHTML */ { doblock, "ifnotinfo", 9 }, /* TEXICMD_IFNOTINFO */ @@ -388,46 +205,79 @@ static const struct texitok texitoks[TEXICMD__MAX] = { { doblock, "ifnotxml", 8 }, /* TEXICMD_IFNOTXML */ { doblock, "ifplaintext", 11 }, /* TEXICMD_IFPLAINTEXT */ { doignblock, "iftex", 5 }, /* TEXICMD_IFTEX */ - { doignblock, "ifset", 5 }, /* TEXICMD_IFSET */ + { dovalue, "ifset", 5 }, /* TEXICMD_IFSET */ { doignblock, "ifxml", 5 }, /* TEXICMD_IFXML */ { doignblock, "ignore", 6 }, /* TEXICMD_IGNORE */ { doignbracket, "image", 5 }, /* TEXICMD_IMAGE */ { doinclude, "include", 7 }, /* TEXICMD_INCLUDE */ { dodisplay, "indentblock", 11 }, /* TEXICMD_INDENTBLOCK */ + { dolink, "indicateurl", 11 }, /* TEXICMD_INDICATEURL */ + { dolink, "inforef", 7 }, /* TEXICMD_INFOREF */ { doignline, "insertcopying", 13 }, /* TEXICMD_INSERTCOPYING */ { doitem, "item", 4 }, /* TEXICMD_ITEM */ { doitemize, "itemize", 7 }, /* TEXICMD_ITEMIZE */ { doitem, "itemx", 5 }, /* TEXICMD_ITEMX */ - { dofont, "kbd", 3 }, /* TEXICMD_KBD */ + { 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 */ + { 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 */ { 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 */ + { 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 */ + { dosymbol, "pounds", 6 }, /* TEXICMD_POUNDS */ + { doignline, "printindex", 10 }, /* TEXICMD_PRINTINDEX */ { dolink, "pxref", 5 }, /* TEXICMD_PXREF */ + { dosymbol, "questiondown", 12 }, /* TEXICMD_QUESTIONDOWN */ { dosymbol, "?", 1 }, /* TEXICMD_QUESTIONMARK */ { doquotation, "quotation", 9 }, /* TEXICMD_QUOTATION */ - { doignline, "page", 4 }, /* TEXICMD_PAGE */ - { doignline, "paragraphindent", 14 }, /* TEXICMD_PARINDENT */ - { doignline, "printindex", 10 }, /* TEXICMD_PRINTINDEX */ - { dofont, "r", 1 }, /* TEXICMD_R */ + { 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 */ - { dofont, "samp", 4 }, /* TEXICMD_SAMP */ - { dofont, "sansserif", 9 }, /* TEXICMD_SANSSERIF */ + { doaccent, "ringaccent", 10 }, /* TEXICMD_RINGACCENT */ + { doinline, "samp", 4 }, /* TEXICMD_SAMP */ + { doinline, "sansserif", 9 }, /* TEXICMD_SANSSERIF */ { dobracket, "sc", 2 }, /* TEXICMD_SC */ { dosection, "section", 7 }, /* TEXICMD_SECTION */ - { doignline, "set", 3 }, /* TEXICMD_SET */ + { 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 */ - { dofont, "slanted", 7 }, /* TEXICMD_SLANTED */ + { 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 */ @@ -437,29 +287,49 @@ static const struct texitok texitoks[TEXICMD__MAX] = { { dodisplay, "smallindentblock", 16 }, /* TEXICMD_SMALLINDENTBLOCK */ { dosymbol, "{", 1 }, /* TEXICMD_SQUIGGLE_LEFT */ { dosymbol, "}", 1 }, /* TEXICMD_SQUIGGLE_RIGHT */ - { dofont, "strong", 6 }, /* TEXICMD_STRONG */ + { 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 */ { doignline, "subtitle", 8 }, /* TEXICMD_SUBTITLE */ - { doignline, "syncodeindex", 12 }, /* TEXICMD_SYNCODEINDEX */ - { dofont, "t", 1 }, /* TEXICMD_T */ + { doignline, "summarycontents", 15 }, /* TEXICMD_SUMMARYCONTENTS */ + { 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 */ { doignline, "title", 5 }, /* TEXICMD_TITLE */ { dobracket, "titlefont", 9 }, /* TEXICMD_TITLEFONT */ { doignblock, "titlepage", 9 }, /* TEXICMD_TITLEPAGE */ { dotop, "top", 3 }, /* TEXICMD_TOP */ + { 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 */ { 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, "verb", 4 }, /* TEXICMD_VERB */ + { doverbatim, "verbatim", 8 }, /* TEXICMD_VERBATIM */ { doverbinclude, "verbatiminclude", 15 }, /* TEXICMD_VERBATIMINCLUDE */ { doignline, "vindex", 6 }, /* TEXICMD_VINDEX */ { dosp, "vskip", 5 }, /* TEXICMD_VSKIP */ @@ -469,753 +339,51 @@ static const struct texitok texitoks[TEXICMD__MAX] = { /* TEXICMD__END */ }; -/* - * Unmap the top-most file that we're using. - */ -static void -texifilepop(struct texi *p) -{ - struct texifile *f; +const struct texitok *const texitoks = __texitoks; - assert(p->filepos > 0); - f = &p->files[--p->filepos]; - munmap(f->map, f->mapsz); -} - -/* - * Unmap all files that we're currently using and free all resources - * that we've allocated during the parse. - * The utility should exit(...) after this is called. - */ static void -texiexit(struct texi *p) +dodefindex(struct texi *p, enum texicmd cmd, size_t *pos) { - size_t i; - - if (p->outcol) - putchar('\n'); - - while (p->filepos > 0) - texifilepop(p); - - for (i = 0; i < p->dirsz; i++) - free(p->dirs[i]); - - free(p->dirs); - free(p->subtitle); - free(p->title); -} - -/* - * Fatal error: unmap all files and exit. - * The "errstring" is passed