Annotation of mandoc/mandoc.h, Revision 1.112
1.112 ! schwarze 1: /* $Id: mandoc.h,v 1.111 2013/10/05 20:30:05 schwarze Exp $ */
1.1 kristaps 2: /*
1.46 schwarze 3: * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
1.108 schwarze 4: * Copyright (c) 2012, 2013 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 */
23:
1.22 kristaps 24: /*
25: * Status level. This refers to both internal status (i.e., whilst
26: * running, when warnings/errors are reported) and an indicator of a
27: * threshold of when to halt (when said internal state exceeds the
28: * threshold).
29: */
1.18 schwarze 30: enum mandoclevel {
31: MANDOCLEVEL_OK = 0,
32: MANDOCLEVEL_RESERVED,
1.22 kristaps 33: MANDOCLEVEL_WARNING, /* warnings: syntax, whitespace, etc. */
34: MANDOCLEVEL_ERROR, /* input has been thrown away */
35: MANDOCLEVEL_FATAL, /* input is borked */
36: MANDOCLEVEL_BADARG, /* bad argument in invocation */
37: MANDOCLEVEL_SYSERR, /* system error */
1.18 schwarze 38: MANDOCLEVEL_MAX
39: };
40:
1.22 kristaps 41: /*
42: * All possible things that can go wrong within a parse, be it libroff,
43: * libmdoc, or libman.
44: */
1.1 kristaps 45: enum mandocerr {
46: MANDOCERR_OK,
1.13 schwarze 47:
1.21 kristaps 48: MANDOCERR_WARNING, /* ===== start of warnings ===== */
1.32 kristaps 49:
50: /* related to the prologue */
51: MANDOCERR_NOTITLE, /* no title in document */
52: MANDOCERR_UPPERCASE, /* document title should be all caps */
53: MANDOCERR_BADMSEC, /* unknown manual section */
1.105 schwarze 54: MANDOCERR_BADVOLARCH, /* unknown manual volume or arch */
1.58 schwarze 55: MANDOCERR_NODATE, /* date missing, using today's date */
56: MANDOCERR_BADDATE, /* cannot parse date, using it verbatim */
1.32 kristaps 57: MANDOCERR_PROLOGOOO, /* prologue macros out of order */
58: MANDOCERR_PROLOGREP, /* duplicate prologue macro */
59: MANDOCERR_BADPROLOG, /* macro not allowed in prologue */
60: MANDOCERR_BADBODY, /* macro not allowed in body */
61:
62: /* related to document structure */
1.26 kristaps 63: MANDOCERR_SO, /* .so is fragile, better use ln(1) */
1.32 kristaps 64: MANDOCERR_NAMESECFIRST, /* NAME section must come first */
65: MANDOCERR_BADNAMESEC, /* bad NAME section contents */
1.9 kristaps 66: MANDOCERR_SECOOO, /* sections out of conventional order */
1.32 kristaps 67: MANDOCERR_SECREP, /* duplicate section name */
1.110 schwarze 68: MANDOCERR_SECMSEC, /* section header suited to sections ... */
1.32 kristaps 69:
70: /* related to macros and nesting */
71: MANDOCERR_MACROOBS, /* skipping obsolete macro */
1.35 kristaps 72: MANDOCERR_IGNPAR, /* skipping paragraph macro */
1.104 schwarze 73: MANDOCERR_MOVEPAR, /* moving paragraph macro out of list */
1.54 kristaps 74: MANDOCERR_IGNNS, /* skipping no-space macro */
1.14 schwarze 75: MANDOCERR_SCOPENEST, /* blocks badly nested */
1.32 kristaps 76: MANDOCERR_CHILD, /* child violates parent syntax */
77: MANDOCERR_NESTEDDISP, /* nested displays are not portable */
78: MANDOCERR_SCOPEREP, /* already in literal mode */
1.68 kristaps 79: MANDOCERR_LINESCOPE, /* line scope broken */
1.32 kristaps 80:
81: /* related to missing macro arguments */
82: MANDOCERR_MACROEMPTY, /* skipping empty macro */
1.46 schwarze 83: MANDOCERR_ARGCWARN, /* argument count wrong */
1.32 kristaps 84: MANDOCERR_DISPTYPE, /* missing display type */
85: MANDOCERR_LISTFIRST, /* list type must come first */
86: MANDOCERR_NOWIDTHARG, /* tag lists require a width argument */
87: MANDOCERR_FONTTYPE, /* missing font type */
