=================================================================== RCS file: /cvs/texi2mdoc/main.c,v retrieving revision 1.1.1.1 retrieving revision 1.33 diff -u -p -r1.1.1.1 -r1.33 --- texi2mdoc/main.c 2015/02/16 22:24:43 1.1.1.1 +++ texi2mdoc/main.c 2015/02/23 12:39:59 1.33 @@ -1,4 +1,4 @@ -/* $Id: main.c,v 1.1.1.1 2015/02/16 22:24:43 kristaps Exp $ */ +/* $Id: main.c,v 1.33 2015/02/23 12:39:59 kristaps Exp $ */ /* * Copyright (c) 2015 Kristaps Dzonsons * @@ -21,917 +21,1497 @@ #include #include #include +#include +#include #include #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_A4PAPER, - TEXICMD_ANCHOR, - TEXICMD_AT, - TEXICMD_BYE, - TEXICMD_CHAPTER, - TEXICMD_CINDEX, - TEXICMD_CODE, - TEXICMD_COMMAND, - TEXICMD_COMMENT, - TEXICMD_CONTENTS, - TEXICMD_COPYING, - TEXICMD_COPYRIGHT, - TEXICMD_DETAILMENU, - TEXICMD_DIRCATEGORY, - TEXICMD_DIRENTRY, - TEXICMD_EMAIL, - TEXICMD_EMPH, - TEXICMD_END, - TEXICMD_EXAMPLE, - TEXICMD_FILE, - TEXICMD_I, - TEXICMD_IFHTML, - TEXICMD_IFNOTTEX, - TEXICMD_IFTEX, - TEXICMD_IMAGE, - TEXICMD_ITEM, - TEXICMD_ITEMIZE, - TEXICMD_KBD, - TEXICMD_LATEX, - TEXICMD_MENU, - TEXICMD_NODE, - TEXICMD_QUOTATION, - TEXICMD_PARINDENT, - TEXICMD_REF, - TEXICMD_SAMP, - TEXICMD_SECTION, - TEXICMD_SETCHAPNEWPAGE, - TEXICMD_SETFILENAME, - TEXICMD_SETTITLE, - TEXICMD_SUBSECTION, - TEXICMD_TABLE, - TEXICMD_TEX, - TEXICMD_TEXSYM, - TEXICMD_TITLEFONT, - TEXICMD_TITLEPAGE, - TEXICMD_TOP, - TEXICMD_UNNUMBERED, - TEXICMD_URL, - TEXICMD_VAR, - 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; - -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) */ -}; - -/* - * The main parse structure. - * This keeps any necessary information handy. - */ -struct texi { - struct texifile files[64]; - size_t filepos; - unsigned flags; -#define TEXI_IGN 0x01 /* don't print anything */ -#define TEXI_HEADER (TEXI_IGN | 0x02) /* haven't seen @top yet */ -#define TEXI_LITERAL 0x04 /* output all whitespace */ - size_t outcol; /* column of output */ - int outmacro; /* whether output is in line macro */ - int seenws; /* whitespace has been ignored */ -}; - -#define ismpunct(_x) \ - ('.' == (_x) || \ - ',' == (_x) || \ - ';' == (_x)) - -static void doarg1(struct texi *, enum texicmd, const char *, size_t, size_t *); +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 docommand(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doemph(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 dofile(struct texi *, enum texicmd, const char *, size_t, size_t *); -static void doifnottex(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 doitalic(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 doliteral(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 dosh(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 doverbatim(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 */ + { doignargn, "acronym", 7 }, /* TEXICMD_ACRONYM */ + { doaccent, "'", 1 }, /* TEXICMD_ACUTE */ { doignline, "afourpaper", 10 }, /* TEXICMD_A4PAPER */ { 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 */ + { doinline, "b", 1 }, /* TEXICMD_BOLD */ + { dosymbol, "!", 1 }, /* TEXICMD_BANG */ + { dosymbol, "bullet", 6 }, /* TEXICMD_BULLET */ { dobye, "bye", 3 }, /* TEXICMD_BYE */ - { dosh, "chapter", 7 }, /* TEXICMD_CHAPTER */ + { doignline, "center", 6 }, /* TEXICMD_CENTER */ + { dosection, "chapter", 7 }, /* TEXICMD_CHAPTER */ { doignline, "cindex", 6 }, /* TEXICMD_CINDEX */ - { doliteral, "code", 4 }, /* TEXICMD_CODE */ - { docommand, "command", 7 }, /* TEXICMD_COMMAND */ + { 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 */ + { 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 */ + { dodefn, "deftpx", 6 }, /* TEXICMD_DEFTPX */ + { dodefn, "deftypefn", 9 }, /* TEXICMD_DEFTYPEFN */ + { dodefn, "deftypefnx", 10 }, /* TEXICMD_DEFTYPEFNX */ + { dodefn, "deftypefun", 10 }, /* TEXICMD_DEFTYPEFUN */ + { dodefn, "deftypefunx", 11 }, /* TEXICMD_DEFTYPEFUNX */ + { dodefn, "deftypevar", 10 }, /* TEXICMD_DEFTYPEVAR */ + { dodefn, "deftypevarx", 11 }, /* TEXICMD_DEFTYPEVARX */ + { dodefn, "deftypevr", 9 }, /* TEXICMD_DEFTYPEVR */ + { dodefn, "deftypevrx", 10 }, /* TEXICMD_DEFTYPEVRX */ + { dodefn, "defun", 5 }, /* TEXICMD_DEFUN */ + { dodefn, "defunx", 6 }, /* TEXICMD_DEFUNX */ + { dodefn, "defvar", 6 }, /* TEXICMD_DEFVAR */ + { dodefn, "defvarx", 7 }, /* TEXICMD_DEFVARX */ + { dodefn, "defvr", 5 }, /* TEXICMD_DEFVR */ + { dodefn, "defvrx", 6 }, /* TEXICMD_DEFVRX */ { doignblock, "detailmenu", 10 }, /* TEXICMD_DETAILMENU */ + { doinline, "dfn", 3 }, /* TEXICMD_DFN */ { doignline, "dircategory", 11 }, /* TEXICMD_DIRCATEGORY */ { doignblock, "direntry", 8 }, /* TEXICMD_DIRENTRY */ - { doarg1, "email", 5 }, /* TEXICMD_EMAIL */ - { doemph, "emph", 4 }, /* TEXICMD_EMPH */ + { dodisplay, "display", 7 }, /* TEXICMD_DISPLAY */ + { doignbracket, "dmn", 3 }, /* TEXICMD_DMN */ + { dosymbol, "dots", 4 }, /* TEXICMD_DOTS */ + { dolink, "email", 5 }, /* TEXICMD_EMAIL */ + { doinline, "emph", 4 }, /* TEXICMD_EMPH */ { NULL, "end", 3 }, /* TEXICMD_END */ + { doenumerate, "enumerate", 9 }, /* TEXICMD_ENUMERATE */ + { doinline, "env", 3 }, /* TEXICMD_ENV */ + { dosymbol, "equiv", 5 }, /* TEXICMD_EQUIV */ + { dosymbol, "error", 5 }, /* TEXICMD_ERROR */ { doexample, "example", 7 }, /* TEXICMD_EXAMPLE */ - { dofile, "file", 4 }, /* TEXICMD_FILE */ - { doitalic, "i", 1 }, /* TEXICMD_I */ + { 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 */ + { doignbracket, "footnote", 8 }, /* TEXICMD_FOOTNOTE */ + { dotable, "ftable", 6 }, /* TEXICMD_FTABLE */ + { dodisplay, "format", 6 }, /* TEXICMD_FORMAT */ + { doaccent, "`", 1 }, /* TEXICMD_GRAVE */ + { doblock, "group", 5 }, /* TEXICMD_GROUP */ + { dosection, "heading", 7 }, /* TEXICMD_HEADING */ + { doignline, "headings", 8 }, /* TEXICMD_HEADINGS */ + { doitem, "headitem", 8 }, /* TEXICMD_HEADITEM */ + { doignblock, "html", 4 }, /* TEXICMD_HTML */ + { dosymbol, "-", 1 }, /* TEXICMD_HYPHEN */ + { doinline, "i", 1 }, /* TEXICMD_I */ + { dovalue, "ifclear", 7 }, /* TEXICMD_IFCLEAR */ + { doignblock, "ifdocbook", 9 }, /* TEXICMD_IFDOCBOOK */ { doignblock, "ifhtml", 6 }, /* TEXICMD_IFHTML */ - { doifnottex, "ifnottex", 8 }, /* TEXICMD_IFNOTTEX */ + { doblock, "ifinfo", 6 }, /* TEXICMD_IFINFO */ + { doblock, "ifnotdocbook", 12 }, /* TEXICMD_IFNOTDOCBOOK */ + { doblock, "ifnothtml", 9 }, /* TEXICMD_IFNOTHTML */ + { doblock, "ifnotinfo", 9 }, /* TEXICMD_IFNOTINFO */ + { doignblock, "ifnotplaintext", 14 }, /* TEXICMD_IFNOTPLAINTEXT */ + { doblock, "ifnottex", 8 }, /* TEXICMD_IFNOTTEX */ + { doblock, "ifnotxml", 8 }, /* TEXICMD_IFNOTXML */ + { doblock, "ifplaintext", 11 }, /* TEXICMD_IFPLAINTEXT */ { doignblock, "iftex", 5 }, /* TEXICMD_IFTEX */ + { 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 */ + { doignline, "", 0 }, /* TEXICMD_USER_INDEX */ + { doignline, "insertcopying", 13 }, /* TEXICMD_INSERTCOPYING */ { doitem, "item", 4 }, /* TEXICMD_ITEM */ { doitemize, "itemize", 7 }, /* TEXICMD_ITEMIZE */ - { doliteral, "kbd", 3 }, /* TEXICMD_KBD */ + { doitem, "itemx", 5 }, /* TEXICMD_ITEMX */ + { doinline, "kbd", 3 }, /* TEXICMD_KBD */ + { dobracket, "key", 3 }, /* TEXICMD_KEY */ + { doignline, "kindex", 6 }, /* TEXICMD_KINDEX */ { dosymbol, "LaTeX", 5 }, /* TEXICMD_LATEX */ + { dosecoffs, "lowersections", 13 }, /* TEXICMD_LOWERSECTIONS */ + { domacro, "macro", 5 }, /* TEXICMD_MACRO */ + { 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 */ + { doinline, "option", 6 }, /* TEXICMD_OPTION */ + { dolink, "pxref", 5 }, /* TEXICMD_PXREF */ + { dosymbol, "?", 1 }, /* TEXICMD_QUESTIONMARK */ { doquotation, "quotation", 9 }, /* TEXICMD_QUOTATION */ - { doignline, "paragraphindent", 14 }, /* TEXICMD_PARINDENT */ + { doignline, "page", 4 }, /* TEXICMD_PAGE */ + { doignline, "paragraphindent", 15 }, /* TEXICMD_PARINDENT */ + { dosymbol, ".", 1 }, /* TEXICMD_PERIOD */ + { doignline, "pindex", 6 }, /* TEXICMD_PINDEX */ + { doignline, "printindex", 10 }, /* TEXICMD_PRINTINDEX */ + { doinline, "r", 1 }, /* TEXICMD_R */ + { dosecoffs, "raisesections", 13 }, /* TEXICMD_RAISESECTIONS */ { dobracket, "ref", 3 }, /* TEXICMD_REF */ - { doliteral, "samp", 4 }, /* TEXICMD_SAMP */ + { doignline, "refill", 6 }, /* TEXICMD_REFILL */ + { dosymbol, "result", 6 }, /* TEXICMD_RESULT */ + { doinline, "samp", 4 }, /* TEXICMD_SAMP */ + { doinline, "sansserif", 9 }, /* TEXICMD_SANSSERIF */ + { dobracket, "sc", 2 }, /* TEXICMD_SC */ { dosection, "section", 7 }, /* TEXICMD_SECTION */ + { dovalue, "set", 3 }, /* TEXICMD_SET */ { doignline, "setchapternewpage", 17 }, /* TEXICMD_SETCHAPNEWPAGE */ { doignline, "setfilename", 11 }, /* TEXICMD_SETFILENAME */ - { doignline, "settitle", 8 }, /* TEXICMD_SETTITLE */ + { dotitle, "settitle", 8 }, /* TEXICMD_SETTITLE */ + { doignline, "shortcontents", 13 }, /* TEXICMD_SHORTCONTENTS */ + { doinline, "slanted", 7 }, /* TEXICMD_SLANTED */ + { dosp, "sp", 2 }, /* TEXICMD_SP */ + { dosymbol, " ", 1 }, /* TEXICMD_SPACE */ + { doignline, "smallbook", 9 }, /* TEXICMD_SMALLBOOK */ + { dodisplay, "smalldisplay", 12 }, /* TEXICMD_SMALLDISPLAY */ + { doexample, "smallexample", 12 }, /* TEXICMD_SMALLEXAMPLE */ + { dodisplay, "smallformat", 11 }, /* TEXICMD_SMALLFORMAT */ + { dodisplay, "smallindentblock", 16 }, /* TEXICMD_SMALLINDENTBLOCK */ + { dosymbol, "{", 1 }, /* TEXICMD_SQUIGGLE_LEFT */ + { dosymbol, "}", 1 }, /* TEXICMD_SQUIGGLE_RIGHT */ + { 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, "summarycontents", 15 }, /* TEXICMD_SUMMARYCONTENTS */ + { doignline, "synindex", 8 }, /* TEXICMD_SYNINDEX */ + { doignline, "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 */ + { 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 */ - { dosh, "unnumbered", 10 }, /* TEXICMD_UNNUMBERED */ - { doarg1, "url", 3 }, /* TEXICMD_URL */ - { doliteral, "var", 3 }, /* TEXICMD_VAR */ + { 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 */ + { 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 */ + { dotable, "vtable", 6 }, /* TEXICMD_VTABLE */ + { dobracket, "w", 1 }, /* TEXICMD_W */ + { dolink, "xref", 4 }, /* TEXICMD_XREF */ + /* TEXICMD__END */ }; +const struct texitok *const texitoks = __texitoks; + static void -texifilepop(struct texi *p) +dodefindex(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { - struct texifile *f; + size_t start, end; + char *cp; - assert(p->filepos > 0); - f = &p->files[--p->filepos]; - munmap(f->map, f->mapsz); -} + while (*pos < sz && isws(buf[*pos])) + advance(p, buf, pos); -static void -texiexit(struct texi *p) -{ + start = end = *pos; + while (end < sz && ! ismspace(buf[end])) + end++; - while (p->filepos > 0) - texifilepop(p); + if (start == end) { + advanceeoln(p, buf, sz, pos, 1); + 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; } static void -texifatal(struct texi *p, const char *errstring) +dodefn(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { + const char *blk; - perror(errstring); - texiexit(p); - exit(EXIT_FAILURE); + blk = NULL; + switch (cmd) { + case (TEXICMD_DEFFN): + case (TEXICMD_DEFTP): + case (TEXICMD_DEFTYPEFN): + case (TEXICMD_DEFTYPEFUN): + case (TEXICMD_DEFTYPEVAR): + case (TEXICMD_DEFTYPEVR): + case (TEXICMD_DEFUN): + case (TEXICMD_DEFVAR): + case (TEXICMD_DEFVR): + blk = texitoks[cmd].tok; + break; + default: + break; + } + + if (p->ign) { + NULL != blk ? + parseto(p, buf, sz, pos, blk) : + parseeoln(p, buf, sz, pos); + return; + } + + if (NULL != blk) + texivspace(p); + + switch (cmd) { + case (TEXICMD_DEFMAC): + case (TEXICMD_DEFMACX): + texiputchars(p, "Macro"); + break; + case (TEXICMD_DEFTYPEVAR): + case (TEXICMD_DEFTYPEVARX): + case (TEXICMD_DEFVAR): + case (TEXICMD_DEFVARX): + texiputchars(p, "Variable"); + break; + case (TEXICMD_DEFTYPEFUN): + case (TEXICMD_DEFTYPEFUNX): + case (TEXICMD_DEFUN): + case (TEXICMD_DEFUNX): + texiputchars(p, "Function"); + break; + default: + parselinearg(p, buf, sz, pos); + break; + } + + texiputchars(p, ":\n"); + + switch (cmd) { + case (TEXICMD_DEFMAC): + case (TEXICMD_DEFMACX): + teximacroopen(p, "Dv"); + while (parselinearg(p, buf, sz, pos)) + /* Spin. */ ; + teximacroclose(p); + break; + case (TEXICMD_DEFFN): + case (TEXICMD_DEFFNX): + case (TEXICMD_DEFUN): + case (TEXICMD_DEFUNX): + teximacroopen(p, "Fo"); + parselinearg(p, buf, sz, pos); + teximacroclose(p); + teximacroopen(p, "Fa"); + while (parselinearg(p, buf, sz, pos)) + /* Spin. */ ; + teximacroclose(p); + teximacro(p, "Fc"); + break; + case (TEXICMD_DEFTYPEFUN): + case (TEXICMD_DEFTYPEFUNX): + case (TEXICMD_DEFTYPEFN): + case (TEXICMD_DEFTYPEFNX): + teximacroopen(p, "Ft"); + parselinearg(p, buf, sz, pos); + teximacroclose(p); + teximacroopen(p, "Fo"); + parselinearg(p, buf, sz, pos); + teximacroclose(p); + teximacroopen(p, "Fa"); + while (parselinearg(p, buf, sz, pos)) + /* Spin. */ ; + teximacroclose(p); + teximacro(p, "Fc"); + break; + case (TEXICMD_DEFTP): + case (TEXICMD_DEFTPX): + case (TEXICMD_DEFTYPEVAR): + case (TEXICMD_DEFTYPEVARX): + case (TEXICMD_DEFTYPEVR): + case (TEXICMD_DEFTYPEVRX): + teximacroopen(p, "Vt"); + while (parselinearg(p, buf, sz, pos)) + /* Spin. */ ; + teximacroclose(p); + break; + case (TEXICMD_DEFVAR): + case (TEXICMD_DEFVARX): + case (TEXICMD_DEFVR): + case (TEXICMD_DEFVRX): + teximacroopen(p, "Va"); + while (parselinearg(p, buf, sz, pos)) + /* Spin. */ ; + teximacroclose(p); + break; + default: + abort(); + } + + texivspace(p); + if (NULL != blk) + parseto(p, buf, sz, pos, blk); } -/* - * 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, ...) +domacro(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { - va_list ap; + size_t start, end, endtoksz, len; + struct teximacro m; + const char *endtok, *blk; - fprintf(stderr, "%s:%zu:%zu: ", - 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); + memset(&m, 0, sizeof(struct teximacro)); + + while (*pos < sz && isws(buf[*pos])) + advance(p, buf, pos); + + for (start = end = *pos; end < sz; end++) + if (ismspace(buf[end]) || '{' == buf[end]) + break; + + if (start == end) + texierr(p, "zero-length macro name"); + + advanceto(p, buf, pos, end); + + m.key = malloc(end - start + 1); + if (NULL == m.key) + texiabort(p, NULL); + memcpy(m.key, &buf[start], end - start); + m.key[end - start] = '\0'; + + m.args = argparse(p, buf, sz, pos, &m.argsz, 0); + advanceeoln(p, buf, sz, pos, 0); + + start = *pos; + endtok = "\n@end macro\n"; + endtoksz = strlen(endtok); + blk = memmem(&buf[start], sz, 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]; + m.value = malloc(len + 1); + if (NULL == m.value) + texiabort(p, NULL); + memcpy(m.value, &buf[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, buf, sz, pos, 1); } -/* - * Put a single data character. - * This MUST NOT be a mdoc(7) command: it should be free text that's - * outputted to the screen. - */ static void -texiputchar(struct texi *p, char c) +doignblock(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { + char end[32]; + const char *term; + size_t endsz, endpos; - if (TEXI_IGN & p->flags) - return; + /* + * We want to completely ignore everything in these blocks, so + * simply jump to the @end block. + */ + endsz = snprintf(end, sizeof(end), + "\n@end %s\n", texitoks[cmd].tok); + assert(endsz < sizeof(end)); - putchar(c); - if ('\n' == c) { - p->outcol = 0; - p->outmacro = 0; - p->seenws = 0; - } else - p->outcol++; + /* + * 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. + */ + 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); + + /* Only do this if we're not already at the end. */ + if (endpos < sz) + advanceto(p, buf, pos, endpos + endsz); } -/* - * Put multiple characters (see texiputchar()). - */ static void -texiputchars(struct texi *p, const char *s) +doblock(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { - - while ('\0' != *s) - texiputchar(p, *s++); + + parseto(p, buf, sz, pos, texitoks[cmd].tok); } -/* - * Put an mdoc(7) command without the trailing newline. - * This should ONLY be used for mdoc(7) commands! - */ static void -texifputs(struct texi *p, const char *s) +doinline(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { - int rc; + const char *macro = NULL; - if (TEXI_IGN & p->flags) + 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; + case (TEXICMD_ENV): + macro = "Ev"; + break; + case (TEXICMD_FILE): + macro = "Pa"; + break; + case (TEXICMD_OPTION): + macro = "Op"; + break; + case (TEXICMD_VAR): + macro = "Va"; + break; + default: + break; + } + + if (NULL == macro || p->literal || TEXILIST_TABLE == p->list) { + parsebracket(p, buf, sz, pos); return; - if (p->outcol) - texiputchar(p, '\n'); - if (EOF != (rc = fputs(s, stdout))) - p->outcol += rc; + } + + teximacroopen(p, macro); + p->seenws = 0; + parsebracket(p, buf, sz, pos); + texipunctuate(p, buf, sz, pos); + teximacroclose(p); } -/* - * Put an mdoc(7) command with the trailing newline. - * This should ONLY be used for mdoc(7) commands! - */ static void -teximacro(struct texi *p, const char *s) +doverb(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { + char delim; - if (TEXI_IGN & p->flags) + while (*pos < sz && isws(buf[*pos])) + advance(p, buf, pos); + if (*pos == sz || '{' != buf[*pos]) return; - if (p->outcol) - texiputchar(p, '\n'); - puts(s); - p->outcol = 0; + advance(p, buf, pos); + if (*pos == sz) + return; + + delim = buf[*pos]; + advance(p, buf, pos); + /* Make sure we flush out our initial whitespace... */ + if (p->seenws && p->outcol && 0 == p->literal) + texiputchar(p, ' '); p->seenws = 0; + /* Read until we see the delimiter then end-brace. */ + while (*pos < sz - 1) { + if (buf[*pos] == delim && buf[*pos + 1] == '}') + break; + texiputchar(p, buf[*pos]); + advance(p, buf, pos); + } + if (*pos == sz - 1) + return; + /* Make sure we read after the end-brace. */ + assert(delim == buf[*pos]); + advance(p, buf, pos); + assert('}' == buf[*pos]); + advance(p, buf, pos); } -/* - * Advance by a single byte in the input stream. - */ static void -advance(struct texi *p, const char *buf, size_t *pos) +doverbatim(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { + const char *end, *term; + size_t endsz, endpos; - if ('\n' == buf[*pos]) { - p->files[p->filepos - 1].line++; - p->files[p->filepos - 1].col = 0; - } else - p->files[p->filepos - 1].col++; + advanceeoln(p, buf, sz, pos, 1); - (*pos)++; -} + /* We end at exactly this token. */ + end = "\n@end verbatim\n"; + endsz = strlen(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 (TEXI_LITERAL & p->flags) { - while (*pos < sz && isspace(buf[*pos])) { - texiputchar(p, buf[*pos]); - advance(p, buf, pos); - } - return(*pos); - } + /* + * 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]; - while (*pos < sz && isspace(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]) - teximacro(p, ".Pp"); + teximacro(p, "Bd -literal -offset indent"); + assert(endpos <= sz); + 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); } - return(*pos); + teximacro(p, "Ed"); + advanceto(p, buf, pos, endpos + endsz); } -/* - * Advance to the EOLN in the input stream. - */ -static size_t -advanceeoln(struct texi *p, const char *buf, size_t sz, size_t *pos) +static void +doverbinclude(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { + char fname[PATH_MAX], path[PATH_MAX]; + int rc; + size_t i, end; + const char *v; + enum texicmd type; - while (*pos < sz && '\n' != buf[*pos]) + while (*pos < sz && ' ' == buf[*pos]) advance(p, buf, pos); - return(*pos); + + for (i = 0; *pos < sz && '\n' != buf[*pos]; ) { + if (i == sizeof(fname) - 1) + break; + if ('@' != buf[*pos]) { + fname[i++] = buf[*pos]; + advance(p, buf, pos); + continue; + } + type = texicmd(p, buf, *pos, sz, &end, NULL); + advanceto(p, buf, pos, end); + if (TEXICMD_VALUE != type) + texierr(p, "unknown verbatiminclude command"); + v = valueblookup(p, buf, sz, 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]) + texierr(p, "path too long"); + else if ('/' == fname[0]) + texierr(p, "no absolute paths"); + fname[i] = '\0'; + + if (strstr(fname, "../") || strstr(fname, "/..")) + texierr(p, "insecure path"); + + rc = snprintf(path, sizeof(path), + "%s/%s", p->dirs[0], fname); + if (rc < 0) + texierr(p, "couldn't format path"); + else if ((size_t)rc >= sizeof(path)) + texierr(p, "path too long"); + + parsefile(p, path, 0); } -/* - * 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) +doinclude(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { + char fname[PATH_MAX], path[PATH_MAX]; + size_t i, end; + int rc; + const char *v; + enum texicmd type; - assert(*pos <= end); - while (*pos < end) + while (*pos < sz && ' ' == buf[*pos]) 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) -{ + /* Read in the filename. */ + for (i = 0; *pos < sz && '\n' != buf[*pos]; ) { + if (i == sizeof(fname) - 1) + break; + if ('@' != buf[*pos]) { + fname[i++] = buf[*pos]; + advance(p, buf, pos); + continue; + } + type = texicmd(p, buf, *pos, sz, &end, NULL); + advanceto(p, buf, pos, end); + if (TEXICMD_VALUE != type) + texierr(p, "unknown include command"); + v = valueblookup(p, buf, sz, pos); + if (NULL == v) + continue; + while ('\0' != *v) { + if (i == sizeof(fname) - 1) + break; + fname[i++] = *v++; + } + if ('\0' != *v) + break; + } - /* - * XXX: if we're in literal mode, then we shouldn't do any - * reflowing of text here. - */ - if (p->outcol > 72 && ! (TEXI_LITERAL & p->flags)) - texiputchar(p, '\n'); + if (i == 0) + texierr(p, "path too short"); + else if ('\n' != buf[*pos]) + texierr(p, "path too long"); + else if ('/' == fname[0]) + texierr(p, "no absolute paths"); + fname[i] = '\0'; - if (p->seenws && p->outcol && ! (TEXI_LITERAL & p->flags)) - texiputchar(p, ' '); + if (strstr(fname, "../") || strstr(fname, "/..")) + texierr(p, "insecure path"); - p->seenws = 0; + for (i = 0; i < p->dirsz; i++) { + rc = snprintf(path, sizeof(path), + "%s/%s", p->dirs[i], fname); + if (rc < 0) + texierr(p, "couldn't format path"); + else if ((size_t)rc >= sizeof(path)) + texierr(p, "path too long"); + else if (-1 == access(path, R_OK)) + continue; - while (*pos < sz && ! isspace(buf[*pos])) { - switch (buf[*pos]) { - case ('@'): - case ('}'): - case ('{'): - 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); + parsefile(p, path, 1); + return; } + + texierr(p, "couldn't find %s in includes", fname); } -static enum texicmd -texicmd(struct texi *p, const char *buf, - size_t pos, size_t sz, size_t *end) +static void +dobracket(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { - size_t i, len; - assert('@' == buf[pos]); - for (*end = ++pos; *end < sz && ! isspace(buf[*end]); (*end)++) - if ('@' == buf[*end] || '{' == buf[*end]) - break; + parsebracket(p, buf, sz, pos); +} - 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); +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): + teximacro(p, "Bd -filled"); + break; + default: + teximacro(p, "Bd -filled -offset indent"); + break; } - texiwarn(p, "bad command: %.*s", (int)len, &buf[pos]); - return(TEXICMD__MAX); + p->seenvs = 1; + /* FIXME: ignore and parseeoln. */ + advanceeoln(p, buf, sz, pos, 1); + parseto(p, buf, sz, pos, texitoks[cmd].tok); + teximacro(p, "Ed"); } static void -parseeof(struct texi *p, const char *buf, size_t sz) +doexample(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { - size_t pos = 0; - enum texicmd cmd; - size_t end; - while ((pos = advancenext(p, buf, sz, &pos)) < sz) { - switch (buf[pos]) { - case ('}'): - texiwarn(p, "unexpected \"}\""); - advance(p, buf, &pos); - continue; - case ('{'): - 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); - } + 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); + p->literal--; + teximacro(p, "Ed"); } static void -parsebracket(struct texi *p, const char *buf, size_t sz, size_t *pos) +dobye(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { - size_t end; - enum texicmd cmd; - if (*pos == sz || '{' != buf[*pos]) - return; - advance(p, buf, pos); + texiexit(p); + exit(EXIT_SUCCESS); +} - while ((*pos = advancenext(p, buf, sz, pos)) < sz) { - switch (buf[*pos]) { - case ('}'): - advance(p, buf, pos); - return; - case ('{'): - texiwarn(p, "unexpected \"{\""); - advance(p, buf, pos); - continue; - case ('@'): - break; - default: - texiword(p, buf, sz, pos); - continue; - } +static void +dotitle(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) +{ + size_t start, end; - 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); - } + 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); + free(p->subtitle); + p->subtitle = malloc(end - start + 1); + if (NULL == p->subtitle) + texiabort(p, NULL); + memcpy(p->subtitle, &buf[start], end - start); + p->subtitle[end - start] = '\0'; } static void -parseto(struct texi *p, const char *buf, - size_t sz, size_t *pos, const char *endtoken) +doaccent(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { - size_t end; - enum texicmd cmd; - size_t endtoksz; - endtoksz = strlen(endtoken); - assert(endtoksz > 0); - - while ((*pos = advancenext(p, buf, sz, pos)) < sz) { + if (*pos == sz) + return; + advance(p, buf, pos); + switch (cmd) { + case (TEXICMD_ACUTE): switch (buf[*pos]) { - case ('}'): - texiwarn(p, "unexpected \"}\""); - advance(p, buf, pos); - continue; - case ('{'): - texiwarn(p, "unexpected \"{\""); - advance(p, buf, pos); - continue; - case ('@'): + 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: - texiword(p, buf, sz, pos); - continue; + texiputchar(p, buf[*pos]); } - - cmd = texicmd(p, buf, *pos, sz, &end); - advanceto(p, buf, pos, end); - if (TEXICMD_END == cmd) { - while (*pos < sz && ' ' == buf[*pos]) - advance(p, buf, pos); - /* - * FIXME: skip tabs and also check the full - * word, not just its initial substring! - */ - if (sz - *pos >= endtoksz && 0 == strncmp - (&buf[*pos], endtoken, endtoksz)) { - advanceeoln(p, buf, sz, pos); - break; - } - texiwarn(p, "unexpected \"end\""); - advanceeoln(p, buf, sz, pos); - continue; - } else if (TEXICMD__MAX != cmd) - if (NULL != texitoks[cmd].fp) - (*texitoks[cmd].fp)(p, cmd, buf, sz, pos); + break; + case (TEXICMD_CIRCUMFLEX): + switch (buf[*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]); + } + break; + case (TEXICMD_GRAVE): + switch (buf[*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]); + } + break; + case (TEXICMD_TILDE): + switch (buf[*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]); + } + break; + case (TEXICMD_UMLAUT): + switch (buf[*pos]) { + case ('a'): case ('A'): + case ('e'): case ('E'): + case ('i'): case ('I'): + case ('o'): case ('O'): + case ('u'): case ('U'): + case ('y'): + texiputchars(p, "\\(:"); + texiputchar(p, buf[*pos]); + break; + default: + texiputchar(p, buf[*pos]); + } + break; + default: + abort(); } } static void -doignblock(struct texi *p, enum texicmd cmd, +dosymbol(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { - unsigned int sv = p->flags; - const char *blockname; - - p->flags |= TEXI_IGN; + + if (p->seenws && p->outcol && 0 == p->literal) { + texiputchar(p, ' '); + p->seenws = 0; + } + switch (cmd) { - case (TEXICMD_COPYING): - blockname = "copying"; + case (TEXICMD_ASTERISK): + case (TEXICMD_NEWLINE): + case (TEXICMD_SPACE): + case (TEXICMD_TABSYM): + texiputchar(p, ' '); break; - case (TEXICMD_DETAILMENU): - blockname = "detailmenu"; + case (TEXICMD_AT): + texiputchar(p, '@'); break; - case (TEXICMD_DIRENTRY): - blockname = "direntry"; + case (TEXICMD_BANG): + texiputchar(p, '!'); break; - case (TEXICMD_IFHTML): - blockname = "ifhtml"; + case (TEXICMD_BULLET): + texiputchars(p, "\\(bu"); break; - case (TEXICMD_IFTEX): - blockname = "iftex"; + case (TEXICMD_COPYRIGHT): + texiputchars(p, "\\(co"); break; - case (TEXICMD_MENU): - blockname = "menu"; + case (TEXICMD_DOTS): + texiputchars(p, "..."); break; - case (TEXICMD_TEX): - blockname = "tex"; + case (TEXICMD_EQUIV): + texiputchars(p, "\\(=="); break; - case (TEXICMD_TITLEPAGE): - blockname = "titlepage"; + case (TEXICMD_ERROR): + texiputchars(p, "error\\(->"); break; + case (TEXICMD_EXPANSION): + texiputchars(p, "\\(->"); + break; + case (TEXICMD_LATEX): + texiputchars(p, "LaTeX"); + break; + case (TEXICMD_MINUS): + texiputchars(p, "\\-"); + break; + case (TEXICMD_PERIOD): + texiputchar(p, '.'); + break; + case (TEXICMD_QUESTIONMARK): + texiputchar(p, '?'); + break; + case (TEXICMD_RESULT): + texiputchars(p, "\\(rA"); + break; + case (TEXICMD_SQUIGGLE_LEFT): + texiputchars(p, "{"); + break; + case (TEXICMD_SQUIGGLE_RIGHT): + texiputchars(p, "}"); + break; + case (TEXICMD_TEXSYM): + texiputchars(p, "TeX"); + break; + case (TEXICMD_COLON): + case (TEXICMD_HYPHEN): + break; default: abort(); } - parseto(p, buf, sz, pos, blockname); - p->flags = sv; + + if (texitoks[cmd].len > 1) + doignbracket(p, cmd, buf, sz, pos); } static void -doifnottex(struct texi *p, enum texicmd cmd, +doquotation(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { - parseto(p, buf, sz, pos, "ifnottex"); + teximacro(p, "Qo"); + parseto(p, buf, sz, pos, "quotation"); + teximacro(p, "Qc"); } static void -doinline(struct texi *p, const char *buf, - size_t sz, size_t *pos, const char *macro) +domath(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { + size_t nest; - if ( ! p->outmacro) - texifputs(p, "."); - texiputchars(p, macro); - texiputchar(p, ' '); - p->seenws = 0; - p->outmacro++; - parsebracket(p, buf, sz, pos); - p->outmacro--; - if (*pos < sz - 1 && - ismpunct(buf[*pos]) && - isspace(buf[*pos + 1])) { + /* + * 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 ( ! p->outmacro) - texiputchar(p, '\n'); + if (*pos == sz) + return; + assert('}' == buf[*pos]); + advance(p, buf, pos); } static void -doitalic(struct texi *p, enum texicmd cmd, +dovalue(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { + size_t start, end, i; + char *key, *val; + const char *cp; - texiputchars(p, "\\fI"); - parsebracket(p, buf, sz, pos); - texiputchars(p, "\\fP"); -} + 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])) + break; + /* We don't allow empty keys. */ + if (start == end) + return; + advanceto(p, buf, pos, end); -static void -doliteral(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) -{ + key = malloc(end - start + 1); + if (NULL == key) + texiabort(p, NULL); + memcpy(key, &buf[start], end - start); + key[end - start] = '\0'; - if (TEXI_LITERAL & p->flags) - parsebracket(p, buf, sz, pos); - else - doinline(p, buf, sz, pos, "Li"); + while (*pos < sz && isws(buf[*pos])) + advance(p, buf, pos); + for (start = end = *pos; end < sz; end++) + if ('\n' == buf[end]) + break; + /* We do allow empty values. */ + advanceeoln(p, buf, sz, pos, 1); + + val = malloc(end - start + 1); + if (NULL == val) + texiabort(p, NULL); + memcpy(val, &buf[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; + parsemembuf(p, cp, strlen(cp)); + p->valstackpos--; + } else + texiputchars(p, "{No value}"); + } else if (TEXICMD_IFCLEAR == cmd) { + if (NULL != valuellookup(p, buf, sz, pos)) + doignblock(p, cmd, buf, sz, pos); + else + parseto(p, buf, sz, pos, texitoks[cmd].tok); + } else if (TEXICMD_IFSET == cmd) { + if (NULL == valuellookup(p, buf, sz, pos)) + doignblock(p, cmd, buf, sz, pos); + else + parseto(p, buf, sz, pos, texitoks[cmd].tok); + } else if (TEXICMD_CLEAR == cmd) + valuelclear(p, buf, sz, pos); } static void -doemph(struct texi *p, enum texicmd cmd, +dolink(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { + int c; - if (TEXI_LITERAL & p->flags) - doitalic(p, cmd, buf, sz, pos); - else - doinline(p, buf, sz, pos, "Em"); + switch (cmd) { + case (TEXICMD_EMAIL): + teximacroopen(p, "Mt"); + break; + case (TEXICMD_UREF): + case (TEXICMD_URL): + teximacroopen(p, "Lk"); + break; + case (TEXICMD_XREF): + texiputchars(p, "See Section"); + teximacroopen(p, "Qq"); + break; + case (TEXICMD_PXREF): + texiputchars(p, "see Section"); + teximacroopen(p, "Qq"); + break; + default: + abort(); + } + + c = parsearg(p, buf, sz, pos, 0); + p->ign++; + while (c > 0) + c = parsearg(p, buf, sz, pos, 1); + p->ign--; + + texipunctuate(p, buf, sz, pos); + teximacroclose(p); } static void -docommand(struct texi *p, enum texicmd cmd, +doignargn(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { + int c; - doinline(p, buf, sz, pos, "Xr"); + c = parsearg(p, buf, sz, pos, 0); + p->ign++; + while (c > 0) + c = parsearg(p, buf, sz, pos, 1); + p->ign--; } -static void -dobracket(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +/* + * 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) { - parsebracket(p, buf, sz, pos); + 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 -dofile(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dosubsubsection(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { + int sec; - if (TEXI_LITERAL & p->flags) - parsebracket(p, buf, sz, pos); - else - doinline(p, buf, sz, pos, "Pa"); + 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 -doexample(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dosubsection(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { - unsigned int sv; + int sec; - teximacro(p, ".Bd -literal"); - advanceeoln(p, buf, sz, pos); - if ('\n' == buf[*pos]) - advance(p, buf, pos); - sv = p->flags; - p->flags |= TEXI_LITERAL; - parseto(p, buf, sz, pos, "example"); - p->flags = sv; - teximacro(p, ".Ed"); + 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 -dobye(struct texi *p, enum texicmd cmd, +dosecoffs(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { - texiexit(p); - exit(EXIT_SUCCESS); + if (TEXICMD_RAISESECTIONS == cmd) + p->secoffs++; + else + p->secoffs--; } static void -dosymbol(struct texi *p, enum texicmd cmd, +dosection(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { + int sec; switch (cmd) { - case (TEXICMD_AT): - texiputchars(p, "@"); + case (TEXICMD_APPENDIX): + case (TEXICMD_CHAPTER): + case (TEXICMD_TOP): + case (TEXICMD_UNNUMBERED): + sec = sectioner(p, 0); break; - case (TEXICMD_COPYRIGHT): - texiputchars(p, "\\(co"); + case (TEXICMD_APPENDIXSEC): + case (TEXICMD_HEADING): + case (TEXICMD_SECTION): + case (TEXICMD_UNNUMBEREDSEC): + sec = sectioner(p, 1); break; - case (TEXICMD_LATEX): - texiputchars(p, "LaTeX"); - break; - case (TEXICMD_TEXSYM): - texiputchars(p, "TeX"); - break; default: abort(); } - doignbracket(p, cmd, buf, sz, pos); -} + 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]); -static void -doquotation(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) -{ - - teximacro(p, ".Qo"); - parseto(p, buf, sz, pos, "quotation"); - teximacro(p, ".Qc"); + teximacroopen(p, sects[sec]); + parseeoln(p, buf, sz, pos); + teximacroclose(p); + p->seenvs = 1; } static void -doarg1(struct texi *p, enum texicmd cmd, +dosp(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { - if (*pos == sz || '{' != buf[*pos]) - return; - advance(p, buf, pos); - if ( ! p->outmacro) - texifputs(p, "."); - switch (cmd) { - case (TEXICMD_EMAIL): - texiputchars(p, "Lk "); - break; - case (TEXICMD_URL): - texiputchars(p, "Mt "); - break; - default: - abort(); - } - while (*pos < sz && '}' != buf[*pos] && ',' != buf[*pos]) { - texiputchar(p, buf[*pos]); - advance(p, buf, pos); - } - while (*pos < sz && '}' != buf[*pos]) - advance(p, buf, pos); - if (*pos < sz) - advance(p, buf, pos); - if (*pos < sz - 1 && - ismpunct(buf[*pos]) && - isspace(buf[*pos + 1])) { - texiputchar(p, ' '); - texiputchar(p, buf[*pos]); - advance(p, buf, pos); - } - if ( ! p->outmacro) - texiputchar(p, '\n'); + texivspace(p); + /* FIXME: ignore and parseeoln. */ + advanceeoln(p, buf, sz, pos, 1); } static void -dosubsection(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) +dotop(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) { + const char *cp; + time_t t; + char date[32]; - if (TEXI_IGN & p->flags) { - advanceeoln(p, buf, sz, pos); - return; - } - while (*pos < sz && ' ' == buf[*pos]) - advance(p, buf, pos); - texifputs(p, ".Pp"); - while (*pos < sz && '\n' != buf[*pos]) { - texiputchar(p, buf[*pos]); - advance(p, buf, pos); - } - texifputs(p, ".Pp"); -} + if (--p->ign) + texierr(p, "@top command while ignoring (%d)", p->ign); -static void -dosection(struct texi *p, enum texicmd cmd, - const char *buf, size_t sz, size_t *pos) -{ + /* + * 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)); - if (TEXI_IGN & p->flags) { - advanceeoln(p, buf, sz, pos); - return; - } - while (*pos < sz && ' ' == buf[*pos]) - advance(p, buf, pos); - texifputs(p, ".Ss "); - while (*pos < sz && '\n' != buf[*pos]) { - texiputchar(p, buf[*pos]); - advance(p, buf, pos); - } - texiputchar(p, '\n'); + 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); } static void -dosh(struct texi *p, enum texicmd cmd, +doitem(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { - if (TEXI_IGN & p->flags) { - advanceeoln(p, buf, sz, pos); + /* Multitable is using raw tbl(7). */ + if (TEXILIST_TABLE == p->list) { + texiputchar(p, '\n'); return; + } + + if (p->outmacro) + texierr(p, "item in open line scope!?"); + else if (p->literal) + texierr(p, "item in a literal scope!?"); + + switch (p->list) { + case (TEXILIST_ITEM): + teximacroopen(p, "It"); + break; + case (TEXILIST_NOITEM): + teximacro(p, "It"); + break; + default: + texivspace(p); + break; } - while (*pos < sz && ' ' == buf[*pos]) - advance(p, buf, pos); - texifputs(p, ".Sh "); - while (*pos < sz && '\n' != buf[*pos]) { - texiputchar(p, toupper(buf[*pos])); - advance(p, buf, pos); - } - texiputchar(p, '\n'); + + /* Trick so we don't start with Pp. */ + p->seenvs = 1; + parseeoln(p, buf, sz, pos); + + if (TEXILIST_ITEM == p->list) + teximacroclose(p); + else if (p->outcol > 0) + texiputchar(p, '\n'); } static void -dotop(struct texi *p, enum texicmd cmd, +dotab(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { - p->flags &= ~TEXI_HEADER; - advanceeoln(p, buf, sz, pos); - teximacro(p, ".Dd $Mdocdate: February 16 2015 $"); - teximacro(p, ".Dt SOMETHING 7"); - teximacro(p, ".Os"); - teximacro(p, ".Sh NAME"); - teximacro(p, ".Nm Something"); - teximacro(p, ".Nd Something"); + /* This command is only useful in @multitable. */ + if (TEXILIST_TABLE == p->list) + texiputchar(p, '\t'); } static void -doitem(struct texi *p, enum texicmd