[BACK]Return to main.c CVS log [TXT][DIR] Up to [cvsweb.bsd.lv] / texi2mdoc

Annotation of texi2mdoc/main.c, Revision 1.14

1.14    ! kristaps    1: /*     $Id: main.c,v 1.13 2015/02/19 07:50:51 kristaps Exp $ */
1.1       kristaps    2: /*
                      3:  * Copyright (c) 2015 Kristaps Dzonsons <kristaps@bsd.lv>
                      4:  *
                      5:  * Permission to use, copy, modify, and distribute this software for any
                      6:  * purpose with or without fee is hereby granted, provided that the above
                      7:  * copyright notice and this permission notice appear in all copies.
                      8:  *
                      9:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
                     10:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
                     11:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
                     12:  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
                     13:  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
                     14:  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
                     15:  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
                     16:  */
                     17: #include <sys/mman.h>
                     18: #include <sys/stat.h>
                     19:
                     20: #include <assert.h>
                     21: #include <ctype.h>
                     22: #include <fcntl.h>
                     23: #include <getopt.h>
1.2       kristaps   24: #include <libgen.h>
                     25: #include <limits.h>
1.1       kristaps   26: #include <stdarg.h>
                     27: #include <stdio.h>
                     28: #include <stdlib.h>
                     29: #include <string.h>
1.10      kristaps   30: #include <time.h>
1.6       kristaps   31: #include <unistd.h>
1.1       kristaps   32:
                     33: /*
                     34:  * This defines each one of the Texinfo commands that we understand.
                     35:  * Obviously this only refers to native commands; overriden names are a
                     36:  * different story.
                     37:  */
                     38: enum   texicmd {
1.2       kristaps   39:        TEXICMD_ACRONYM,
1.1       kristaps   40:        TEXICMD_A4PAPER,
                     41:        TEXICMD_ANCHOR,
1.2       kristaps   42:        TEXICMD_APPENDIX,
                     43:        TEXICMD_APPENDIXSEC,
1.3       kristaps   44:        TEXICMD_ASTERISK,
1.1       kristaps   45:        TEXICMD_AT,
1.3       kristaps   46:        TEXICMD_AUTHOR,
1.12      kristaps   47:        TEXICMD_B,
1.3       kristaps   48:        TEXICMD_BANG,
1.7       kristaps   49:        TEXICMD_BULLET,
1.1       kristaps   50:        TEXICMD_BYE,
1.5       kristaps   51:        TEXICMD_CENTER,
1.1       kristaps   52:        TEXICMD_CHAPTER,
                     53:        TEXICMD_CINDEX,
1.3       kristaps   54:        TEXICMD_CITE,
1.1       kristaps   55:        TEXICMD_CODE,
1.3       kristaps   56:        TEXICMD_COLON,
1.1       kristaps   57:        TEXICMD_COMMAND,
                     58:        TEXICMD_COMMENT,
1.2       kristaps   59:        TEXICMD_COMMENT_LONG,
1.1       kristaps   60:        TEXICMD_CONTENTS,
                     61:        TEXICMD_COPYING,
                     62:        TEXICMD_COPYRIGHT,
1.3       kristaps   63:        TEXICMD_DEFTYPEFN,
                     64:        TEXICMD_DEFTYPEFNX,
                     65:        TEXICMD_DEFTYPEFUN,
                     66:        TEXICMD_DEFTYPEFUNX,
                     67:        TEXICMD_DEFTYPEVAR,
                     68:        TEXICMD_DEFTYPEVR,
1.1       kristaps   69:        TEXICMD_DETAILMENU,
1.3       kristaps   70:        TEXICMD_DFN,
1.1       kristaps   71:        TEXICMD_DIRCATEGORY,
                     72:        TEXICMD_DIRENTRY,
1.3       kristaps   73:        TEXICMD_DISPLAY,
1.2       kristaps   74:        TEXICMD_DOTS,
1.1       kristaps   75:        TEXICMD_EMAIL,
                     76:        TEXICMD_EMPH,
                     77:        TEXICMD_END,
1.2       kristaps   78:        TEXICMD_ENUMERATE,
1.3       kristaps   79:        TEXICMD_ENV,
1.1       kristaps   80:        TEXICMD_EXAMPLE,
                     81:        TEXICMD_FILE,
1.3       kristaps   82:        TEXICMD_GROUP,
1.2       kristaps   83:        TEXICMD_HEADING,
1.3       kristaps   84:        TEXICMD_HEADINGS,
                     85:        TEXICMD_HYPHEN,
1.1       kristaps   86:        TEXICMD_I,
1.3       kristaps   87:        TEXICMD_IFCLEAR,
1.14    ! kristaps   88:        TEXICMD_IFDOCBOOK,
1.1       kristaps   89:        TEXICMD_IFHTML,
1.3       kristaps   90:        TEXICMD_IFINFO,
1.14    ! kristaps   91:        TEXICMD_IFNOTDOCBOOK,
        !            92:        TEXICMD_IFNOTHTML,
        !            93:        TEXICMD_IFNOTINFO,
        !            94:        TEXICMD_IFNOTPLAINTEXT,
1.1       kristaps   95:        TEXICMD_IFNOTTEX,
1.14    ! kristaps   96:        TEXICMD_IFNOTXML,
        !            97:        TEXICMD_IFPLAINTEXT,
1.1       kristaps   98:        TEXICMD_IFTEX,
1.3       kristaps   99:        TEXICMD_IFSET,
1.14    ! kristaps  100:        TEXICMD_IFXML,
1.1       kristaps  101:        TEXICMD_IMAGE,
1.2       kristaps  102:        TEXICMD_INCLUDE,
1.13      kristaps  103:        TEXICMD_INDENTBLOCK,
1.5       kristaps  104:        TEXICMD_INSERTCOPYING,
1.1       kristaps  105:        TEXICMD_ITEM,
                    106:        TEXICMD_ITEMIZE,
                    107:        TEXICMD_KBD,
                    108:        TEXICMD_LATEX,
1.3       kristaps  109:        TEXICMD_MATH,
1.1       kristaps  110:        TEXICMD_MENU,
1.3       kristaps  111:        TEXICMD_NEWLINE,
1.1       kristaps  112:        TEXICMD_NODE,
1.3       kristaps  113:        TEXICMD_NOINDENT,
1.8       kristaps  114:        TEXICMD_PXREF,
1.3       kristaps  115:        TEXICMD_QUESTIONMARK,
1.1       kristaps  116:        TEXICMD_QUOTATION,
1.3       kristaps  117:        TEXICMD_PAGE,
1.1       kristaps  118:        TEXICMD_PARINDENT,
1.2       kristaps  119:        TEXICMD_PRINTINDEX,
1.12      kristaps  120:        TEXICMD_R,
1.1       kristaps  121:        TEXICMD_REF,
                    122:        TEXICMD_SAMP,
1.12      kristaps  123:        TEXICMD_SANSSERIF,
1.7       kristaps  124:        TEXICMD_SC,
1.1       kristaps  125:        TEXICMD_SECTION,
1.3       kristaps  126:        TEXICMD_SET,
1.1       kristaps  127:        TEXICMD_SETCHAPNEWPAGE,
                    128:        TEXICMD_SETFILENAME,
                    129:        TEXICMD_SETTITLE,
1.12      kristaps  130:        TEXICMD_SLANTED,
1.3       kristaps  131:        TEXICMD_SP,
                    132:        TEXICMD_SPACE,
1.12      kristaps  133:        TEXICMD_SMALLDISPLAY,
1.3       kristaps  134:        TEXICMD_SMALLEXAMPLE,
1.13      kristaps  135:        TEXICMD_SMALLINDENTBLOCK,
1.3       kristaps  136:        TEXICMD_SQUIGGLE_LEFT,
                    137:        TEXICMD_SQUIGGLE_RIGHT,
1.8       kristaps  138:        TEXICMD_STRONG,
1.1       kristaps  139:        TEXICMD_SUBSECTION,
1.3       kristaps  140:        TEXICMD_SUBTITLE,
1.12      kristaps  141:        TEXICMD_T,
1.3       kristaps  142:        TEXICMD_TAB,
1.1       kristaps  143:        TEXICMD_TABLE,
                    144:        TEXICMD_TEX,
                    145:        TEXICMD_TEXSYM,
1.3       kristaps  146:        TEXICMD_TITLE,
1.1       kristaps  147:        TEXICMD_TITLEFONT,
                    148:        TEXICMD_TITLEPAGE,
                    149:        TEXICMD_TOP,
                    150:        TEXICMD_UNNUMBERED,
1.2       kristaps  151:        TEXICMD_UNNUMBEREDSEC,
1.3       kristaps  152:        TEXICMD_UREF,
1.1       kristaps  153:        TEXICMD_URL,
                    154:        TEXICMD_VAR,
1.9       kristaps  155:        TEXICMD_VSKIP,
1.3       kristaps  156:        TEXICMD_W,
1.8       kristaps  157:        TEXICMD_XREF,
1.1       kristaps  158:        TEXICMD__MAX
                    159: };
                    160:
                    161: /*
                    162:  * The file currently being parsed.
                    163:  * This keeps track of our location within that file.
                    164:  */
                    165: struct texifile {
                    166:        const char      *name; /* name of the file */
                    167:        size_t           line; /* current line (from zero) */
                    168:        size_t           col; /* current column in line (from zero) */
                    169:        char            *map; /* mmap'd file */
                    170:        size_t           mapsz; /* size of mmap */
                    171: };
                    172:
                    173: struct texi;
                    174:
1.2       kristaps  175: /*
                    176:  * Callback for functions implementing texi commands.
                    177:  */
1.1       kristaps  178: typedef        void (*texicmdfp)(struct texi *,
                    179:        enum texicmd, const char *, size_t, size_t *);
                    180:
                    181: /*
                    182:  * Describes Texinfo commands, whether native or overriden.
                    183:  */
                    184: struct texitok {
                    185:        texicmdfp        fp; /* callback (or NULL if none) */
                    186:        const char      *tok; /* name of the token */
                    187:        size_t           len; /* strlen(tok) */
                    188: };
                    189:
1.3       kristaps  190: enum   texilist {
                    191:        TEXILIST_NONE = 0,
                    192:        TEXILIST_ITEM,
                    193:        TEXILIST_NOITEM,
                    194: };
                    195:
1.1       kristaps  196: /*
                    197:  * The main parse structure.
                    198:  * This keeps any necessary information handy.
                    199:  */
                    200: struct texi {
1.5       kristaps  201:        struct texifile   files[64]; /* stack of open files */
                    202:        size_t            filepos; /* number of open files */
                    203:        size_t            outcol; /* column in output line */
                    204:        char            **dirs; /* texi directories */
                    205:        size_t            dirsz; /* number of texi directories */
1.11      kristaps  206:        char             *title; /* title of document */
                    207:        char             *subtitle; /* subtitle of document */
                    208:        /*
                    209:         * The following control what we output to the screen.
