Annotation of mandoc/mandoc.h, Revision 1.118
1.118 ! schwarze 1: /* $Id: mandoc.h,v 1.117 2014/03/19 22:20:43 schwarze Exp $ */
1.1 kristaps 2: /*
1.46 schwarze 3: * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.118 ! schwarze 4: * Copyright (c) 2010-2014 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
9: *
10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17: */
18: #ifndef MANDOC_H
19: #define MANDOC_H
1.7 kristaps 20:
21: #define ASCII_NBRSP 31 /* non-breaking space */
22: #define ASCII_HYPH 30 /* breakable hyphen */
1.115 schwarze 23: #define ASCII_BREAK 29 /* breakable zero-width space */
1.7 kristaps 24:
1.22 kristaps 25: /*
26: * Status level. This refers to both internal status (i.e., whilst
27: * running, when warnings/errors are reported) and an indicator of a
28: * threshold of when to halt (when said internal state exceeds the
29: * threshold).
30: */
1.18 schwarze 31: enum mandoclevel {
32: MANDOCLEVEL_OK = 0,
33: MANDOCLEVEL_RESERVED,
1.22 kristaps 34: MANDOCLEVEL_WARNING, /* warnings: syntax, whitespace, etc. */
35: MANDOCLEVEL_ERROR, /* input has been thrown away */
36: MANDOCLEVEL_FATAL, /* input is borked */
37: MANDOCLEVEL_BADARG, /* bad argument in invocation */
38: MANDOCLEVEL_SYSERR, /* system error */
1.18 schwarze 39: MANDOCLEVEL_MAX
40: };
41:
1.22 kristaps 42: /*
43: * All possible things that can go wrong within a parse, be it libroff,
44: * libmdoc, or libman.
45: */
1.1 kristaps 46: enum mandocerr {
47: MANDOCERR_OK,
1.13 schwarze 48:
1.21 kristaps 49: MANDOCERR_WARNING, /* ===== start of warnings ===== */
1.32 kristaps 50:
51: /* related to the prologue */
52: MANDOCERR_NOTITLE, /* no title in document */
53: MANDOCERR_UPPERCASE, /* document title should be all caps */
54: MANDOCERR_BADMSEC, /* unknown manual section */
1.105 schwarze 55: MANDOCERR_BADVOLARCH, /* unknown manual volume or arch */
1.58 schwarze 56: MANDOCERR_NODATE, /* date missing, using today's date */
57: MANDOCERR_BADDATE, /* cannot parse date, using it verbatim */
1.32 kristaps 58: MANDOCERR_PROLOGOOO, /* prologue macros out of order */
59: MANDOCERR_PROLOGREP, /* duplicate prologue macro */
60: MANDOCERR_BADPROLOG, /* macro not allowed in prologue */
61: MANDOCERR_BADBODY, /* macro not allowed in body */
62:
63: /* related to document structure */
1.26 kristaps 64: MANDOCERR_SO, /* .so is fragile, better use ln(1) */
1.32 kristaps 65: MANDOCERR_NAMESECFIRST, /* NAME section must come first */
66: MANDOCERR_BADNAMESEC, /* bad NAME section contents */
1.9 kristaps 67: MANDOCERR_SECOOO, /* sections out of conventional order */
1.32 kristaps 68: MANDOCERR_SECREP, /* duplicate section name */
1.110 schwarze 69: MANDOCERR_SECMSEC, /* section header suited to sections ... */
1.32 kristaps 70:
71: /* related to macros and nesting */
72: MANDOCERR_MACROOBS, /* skipping obsolete macro */
1.35 kristaps 73: MANDOCERR_IGNPAR, /* skipping paragraph macro */
1.104 schwarze 74: MANDOCERR_MOVEPAR, /* moving paragraph macro out of list */
1.54 kristaps 75: MANDOCERR_IGNNS, /* skipping no-space macro */
1.14 schwarze 76: MANDOCERR_SCOPENEST, /* blocks badly nested */
1.32 kristaps 77: MANDOCERR_CHILD, /* child violates parent syntax */
78: MANDOCERR_NESTEDDISP, /* nested displays are not portable */
79: MANDOCERR_SCOPEREP, /* already in literal mode */
1.68 kristaps 80: MANDOCERR_LINESCOPE, /* line scope broken */
1.32 kristaps 81:
82: /* related to missing macro arguments */
83: MANDOCERR_MACROEMPTY, /* skipping empty macro */
1.46 schwarze 84: MANDOCERR_ARGCWARN, /* argument count wrong */