cmd, +domultitable(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { - size_t end; + enum texilist sv = p->list; + int svliteral = p->literal; + enum texicmd type; + size_t i, end, columns; - /* See if we have arguments... */ - for (end = *pos; end < sz; end++) - if (' ' != buf[end] && '\t' != buf[end]) - break; + 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; - /* If we have arguments, print them too. */ - if ('\n' != buf[end]) { - texifputs(p, ".It"); - /* FIXME: process commands. */ + /* Advance to the first argument... */ + while (*pos < sz && isws(buf[*pos])) + advance(p, buf, pos); + + /* Make sure we don't print anything when scanning. */ + p->ign++; + if ('@' == buf[*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); + if (TEXICMD_COLUMNFRACTIONS != type) + texierr(p, "unknown multitable command"); while (*pos < sz && '\n' != buf[*pos]) { - texiputchar(p, buf[*pos]); - advance(p, buf, pos); + while (*pos < sz && isws(buf[*pos])) + advance(p, buf, pos); + while (*pos < sz && ! isws(buf[*pos])) { + if ('\n' == buf[*pos]) + break; + advance(p, buf, pos); + } + columns++; } - texiputchar(p, '\n'); - } else - teximacro(p, ".It"); + } else + /* + * We have arguments. + * We could parse these, but it's easier to just let + * tbl(7) figure it out. + * So use this only to count arguments. + */ + while (parselinearg(p, buf, sz, pos) > 0) + columns++; + p->ign--; + + /* Left-justify each table entry. */ + for (i = 0; i < columns; i++) { + if (i > 0) + texiputchar(p, ' '); + texiputchar(p, 'l'); + } + texiputchars(p, ".\n"); + p->outmacro++; + parseto(p, buf, sz, 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) { + enum texilist sv = p->list; - teximacro(p, ".Bl -tag -width Ds"); - parseto(p, buf, sz, pos, "table"); - teximacro(p, ".El"); + 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); + teximacro(p, "El"); + p->list = sv; } static void +doenumerate(struct texi *p, enum texicmd cmd, + const char *buf, size_t sz, size_t *pos) +{ + enum texilist sv = p->list; + + 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"); + teximacro(p, "El"); + p->list = sv; +} + +static void doitemize(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { + enum texilist sv = p->list; - teximacro(p, ".Bl -bullet"); + 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"); - teximacro(p, ".El"); + teximacro(p, "El"); + p->list = sv; } static void doignbracket(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { - unsigned int sv = p->flags; - p->flags |= TEXI_IGN; + p->ign++; parsebracket(p, buf, sz, pos); - p->flags = sv; + p->ign--; } static void @@ -939,52 +1519,55 @@ doignline(struct texi *p, enum texicmd cmd, const char *buf, size_t sz, size_t *pos) { - advanceeoln(p, buf, sz, pos); - if (*pos < sz) - advance(p, buf, pos); + /* FIXME: ignore and parseeoln. */ + advanceeoln(p, buf, sz, pos, 1); } -static int -parsefile(struct texi *p, const char *fname) +/* + * Parse colon-separated directories from "cp" (if not NULL) and returns + * the array of pointers. + * Prepends "base" to the array. + * This does NOT sanitise the directories! + */ +static char ** +parsedirs(struct texi *p, const char *base, const char *cp, size_t *sz) { - struct texifile *f; - int fd; - struct stat st; + char *tok, *str, *tofree; + const char *cpp; + size_t i; + char **dirs; - assert(p->filepos < 64); - f = &p->files[p->filepos]; - memset(f, 0, sizeof(struct texifile)); + *sz = NULL != (cpp = cp) ? 2 : 1; + if (*sz > 1) + for ( ; NULL != (cpp = strchr(cpp, ':')); (*sz)++) + cpp++; - f->name = fname; - if (-1 == (fd = open(fname, O_RDONLY, 0))) { - texifatal(p, fname); - } else if (-1 == fstat(fd, &st)) { - close(fd); - texifatal(p, fname); - } + if (NULL == (dirs = calloc(*sz, sizeof(char *)))) + texiabort(p, NULL); + else if (NULL == (dirs[0] = strdup(base))) + texiabort(p, NULL); - f->mapsz = st.st_size; - f->map = mmap(NULL, f->mapsz, - PROT_READ, MAP_SHARED, fd, 0); - close(fd); + if (NULL == cp) + return(dirs); + + if (NULL == (tofree = tok = str = strdup(cp))) + texiabort(p, NULL); - if (MAP_FAILED == f->map) { - texifatal(p, fname); - return(0); - } + for (i = 1; NULL != (tok = strsep(&str, ":")); i++) + if (NULL == (dirs[i] = strdup(tok))) + texiabort(p, NULL); - p->filepos++; - parseeof(p, f->map, f->mapsz); - texifilepop(p); - return(1); + free(tofree); + return(dirs); } int main(int argc, char *argv[]) { struct texi texi; - int c, rc; - const char *progname; + int c; + char *path, *dir; + const char *progname, *Idir, *cp; progname = strrchr(argv[0], '/'); if (progname == NULL) @@ -992,8 +1575,14 @@ main(int argc, char *argv[]) else ++progname; - while (-1 != (c = getopt(argc, argv, ""))) + memset(&texi, 0, sizeof(struct texi)); + Idir = NULL; + + while (-1 != (c = getopt(argc, argv, "I:"))) switch (c) { + case ('I'): + Idir = optarg; + break; default: goto usage; } @@ -1002,12 +1591,30 @@ main(int argc, char *argv[]) if (0 == (argc -= optind)) goto usage; - memset(&texi, 0, sizeof(struct texi)); - texi.flags = TEXI_HEADER; - rc = parsefile(&texi, argv[0]); - return(rc ? EXIT_SUCCESS : EXIT_FAILURE); + if (NULL == (path = strdup(argv[0]))) + texiabort(&texi, NULL); + else if (NULL == (dir = dirname(path))) + texiabort(&texi, NULL); + free(path); + + 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); usage: - fprintf(stderr, "usage: %s file\n", progname); + fprintf(stderr, "usage: %s [-Idirs] file\n", progname); return(EXIT_FAILURE); }