                    210:         * The complexity is required to accomodate for mdoc(7).
                    211:         */
1.8       kristaps  212:        enum texilist     list; /* current list (set recursively) */
                    213:        int               outmacro; /* if >0, output is in line macro */
                    214:        int               seenws; /* ws has been seen (and ignored) */
1.11      kristaps  215:        int               seenvs; /* newline has been Pp'd */
1.8       kristaps  216:        int               ign; /* if >0, don't print anything */
                    217:        int               literal; /* if >0, literal context */
1.1       kristaps  218: };
                    219:
1.8       kristaps  220: /* Texi disregards spaces and tabs. */
1.2       kristaps  221: #define        isws(_x) \
                    222:        (' ' == (_x) || '\t' == (_x))
1.9       kristaps  223: #define        ismspace(_x) \
1.10      kristaps  224:        (isws((_x)) || '\n' == (_x))
1.1       kristaps  225:
1.3       kristaps  226: static void doblock(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.1       kristaps  227: static void dobracket(struct texi *, enum texicmd, const char *, size_t, size_t *);
                    228: static void dobye(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.3       kristaps  229: static void dodeftypefun(struct texi *, enum texicmd, const char *, size_t, size_t *);
                    230: static void dodeftypevar(struct texi *, enum texicmd, const char *, size_t, size_t *);
                    231: static void dodisplay(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.2       kristaps  232: static void doenumerate(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.1       kristaps  233: static void doexample(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.12      kristaps  234: static void dofont(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.8       kristaps  235: static void doignargn(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.1       kristaps  236: static void doignblock(struct texi *, enum texicmd, const char *, size_t, size_t *);
                    237: static void doignbracket(struct texi *, enum texicmd, const char *, size_t, size_t *);
                    238: static void doignline(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.12      kristaps  239: static void doinline(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.2       kristaps  240: static void doinclude(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.1       kristaps  241: static void doitem(struct texi *, enum texicmd, const char *, size_t, size_t *);
                    242: static void doitemize(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.8       kristaps  243: static void dolink(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.3       kristaps  244: static void domath(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.1       kristaps  245: static void doquotation(struct texi *, enum texicmd, const char *, size_t, size_t *);
                    246: static void dotable(struct texi *, enum texicmd, const char *, size_t, size_t *);
                    247: static void dotop(struct texi *, enum texicmd, const char *, size_t, size_t *);
                    248: static void dosection(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.3       kristaps  249: static void dosp(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.1       kristaps  250: static void dosubsection(struct texi *, enum texicmd, const char *, size_t, size_t *);
                    251: static void dosymbol(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.10      kristaps  252: static void dotitle(struct texi *, enum texicmd, const char *, size_t, size_t *);
1.1       kristaps  253:
                    254: static const struct texitok texitoks[TEXICMD__MAX] = {
1.8       kristaps  255:        { doignargn, "acronym", 7 }, /* TEXICMD_ACRONYM */
1.1       kristaps  256:        { doignline, "afourpaper", 10 }, /* TEXICMD_A4PAPER */
                    257:        { doignbracket, "anchor", 6 }, /* TEXICMD_ANCHOR */
1.12      kristaps  258:        { dosection, "appendix", 8 }, /* TEXICMD_APPENDIX */
                    259:        { dosection, "appendixsec", 11 }, /* TEXICMD_APPENDIXSEC */
1.3       kristaps  260:        { dosymbol, "*", 1 }, /* TEXICMD_ASTERISK */
1.1       kristaps  261:        { dosymbol, "@", 1 }, /* TEXICMD_AT */
1.3       kristaps  262:        { doignline, "author", 6 }, /* TEXICMD_AUTHOR */
1.12      kristaps  263:        { dofont, "b", 1 }, /* TEXICMD_BOLD */
1.3       kristaps  264:        { dosymbol, "!", 1 }, /* TEXICMD_BANG */
1.7       kristaps  265:        { dosymbol, "bullet", 6 }, /* TEXICMD_BULLET */
1.1       kristaps  266:        { dobye, "bye", 3 }, /* TEXICMD_BYE */
1.12      kristaps  267:        { doignline, "center", 6 }, /* TEXICMD_CENTER */
                    268:        { dosection, "chapter", 7 }, /* TEXICMD_CHAPTER */
1.1       kristaps  269:        { doignline, "cindex", 6 }, /* TEXICMD_CINDEX */
1.12      kristaps  270:        { dofont, "code", 4 }, /* TEXICMD_CODE */
                    271:        { dofont, "cite", 4 }, /* TEXICMD_CITE */
1.3       kristaps  272:        { dosymbol, ":", 1 }, /* TEXICMD_COLON */
1.12      kristaps  273:        { doinline, "command", 7 }, /* TEXICMD_COMMAND */
1.1       kristaps  274:        { doignline, "c", 1 }, /* TEXICMD_COMMENT */
1.2       kristaps  275:        { doignline, "comment", 7 }, /* TEXICMD_COMMENT_LONG */
1.1       kristaps  276:        { doignline, "contents", 8 }, /* TEXICMD_CONTENTS */
                    277:        { doignblock, "copying", 7 }, /* TEXICMD_COPYING */
                    278:        { dosymbol, "copyright", 9 }, /* TEXICMD_COPYRIGHT */
1.3       kristaps  279:        { dodeftypefun, "deftypefn", 9 }, /* TEXICMD_DEFTYPEFN */
                    280:        { dodeftypefun, "deftypefnx", 10 }, /* TEXICMD_DEFTYPEFNX */
                    281:        { dodeftypefun, "deftypefun", 10 }, /* TEXICMD_DEFTYPEFUN */
                    282:        { dodeftypefun, "deftypefunx", 11 }, /* TEXICMD_DEFTYPEFUNX */
                    283:        { dodeftypevar, "deftypevar", 10 }, /* TEXICMD_DEFTYPEVAR */
                    284:        { dodeftypevar, "deftypevr", 9 }, /* TEXICMD_DEFTYPEVR */
1.1       kristaps  285:        { doignblock, "detailmenu", 10 }, /* TEXICMD_DETAILMENU */
1.12      kristaps  286:        { dofont, "dfn", 3 }, /* TEXICMD_DFN */
1.1       kristaps  287:        { doignline, "dircategory", 11 }, /* TEXICMD_DIRCATEGORY */
                    288:        { doignblock, "direntry", 8 }, /* TEXICMD_DIRENTRY */
1.3       kristaps  289:        { dodisplay, "display", 7 }, /* TEXICMD_DISPLAY */
1.2       kristaps  290:        { dosymbol, "dots", 4 }, /* TEXICMD_DOTS */
1.8       kristaps  291:        { dolink, "email", 5 }, /* TEXICMD_EMAIL */
1.12      kristaps  292:        { dofont, "emph", 4 }, /* TEXICMD_EMPH */
1.1       kristaps  293:        { NULL, "end", 3 }, /* TEXICMD_END */
1.2       kristaps  294:        { doenumerate, "enumerate", 9 }, /* TEXICMD_ENUMERATE */
1.12      kristaps  295:        { doinline, "env", 3 }, /* TEXICMD_ENV */
1.1       kristaps  296:        { doexample, "example", 7 }, /* TEXICMD_EXAMPLE */
1.12      kristaps  297:        { doinline, "file", 4 }, /* TEXICMD_FILE */
1.3       kristaps  298:        { doblock, "group", 5 }, /* TEXICMD_GROUP */
1.2       kristaps  299:        { dosection, "heading", 7 }, /* TEXICMD_HEADING */
1.3       kristaps  300:        { doignline, "headings", 8 }, /* TEXICMD_HEADINGS */
                    301:        { dosymbol, "-", 1 }, /* TEXICMD_HYPHEN */
1.12      kristaps  302:        { dofont, "i", 1 }, /* TEXICMD_I */
1.3       kristaps  303:        { doignblock, "ifclear", 7 }, /* TEXICMD_IFCLEAR */
1.14    ! kristaps  304:        { doignblock, "ifdocbook", 9 }, /* TEXICMD_IFDOCBOOK */
1.1       kristaps  305:        { doignblock, "ifhtml", 6 }, /* TEXICMD_IFHTML */
1.3       kristaps  306:        { doignblock, "ifinfo", 6 }, /* TEXICMD_IFINFO */
1.14    ! kristaps  307:        { doblock, "ifnotdocbook", 12 }, /* TEXICMD_IFNOTDOCBOOK */
        !           308:        { doblock, "ifnothtml", 9 }, /* TEXICMD_IFNOTHTML */
        !           309:        { doblock, "ifnotinfo", 9 }, /* TEXICMD_IFNOTINFO */
        !           310:        { doignblock, "ifnotplaintext", 14 }, /* TEXICMD_IFNOTPLAINTEXT */
1.3       kristaps  311:        { doblock, "ifnottex", 8 }, /* TEXICMD_IFNOTTEX */
1.14    ! kristaps  312:        { doblock, "ifnotxml", 8 }, /* TEXICMD_IFNOTXML */
        !           313:        { doblock, "ifplaintext", 11 }, /* TEXICMD_IFPLAINTEXT */
1.1       kristaps  314:        { doignblock, "iftex", 5 }, /* TEXICMD_IFTEX */
1.3       kristaps  315:        { doignblock, "ifset", 5 }, /* TEXICMD_IFSET */
1.14    ! kristaps  316:        { doignblock, "ifxml", 5 }, /* TEXICMD_IFXML */
1.1       kristaps  317:        { doignbracket, "image", 5 }, /* TEXICMD_IMAGE */
1.2       kristaps  318:        { doinclude, "include", 7 }, /* TEXICMD_INCLUDE */
1.13      kristaps  319:        { dodisplay, "indentblock", 11 }, /* TEXICMD_INDENTBLOCK */
1.5       kristaps  320:        { doignline, "insertcopying", 13 }, /* TEXICMD_INSERTCOPYING */
1.1       kristaps  321:        { doitem, "item", 4 }, /* TEXICMD_ITEM */
                    322:        { doitemize, "itemize", 7 }, /* TEXICMD_ITEMIZE */
1.12      kristaps  323:        { dofont, "kbd", 3 }, /* TEXICMD_KBD */
1.1       kristaps  324:        { dosymbol, "LaTeX", 5 }, /* TEXICMD_LATEX */
1.3       kristaps  325:        { domath, "math", 4 }, /* TEXICMD_MATH */
1.1       kristaps  326:        { doignblock, "menu", 4 }, /* TEXICMD_MENU */
1.3       kristaps  327:        { dosymbol, "\n", 1 }, /* TEXICMD_NEWLINE */
1.1       kristaps  328:        { doignline, "node", 4 }, /* TEXICMD_NODE */
1.3       kristaps  329:        { doignline, "noindent", 8 }, /* TEXICMD_NOINDENT */
1.8       kristaps  330:        { dolink, "pxref", 5 }, /* TEXICMD_PXREF */
1.3       kristaps  331:        { dosymbol, "?", 1 }, /* TEXICMD_QUESTIONMARK */
1.1       kristaps  332:        { doquotation, "quotation", 9 }, /* TEXICMD_QUOTATION */
1.3       kristaps  333:        { doignline, "page", 4 }, /* TEXICMD_PAGE */
                    334:        { doignline, "paragraphindent", 14 }, /* TEXICMD_PARINDENT */