1.32 kristaps 85: MANDOCERR_DISPTYPE, /* missing display type */
86: MANDOCERR_LISTFIRST, /* list type must come first */
87: MANDOCERR_NOWIDTHARG, /* tag lists require a width argument */
88: MANDOCERR_FONTTYPE, /* missing font type */
1.53 kristaps 89: MANDOCERR_WNOSCOPE, /* skipping end of block that is not open */
1.13 schwarze 90:
1.32 kristaps 91: /* related to bad macro arguments */
92: MANDOCERR_IGNARGV, /* skipping argument */
93: MANDOCERR_ARGVREP, /* duplicate argument */
94: MANDOCERR_DISPREP, /* duplicate display type */
95: MANDOCERR_LISTREP, /* duplicate list type */
96: MANDOCERR_BADATT, /* unknown AT&T UNIX version */
1.5 kristaps 97: MANDOCERR_BADBOOL, /* bad Boolean value */
1.31 kristaps 98: MANDOCERR_BADFONT, /* unknown font */
1.32 kristaps 99: MANDOCERR_BADSTANDARD, /* unknown standard specifier */
100: MANDOCERR_BADWIDTH, /* bad width argument */
101:
102: /* related to plain text */
103: MANDOCERR_NOBLANKLN, /* blank line in non-literal context */
104: MANDOCERR_BADTAB, /* tab in non-literal context */
105: MANDOCERR_EOLNSPACE, /* end of line whitespace */
1.5 kristaps 106: MANDOCERR_BADCOMMENT, /* bad comment style */
1.32 kristaps 107: MANDOCERR_BADESCAPE, /* unknown escape sequence */
108: MANDOCERR_BADQUOTE, /* unterminated quoted string */
1.79 kristaps 109:
110: /* related to equations */
111: MANDOCERR_EQNQUOTE, /* unexpected literal in equation */
1.32 kristaps 112:
113: MANDOCERR_ERROR, /* ===== start of errors ===== */
1.80 kristaps 114:
115: /* related to equations */
1.84 kristaps 116: MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
117: MANDOCERR_EQNSCOPE, /* equation scope open on exit */
1.89 kristaps 118: MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
119: MANDOCERR_EQNEOF, /* unexpected end of equation */
120: MANDOCERR_EQNSYNT, /* equation syntax error */
1.32 kristaps 121:
1.42 kristaps 122: /* related to tables */
1.36 kristaps 123: MANDOCERR_TBL, /* bad table syntax */
124: MANDOCERR_TBLOPT, /* bad table option */
1.37 kristaps 125: MANDOCERR_TBLLAYOUT, /* bad table layout */
126: MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
1.39 kristaps 127: MANDOCERR_TBLNODATA, /* no table data cells specified */
1.47 kristaps 128: MANDOCERR_TBLIGNDATA, /* ignore data in cell */
1.48 kristaps 129: MANDOCERR_TBLBLOCK, /* data block still open */
1.50 kristaps 130: MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
1.42 kristaps 131:
1.33 schwarze 132: MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
1.32 kristaps 133: MANDOCERR_BADCHAR, /* skipping bad character */
1.52 schwarze 134: MANDOCERR_NAMESC, /* escaped character not allowed in a name */
1.103 schwarze 135: MANDOCERR_NONAME, /* manual name not yet set */
1.32 kristaps 136: MANDOCERR_NOTEXT, /* skipping text before the first section header */
137: MANDOCERR_MACRO, /* skipping unknown macro */
1.25 kristaps 138: MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */
1.5 kristaps 139: MANDOCERR_ARGCOUNT, /* argument count wrong */
1.106 schwarze 140: MANDOCERR_STRAYTA, /* skipping column outside column list */
1.32 kristaps 141: MANDOCERR_NOSCOPE, /* skipping end of block that is not open */
1.30 kristaps 142: MANDOCERR_SCOPEBROKEN, /* missing end of block */
1.19 schwarze 143: MANDOCERR_SCOPEEXIT, /* scope open on exit */
1.28 kristaps 144: MANDOCERR_UNAME, /* uname(3) system call failed */
1.5 kristaps 145: /* FIXME: merge following with MANDOCERR_ARGCOUNT */
146: MANDOCERR_NOARGS, /* macro requires line argument(s) */