to perror(3). - */ -static void -texiabort(struct texi *p, const char *errstring) -{ - - perror(errstring); - texiexit(p); - exit(EXIT_FAILURE); -} - -/* - * Print a generic warning message (to stderr) tied to our current - * location in the parse sequence. - */ -static void -texiwarn(const struct texi *p, const char *fmt, ...) -{ - va_list ap; - - fprintf(stderr, "%s:%zu:%zu: warning: ", - p->files[p->filepos - 1].name, - p->files[p->filepos - 1].line + 1, - p->files[p->filepos - 1].col + 1); - va_start(ap, fmt); - vfprintf(stderr, fmt, ap); - va_end(ap); - fputc('\n', stderr); -} - -/* - * Print an error message (to stderr) tied to our current location in - * the parse sequence, invoke texiexit(), then die. - */ -static void -texierr(struct texi *p, const char *fmt, ...) -{ - va_list ap; - - fprintf(stderr, "%s:%zu:%zu: error: ", - p->files[p->filepos - 1].name, - p->files[p->filepos - 1].line + 1, - p->files[p->filepos - 1].col + 1); - va_start(ap, fmt); - vfprintf(stderr, fmt, ap); - va_end(ap); - fputc('\n', stderr); - texiexit(p); - exit(EXIT_FAILURE); -} - -/* - * Put a single data character to the output if we're not ignoring. - * Adjusts our output status. - * This shouldn't be called for macros: just for ordinary text. - */ -static void -texiputchar(struct texi *p, char c) -{ - - if (p->ign) - return; - - if ('.' == c && 0 == p->outcol) - fputs("\\&", stdout); - - putchar(c); - p->seenvs = 0; - if ('\n' == c) { - p->outcol = 0; - p->seenws = 0; - } else - p->outcol++; -} - -/* - * Put multiple characters (see texiputchar()). - * This shouldn't be called for macros: just for ordinary text. - */ -static void -texiputchars(struct texi *p, const char *s) -{ - - while ('\0' != *s) - texiputchar(p, *s++); -} - -/* - * Close an mdoc(7) macro opened with teximacroopen(). - * If there are no more macros on the line, prints a newline. - */ -static void -teximacroclose(struct texi *p) -{ - - if (p->ign) - return; - - if (0 == --p->outmacro) { - putchar('\n'); - p->outcol = p->seenws = 0; - } -} - -/* - * Open a mdoc(7) macro. - * This is used for line macros, e.g., Qq [foo bar baz]. - * It can be invoked for nested macros, e.g., Qq Li foo . - */ -static void -teximacroopen(struct texi *p, const char *s) -{ - int rc; - - if (p->ign) - return; - - if (p->outcol && 0 == p->outmacro) { - putchar('\n'); - p->outcol = 0; - } - - if (0 == p->outmacro) - putchar('.'); - else - putchar(' '); - - if (EOF != (rc = fputs(s, stdout))) - p->outcol += rc; - - putchar(' '); - p->outcol++; - p->outmacro++; - p->seenws = 0; -} - -/* - * Put a stadnalone mdoc(7) command with the trailing newline. - */ -static void -teximacro(struct texi *p, const char *s) -{ - - if (p->ign) - return; - - if (p->outmacro) - texierr(p, "\"%s\" in open line scope!?", s); - if (p->literal) - texierr(p, "\"%s\" in a literal scope!?", s); - - if (p->outcol) - putchar('\n'); - - putchar('.'); - puts(s); - p->outcol = p->seenws = 0; -} - -static void -texivspace(struct texi *p) -{ - - if (p->seenvs) - return; - teximacro(p, "Pp"); - p->seenvs = 1; -} - -/* - * Advance by a single byte in the input stream. - */ -static void -advance(struct texi *p, const char *buf, size_t *pos) -{ - - if ('\n' == buf[*pos]) { - p->files[p->filepos - 1].line++; - p->files[p->filepos - 1].col = 0; - } else - p->files[p->filepos - 1].col++; - - (*pos)++; -} - -static void -texipunctuate(struct texi *p, const char *buf, size_t sz, size_t *pos) -{ size_t start, end; + char *cp; - if (1 != p->outmacro) - return; + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); - for (start = end = *pos; end < sz; end++) { - switch (buf[end]) { - case (','): - case (')'): - case ('.'): - case ('"'): - case (':'): - case ('!'): - case ('?'): - continue; - default: - break; - } - break; - } - if (end == *pos) - return; - if (end + 1 == sz || ' ' == buf[end] || '\n' == buf[end]) { - for ( ; start < end; start++) { - texiputchar(p, ' '); - texiputchar(p, buf[start]); - advance(p, buf, pos); - } - } -} + start = end = *pos; + while (end < BUFSZ(p) && ! ismspace(BUF(p)[end])) + end++; -/* - * Advance to the next non-whitespace word in the input stream. - * If we're in literal mode, then print all of the whitespace as we're - * doing so. - */ -static size_t -advancenext(struct texi *p, const char *buf, size_t sz, size_t *pos) -{ - - if (p->literal) { - while (*pos < sz && ismspace(buf[*pos])) { - if (*pos && '\n' == buf[*pos] && - '\\' == buf[*pos - 1]) - texiputchar(p, 'e'); - texiputchar(p, buf[*pos]); - advance(p, buf, pos); - } - return(*pos); - } - - while (*pos < sz && ismspace(buf[*pos])) { - p->seenws = 1; - /* - * If it looks like we've printed a double-line, then - * output a paragraph. - * FIXME: this is stupid. - */ - if (*pos && '\n' == buf[*pos] && '\n' == buf[*pos - 1]) - texivspace(p); - advance(p, buf, pos); - } - return(*pos); -} - -/* - * Advance to the EOLN in the input stream. - */ -static size_t -advanceeoln(struct texi *p, const char *buf, - size_t sz, size_t *pos, int consumenl) -{ - - /* FIXME: disregards @NEWLINE. */ - while (*pos < sz && '\n' != buf[*pos]) - advance(p, buf, pos); - if (*pos < sz && consumenl) - advance(p, buf, pos); - return(*pos); -} - -/* - * Advance to position "end", which is an absolute position in the - * current buffer greater than or equal to the current position. - */ -static void -advanceto(struct texi *p, const char *buf, size_t *pos, size_t end) -{ - - assert(*pos <= end); - while (*pos < end) - advance(p, buf, pos); -} - -/* - * Output a free-form word in the input stream, progressing to the next - * command or white-space. - * This also will advance the input stream. - */ -static void -texiword(struct texi *p, const char *buf, - size_t sz, size_t *pos, char extra) -{ - - if (p->seenws && 0 == p->outmacro && p->outcol > 72 && 0 == p->literal) - texiputchar(p, '\n'); - /* FIXME: abstract this: we use it elsewhere. */ - if (p->seenws && p->outcol && 0 == p->literal) - texiputchar(p, ' '); - - p->seenws = 0; - - while (*pos < sz && ! ismspace(buf[*pos])) { - switch (buf[*pos]) { - case ('@'): - case ('}'): - case ('{'): - return; - } - if ('\0' != extra && buf[*pos] == extra) - return; - if (*pos < sz - 1 && - '`' == buf[*pos] && - '`' == buf[*pos + 1]) { - texiputchars(p, "\\(lq"); - advance(p, buf, pos); - } else if (*pos < sz - 1 && - '\'' == buf[*pos] && - '\'' == buf[*pos + 1]) { - texiputchars(p, "\\(rq"); - advance(p, buf, pos); - } else - texiputchar(p, buf[*pos]); - advance(p, buf, pos); - } -} - -/* - * Look up the command at position "pos" in the buffer, returning it (or - * TEXICMD__MAX if none found) and setting "end" to be the absolute - * index after the command name. - */ -static enum texicmd -texicmd(struct texi *p, const char *buf, - size_t pos, size_t sz, size_t *end) -{ - size_t i, len; - - assert('@' == buf[pos]); - - if ((*end = pos) == sz) - return(TEXICMD__MAX); - else if ((*end = ++pos) == sz) - return(TEXICMD__MAX); - - /* Alphabetic commands are special. */ - if ( ! isalpha(buf[pos])) { - if ((*end = pos + 1) == sz) - return(TEXICMD__MAX); - for (i = 0; i < TEXICMD__MAX; i++) { - if (1 != texitoks[i].len) - continue; - if (0 == strncmp(texitoks[i].tok, &buf[pos], 1)) - return(i); - } - texiwarn(p, "bad command: @%c", buf[pos]); - return(TEXICMD__MAX); - } - - for (*end = pos; *end < sz && ! ismspace(buf[*end]); (*end)++) - if ((*end > pos && ('@' == buf[*end] || - '{' == buf[*end] || '}' == buf[*end]))) - break; - - len = *end - pos; - for (i = 0; i < TEXICMD__MAX; i++) { - if (len != texitoks[i].len) - continue; - if (0 == strncmp(texitoks[i].tok, &buf[pos], len)) - return(i); - } - - texiwarn(p, "bad command: @%.*s", (int)len, &buf[pos]); - return(TEXICMD__MAX); -} - -/* - * Parse an argument from a bracketed command, e.g., @url{foo, baz}. - * Num should be set to the argument we're currently parsing, although - * it suffixes for it to be zero or non-zero. - * This will return 1 if there are more arguments, 0 otherwise. - * This will stop (returning 0) in the event of EOF or if we're not at a - * bracket for the zeroth parse. - */ -static int -parsearg(struct texi *p, const char *buf, - size_t sz, size_t *pos, size_t num) -{ - size_t end; - enum texicmd cmd; - - while (*pos < sz && ismspace(buf[*pos])) - advance(p, buf, pos); - if (*pos == sz || (0 == num && '{' != buf[*pos])) - return(0); - if (0 == num) - advance(p, buf, pos); - - while ((*pos = advancenext(p, buf, sz, pos)) < sz) { - switch (buf[*pos]) { - case (','): - advance(p, buf, pos); - return(1); - case ('}'): - advance(p, buf, pos); - return(0); - case ('{'): - if (0 == p->ign) - texiwarn(p, "unexpected \"{\""); - advance(p, buf, pos); - continue; - case ('@'): - break; - default: - texiword(p, buf, sz, pos, ','); - continue; - } - - cmd = texicmd(p, buf, *pos, sz, &end); - advanceto(p, buf, pos, end); - if (TEXICMD__MAX == cmd) - continue; - if (NULL != texitoks[cmd].fp) - (*texitoks[cmd].fp)(p, cmd, buf, sz, pos); - } - return(0); -} - -/* - * Parse until the end of a bracketed statement, e.g., @foo{bar baz}. - * This will stop in the event of EOF or