1.2       kristaps  335:        { doignline, "printindex", 10 }, /* TEXICMD_PRINTINDEX */
1.12      kristaps  336:        { dofont, "r", 1 }, /* TEXICMD_R */
1.1       kristaps  337:        { dobracket, "ref", 3 }, /* TEXICMD_REF */
1.12      kristaps  338:        { dofont, "samp", 4 }, /* TEXICMD_SAMP */
                    339:        { dofont, "sansserif", 9 }, /* TEXICMD_SANSSERIF */
1.7       kristaps  340:        { dobracket, "sc", 2 }, /* TEXICMD_SC */
1.1       kristaps  341:        { dosection, "section", 7 }, /* TEXICMD_SECTION */
1.3       kristaps  342:        { doignline, "set", 3 }, /* TEXICMD_SET */
1.1       kristaps  343:        { doignline, "setchapternewpage", 17 }, /* TEXICMD_SETCHAPNEWPAGE */
                    344:        { doignline, "setfilename", 11 }, /* TEXICMD_SETFILENAME */
1.10      kristaps  345:        { dotitle, "settitle", 8 }, /* TEXICMD_SETTITLE */
1.12      kristaps  346:        { dofont, "slanted", 7 }, /* TEXICMD_SLANTED */
1.3       kristaps  347:        { dosp, "sp", 2 }, /* TEXICMD_SP */
                    348:        { dosymbol, " ", 1 }, /* TEXICMD_SPACE */
1.12      kristaps  349:        { dodisplay, "smalldisplay", 12 }, /* TEXICMD_SMALLDISPLAY */
1.3       kristaps  350:        { doexample, "smallexample", 12 }, /* TEXICMD_SMALLEXAMPLE */
1.13      kristaps  351:        { dodisplay, "smallindentblock", 16 }, /* TEXICMD_SMALLINDENTBLOCK */
1.3       kristaps  352:        { dosymbol, "{", 1 }, /* TEXICMD_SQUIGGLE_LEFT */
                    353:        { dosymbol, "}", 1 }, /* TEXICMD_SQUIGGLE_RIGHT */
1.12      kristaps  354:        { dofont, "strong", 6 }, /* TEXICMD_STRONG */
1.1       kristaps  355:        { dosubsection, "subsection", 10 }, /* TEXICMD_SUBSECTION */
1.3       kristaps  356:        { doignline, "subtitle", 8 }, /* TEXICMD_SUBTITLE */
1.12      kristaps  357:        { dofont, "t", 1 }, /* TEXICMD_T */
1.3       kristaps  358:        { dosymbol, "\t", 1 }, /* TEXICMD_TAB */
1.1       kristaps  359:        { dotable, "table", 5 }, /* TEXICMD_TABLE */
                    360:        { doignblock, "tex", 3 }, /* TEXICMD_TEX */
                    361:        { dosymbol, "TeX", 3 }, /* TEXICMD_TEXSYM */
1.3       kristaps  362:        { doignline, "title", 5 }, /* TEXICMD_TITLE */
1.1       kristaps  363:        { dobracket, "titlefont", 9 }, /* TEXICMD_TITLEFONT */
                    364:        { doignblock, "titlepage", 9 }, /* TEXICMD_TITLEPAGE */
                    365:        { dotop, "top", 3 }, /* TEXICMD_TOP */
1.12      kristaps  366:        { dosection, "unnumbered", 10 }, /* TEXICMD_UNNUMBERED */
1.2       kristaps  367:        { dosection, "unnumberedsec", 13 }, /* TEXICMD_UNNUMBEREDSEC */
1.8       kristaps  368:        { dolink, "uref", 4 }, /* TEXICMD_UREF */
                    369:        { dolink, "url", 3 }, /* TEXICMD_URL */
1.12      kristaps  370:        { doinline, "var", 3 }, /* TEXICMD_VAR */
1.9       kristaps  371:        { dosp, "vskip", 5 }, /* TEXICMD_VSKIP */
1.3       kristaps  372:        { dobracket, "w", 1 }, /* TEXICMD_W */
1.8       kristaps  373:        { dolink, "xref", 4 }, /* TEXICMD_XREF */
1.1       kristaps  374: };
                    375:
1.2       kristaps  376: /*
                    377:  * Unmap the top-most file that we're using.
                    378:  */
1.1       kristaps  379: static void
                    380: texifilepop(struct texi *p)
                    381: {
                    382:        struct texifile *f;
                    383:
                    384:        assert(p->filepos > 0);
                    385:        f = &p->files[--p->filepos];
                    386:        munmap(f->map, f->mapsz);
                    387: }
                    388:
1.2       kristaps  389: /*
1.8       kristaps  390:  * Unmap all files that we're currently using and free all resources
                    391:  * that we've allocated during the parse.
1.2       kristaps  392:  * The utility should exit(...) after this is called.
                    393:  */
1.1       kristaps  394: static void
                    395: texiexit(struct texi *p)
                    396: {
1.5       kristaps  397:        size_t   i;
                    398:
                    399:        if (p->outcol)
                    400:                putchar('\n');
1.1       kristaps  401:
                    402:        while (p->filepos > 0)
                    403:                texifilepop(p);
1.5       kristaps  404:
                    405:        for (i = 0; i < p->dirsz; i++)
                    406:                free(p->dirs[i]);
1.10      kristaps  407:
1.5       kristaps  408:        free(p->dirs);
1.10      kristaps  409:        free(p->subtitle);
                    410:        free(p->title);
1.1       kristaps  411: }
                    412:
1.2       kristaps  413: /*
                    414:  * Fatal error: unmap all files and exit.
                    415:  * The "errstring" is passed to perror(3).
                    416:  */
1.1       kristaps  417: static void
1.2       kristaps  418: texiabort(struct texi *p, const char *errstring)
1.1       kristaps  419: {
                    420:
                    421:        perror(errstring);
                    422:        texiexit(p);
                    423:        exit(EXIT_FAILURE);
                    424: }
                    425:
                    426: /*
                    427:  * Print a generic warning message (to stderr) tied to our current
                    428:  * location in the parse sequence.
                    429:  */
                    430: static void
                    431: texiwarn(const struct texi *p, const char *fmt, ...)
                    432: {
                    433:        va_list  ap;
                    434:
1.2       kristaps  435:        fprintf(stderr, "%s:%zu:%zu: warning: ",
1.1       kristaps  436:                p->files[p->filepos - 1].name,
                    437:                p->files[p->filepos - 1].line + 1,
                    438:                p->files[p->filepos - 1].col + 1);
                    439:        va_start(ap, fmt);
                    440:        vfprintf(stderr, fmt, ap);
                    441:        va_end(ap);
                    442:        fputc('\n', stderr);
                    443: }
                    444:
1.8       kristaps  445: /*
                    446:  * Print an error message (to stderr) tied to our current location in
                    447:  * the parse sequence, invoke texiexit(), then die.
                    448:  */
1.2       kristaps  449: static void
                    450: texierr(struct texi *p, const char *fmt, ...)
                    451: {
                    452:        va_list  ap;
                    453:
                    454:        fprintf(stderr, "%s:%zu:%zu: error: ",
                    455:                p->files[p->filepos - 1].name,
                    456:                p->files[p->filepos - 1].line + 1,
                    457:                p->files[p->filepos - 1].col + 1);
                    458:        va_start(ap, fmt);
                    459:        vfprintf(stderr, fmt, ap);
                    460:        va_end(ap);
                    461:        fputc('\n', stderr);
                    462:        texiexit(p);
                    463:        exit(EXIT_FAILURE);
                    464: }
                    465:
1.1       kristaps  466: /*
1.8       kristaps  467:  * Put a single data character to the output if we're not ignoring.
                    468:  * Adjusts our output status.
1.1       kristaps  469:  */
                    470: static void
                    471: texiputchar(struct texi *p, char c)
                    472: {
                    473:
1.3       kristaps  474:        if (p->ign)
1.1       kristaps  475:                return;
                    476:        putchar(c);
1.11      kristaps  477:        p->seenvs = 0;
1.1       kristaps  478:        if ('\n' == c) {
                    479:                p->outcol = 0;
                    480:                p->seenws = 0;
                    481:        } else
                    482:                p->outcol++;
                    483: }
                    484:
                    485: /*
                    486:  * Put multiple characters (see texiputchar()).
                    487:  */
                    488: static void
                    489: texiputchars(struct texi *p, const char *s)
                    490: {
                    491:
                    492:        while ('\0' != *s)
                    493:                texiputchar(p, *s++);
                    494: }
                    495:
                    496: /*
1.8       kristaps  497:  * Close an mdoc(7) macro opened with teximacroopen().
                    498:  * If there are no more macros on the line, prints a newline.
1.1       kristaps  499:  */
                    500: static void
1.3       kristaps  501: teximacroclose(struct texi *p)
                    502: {
                    503:
1.8       kristaps  504:        /* FIXME: punctuation. */
1.5       kristaps  505:        if (0 == --p->outmacro)
                    506:                texiputchar(p, '\n');
1.3       kristaps  507: }
                    508:
                    509: /*
1.8       kristaps  510:  * Open a mdoc(7) macro.
                    511:  * This is used for line macros, e.g., Qq [foo bar baz].
                    512:  * It can be invoked for nested macros, e.g., Qq Li foo .
1.3       kristaps  513:  */
                    514: static void
                    515: teximacroopen(struct texi *p, const char *s)
1.1       kristaps  516: {
                    517:
1.5       kristaps  518:        if (p->outcol && 0 == p->outmacro)
                    519:                texiputchar(p, '\n');
                    520:        if (0 == p->outmacro)
                    521:                texiputchar(p, '.');
                    522:        else
                    523:                texiputchar(p, ' ');
                    524:        texiputchars(p, s);
                    525:        texiputchar(p, ' ');
1.3       kristaps  526:        p->outmacro++;
1.5       kristaps  527:        p->seenws = 0;
1.1       kristaps  528: }
                    529:
                    530: /*
1.8       kristaps  531:  * Put a stadnalone mdoc(7) command with the trailing newline.