147: MANDOCERR_NOBODY, /* macro requires body argument(s) */
148: MANDOCERR_NOARGV, /* macro requires argument(s) */
1.108 schwarze 149: MANDOCERR_NUMERIC, /* request requires a numeric argument */
1.10 kristaps 150: MANDOCERR_LISTTYPE, /* missing list type */
1.5 kristaps 151: MANDOCERR_ARGSLOST, /* line argument(s) will be lost */
152: MANDOCERR_BODYLOST, /* body argument(s) will be lost */
1.13 schwarze 153:
1.21 kristaps 154: MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
1.32 kristaps 155:
1.113 schwarze 156: MANDOCERR_TOOLARGE, /* input too large */
1.66 kristaps 157: MANDOCERR_NOTMANUAL, /* manual isn't really a manual */
1.8 kristaps 158: MANDOCERR_COLUMNS, /* column syntax is inconsistent */
1.32 kristaps 159: MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
1.5 kristaps 160: MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */
161: MANDOCERR_SYNTCHILD, /* child violates parent syntax */
162: MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
1.26 kristaps 163: MANDOCERR_SOPATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
1.5 kristaps 164: MANDOCERR_NODOCBODY, /* no document body */
165: MANDOCERR_NODOCPROLOG, /* no document prologue */
1.18 schwarze 166: MANDOCERR_MEM, /* static buffer exhausted */
1.113 schwarze 167:
168: /* ===== system errors ===== */
169:
170: MANDOCERR_SYSOPEN, /* cannot open file */
171: MANDOCERR_SYSSTAT, /* cannot stat file */
172: MANDOCERR_SYSREAD, /* cannot read file */
173:
1.1 kristaps 174: MANDOCERR_MAX
1.43 kristaps 175: };
176:
1.107 schwarze 177: struct tbl_opts {
1.43 kristaps 178: char tab; /* cell-separator */
179: char decimal; /* decimal point */
180: int linesize;
181: int opts;
182: #define TBL_OPT_CENTRE (1 << 0)
183: #define TBL_OPT_EXPAND (1 << 1)
184: #define TBL_OPT_BOX (1 << 2)
185: #define TBL_OPT_DBOX (1 << 3)
186: #define TBL_OPT_ALLBOX (1 << 4)
187: #define TBL_OPT_NOKEEP (1 << 5)
188: #define TBL_OPT_NOSPACE (1 << 6)
1.45 kristaps 189: int cols; /* number of columns */
1.38 kristaps 190: };
191:
1.41 kristaps 192: /*
193: * The head of a table specifies all of its columns. When formatting a
194: * tbl_span, iterate over these and plug in data from the tbl_span when
195: * appropriate, using tbl_cell as a guide to placement.
196: */
197: struct tbl_head {
1.45 kristaps 198: int ident; /* 0 <= unique id < cols */
1.101 schwarze 199: int vert; /* width of preceding vertical line */
1.41 kristaps 200: struct tbl_head *next;
201: struct tbl_head *prev;
202: };
203:
1.38 kristaps 204: enum tbl_cellt {
205: TBL_CELL_CENTRE, /* c, C */
206: TBL_CELL_RIGHT, /* r, R */
207: TBL_CELL_LEFT, /* l, L */
208: TBL_CELL_NUMBER, /* n, N */
209: TBL_CELL_SPAN, /* s, S */
210: TBL_CELL_LONG, /* a, A */
211: TBL_CELL_DOWN, /* ^ */
212: TBL_CELL_HORIZ, /* _, - */
213: TBL_CELL_DHORIZ, /* = */
214: TBL_CELL_MAX
215: };
216:
217: /*
218: * A cell in a layout row.
219: */
220: struct tbl_cell {
221: struct tbl_cell *next;
1.101 schwarze 222: int vert; /* width of preceding vertical line */
1.38 kristaps 223: enum tbl_cellt pos;
1.51 kristaps 224: size_t spacing;
1.38 kristaps 225: int flags;
226: #define TBL_CELL_TALIGN (1 << 0) /* t, T */
227: #define TBL_CELL_BALIGN (1 << 1) /* d, D */
228: #define TBL_CELL_BOLD (1 << 2) /* fB, B, b */
229: #define TBL_CELL_ITALIC (1 << 3) /* fI, I, i */
230: #define TBL_CELL_EQUAL (1 << 4) /* e, E */
231: #define TBL_CELL_UP (1 << 5) /* u, U */
232: #define TBL_CELL_WIGN (1 << 6) /* z, Z */
1.41 kristaps 233: struct tbl_head *head;
1.38 kristaps 234: };
235:
236: /*
237: * A layout row.