if we're not at a bracket. - */ -static void -parsebracket(struct texi *p, const char *buf, size_t sz, size_t *pos) -{ - size_t end; - enum texicmd cmd; - - while (*pos < sz && ismspace(buf[*pos])) - advance(p, buf, pos); - - if (*pos == sz || '{' != buf[*pos]) + if (start == end) { + advanceeoln(p, pos, 1); return; - advance(p, buf, pos); + } else if (NULL == (cp = malloc(end - start + 1))) + texiabort(p, NULL); - while ((*pos = advancenext(p, buf, sz, pos)) < sz) { - switch (buf[*pos]) { - case ('}'): - advance(p, buf, pos); - return; - case ('{'): - if (0 == p->ign) - texiwarn(p, "unexpected \"{\""); - advance(p, buf, pos); - continue; - case ('@'): - break; - default: - texiword(p, buf, sz, pos, '\0'); - continue; - } + memcpy(cp, &BUF(p)[start], end - start); + cp[end - start] = '\0'; - cmd = texicmd(p, buf, *pos, sz, &end); - advanceto(p, buf, pos, end); - if (TEXICMD__MAX == cmd) - continue; - if (NULL != texitoks[cmd].fp) - (*texitoks[cmd].fp)(p, cmd, buf, sz, pos); - } -} + p->indexs = realloc(p->indexs, + sizeof(char *) * (p->indexsz + 1)); -/* - * This should be invoked when we're on a macro line and want to process - * to the end of the current input line, doing all of our macros along - * the way. - */ -static void -parseeoln(struct texi *p, const char *buf, size_t sz, size_t *pos) -{ - size_t end; - enum texicmd cmd; - - while (*pos < sz && '\n' != buf[*pos]) { - while (*pos < sz && isws(buf[*pos])) { - p->seenws = 1; - if (p->literal) - texiputchar(p, buf[*pos]); - advance(p, buf, pos); - } - switch (buf[*pos]) { - case ('}'): - if (0 == p->ign) - texiwarn(p, "unexpected \"}\""); - advance(p, buf, pos); - continue; - case ('{'): - if (0 == p->ign) - texiwarn(p, "unexpected \"{\""); - advance(p, buf, pos); - continue; - case ('@'): - break; - default: - texiword(p, buf, sz, pos, '\0'); - continue; - } - - cmd = texicmd(p, buf, *pos, sz, &end); - advanceto(p, buf, pos, end); - if (TEXICMD__MAX == cmd) - continue; - if (NULL != texitoks[cmd].fp) - (*texitoks[cmd].fp)(p, cmd, buf, sz, pos); - } + if (NULL == p->indexs) + texiabort(p, NULL); + p->indexs[p->indexsz++] = cp; } -/* - * Parse a single word or command. - * This will return immediately at the EOF. - */ static void -parsesingle(struct texi *p, const char *buf, size_t sz, size_t *pos) +dodefn(struct texi *p, enum texicmd cmd, size_t *pos) { - size_t end; - enum texicmd cmd; - - if ((*pos = advancenext(p, buf, sz, pos)) >= sz) - return; - - switch (buf[*pos]) { - case ('}'): - if (0 == p->ign) - texiwarn(p, "unexpected \"}\""); - advance(p, buf, pos); - return; - case ('{'): - if (0 == p->ign) - texiwarn(p, "unexpected \"{\""); - advance(p, buf, pos); - return; - case ('@'): - break; - default: - texiword(p, buf, sz, pos, '\0'); - return; - } - - cmd = texicmd(p, buf, *pos, sz, &end); - advanceto(p, buf, pos, end); - if (TEXICMD__MAX == cmd) - return; - if (NULL != texitoks[cmd].fp) - (*texitoks[cmd].fp)(p, cmd, buf, sz, pos); -} - -static int -parselinearg(struct texi *p, const char *buf, size_t sz, size_t *pos) -{ - - while (*pos < sz && isws(buf[*pos])) { - p->seenws = 1; - advance(p, buf, pos); - } - - if (*pos < sz && '{' == buf[*pos]) - parsebracket(p, buf, sz, pos); - else if ('\n' != buf[*pos]) - parsesingle(p, buf, sz, pos); - else - return(0); - - return(1); -} - -/* - * Parse til the end of the buffer. - */ -static void -parseeof(struct texi *p, const char *buf, size_t sz) -{ - size_t pos; - - for (pos = 0; pos < sz; ) - parsesingle(p, buf, sz, &pos); -} - -/* - * Parse a block sequence until we have the "@end endtoken" command - * invocation. - * This will return immediately at EOF. - */ -static void -parseto(struct texi *p, const char *buf, - size_t sz, size_t *pos, const char *endtoken) -{ - size_t end; - enum texicmd cmd; - size_t endtoksz; - - endtoksz = strlen(endtoken); - assert(endtoksz > 0); - - while ((*pos = advancenext(p, buf, sz, pos)) < sz) { - switch (buf[*pos]) { - case ('}'): - if (0 == p->ign) - texiwarn(p, "unexpected \"}\""); - advance(p, buf, pos); - continue; - case ('{'): - if (0 == p->ign) - texiwarn(p, "unexpected \"{\""); - advance(p, buf, pos); - continue; - case ('@'): - break; - default: - texiword(p, buf, sz, pos, '\0'); - continue; - } - - cmd = texicmd(p, buf, *pos, sz, &end); - advanceto(p, buf, pos, end); - if (TEXICMD_END == cmd) { - while (*pos < sz && isws(buf[*pos])) - advance(p, buf, pos); - /* - * FIXME: check the full word, not just its - * initial substring! - */ - if (sz - *pos >= endtoksz && 0 == strncmp - (&buf[*pos], endtoken, endtoksz)) { - advanceeoln(p, buf, sz, pos, 0); - break; - } - if (0 == p->ign) - texiwarn(p, "unexpected \"end\""); - advanceeoln(p, buf, sz, pos, 0); - continue; - } else if (TEXICMD__MAX != cmd) - if (NULL != texitoks[cmd].fp) - (*texitoks[cmd].fp)(p, cmd, buf, sz, pos); - } -} - -/* - * Memory-map the file "fname" and begin parsing it. - * This can be called in a nested context. - */ -static void -parsefile(struct texi *p, const char *fname, int parse) -{ - struct texifile *f; - int fd; - struct stat st; - size_t i; - - assert(p->filepos < 64); - f = &p->files[p->filepos]; - memset(f, 0, sizeof(struct texifile)); - - f->name = fname; - if (-1 == (fd = open(fname, O_RDONLY, 0))) { - texiabort(p, fname); - } else if (-1 == fstat(fd, &st)) { - close(fd); - texiabort(p, fname); - } - - f->mapsz = st.st_size; - f->map = mmap(NULL, f->mapsz, - PROT_READ, MAP_SHARED, fd, 0); - close(fd); - - if (MAP_FAILED == f->map) - texiabort(p, fname); - - p->filepos++; - if ( ! parse) { - /* - * We're printing verbatim output. - * Make sure it doesn't get interpreted as mdoc by - * escaping escapes and making sure leading dots don't - * trigger mdoc(7) expansion. - */ - for (i = 0; i < f->mapsz; i++) { - if (i > 0 && '.' == f->map[i]) - if ('\n' == f->map[i - 1]) - fputs("\\&", stdout); - putchar(f->map[i]); - if ('\\' == f->map[i]) - putchar('e'); - } - } else - parseeof(p, f->map, f->mapsz); - texifilepop(p); -} - -static void -dodefn(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, 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): @@ -1229,8 +397,8 @@ 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; } @@ -1238,6 +406,10 @@ dodefn(struct texi *p, enum texicmd cmd, texivspace(p); switch (cmd) { + case (TEXICMD_DEFTYPEMETHOD): + case (TEXICMD_DEFTYPEMETHODX): + texiputchars(p, "Method"); + break; case (TEXICMD_DEFMAC): case (TEXICMD_DEFMACX): texiputchars(p, "Macro"); @@ -1255,17 +427,18 @@ dodefn(struct texi *p, enum texicmd cmd, texiputchars(p, "Function"); break; default: - parselinearg(p, buf, sz, pos); + parselinearg(p, pos); break; } - texiputchars(p, ":\n"); + 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; @@ -1274,10 +447,10 @@ 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"); @@ -1286,14 +459,16 @@ 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"); @@ -1305,7 +480,7 @@ 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; @@ -1314,7 +489,7 @@ 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; @@ -1322,36 +497,214 @@ dodefn(struct texi *p, enum texicmd cmd, abort(); } - texivspace(p); - if (NULL != blk) - parseto(p, buf, sz, pos, blk); + 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; + } + teximacro(p, "Bd -filled -offset indent"); + p->seenvs = 1; + parseto(p, pos, blk); + teximacro(p, "Ed"); } static void -doignblock(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) { - - p->ign++; - parseto(p, buf, sz, pos, texitoks[cmd].tok); - p->ign--; + size_t start, end, endtoksz, len; + struct teximacro m; + const char *endtok, *blk; + + memset(&m, 0, sizeof(struct teximacro)); + + while (*pos < BUFSZ(p) && isws(BUF(p)[*pos])) + advance(p, pos); + + 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, pos, end); + + m.key = malloc(end - start + 1); + if (NULL == m.key) + texiabort(p, NULL); + memcpy(m.key, &BUF(p)[start], end - start); + m.key[end - start] = '\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 = "@end macro\n"; + endtoksz = strlen(endtok); + blk = memmem(&BUF(p)[start], BUFSZ(p) - start, endtok, endtoksz); + if (NULL == blk) + texierr(p, "unterminated macro body"); + /* 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(p)[start], len); + m.value[len] = '\0'; + + p->macros = realloc + (p->macros, + (p->macrosz + 1) * + sizeof(struct teximacro)); + if (NULL == p->macros) + texiabort(p, NULL); + + p->macros[p->macrosz++] = m; + advanceeoln(p, pos, 1); } static void -doblock(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], start[32]; + const char *endt, *startt; + size_t esz, ssz, newpos, stack; + + /* + * 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. + */ + 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; + + /* + * 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. + */ + 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; + } + + 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, 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; + const char *macro = NULL; switch (cmd) { + case (TEXICMD_CODE): + case (TEXICMD_KBD): + case (TEXICMD_SAMP): + case (TEXICMD_T): + macro = "Li"; + break; + case (TEXICMD_CITE): + case (TEXICMD_DFN): + case (TEXICMD_EMPH): + case (TEXICMD_I): + case (TEXICMD_SLANTED): + macro = "Em"; + break; + case (TEXICMD_B): + case (TEXICMD_STRONG): + macro = "Sy"; + break; case (TEXICMD_COMMAND): macro = "Xr"; break; @@ -1368,43 +721,133 @@ doinline(struct texi *p, enum texicmd cmd, macro = "Va"; break; default: - abort(); + break; } - if (p->literal) { - parsebracket(p, buf, sz, pos); + if (NULL == macro || p->literal || TEXILIST_TABLE == p->list) { + parsebracket(p, pos, 0); return; } teximacroopen(p, macro); p->seenws = 0; - parsebracket(p, buf, sz, pos); - texipunctuate(p, buf, sz, pos); + parsebracket(p, pos, 0); + texipunctuate(p, pos); teximacroclose(p); } static void -doverbinclude(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 fname[PATH_MAX], path[PATH_MAX]; - int rc; - size_t i; + char delim; + size_t start; - while (*pos < sz && ' ' == buf[*pos]) - advance(p, buf, pos); + 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; - /* Read in the filename. */ - for (i = 0; *pos < sz && '\n' != buf[*pos]; i++) { + 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 +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, "unexpected end of file"); + return; + } + + term = memmem(&BUF(p)[*pos], BUFSZ(p) - *pos, end, endsz); + if (NULL == term) { + texiwarn(p, "unterminated verbatim block"); + endpos = BUFSZ(p); + } else + endpos = *pos + (size_t)(term - &BUF(p)[*pos]); + + assert(endpos <= BUFSZ(p)); + assert('\n' == BUF(p)[*pos]); + advance(p, pos); + teximacro(p, "Bd -literal -offset indent"); + while (*pos < endpos) { + texiputchar(p, BUF(p)[*pos]); + advance(p, pos); + } + teximacro(p, "Ed"); + if (*pos < BUFSZ(p)) + advanceto(p, pos, endpos + endsz); +} + +static void +doverbinclude(struct texi *p, enum texicmd cmd, size_t *pos) +{ + char fname[PATH_MAX], path[PATH_MAX]; + int rc; + size_t i, end; + const char *v; + enum texicmd type; + + while (*pos < BUFSZ(p) && ' ' == BUF(p)[*pos]) + advance(p, pos); + + for (i = 0; *pos < BUFSZ(p) && '\n' != BUF(p)[*pos]; ) { if (i == sizeof(fname) - 1) break; - fname[i] = buf[*pos]; - advance(p, buf, pos); + if ('@' != BUF(p)[*pos]) { + fname[i++] = BUF(p)[*pos]; + advance(p, pos); + continue; + } + type = texicmd(p, *pos, &end, NULL); + advanceto(p, pos, end); + if (TEXICMD_VALUE != type) + texierr(p, "unknown verbatiminclude command"); + v = valueblookup(p, pos); + if (NULL == v) + continue; + while ('\0' != *v) { + if (i == sizeof(fname) - 1) + break; + fname[i++] = *v++; + } + if ('\0' != *v) + break; } 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"); @@ -1424,27 +867,45 @@ 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; - int rc; + char fname[PATH_MAX], path[PATH_MAX]; + size_t i, end; + int rc; + 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]; i++) { + for (i = 0; *pos < BUFSZ(p) && '\n' != BUF(p)[*pos]; ) { if (i == sizeof(fname) - 1) break; - fname[i] = buf[*pos]; - advance(p, buf, pos); + if ('@' != BUF(p)[*pos]) { + fname[i++] = BUF(p)[*pos]; + advance(p, pos); + continue; + } + type = texicmd(p, *pos, &end, NULL); + advanceto(p, pos, end); + if (TEXICMD_VALUE != type) + texierr(p, "unknown include command"); + v = valueblookup(p, pos); + if (NULL == v) + continue; + while ('\0' != *v) { + if (i == sizeof(fname) - 1) + break; + fname[i++] = *v++; + } + if ('\0' != *v) + break; } 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"); @@ -1471,57 +932,18 @@ doinclude(struct texi *p, enum texicmd cmd, } static void -dofont(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) { - const char *font; - switch (cmd) { - case (TEXICMD_B): - case (TEXICMD_STRONG): - font = "\\fB"; - break; - case (TEXICMD_CITE): - case (TEXICMD_DFN): - case (TEXICMD_EMPH): - case (TEXICMD_I): - case (TEXICMD_SLANTED): - font = "\\fI"; - break; - case (TEXICMD_CODE): - case (TEXICMD_KBD): - case (TEXICMD_R): - case (TEXICMD_SAMP): - case (TEXICMD_SANSSERIF): - case (TEXICMD_T): - font = "\\fR"; - break; - default: - abort(); - } - - if (p->seenws) { - texiputchar(p, ' '); - p->seenws = 0; - } - texiputchars(p, font); - parsebracket(p, buf, sz, pos); - texiputchars(p, "\\fP"); + parsebracket(p, pos, 0); } static void -dobracket(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) { - parsebracket(p, buf, sz, pos); -} + advanceeoln(p, pos, 1); -static void -dodisplay(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) -{ - switch (cmd) { case (TEXICMD_FORMAT): case (TEXICMD_SMALLFORMAT): @@ -1533,29 +955,25 @@ dodisplay(struct texi *p, enum texicmd cmd, } p->seenvs = 1; - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, buf, sz, pos, 1); - parseto(p, buf, sz, pos, texitoks[cmd].tok); + parseto(p, pos, texitoks[cmd].tok); teximacro(p, "Ed"); } 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) { + advanceeoln(p, pos, 1); + teximacro(p, "Bd -literal -offset indent"); - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, buf, sz, pos, 1); p->literal++; - parseto(p, buf, sz, pos, texitoks[cmd].tok); + parseto(p, pos, texitoks[cmd].tok); p->literal--; teximacro(p, "Ed"); } 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); @@ -1563,87 +981,147 @@ 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++; + 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); - memcpy(p->subtitle, &buf[start], end - start); - p->subtitle[end - start] = '\0'; + if (NULL == (p->subtitle = malloc(*pos - start + 1))) + texiabort(p, NULL); + 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'): @@ -1651,28 +1129,64 @@ 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): @@ -1688,27 +1202,129 @@ dosymbol(struct texi *p, enum texicmd cmd, 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): + texiputchars(p, "\\(=="); + break; 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; @@ -1718,67 +1334,109 @@ 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; default: - texiwarn(p, "sym: %d", 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 void -domath(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +domath(struct texi *p, enum texicmd cmd, size_t *pos) { - size_t nest; - /* - * 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]) - 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) - return; - assert('}' == buf[*pos]); - advance(p, buf, pos); + parsebracket(p, pos, 1); } static void -dolink(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +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 < 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, pos, end); + + key = malloc(end - start + 1); + if (NULL == key) + texiabort(p, NULL); + memcpy(key, &BUF(p)[start], end - start); + key[end - start] = '\0'; + + 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, pos, 1); + + val = malloc(end - start + 1); + if (NULL == val) + texiabort(p, NULL); + 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, pos))) + texisplice(p, cp, strlen(cp), *pos); + else + texiputchars(p, "{No value}"); + } else if (TEXICMD_IFCLEAR == cmd) { + if (NULL != valuellookup(p, pos)) + doignblock(p, cmd, pos); + else + parseto(p, pos, texitoks[cmd].tok); + } else if (TEXICMD_IFSET == cmd) { + if (NULL == valuellookup(p, pos)) + doignblock(p, cmd, pos); + else + parseto(p, pos, texitoks[cmd].tok); + } else if (TEXICMD_CLEAR == cmd) + valuelclear(p, pos); +} + +static void +dolink(struct texi *p, enum texicmd cmd, size_t *pos) +{ int c; switch (cmd) { @@ -1787,147 +1445,179 @@ 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--; } +/* + * Sections can be made subsections and so on by way of the + * @raiseections and @lowersections commands. + * Perform this check here and return the actual section number adjusted + * to the raise level. + */ +static int +sectioner(struct texi *p, int sec) +{ + + if ((sec -= p->secoffs) < 0) { + texiwarn(p, "section below minimum, clamping"); + return(0); + } else if (sec >= SECTSZ) { + texiwarn(p, "section above maximum, clamping"); + return(SECTSZ - 1); + } + return(sec); +} + static void -dosubsection(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dosubsubsection(struct texi *p, enum texicmd cmd, size_t *pos) { + int sec; - if (p->outmacro) - texierr(p, "\"Em\" in open line scope!?"); - else if (p->literal) - texierr(p, "\"Em\" in a literal scope!?"); + sec = sectioner(p, 3); + /* We don't have a subsubsubsection, so make one up. */ texivspace(p); - teximacroopen(p, "Em"); - parseeoln(p, buf, sz, pos); + teximacroopen(p, sects[sec]); + parseeoln(p, pos); teximacroclose(p); texivspace(p); } static void -dosection(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dosubsection(struct texi *p, enum texicmd cmd, size_t *pos) { - const char *blk; + 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. */ + if (sec > 1) + texivspace(p); + teximacroopen(p, sects[sec]); + parseeoln(p, pos); + teximacroclose(p); + if (sec > 1) + texivspace(p); +} + +static void +dosecoffs(struct texi *p, enum texicmd cmd, size_t *pos) +{ + + if (TEXICMD_RAISESECTIONS == cmd) + p->secoffs++; + else + p->secoffs--; +} + +static void +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): - blk = "Sh"; + sec = sectioner(p, 0); break; case (TEXICMD_APPENDIXSEC): case (TEXICMD_HEADING): case (TEXICMD_SECTION): case (TEXICMD_UNNUMBEREDSEC): - blk = "Ss"; + sec = sectioner(p, 1); break; default: abort(); } if (p->outmacro) - texierr(p, "%s in open line scope!?", blk); + texierr(p, "\"%s\" in open line scope!?", sects[sec]); else if (p->literal) - texierr(p, "%s in a literal scope!?", blk); + texierr(p, "\"%s\" in a literal scope!?", sects[sec]); - teximacroopen(p, blk); - parseeoln(p, buf, sz, pos); + if (0 == sec && NULL != p->chapters) { + teximdocclose(p, 0); + teximdocopen(p, pos); + } + + teximacroopen(p, sects[sec]); + parseeoln(p, pos); teximacroclose(p); p->seenvs = 1; } 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]; - /* - * 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)); - - p->ign--; - 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). */ @@ -1955,7 +1645,7 @@ 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); @@ -1964,8 +1654,7 @@ 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. */ @@ -1974,41 +1663,46 @@ dotab(struct texi *p, enum texicmd cmd, } 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; 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); - advanceto(p, buf, pos, end); + type = texicmd(p, *pos, &end, NULL); + advanceto(p, pos, end); if (TEXICMD_COLUMNFRACTIONS != type) - texierr(p, "unknown multitable type"); - 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]) + texierr(p, "unknown multitable command"); + 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++; } @@ -2019,7 +1713,7 @@ 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--; @@ -2031,120 +1725,124 @@ domultitable(struct texi *p, enum texicmd cmd, } texiputchars(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; } 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); + parseto(p, pos, texitoks[cmd].tok); teximacro(p, "El"); 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"); + parseto(p, pos, texitoks[cmd].tok); teximacro(p, "El"); 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; - p->list = TEXILIST_ITEM; + 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"); + parseto(p, pos, texitoks[cmd].tok); teximacro(p, "El"); 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) +doignline(struct texi *p, enum texicmd cmd, size_t *pos) { - /* FIXME: ignore and parseeoln. */ - advanceeoln(p, buf, sz, pos, 1); + 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 ** -parsedirs(const char *base, const char *cp, size_t *sz) +parsedirs(struct texi *p, const char *base, const char *cp, size_t *sz) { 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++; - dirs = calloc(*sz, sizeof(char *)); - if (NULL == dirs) { - perror(NULL); - exit(EXIT_FAILURE); - } else if (NULL == (dirs[0] = strdup(base))) { - perror(NULL); - exit(EXIT_FAILURE); - } - + if (0 == *sz) + return(NULL); + if (NULL == (dirs = calloc(*sz, sizeof(char *)))) + texiabort(p, NULL); + if (NULL != base && NULL == (dirs[i++] = strdup(base))) + texiabort(p, NULL); if (NULL == cp) return(dirs); - - if (NULL == (tofree = tok = str = strdup(cp))) { - perror(NULL); - exit(EXIT_FAILURE); - } + if (NULL == (tofree = tok = str = strdup(cp))) + texiabort(p, NULL); - for (i = 1; NULL != (tok = strsep(&str, ":")); i++) - if (NULL == (dirs[i] = strdup(tok))) { - perror(NULL); - exit(EXIT_FAILURE); - } + for ( ; NULL != (tok = strsep(&str, ":")); i++) + if (NULL == (dirs[i] = strdup(tok))) + texiabort(p, NULL); free(tofree); return(dirs); @@ -2155,7 +1853,7 @@ 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], '/'); @@ -2165,10 +1863,15 @@ main(int argc, char *argv[]) ++progname; memset(&texi, 0, sizeof(struct texi)); + texi.ign = 1; + texi.outfile = stdout; 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; @@ -2177,37 +1880,33 @@ main(int argc, char *argv[]) } argv += optind; - if (0 == (argc -= optind)) - goto usage; + argc -= optind; - if (NULL == (path = strdup(argv[0]))) { - perror(NULL); - exit(EXIT_FAILURE); - } else if (NULL == (dir = dirname(path))) { - perror(argv[0]); - free(path); - exit(EXIT_FAILURE); + 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); } - free(path); - if (NULL != (cp = strrchr(argv[0], '/'))) - texi.title = strdup(cp + 1); - else - texi.title = strdup(argv[0]); - - if (NULL == texi.title) { - perror(NULL); - exit(EXIT_FAILURE); - } else if (NULL != (path = strchr(texi.title, '.'))) - *path = '\0'; - - texi.ign = 1; - texi.dirs = parsedirs(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); }