1.1       kristaps  532:  */
                    533: static void
                    534: teximacro(struct texi *p, const char *s)
                    535: {
                    536:
1.4       kristaps  537:        if (p->outmacro)
                    538:                texierr(p, "\"%s\" in open line scope!?", s);
                    539:        else if (p->literal)
                    540:                texierr(p, "\"%s\" in a literal scope!?", s);
                    541:
1.1       kristaps  542:        if (p->outcol)
                    543:                texiputchar(p, '\n');
1.5       kristaps  544:
                    545:        texiputchar(p, '.');
                    546:        texiputchars(p, s);
                    547:        texiputchar(p, '\n');
1.1       kristaps  548: }
                    549:
1.11      kristaps  550: static void
                    551: texivspace(struct texi *p)
                    552: {
                    553:
                    554:        if (p->seenvs)
                    555:                return;
                    556:        teximacro(p, "Pp");
                    557:        p->seenvs = 1;
                    558: }
                    559:
1.1       kristaps  560: /*
                    561:  * Advance by a single byte in the input stream.
                    562:  */
                    563: static void
                    564: advance(struct texi *p, const char *buf, size_t *pos)
                    565: {
                    566:
                    567:        if ('\n' == buf[*pos]) {
                    568:                p->files[p->filepos - 1].line++;
                    569:                p->files[p->filepos - 1].col = 0;
                    570:        } else
                    571:                p->files[p->filepos - 1].col++;
                    572:
                    573:        (*pos)++;
                    574: }
                    575:
1.11      kristaps  576: static void
                    577: texipunctuate(struct texi *p, const char *buf, size_t sz, size_t *pos)
                    578: {
                    579:        size_t   start, end;
                    580:
                    581:        if (1 != p->outmacro)
                    582:                return;
                    583:
                    584:        for (start = end = *pos; end < sz; end++) {
                    585:                switch (buf[end]) {
                    586:                case (','):
                    587:                case (')'):
                    588:                case ('.'):
                    589:                case ('"'):
                    590:                case (':'):
                    591:                case ('!'):
                    592:                case ('?'):
                    593:                        continue;
                    594:                default:
                    595:                        break;
                    596:                }
                    597:                break;
                    598:        }
                    599:        if (end == *pos)
                    600:                return;
                    601:        if (end + 1 == sz || ' ' == buf[end] || '\n' == buf[end]) {
                    602:                for ( ; start < end; start++) {
                    603:                        texiputchar(p, ' ');
                    604:                        texiputchar(p, buf[start]);
                    605:                        advance(p, buf, pos);
                    606:                }
                    607:        }
                    608: }
                    609:
1.1       kristaps  610: /*
                    611:  * Advance to the next non-whitespace word in the input stream.
                    612:  * If we're in literal mode, then print all of the whitespace as we're
                    613:  * doing so.
                    614:  */
                    615: static size_t
                    616: advancenext(struct texi *p, const char *buf, size_t sz, size_t *pos)
                    617: {
                    618:
1.3       kristaps  619:        if (p->literal) {
1.9       kristaps  620:                while (*pos < sz && ismspace(buf[*pos])) {
1.5       kristaps  621:                        if (*pos && '\n' == buf[*pos] &&
                    622:                                '\\' == buf[*pos - 1])
                    623:                                texiputchar(p, 'e');
1.1       kristaps  624:                        texiputchar(p, buf[*pos]);
                    625:                        advance(p, buf, pos);
                    626:                }
                    627:                return(*pos);
                    628:        }
                    629:
1.9       kristaps  630:        while (*pos < sz && ismspace(buf[*pos])) {
1.1       kristaps  631:                p->seenws = 1;
                    632:                /*
                    633:                 * If it looks like we've printed a double-line, then
                    634:                 * output a paragraph.
                    635:                 * FIXME: this is stupid.
                    636:                 */
1.11      kristaps  637:                if (*pos && '\n' == buf[*pos] && '\n' == buf[*pos - 1])
                    638:                        texivspace(p);
1.1       kristaps  639:                advance(p, buf, pos);
                    640:        }
                    641:        return(*pos);
                    642: }
                    643:
                    644: /*
                    645:  * Advance to the EOLN in the input stream.
                    646:  */
                    647: static size_t
1.3       kristaps  648: advanceeoln(struct texi *p, const char *buf,
                    649:        size_t sz, size_t *pos, int consumenl)
1.1       kristaps  650: {
                    651:
1.8       kristaps  652:        /* FIXME: disregards @NEWLINE. */
1.1       kristaps  653:        while (*pos < sz && '\n' != buf[*pos])
                    654:                advance(p, buf, pos);
1.3       kristaps  655:        if (*pos < sz && consumenl)
                    656:                advance(p, buf, pos);
1.1       kristaps  657:        return(*pos);
                    658: }
                    659:
                    660: /*
                    661:  * Advance to position "end", which is an absolute position in the
                    662:  * current buffer greater than or equal to the current position.
                    663:  */
                    664: static void
                    665: advanceto(struct texi *p, const char *buf, size_t *pos, size_t end)
                    666: {
                    667:
                    668:        assert(*pos <= end);
                    669:        while (*pos < end)
                    670:                advance(p, buf, pos);
                    671: }
                    672:
                    673: /*
                    674:  * Output a free-form word in the input stream, progressing to the next
                    675:  * command or white-space.
                    676:  * This also will advance the input stream.
                    677:  */
                    678: static void
1.8       kristaps  679: texiword(struct texi *p, const char *buf,
                    680:        size_t sz, size_t *pos, char extra)
1.1       kristaps  681: {
                    682:
1.3       kristaps  683:        if (0 == p->outmacro && p->outcol > 72 && 0 == p->literal)
1.1       kristaps  684:                texiputchar(p, '\n');
1.8       kristaps  685:        /* FIXME: abstract this: we use it elsewhere. */
1.3       kristaps  686:        if (p->seenws && p->outcol && 0 == p->literal)
1.1       kristaps  687:                texiputchar(p, ' ');
                    688:
                    689:        p->seenws = 0;
                    690:
1.9       kristaps  691:        while (*pos < sz && ! ismspace(buf[*pos])) {
1.1       kristaps  692:                switch (buf[*pos]) {
                    693:                case ('@'):
                    694:                case ('}'):
                    695:                case ('{'):
                    696:                        return;
                    697:                }
1.8       kristaps  698:                if ('\0' != extra && buf[*pos] == extra)
                    699:                        return;
1.1       kristaps  700:                if (*pos < sz - 1 &&
                    701:                         '`' == buf[*pos] &&
                    702:                         '`' == buf[*pos + 1]) {
                    703:                        texiputchars(p, "\\(lq");
                    704:                        advance(p, buf, pos);
                    705:                } else if (*pos < sz - 1 &&
                    706:                         '\'' == buf[*pos] &&
                    707:                         '\'' == buf[*pos + 1]) {
                    708:                        texiputchars(p, "\\(rq");
                    709:                        advance(p, buf, pos);
                    710:                } else
                    711:                        texiputchar(p, buf[*pos]);
                    712:                advance(p, buf, pos);
                    713:        }
                    714: }
                    715:
1.8       kristaps  716: /*
                    717:  * Look up the command at position "pos" in the buffer, returning it (or
                    718:  * TEXICMD__MAX if none found) and setting "end" to be the absolute
                    719:  * index after the command name.
                    720:  */
1.1       kristaps  721: static enum texicmd
                    722: texicmd(struct texi *p, const char *buf,
                    723:        size_t pos, size_t sz, size_t *end)
                    724: {
                    725:        size_t   i, len;
                    726:
                    727:        assert('@' == buf[pos]);
1.3       kristaps  728:
1.9       kristaps  729:        if ((*end = pos) == sz)
                    730:                return(TEXICMD__MAX);
                    731:        else if ((*end = ++pos) == sz)
1.3       kristaps  732:                return(TEXICMD__MAX);
                    733:
                    734:        /* Alphabetic commands are special. */
                    735:        if ( ! isalpha(buf[pos])) {
1.9       kristaps  736:                if ((*end = pos + 1) == sz)
                    737:                        return(TEXICMD__MAX);
1.3       kristaps  738:                for (i = 0; i < TEXICMD__MAX; i++) {
                    739:                        if (1 != texitoks[i].len)
                    740:                                continue;
                    741:                        if (0 == strncmp(texitoks[i].tok, &buf[pos], 1))
                    742:                                return(i);
                    743:                }
                    744:                texiwarn(p, "bad command: @%c", buf[pos]);
                    745:                return(TEXICMD__MAX);
                    746:        }
                    747:
1.9       kristaps  748:        for (*end = pos; *end < sz && ! ismspace(buf[*end]); (*end)++)
1.3       kristaps  749:                if ((*end > pos && ('@' == buf[*end] ||
                    750:                          '{' == buf[*end] || '}' == buf[*end])))
1.1       kristaps  751:                        break;
                    752:
                    753:        len = *end - pos;
                    754:        for (i = 0; i < TEXICMD__MAX; i++) {
                    755:                if (len != texitoks[i].len)
                    756:                        continue;
                    757:                if (0 == strncmp(texitoks[i].tok, &buf[pos], len))
                    758:                        return(i);
                    759:        }
                    760:
1.3       kristaps  761:        texiwarn(p, "bad command: @%.*s", (int)len, &buf[pos]);
1.1       kristaps  762:        return(TEXICMD__MAX);
                    763: }
                    764:
1.8       kristaps  765: /*
                    766:  * Parse an argument from a bracketed command, e.g., @url{foo, baz}.
                    767:  * Num should be set to the argument we're currently parsing, although
                    768:  * it suffixes for it to be zero or non-zero.
                    769:  * This will return 1 if there are more arguments, 0 otherwise.
                    770:  * This will stop (returning 0) in the event of EOF or if we're not at a
                    771:  * bracket for the zeroth parse.
                    772:  */
                    773: static int
                    774: parsearg(struct texi *p, const char *buf,
                    775:        size_t sz, size_t *pos, size_t num)
                    776: {
                    777:        size_t           end;
                    778:        enum texicmd     cmd;
                    779:
1.9       kristaps  780:        while (*pos < sz && ismspace(buf[*pos]))
1.8       kristaps  781:                advance(p, buf, pos);
                    782:        if (*pos == sz || (0 == num && '{' != buf[*pos]))
                    783:                return(0);
                    784:        if (0 == num)
                    785:                advance(p, buf, pos);
                    786:
                    787:        while ((*pos = advancenext(p, buf, sz, pos)) < sz) {
                    788:                switch (buf[*pos]) {
                    789:                case (','):
                    790:                        advance(p, buf, pos);
                    791:                        return(1);
                    792:                case ('}'):
                    793:                        advance(p, buf, pos);
                    794:                        return(0);
                    795:                case ('{'):
                    796:                        if (0 == p->ign)
                    797:                                texiwarn(p, "unexpected \"{\"");
                    798:                        advance(p, buf, pos);
                    799:                        continue;
                    800:                case ('@'):
                    801:                        break;
                    802:                default:
                    803:                        texiword(p, buf, sz, pos, ',');
                    804:                        continue;
                    805:                }
                    806:
                    807:                cmd = texicmd(p, buf, *pos, sz, &end);
                    808:                advanceto(p, buf, pos, end);
                    809:                if (TEXICMD__MAX == cmd)
                    810:                        continue;
                    811:                if (NULL != texitoks[cmd].fp)
                    812:                        (*texitoks[cmd].fp)(p, cmd, buf, sz, pos);
                    813:        }
                    814:        return(0);
                    815: }
                    816:
                    817: /*
                    818:  * Parse until the end of a bracketed statement, e.g., @foo{bar baz}.