238: */
239: struct tbl_row {
240: struct tbl_row *next;
241: struct tbl_cell *first;
242: struct tbl_cell *last;
243: };
244:
1.40 kristaps 245: enum tbl_datt {
1.51 kristaps 246: TBL_DATA_NONE, /* has no data */
247: TBL_DATA_DATA, /* consists of data/string */
248: TBL_DATA_HORIZ, /* horizontal line */
249: TBL_DATA_DHORIZ, /* double-horizontal line */
250: TBL_DATA_NHORIZ, /* squeezed horizontal line */
251: TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
1.40 kristaps 252: };
253:
1.38 kristaps 254: /*
255: * A cell within a row of data. The "string" field contains the actual
256: * string value that's in the cell. The rest is layout.
257: */
258: struct tbl_dat {
1.51 kristaps 259: struct tbl_cell *layout; /* layout cell */
1.50 kristaps 260: int spans; /* how many spans follow */
1.38 kristaps 261: struct tbl_dat *next;
1.51 kristaps 262: char *string; /* data (NULL if not TBL_DATA_DATA) */
1.40 kristaps 263: enum tbl_datt pos;
264: };
265:
266: enum tbl_spant {
267: TBL_SPAN_DATA, /* span consists of data */
268: TBL_SPAN_HORIZ, /* span is horizontal line */
269: TBL_SPAN_DHORIZ /* span is double horizontal line */
1.38 kristaps 270: };
271:
272: /*
273: * A row of data in a table.
274: */
275: struct tbl_span {
1.107 schwarze 276: struct tbl_opts *opts;
1.44 kristaps 277: struct tbl_head *head;
1.51 kristaps 278: struct tbl_row *layout; /* layout row */
1.38 kristaps 279: struct tbl_dat *first;
280: struct tbl_dat *last;
1.56 kristaps 281: int line; /* parse line */
1.44 kristaps 282: int flags;
283: #define TBL_SPAN_FIRST (1 << 0)
284: #define TBL_SPAN_LAST (1 << 1)
1.40 kristaps 285: enum tbl_spant pos;
1.38 kristaps 286: struct tbl_span *next;
1.55 kristaps 287: };
288:
1.83 kristaps 289: enum eqn_boxt {
1.84 kristaps 290: EQN_ROOT, /* root of parse tree */
291: EQN_TEXT, /* text (number, variable, whatever) */
1.90 kristaps 292: EQN_SUBEXPR, /* nested `eqn' subexpression */
1.92 kristaps 293: EQN_LIST, /* subexpressions list */
294: EQN_MATRIX /* matrix subexpression */
1.83 kristaps 295: };
296:
1.85 kristaps 297: enum eqn_markt {
298: EQNMARK_NONE = 0,
299: EQNMARK_DOT,
300: EQNMARK_DOTDOT,
301: EQNMARK_HAT,
302: EQNMARK_TILDE,
303: EQNMARK_VEC,
304: EQNMARK_DYAD,
305: EQNMARK_BAR,
306: EQNMARK_UNDER,
307: EQNMARK__MAX
308: };
309:
1.86 kristaps 310: enum eqn_fontt {
311: EQNFONT_NONE = 0,
312: EQNFONT_ROMAN,
313: EQNFONT_BOLD,
1.93 kristaps 314: EQNFONT_FAT,
1.86 kristaps 315: EQNFONT_ITALIC,
316: EQNFONT__MAX
317: };
318:
1.87 kristaps 319: enum eqn_post {
320: EQNPOS_NONE = 0,
321: EQNPOS_OVER,
322: EQNPOS_SUP,
323: EQNPOS_SUB,
324: EQNPOS_TO,
325: EQNPOS_FROM,
326: EQNPOS__MAX
327: };
328:
1.89 kristaps 329: enum eqn_pilet {
330: EQNPILE_NONE = 0,
1.95 kristaps 331: EQNPILE_PILE,
1.89 kristaps 332: EQNPILE_CPILE,
333: EQNPILE_RPILE,
334: EQNPILE_LPILE,
1.95 kristaps 335: EQNPILE_COL,
1.92 kristaps 336: EQNPILE_CCOL,
337: EQNPILE_RCOL,
338: EQNPILE_LCOL,
1.89 kristaps 339: EQNPILE__MAX
340: };
341:
1.86 kristaps 342: /*
1.84 kristaps 343: * A "box" is a parsed mathematical expression as defined by the eqn.7
344: * grammar.