                    819:  * This will stop in the event of EOF or if we're not at a bracket.
                    820:  */
1.1       kristaps  821: static void
                    822: parsebracket(struct texi *p, const char *buf, size_t sz, size_t *pos)
                    823: {
                    824:        size_t           end;
                    825:        enum texicmd     cmd;
                    826:
1.9       kristaps  827:        while (*pos < sz && ismspace(buf[*pos]))
1.3       kristaps  828:                advance(p, buf, pos);
                    829:
1.1       kristaps  830:        if (*pos == sz || '{' != buf[*pos])
                    831:                return;
                    832:        advance(p, buf, pos);
                    833:
                    834:        while ((*pos = advancenext(p, buf, sz, pos)) < sz) {
                    835:                switch (buf[*pos]) {
                    836:                case ('}'):
                    837:                        advance(p, buf, pos);
                    838:                        return;
                    839:                case ('{'):
1.3       kristaps  840:                        if (0 == p->ign)
                    841:                                texiwarn(p, "unexpected \"{\"");
                    842:                        advance(p, buf, pos);
                    843:                        continue;
                    844:                case ('@'):
                    845:                        break;
                    846:                default:
1.8       kristaps  847:                        texiword(p, buf, sz, pos, '\0');
1.3       kristaps  848:                        continue;
                    849:                }
                    850:
                    851:                cmd = texicmd(p, buf, *pos, sz, &end);
                    852:                advanceto(p, buf, pos, end);
                    853:                if (TEXICMD__MAX == cmd)
                    854:                        continue;
                    855:                if (NULL != texitoks[cmd].fp)
                    856:                        (*texitoks[cmd].fp)(p, cmd, buf, sz, pos);
                    857:        }
                    858: }
                    859:
                    860: /*
                    861:  * This should be invoked when we're on a macro line and want to process
                    862:  * to the end of the current input line, doing all of our macros along
                    863:  * the way.
                    864:  */
                    865: static void
                    866: parseeoln(struct texi *p, const char *buf, size_t sz, size_t *pos)
                    867: {
                    868:        size_t           end;
                    869:        enum texicmd     cmd;
                    870:
                    871:        assert(0 == p->literal);
                    872:
                    873:        while (*pos < sz && '\n' != buf[*pos]) {
                    874:                while (*pos < sz && isws(buf[*pos])) {
                    875:                        p->seenws = 1;
                    876:                        advance(p, buf, pos);
                    877:                }
                    878:                switch (buf[*pos]) {
                    879:                case ('}'):
                    880:                        if (0 == p->ign)
                    881:                                texiwarn(p, "unexpected \"}\"");
                    882:                        advance(p, buf, pos);
                    883:                        continue;
                    884:                case ('{'):
                    885:                        if (0 == p->ign)
                    886:                                texiwarn(p, "unexpected \"{\"");
1.1       kristaps  887:                        advance(p, buf, pos);
                    888:                        continue;
                    889:                case ('@'):
                    890:                        break;
                    891:                default:
1.8       kristaps  892:                        texiword(p, buf, sz, pos, '\0');
1.1       kristaps  893:                        continue;
                    894:                }
                    895:
                    896:                cmd = texicmd(p, buf, *pos, sz, &end);
                    897:                advanceto(p, buf, pos, end);
                    898:                if (TEXICMD__MAX == cmd)
                    899:                        continue;
                    900:                if (NULL != texitoks[cmd].fp)
                    901:                        (*texitoks[cmd].fp)(p, cmd, buf, sz, pos);
                    902:        }
                    903: }
                    904:
1.8       kristaps  905: /*
                    906:  * Parse a single word or command.
                    907:  * This will return immediately at the EOF.
                    908:  */
1.1       kristaps  909: static void
1.3       kristaps  910: parsesingle(struct texi *p, const char *buf, size_t sz, size_t *pos)
                    911: {
                    912:        size_t           end;
                    913:        enum texicmd     cmd;
                    914:
                    915:        if ((*pos = advancenext(p, buf, sz, pos)) >= sz)
                    916:                return;
                    917:
                    918:        switch (buf[*pos]) {
                    919:        case ('}'):
                    920:                if (0 == p->ign)
                    921:                        texiwarn(p, "unexpected \"}\"");
                    922:                advance(p, buf, pos);
                    923:                return;
                    924:        case ('{'):
                    925:                if (0 == p->ign)
                    926:                        texiwarn(p, "unexpected \"{\"");
                    927:                advance(p, buf, pos);
                    928:                return;
                    929:        case ('@'):
                    930:                break;
                    931:        default:
1.8       kristaps  932:                texiword(p, buf, sz, pos, '\0');
1.3       kristaps  933:                return;
                    934:        }
                    935:
                    936:        cmd = texicmd(p, buf, *pos, sz, &end);
                    937:        advanceto(p, buf, pos, end);
                    938:        if (TEXICMD__MAX == cmd)
                    939:                return;
                    940:        if (NULL != texitoks[cmd].fp)
                    941:                (*texitoks[cmd].fp)(p, cmd, buf, sz, pos);
                    942: }
                    943:
1.8       kristaps  944: /*
                    945:  * Parse til the end of the buffer.
                    946:  */
1.3       kristaps  947: static void
1.7       kristaps  948: parseeof(struct texi *p, const char *buf, size_t sz)
                    949: {
                    950:        size_t   pos;
                    951:
                    952:        for (pos = 0; pos < sz; )
                    953:                parsesingle(p, buf, sz, &pos);
                    954: }
                    955:
1.8       kristaps  956: /*
                    957:  * Parse a block sequence until we have the "@end endtoken" command
                    958:  * invocation.
                    959:  * This will return immediately at EOF.
                    960:  */
1.7       kristaps  961: static void
1.1       kristaps  962: parseto(struct texi *p, const char *buf,
                    963:        size_t sz, size_t *pos, const char *endtoken)
                    964: {
                    965:        size_t           end;
                    966:        enum texicmd     cmd;
                    967:        size_t           endtoksz;
                    968:
                    969:        endtoksz = strlen(endtoken);
                    970:        assert(endtoksz > 0);
                    971:
                    972:        while ((*pos = advancenext(p, buf, sz, pos)) < sz) {
                    973:                switch (buf[*pos]) {
                    974:                case ('}'):
1.3       kristaps  975:                        if (0 == p->ign)
                    976:                                texiwarn(p, "unexpected \"}\"");
1.1       kristaps  977:                        advance(p, buf, pos);
                    978:                        continue;
                    979:                case ('{'):
1.3       kristaps  980:                        if (0 == p->ign)
                    981:                                texiwarn(p, "unexpected \"{\"");
1.1       kristaps  982:                        advance(p, buf, pos);
                    983:                        continue;
                    984:                case ('@'):
                    985:                        break;
                    986:                default:
1.8       kristaps  987:                        texiword(p, buf, sz, pos, '\0');
1.1       kristaps  988:                        continue;
                    989:                }
                    990:
                    991:                cmd = texicmd(p, buf, *pos, sz, &end);
                    992:                advanceto(p, buf, pos, end);
                    993:                if (TEXICMD_END == cmd) {
1.2       kristaps  994:                        while (*pos < sz && isws(buf[*pos]))
1.1       kristaps  995:                                advance(p, buf, pos);
                    996:                        /*
1.8       kristaps  997:                         * FIXME: check the full word, not just its
                    998:                         * initial substring!
1.1       kristaps  999:                         */
                   1000:                        if (sz - *pos >= endtoksz && 0 == strncmp
                   1001:                                 (&buf[*pos], endtoken, endtoksz)) {
1.3       kristaps 1002:                                advanceeoln(p, buf, sz, pos, 0);
1.1       kristaps 1003:                                break;
                   1004:                        }
1.3       kristaps 1005:                        if (0 == p->ign)
                   1006:                                texiwarn(p, "unexpected \"end\"");
                   1007:                        advanceeoln(p, buf, sz, pos, 0);
1.1       kristaps 1008:                        continue;
                   1009:                } else if (TEXICMD__MAX != cmd)
                   1010:                        if (NULL != texitoks[cmd].fp)
                   1011:                                (*texitoks[cmd].fp)(p, cmd, buf, sz, pos);
                   1012:        }
                   1013: }
                   1014:
1.8       kristaps 1015: /*
                   1016:  * Memory-map the file "fname" and begin parsing it.
                   1017:  * This can be called in a nested context.
                   1018:  */
1.1       kristaps 1019: static void
1.2       kristaps 1020: parsefile(struct texi *p, const char *fname)
                   1021: {
                   1022:        struct texifile  *f;
                   1023:        int               fd;
                   1024:        struct stat       st;
                   1025:
                   1026:        assert(p->filepos < 64);
                   1027:        f = &p->files[p->filepos];
                   1028:        memset(f, 0, sizeof(struct texifile));
                   1029:
                   1030:        f->name = fname;
                   1031:        if (-1 == (fd = open(fname, O_RDONLY, 0))) {
                   1032:                texiabort(p, fname);
                   1033:        } else if (-1 == fstat(fd, &st)) {
                   1034:                close(fd);
                   1035:                texiabort(p, fname);
                   1036:        }
                   1037:
                   1038:        f->mapsz = st.st_size;
                   1039:        f->map = mmap(NULL, f->mapsz,
                   1040:                PROT_READ, MAP_SHARED, fd, 0);
                   1041:        close(fd);
                   1042:
                   1043:        if (MAP_FAILED == f->map)
                   1044:                texiabort(p, fname);
                   1045:
                   1046:        p->filepos++;
                   1047:        parseeof(p, f->map, f->mapsz);
                   1048:        texifilepop(p);
                   1049: }
                   1050:
                   1051: static void
1.3       kristaps 1052: dodeftypevar(struct texi *p, enum texicmd cmd,
                   1053:        const char *buf, size_t sz, size_t *pos)
                   1054: {
                   1055:        const char      *blk;
                   1056:
                   1057:        blk = TEXICMD_DEFTYPEVR == cmd ?
                   1058:                "deftypevr" : "deftypevar";
                   1059:
                   1060:        if (p->ign) {
                   1061:                parseto(p, buf, sz, pos, blk);
                   1062:                return;
                   1063:        }
                   1064:
1.11      kristaps 1065:        texivspace(p);
1.3       kristaps 1066:        if (TEXICMD_DEFTYPEVR == cmd) {
                   1067:                parsebracket(p, buf, sz, pos);
                   1068:                texiputchars(p, ":\n");
                   1069:        }
1.5       kristaps 1070:        teximacroopen(p, "Vt");
1.4       kristaps 1071:        parseeoln(p, buf, sz, pos);
1.3       kristaps 1072:        teximacroclose(p);
1.11      kristaps 1073:        texivspace(p);
1.3       kristaps 1074:        parseto(p, buf, sz, pos, blk);
                   1075: }
                   1076:
                   1077: static void
                   1078: dodeftypefun(struct texi *p, enum texicmd cmd,
                   1079:        const char *buf, size_t sz, size_t *pos)
                   1080: {
                   1081:        const char      *blk;
                   1082:
1.5       kristaps 1083:        blk = NULL;
1.3       kristaps 1084:        switch (cmd) {
                   1085:        case (TEXICMD_DEFTYPEFN):
                   1086:        case (TEXICMD_DEFTYPEFUN):
1.5       kristaps 1087:                blk = texitoks[cmd].tok;
1.3       kristaps 1088:                break;
1.5       kristaps 1089:        default:
1.3       kristaps 1090:                break;
                   1091:        }
                   1092:
                   1093:        if (p->ign) {
                   1094:                if (NULL != blk)
                   1095:                        parseto(p, buf, sz, pos, blk);
                   1096:                return;
                   1097:        }
                   1098:
                   1099:        switch (cmd) {
                   1100:        case (TEXICMD_DEFTYPEFN):
                   1101:        case (TEXICMD_DEFTYPEFUN):
1.11      kristaps 1102:                texivspace(p);
1.3       kristaps 1103:                break;
                   1104:        default:
                   1105:                break;
                   1106:        }
                   1107:        if (TEXICMD_DEFTYPEFN == cmd ||
                   1108:                        TEXICMD_DEFTYPEFNX == cmd) {
                   1109:                parsebracket(p, buf, sz, pos);
                   1110:                texiputchars(p, ":\n");
                   1111:        }
1.5       kristaps 1112:        teximacroopen(p, "Ft");
1.3       kristaps 1113:        parsesingle(p, buf, sz, pos);
                   1114:        teximacroclose(p);
1.5       kristaps 1115:        teximacroopen(p, "Fn");
1.3       kristaps 1116:        parsesingle(p, buf, sz, pos);
                   1117:        teximacroclose(p);
1.5       kristaps 1118:        teximacroopen(p, "Li");
1.4       kristaps 1119:        parseeoln(p, buf, sz, pos);
1.3       kristaps 1120:        teximacroclose(p);
1.11      kristaps 1121:        texivspace(p);
1.3       kristaps 1122:        if (NULL != blk)
                   1123:                parseto(p, buf, sz, pos, blk);
                   1124: }
                   1125:
                   1126: static void
1.1       kristaps 1127: doignblock(struct texi *p, enum texicmd cmd,
                   1128:        const char *buf, size_t sz, size_t *pos)
                   1129: {
                   1130:
1.3       kristaps 1131:        p->ign++;
1.5       kristaps 1132:        parseto(p, buf, sz, pos, texitoks[cmd].tok);
1.3       kristaps 1133:        p->ign--;
1.1       kristaps 1134: }
                   1135:
                   1136: static void
1.3       kristaps 1137: doblock(struct texi *p, enum texicmd cmd,
1.1       kristaps 1138:        const char *buf, size_t sz, size_t *pos)
                   1139: {
                   1140:
1.5       kristaps 1141:        parseto(p, buf, sz, pos, texitoks[cmd].tok);
1.1       kristaps 1142: }
                   1143:
                   1144: static void
1.12      kristaps 1145: doinline(struct texi *p, enum texicmd cmd,
                   1146:        const char *buf, size_t sz, size_t *pos)
1.1       kristaps 1147: {
1.12      kristaps 1148:        const char      *macro;
                   1149:
                   1150:        switch (cmd) {
                   1151:        case (TEXICMD_COMMAND):
                   1152:                macro = "Xr";
                   1153:                break;
                   1154:        case (TEXICMD_ENV):
                   1155:                macro = "Ev";
                   1156:                break;
                   1157:        case (TEXICMD_FILE):
                   1158:                macro = "Pa";
                   1159:                break;
                   1160:        case (TEXICMD_VAR):
                   1161:                macro = "Va";
                   1162:                break;
                   1163:        default:
                   1164:                abort();
                   1165:        }
                   1166:
                   1167:        if (p->literal) {
                   1168:                parsebracket(p, buf, sz, pos);
                   1169:                return;
                   1170:        }
1.1       kristaps 1171:
1.5       kristaps 1172:        teximacroopen(p, macro);
1.1       kristaps 1173:        p->seenws = 0;
                   1174:        parsebracket(p, buf, sz, pos);
1.11      kristaps 1175:        texipunctuate(p, buf, sz, pos);
1.5       kristaps 1176:        teximacroclose(p);
1.1       kristaps 1177: }
                   1178:
                   1179: static void
1.2       kristaps 1180: doinclude(struct texi *p, enum texicmd cmd,
                   1181:        const char *buf, size_t sz, size_t *pos)
                   1182: {
                   1183:        char     fname[PATH_MAX], path[PATH_MAX];
                   1184:        size_t   i;
                   1185:        int      rc;
                   1186:
                   1187:        while (*pos < sz && ' ' == buf[*pos])
                   1188:                advance(p, buf, pos);
                   1189:
                   1190:        /* Read in the filename. */
                   1191:        for (i = 0; *pos < sz && '\n' != buf[*pos]; i++) {
                   1192:                if (i == sizeof(fname) - 1)
                   1193:                        break;
                   1194:                fname[i] = buf[*pos];
                   1195:                advance(p, buf, pos);
                   1196:        }
                   1197:
                   1198:        if (i == 0)
                   1199:                texierr(p, "path too short");
                   1200:        else if ('\n' != buf[*pos])
                   1201:                texierr(p, "path too long");
                   1202:        else if ('/' == fname[0])
                   1203:                texierr(p, "no absolute paths");
                   1204:        fname[i] = '\0';
                   1205:
                   1206:        if (strstr(fname, "../") || strstr(fname, "/.."))
                   1207:                texierr(p, "insecure path");
                   1208:
1.5       kristaps 1209:        for (i = 0; i < p->dirsz; i++) {
                   1210:                rc = snprintf(path, sizeof(path),
                   1211:                        "%s/%s", p->dirs[i], fname);
                   1212:                if (rc < 0)
                   1213:                        texierr(p, "couldn't format path");
                   1214:                else if ((size_t)rc >= sizeof(path))
                   1215:                        texierr(p, "path too long");
                   1216:                else if (-1 == access(path, R_OK))
                   1217:                        continue;
                   1218:
                   1219:                parsefile(p, path);
                   1220:                return;
                   1221:        }
1.2       kristaps 1222:
1.5       kristaps 1223:        texierr(p, "couldn't find %s in includes", fname);
1.2       kristaps 1224: }
                   1225:
                   1226: static void
1.12      kristaps 1227: dofont(struct texi *p, enum texicmd cmd,
1.1       kristaps 1228:        const char *buf, size_t sz, size_t *pos)
                   1229: {
1.12      kristaps 1230:        const char      *font;
                   1231:
                   1232:        switch (cmd) {
                   1233:        case (TEXICMD_B):
                   1234:        case (TEXICMD_STRONG):
                   1235:                font = "\\fB";
                   1236:                break;
                   1237:        case (TEXICMD_CITE):
                   1238:        case (TEXICMD_DFN):
                   1239:        case (TEXICMD_EMPH):
                   1240:        case (TEXICMD_I):
                   1241:        case (TEXICMD_SLANTED):
                   1242:                font = "\\fI";
                   1243:                break;
                   1244:        case (TEXICMD_CODE):
                   1245:        case (TEXICMD_KBD):
                   1246:        case (TEXICMD_R):
                   1247:        case (TEXICMD_SAMP):
                   1248:        case (TEXICMD_SANSSERIF):
                   1249:        case (TEXICMD_T):
                   1250:                font = "\\fR";
                   1251:                break;
                   1252:        default:
                   1253:                abort();
                   1254:        }
1.1       kristaps 1255:
1.12      kristaps 1256:        texiputchars(p, font);
1.1       kristaps 1257:        parsebracket(p, buf, sz, pos);
                   1258:        texiputchars(p, "\\fP");
                   1259: }
                   1260:
                   1261: static void
                   1262: dobracket(struct texi *p, enum texicmd cmd,
                   1263:        const char *buf, size_t sz, size_t *pos)
                   1264: {
                   1265:
                   1266:        parsebracket(p, buf, sz, pos);
                   1267: }
                   1268:
                   1269: static void
1.3       kristaps 1270: dodisplay(struct texi *p, enum texicmd cmd,
                   1271:        const char *buf, size_t sz, size_t *pos)
                   1272: {
                   1273:
1.13      kristaps 1274:        teximacro(p, "Bd -filled -offset indent");
1.11      kristaps 1275:        p->seenvs = 1;
1.12      kristaps 1276:        /* FIXME: ignore and parseeoln. */
1.3       kristaps 1277:        advanceeoln(p, buf, sz, pos, 1);
1.13      kristaps 1278:        parseto(p, buf, sz, pos, texitoks[cmd].tok);
1.5       kristaps 1279:        teximacro(p, "Ed");
1.3       kristaps 1280: }
                   1281:
                   1282: static void
1.1       kristaps 1283: doexample(struct texi *p, enum texicmd cmd,
                   1284:        const char *buf, size_t sz, size_t *pos)
                   1285: {
                   1286:
1.5       kristaps 1287:        teximacro(p, "Bd -literal -offset indent");
1.12      kristaps 1288:        /* FIXME: ignore and parseeoln. */
1.3       kristaps 1289:        advanceeoln(p, buf, sz, pos, 1);
                   1290:        p->literal++;
1.13      kristaps 1291:        parseto(p, buf, sz, pos, texitoks[cmd].tok);
1.3       kristaps 1292:        p->literal--;
1.5       kristaps 1293:        teximacro(p, "Ed");
1.1       kristaps 1294: }
                   1295:
                   1296: static void
                   1297: dobye(struct texi *p, enum texicmd cmd,
                   1298:        const char *buf, size_t sz, size_t *pos)
                   1299: {
                   1300:
                   1301:        texiexit(p);
                   1302:        exit(EXIT_SUCCESS);
                   1303: }
                   1304:
                   1305: static void
1.10      kristaps 1306: dotitle(struct texi *p, enum texicmd cmd,
                   1307:        const char *buf, size_t sz, size_t *pos)
                   1308: {
                   1309:        size_t   start, end;
                   1310:
                   1311:        while (*pos < sz && isws(buf[*pos]))
                   1312:                advance(p, buf, pos);
                   1313:        start = end = *pos;
                   1314:        while (end < sz && '\n' != buf[end])