345: */
1.83 kristaps 346: struct eqn_box {
1.88 kristaps 347: int size; /* font size of expression */
348: #define EQN_DEFSIZE INT_MIN
1.84 kristaps 349: enum eqn_boxt type; /* type of node */
1.89 kristaps 350: struct eqn_box *first; /* first child node */
351: struct eqn_box *last; /* last child node */
352: struct eqn_box *next; /* node sibling */
353: struct eqn_box *parent; /* node sibling */
354: char *text; /* text (or NULL) */
355: char *left;
356: char *right;
1.87 kristaps 357: enum eqn_post pos; /* position of next box */
1.86 kristaps 358: enum eqn_markt mark; /* a mark about the box */
359: enum eqn_fontt font; /* font of box */
1.89 kristaps 360: enum eqn_pilet pile; /* equation piling */
1.83 kristaps 361: };
362:
1.87 kristaps 363: /*
364: * An equation consists of a tree of expressions starting at a given
365: * line and position.
366: */
1.55 kristaps 367: struct eqn {
1.94 kristaps 368: char *name; /* identifier (or NULL) */
1.84 kristaps 369: struct eqn_box *root; /* root mathematical expression */
1.82 kristaps 370: int ln; /* invocation line */
1.57 kristaps 371: int pos; /* invocation position */
1.16 kristaps 372: };
1.22 kristaps 373:
1.60 kristaps 374: /*
1.116 schwarze 375: * Parse options.
376: */
377: #define MPARSE_MDOC 1 /* assume -mdoc */
378: #define MPARSE_MAN 2 /* assume -man */
379: #define MPARSE_SO 4 /* honour .so requests */
380: #define MPARSE_QUICK 8 /* abort the parse early */
1.62 kristaps 381:
1.70 kristaps 382: enum mandoc_esc {
383: ESCAPE_ERROR = 0, /* bail! unparsable escape */
384: ESCAPE_IGNORE, /* escape to be ignored */
385: ESCAPE_SPECIAL, /* a regular special character */
386: ESCAPE_FONT, /* a generic font mode */
387: ESCAPE_FONTBOLD, /* bold font mode */
388: ESCAPE_FONTITALIC, /* italic font mode */
1.109 schwarze 389: ESCAPE_FONTBI, /* bold italic font mode */
1.70 kristaps 390: ESCAPE_FONTROMAN, /* roman font mode */
391: ESCAPE_FONTPREV, /* previous font mode */
392: ESCAPE_NUMBERED, /* a numbered glyph */
1.75 kristaps 393: ESCAPE_UNICODE, /* a unicode codepoint */
1.102 schwarze 394: ESCAPE_NOSPACE, /* suppress space if the last on a line */
395: ESCAPE_SKIPCHAR /* skip the next character */
1.70 kristaps 396: };
397:
1.63 kristaps 398: typedef void (*mandocmsg)(enum mandocerr, enum mandoclevel,
399: const char *, int, int, const char *);
1.62 kristaps 400:
401: struct mparse;
1.71 kristaps 402: struct mchars;
1.62 kristaps 403: struct mdoc;
404: struct man;
1.59 kristaps 405:
1.22 kristaps 406: __BEGIN_DECLS
1.62 kristaps 407:
1.112 schwarze 408: enum mandoc_esc mandoc_escape(const char **, const char **, int *);
1.73 kristaps 409: struct mchars *mchars_alloc(void);
1.96 kristaps 410: void mchars_free(struct mchars *);
1.74 kristaps 411: char mchars_num2char(const char *, size_t);
1.76 kristaps 412: int mchars_num2uc(const char *, size_t);
1.98 kristaps 413: int mchars_spec2cp(const struct mchars *,
1.96 kristaps 414: const char *, size_t);
1.98 kristaps 415: const char *mchars_spec2str(const struct mchars *,
1.96 kristaps 416: const char *, size_t, size_t *);
1.116 schwarze 417: struct mparse *mparse_alloc(int, enum mandoclevel, mandocmsg, char *);
1.96 kristaps 418: void mparse_free(struct mparse *);
419: void mparse_keep(struct mparse *);
420: enum mandoclevel mparse_readfd(struct mparse *, int, const char *);
1.99 joerg 421: enum mandoclevel mparse_readmem(struct mparse *, const void *, size_t,
422: const char *);
1.96 kristaps 423: void mparse_reset(struct mparse *);
424: void mparse_result(struct mparse *,
1.117 schwarze 425: struct mdoc **, struct man **, char **);
1.96 kristaps 426: const char *mparse_getkeep(const struct mparse *);
427: const char *mparse_strerror(enum mandocerr);
428: const char *mparse_strlevel(enum mandoclevel);
1.1 kristaps 429:
430: __END_DECLS
431:
432: #endif /*!MANDOC_H*/
CVSweb