                   1315:                end++;
                   1316:        free(p->subtitle);
                   1317:        p->subtitle = malloc(end - start + 1);
                   1318:        memcpy(p->subtitle, &buf[start], end - start);
                   1319:        p->subtitle[end - start] = '\0';
                   1320: }
                   1321:
                   1322: static void
1.1       kristaps 1323: dosymbol(struct texi *p, enum texicmd cmd,
                   1324:        const char *buf, size_t sz, size_t *pos)
                   1325: {
                   1326:
1.3       kristaps 1327:        if (p->seenws && p->outcol && 0 == p->literal) {
                   1328:                texiputchar(p, ' ');
                   1329:                p->seenws = 0;
                   1330:        }
                   1331:
1.1       kristaps 1332:        switch (cmd) {
1.3       kristaps 1333:        case (TEXICMD_ASTERISK):
                   1334:        case (TEXICMD_NEWLINE):
                   1335:        case (TEXICMD_SPACE):
                   1336:        case (TEXICMD_TAB):
                   1337:                texiputchar(p, ' ');
                   1338:                break;
1.1       kristaps 1339:        case (TEXICMD_AT):
1.3       kristaps 1340:                texiputchar(p, '@');
                   1341:                break;
                   1342:        case (TEXICMD_BANG):
                   1343:                texiputchar(p, '!');
1.7       kristaps 1344:                break;
                   1345:        case (TEXICMD_BULLET):
                   1346:                texiputchars(p, "\\(bu");
1.1       kristaps 1347:                break;
                   1348:        case (TEXICMD_COPYRIGHT):
                   1349:                texiputchars(p, "\\(co");
                   1350:                break;
1.2       kristaps 1351:        case (TEXICMD_DOTS):
                   1352:                texiputchars(p, "...");
                   1353:                break;
1.1       kristaps 1354:        case (TEXICMD_LATEX):
                   1355:                texiputchars(p, "LaTeX");
                   1356:                break;
1.3       kristaps 1357:        case (TEXICMD_QUESTIONMARK):
                   1358:                texiputchar(p, '?');
                   1359:                break;
                   1360:        case (TEXICMD_SQUIGGLE_LEFT):
                   1361:                texiputchars(p, "{");
                   1362:                break;
                   1363:        case (TEXICMD_SQUIGGLE_RIGHT):
                   1364:                texiputchars(p, "}");
                   1365:                break;
1.1       kristaps 1366:        case (TEXICMD_TEXSYM):
                   1367:                texiputchars(p, "TeX");
                   1368:                break;
1.3       kristaps 1369:        case (TEXICMD_COLON):
                   1370:        case (TEXICMD_HYPHEN):
                   1371:                break;
1.1       kristaps 1372:        default:
1.5       kristaps 1373:                texiwarn(p, "sym: %d", cmd);
1.1       kristaps 1374:                abort();
                   1375:        }
                   1376:
1.5       kristaps 1377:        if (texitoks[cmd].len > 1)
                   1378:                doignbracket(p, cmd, buf, sz, pos);
1.1       kristaps 1379: }
                   1380:
                   1381: static void
                   1382: doquotation(struct texi *p, enum texicmd cmd,
                   1383:        const char *buf, size_t sz, size_t *pos)
                   1384: {
                   1385:
1.5       kristaps 1386:        teximacro(p, "Qo");
1.1       kristaps 1387:        parseto(p, buf, sz, pos, "quotation");
1.5       kristaps 1388:        teximacro(p, "Qc");
1.1       kristaps 1389: }
                   1390:
1.3       kristaps 1391: static void
                   1392: domath(struct texi *p, enum texicmd cmd,
                   1393:        const char *buf, size_t sz, size_t *pos)
                   1394: {
                   1395:        size_t   nest;
                   1396:
                   1397:        /*
                   1398:         * Math handling is different from everything else.
                   1399:         * We don't allow any subcomponents, and we ignore the rules in
                   1400:         * terms of @-commands.
                   1401:         * This departs from GNU's rules, but whatever.
                   1402:         */
                   1403:        while (*pos < sz && isws(buf[*pos]))
                   1404:                advance(p, buf, pos);
                   1405:        if (*pos == sz || '{' != buf[*pos])
                   1406:                return;
                   1407:        advance(p, buf, pos);
                   1408:        if (p->seenws && p->outcol && 0 == p->literal)
                   1409:                texiputchar(p, ' ');
                   1410:        p->seenws = 0;
                   1411:        for (nest = 1; *pos < sz && nest > 0; ) {
                   1412:                if ('{' == buf[*pos])
                   1413:                        nest++;
                   1414:                else if ('}' == buf[*pos])
                   1415:                        if (0 == --nest)
                   1416:                                continue;
                   1417:                texiputchar(p, buf[*pos]);
                   1418:                advance(p, buf, pos);
                   1419:        }
                   1420:        if (*pos == sz)
                   1421:                return;
                   1422:        assert('}' == buf[*pos]);
                   1423:        advance(p, buf, pos);
                   1424: }
                   1425:
1.1       kristaps 1426: static void
1.8       kristaps 1427: dolink(struct texi *p, enum texicmd cmd,
1.1       kristaps 1428:        const char *buf, size_t sz, size_t *pos)
                   1429: {
1.8       kristaps 1430:        int      c;
1.1       kristaps 1431:
                   1432:        switch (cmd) {
                   1433:        case (TEXICMD_EMAIL):
1.5       kristaps 1434:                teximacroopen(p, "Mt");
1.1       kristaps 1435:                break;
1.3       kristaps 1436:        case (TEXICMD_UREF):
1.1       kristaps 1437:        case (TEXICMD_URL):
1.5       kristaps 1438:                teximacroopen(p, "Lk");
1.1       kristaps 1439:                break;
1.8       kristaps 1440:        case (TEXICMD_XREF):
                   1441:                texiputchars(p, "See Section");
                   1442:                teximacroopen(p, "Qq");
                   1443:                break;
                   1444:        case (TEXICMD_PXREF):
                   1445:                texiputchars(p, "see Section");
                   1446:                teximacroopen(p, "Qq");
                   1447:                break;
1.1       kristaps 1448:        default:
1.8       kristaps 1449:                abort();
1.1       kristaps 1450:        }
1.8       kristaps 1451:
                   1452:        c = parsearg(p, buf, sz, pos, 0);
                   1453:        p->ign++;
                   1454:        while (c > 0)
                   1455:                c = parsearg(p, buf, sz, pos, 1);
                   1456:        p->ign--;
                   1457:
1.11      kristaps 1458:        texipunctuate(p, buf, sz, pos);
1.8       kristaps 1459:        teximacroclose(p);
                   1460: }
                   1461:
                   1462: static void
                   1463: doignargn(struct texi *p, enum texicmd cmd,
                   1464:        const char *buf, size_t sz, size_t *pos)
                   1465: {
                   1466:        int      c;
                   1467:
                   1468:        c = parsearg(p, buf, sz, pos, 0);
                   1469:        p->ign++;
                   1470:        while (c > 0)
                   1471:                c = parsearg(p, buf, sz, pos, 1);
                   1472:        p->ign--;
1.1       kristaps 1473: }
                   1474:
                   1475: static void
                   1476: dosubsection(struct texi *p, enum texicmd cmd,
                   1477:                const char *buf, size_t sz, size_t *pos)
                   1478: {
1.13      kristaps 1479:
                   1480:        if (p->outmacro)
                   1481:                texierr(p, "\"Em\" in open line scope!?");
                   1482:        else if (p->literal)
                   1483:                texierr(p, "\"Em\" in a literal scope!?");
1.1       kristaps 1484:
1.11      kristaps 1485:        texivspace(p);
1.5       kristaps 1486:        teximacroopen(p, "Em");
1.3       kristaps 1487:        parseeoln(p, buf, sz, pos);
1.5       kristaps 1488:        teximacroclose(p);
1.11      kristaps 1489:        texivspace(p);
1.1       kristaps 1490: }
                   1491:
                   1492: static void
                   1493: dosection(struct texi *p, enum texicmd cmd,
                   1494:                const char *buf, size_t sz, size_t *pos)
                   1495: {
1.12      kristaps 1496:        const char      *blk;
                   1497:
                   1498:        switch (cmd) {
                   1499:        case (TEXICMD_APPENDIX):
                   1500:        case (TEXICMD_CHAPTER):
                   1501:        case (TEXICMD_TOP):
                   1502:        case (TEXICMD_UNNUMBERED):
                   1503:                blk = "Sh";
                   1504:                break;
                   1505:        case (TEXICMD_APPENDIXSEC):
                   1506:        case (TEXICMD_HEADING):
                   1507:        case (TEXICMD_SECTION):
                   1508:        case (TEXICMD_UNNUMBEREDSEC):
                   1509:                blk = "Ss";
                   1510:                break;
                   1511:        default:
                   1512:                abort();
                   1513:        }
1.1       kristaps 1514:
1.3       kristaps 1515:        if (p->outmacro)
1.12      kristaps 1516:                texierr(p, "%s in open line scope!?", blk);
1.3       kristaps 1517:        else if (p->literal)
1.12      kristaps 1518:                texierr(p, "%s in a literal scope!?", blk);
1.3       kristaps 1519:
1.12      kristaps 1520:        teximacroopen(p, blk);
1.3       kristaps 1521:        parseeoln(p, buf, sz, pos);
                   1522:        teximacroclose(p);
1.11      kristaps 1523:        p->seenvs = 1;
1.3       kristaps 1524: }
                   1525:
                   1526: static void
                   1527: dosp(struct texi *p, enum texicmd cmd,
                   1528:        const char *buf, size_t sz, size_t *pos)
                   1529: {
                   1530:
1.11      kristaps 1531:        texivspace(p);
1.12      kristaps 1532:        /* FIXME: ignore and parseeoln. */
1.3       kristaps 1533:        advanceeoln(p, buf, sz, pos, 1);
1.1       kristaps 1534: }
                   1535:
                   1536: static void
                   1537: dotop(struct texi *p, enum texicmd cmd,
                   1538:        const char *buf, size_t sz, size_t *pos)
                   1539: {
1.10      kristaps 1540:        const char      *cp;
                   1541:        time_t           t;
                   1542:        char             date[32];
                   1543:
                   1544:        t = time(NULL);
                   1545:        strftime(date, sizeof(date), "%F", localtime(&t));
1.1       kristaps 1546:
1.3       kristaps 1547:        p->ign--;
1.10      kristaps 1548:        teximacroopen(p, "Dd");
                   1549:        texiputchars(p, date);
                   1550:        teximacroclose(p);
                   1551:        teximacroopen(p, "Dt");
                   1552:        for (cp = p->title; '\0' != *cp; cp++)
                   1553:                texiputchar(p, toupper(*cp));
1.11      kristaps 1554:        texiputchars(p, " 7");
1.10      kristaps 1555:        teximacroclose(p);
1.5       kristaps 1556:        teximacro(p, "Os");
                   1557:        teximacro(p, "Sh NAME");
1.10      kristaps 1558:        teximacroopen(p, "Nm");
                   1559:        texiputchars(p, p->title);
                   1560:        teximacroclose(p);
                   1561:        teximacroopen(p, "Nd");
                   1562:        texiputchars(p, NULL != p->subtitle ?
                   1563:                p->subtitle : "Unknown description");
                   1564:        teximacroclose(p);
1.11      kristaps 1565:        p->seenvs = 1;
1.12      kristaps 1566:        dosection(p, cmd, buf, sz, pos);
1.1       kristaps 1567: }
                   1568:
                   1569: static void
                   1570: doitem(struct texi *p, enum texicmd cmd,
                   1571:        const char *buf, size_t sz, size_t *pos)
                   1572: {
                   1573:
1.3       kristaps 1574:        if (p->outmacro)
                   1575:                texierr(p, "item in open line scope!?");
                   1576:        else if (p->literal)
                   1577:                texierr(p, "item in a literal scope!?");
                   1578:
                   1579:        switch (p->list) {
                   1580:        case (TEXILIST_ITEM):
1.5       kristaps 1581:                teximacroopen(p, "It");
1.3       kristaps 1582:                break;
                   1583:        case (TEXILIST_NOITEM):
1.5       kristaps 1584:                teximacro(p, "It");
1.3       kristaps 1585:                break;
                   1586:        default:
1.11      kristaps 1587:                texivspace(p);
1.3       kristaps 1588:                break;
                   1589:        }
1.11      kristaps 1590:        p->seenvs = 1;
1.3       kristaps 1591:
                   1592:        parseeoln(p, buf, sz, pos);
1.1       kristaps 1593:
1.3       kristaps 1594:        if (TEXILIST_ITEM == p->list)
                   1595:                teximacroclose(p);
1.9       kristaps 1596:        else if (p->outcol > 0)
1.1       kristaps 1597:                texiputchar(p, '\n');
                   1598: }
                   1599:
                   1600: static void
                   1601: dotable(struct texi *p, enum texicmd cmd,
                   1602:        const char *buf, size_t sz, size_t *pos)
                   1603: {
1.3       kristaps 1604:        enum texilist   sv = p->list;
                   1605:
                   1606:        p->list = TEXILIST_ITEM;
1.5       kristaps 1607:        teximacro(p, "Bl -tag -width Ds");
1.12      kristaps 1608:        /* FIXME: ignore and parseeoln. */
                   1609:        advanceeoln(p, buf, sz, pos, 1);
1.11      kristaps 1610:        p->seenvs = 1;
1.1       kristaps 1611:        parseto(p, buf, sz, pos, "table");
1.5       kristaps 1612:        teximacro(p, "El");
1.3       kristaps 1613:        p->list = sv;
1.1       kristaps 1614: }
                   1615:
                   1616: static void
1.2       kristaps 1617: doenumerate(struct texi *p, enum texicmd cmd,
                   1618:        const char *buf, size_t sz, size_t *pos)
                   1619: {
1.3       kristaps 1620:        enum texilist    sv = p->list;
1.2       kristaps 1621:
1.3       kristaps 1622:        p->list = TEXILIST_NOITEM;
1.5       kristaps 1623:        teximacro(p, "Bl -enum");
1.11      kristaps 1624:        p->seenvs = 1;
1.12      kristaps 1625:        /* FIXME: ignore and parseeoln. */
                   1626:        advanceeoln(p, buf, sz, pos, 1);
1.2       kristaps 1627:        parseto(p, buf, sz, pos, "enumerate");
1.5       kristaps 1628:        teximacro(p, "El");
1.3       kristaps 1629:        p->list = sv;
1.2       kristaps 1630: }
                   1631:
                   1632: static void
1.1       kristaps 1633: doitemize(struct texi *p, enum texicmd cmd,
                   1634:        const char *buf, size_t sz, size_t *pos)
                   1635: {
1.3       kristaps 1636:        enum texilist   sv = p->list;
1.1       kristaps 1637:
1.3       kristaps 1638:        p->list = TEXILIST_ITEM;
1.5       kristaps 1639:        teximacro(p, "Bl -bullet");
1.11      kristaps 1640:        p->seenvs = 1;
1.12      kristaps 1641:        /* FIXME: ignore and parseeoln. */
                   1642:        advanceeoln(p, buf, sz, pos, 1);
1.1       kristaps 1643:        parseto(p, buf, sz, pos, "itemize");
1.5       kristaps 1644:        teximacro(p, "El");
1.3       kristaps 1645:        p->list = sv;
1.1       kristaps 1646: }
                   1647:
                   1648: static void
                   1649: doignbracket(struct texi *p, enum texicmd cmd,
                   1650:        const char *buf, size_t sz, size_t *pos)
                   1651: {
                   1652:
1.3       kristaps 1653:        p->ign++;
1.1       kristaps 1654:        parsebracket(p, buf, sz, pos);
1.3       kristaps 1655:        p->ign--;
1.1       kristaps 1656: }
                   1657:
                   1658: static void
                   1659: doignline(struct texi *p, enum texicmd cmd,
                   1660:        const char *buf, size_t sz, size_t *pos)
                   1661: {
                   1662:
1.12      kristaps 1663:        /* FIXME: ignore and parseeoln. */
1.3       kristaps 1664:        advanceeoln(p, buf, sz, pos, 1);
1.1       kristaps 1665: }
                   1666:
1.8       kristaps 1667: /*
                   1668:  * Parse colon-separated directories from "cp" (if not NULL) and returns
                   1669:  * the array of pointers.
                   1670:  * Prepends "base" to the array.
                   1671:  * This does NOT sanitise the directories!
                   1672:  */
1.5       kristaps 1673: static char **
                   1674: parsedirs(const char *base, const char *cp, size_t *sz)
                   1675: {
                   1676:        char             *tok, *str, *tofree;
                   1677:        const char       *cpp;
                   1678:        size_t            i;
                   1679:        char            **dirs;
                   1680:
                   1681:        *sz = NULL != (cpp = cp) ? 2 : 1;
                   1682:        if (*sz > 1)
                   1683:                for ( ; NULL != (cpp = strchr(cpp, ':')); (*sz)++)
                   1684:                        cpp++;
                   1685:
                   1686:        dirs = calloc(*sz, sizeof(char *));
                   1687:        if (NULL == dirs) {
                   1688:                perror(NULL);
                   1689:                exit(EXIT_FAILURE);
                   1690:        } else if (NULL == (dirs[0] = strdup(base))) {
                   1691:                perror(NULL);
                   1692:                exit(EXIT_FAILURE);
                   1693:        }
                   1694:
                   1695:        if (NULL == cp)
                   1696:                return(dirs);
                   1697:
                   1698:        if (NULL == (tofree = tok = str = strdup(cp))) {
                   1699:                perror(NULL);
                   1700:                exit(EXIT_FAILURE);
                   1701:        }
                   1702:
                   1703:        for (i = 1; NULL != (tok = strsep(&str, ":")); i++)
                   1704:                if (NULL == (dirs[i] = strdup(tok))) {
                   1705:                        perror(NULL);
                   1706:                        exit(EXIT_FAILURE);
                   1707:                }
                   1708:
                   1709:        free(tofree);
                   1710:        return(dirs);
                   1711: }
                   1712:
1.1       kristaps 1713: int
                   1714: main(int argc, char *argv[])
                   1715: {
                   1716:        struct texi      texi;
1.2       kristaps 1717:        int              c;
                   1718:        char            *path, *dir;
1.10      kristaps 1719:        const char      *progname, *Idir, *cp;
1.1       kristaps 1720:
                   1721:        progname = strrchr(argv[0], '/');
                   1722:        if (progname == NULL)
                   1723:                progname = argv[0];
                   1724:        else
                   1725:                ++progname;
                   1726:
1.10      kristaps 1727:        memset(&texi, 0, sizeof(struct texi));
1.5       kristaps 1728:        Idir = NULL;
1.10      kristaps 1729:
1.5       kristaps 1730:        while (-1 != (c = getopt(argc, argv, "I:")))
1.1       kristaps 1731:                switch (c) {
1.5       kristaps 1732:                case ('I'):
                   1733:                        Idir = optarg;
                   1734:                        break;
1.1       kristaps 1735:                default:
                   1736:                        goto usage;
                   1737:                }
                   1738:
                   1739:        argv += optind;
                   1740:        if (0 == (argc -= optind))
                   1741:                goto usage;
                   1742:
1.2       kristaps 1743:        if (NULL == (path = strdup(argv[0]))) {
                   1744:                perror(NULL);
                   1745:                exit(EXIT_FAILURE);
                   1746:        } else if (NULL == (dir = dirname(path))) {
                   1747:                perror(argv[0]);
                   1748:                free(path);
                   1749:                exit(EXIT_FAILURE);
                   1750:        }
                   1751:        free(path);
                   1752:
1.10      kristaps 1753:        if (NULL != (cp = strrchr(argv[0], '/')))
                   1754:                texi.title = strdup(cp + 1);
                   1755:        else
                   1756:                texi.title = strdup(argv[0]);
                   1757:
                   1758:        if (NULL == texi.title) {
                   1759:                perror(NULL);
                   1760:                exit(EXIT_FAILURE);
                   1761:        } else if (NULL != (path = strchr(texi.title, '.')))
                   1762:                *path = '\0';
                   1763:
1.3       kristaps 1764:        texi.ign = 1;
1.5       kristaps 1765:        texi.dirs = parsedirs(dir, Idir, &texi.dirsz);
1.2       kristaps 1766:        parsefile(&texi, argv[0]);
1.5       kristaps 1767:        /* We shouldn't get here. */
1.2       kristaps 1768:        texiexit(&texi);
                   1769:        return(EXIT_FAILURE);
1.1       kristaps 1770: usage:
1.8       kristaps 1771:        fprintf(stderr, "usage: %s [-Idirs] file\n", progname);
1.1       kristaps 1772:        return(EXIT_FAILURE);
                   1773: }

CVSweb