1.53 kristaps 88: MANDOCERR_WNOSCOPE, /* skipping end of block that is not open */
1.13 schwarze 89:
1.32 kristaps 90: /* related to bad macro arguments */
91: MANDOCERR_IGNARGV, /* skipping argument */
92: MANDOCERR_ARGVREP, /* duplicate argument */
93: MANDOCERR_DISPREP, /* duplicate display type */
94: MANDOCERR_LISTREP, /* duplicate list type */
95: MANDOCERR_BADATT, /* unknown AT&T UNIX version */
1.5 kristaps 96: MANDOCERR_BADBOOL, /* bad Boolean value */
1.31 kristaps 97: MANDOCERR_BADFONT, /* unknown font */
1.32 kristaps 98: MANDOCERR_BADSTANDARD, /* unknown standard specifier */
99: MANDOCERR_BADWIDTH, /* bad width argument */
100:
101: /* related to plain text */
102: MANDOCERR_NOBLANKLN, /* blank line in non-literal context */
103: MANDOCERR_BADTAB, /* tab in non-literal context */
104: MANDOCERR_EOLNSPACE, /* end of line whitespace */
1.5 kristaps 105: MANDOCERR_BADCOMMENT, /* bad comment style */
1.32 kristaps 106: MANDOCERR_BADESCAPE, /* unknown escape sequence */
107: MANDOCERR_BADQUOTE, /* unterminated quoted string */
1.79 kristaps 108:
109: /* related to equations */
110: MANDOCERR_EQNQUOTE, /* unexpected literal in equation */
1.32 kristaps 111:
112: MANDOCERR_ERROR, /* ===== start of errors ===== */
1.80 kristaps 113:
114: /* related to equations */
1.84 kristaps 115: MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
116: MANDOCERR_EQNSCOPE, /* equation scope open on exit */
1.89 kristaps 117: MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
118: MANDOCERR_EQNEOF, /* unexpected end of equation */
119: MANDOCERR_EQNSYNT, /* equation syntax error */
1.32 kristaps 120:
1.42 kristaps 121: /* related to tables */
1.36 kristaps 122: MANDOCERR_TBL, /* bad table syntax */
123: MANDOCERR_TBLOPT, /* bad table option */
1.37 kristaps 124: MANDOCERR_TBLLAYOUT, /* bad table layout */
125: MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
1.39 kristaps 126: MANDOCERR_TBLNODATA, /* no table data cells specified */
1.47 kristaps 127: MANDOCERR_TBLIGNDATA, /* ignore data in cell */
1.48 kristaps 128: MANDOCERR_TBLBLOCK, /* data block still open */
1.50 kristaps 129: MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
1.42 kristaps 130:
1.33 schwarze 131: MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
1.32 kristaps 132: MANDOCERR_BADCHAR, /* skipping bad character */
1.52 schwarze 133: MANDOCERR_NAMESC, /* escaped character not allowed in a name */
1.103 schwarze 134: MANDOCERR_NONAME, /* manual name not yet set */
1.32 kristaps 135: MANDOCERR_NOTEXT, /* skipping text before the first section header */
136: MANDOCERR_MACRO, /* skipping unknown macro */
1.25 kristaps 137: MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */
1.5 kristaps 138: MANDOCERR_ARGCOUNT, /* argument count wrong */
1.106 schwarze 139: MANDOCERR_STRAYTA, /* skipping column outside column list */
1.32 kristaps 140: MANDOCERR_NOSCOPE, /* skipping end of block that is not open */
1.30 kristaps 141: MANDOCERR_SCOPEBROKEN, /* missing end of block */
1.19 schwarze 142: MANDOCERR_SCOPEEXIT, /* scope open on exit */
1.28 kristaps 143: MANDOCERR_UNAME, /* uname(3) system call failed */
1.5 kristaps 144: /* FIXME: merge following with MANDOCERR_ARGCOUNT */
145: MANDOCERR_NOARGS, /* macro requires line argument(s) */
146: MANDOCERR_NOBODY, /* macro requires body argument(s) */
147: MANDOCERR_NOARGV, /* macro requires argument(s) */
1.108 schwarze 148: MANDOCERR_NUMERIC, /* request requires a numeric argument */
1.10 kristaps 149: MANDOCERR_LISTTYPE, /* missing list type */
1.5 kristaps 150: MANDOCERR_ARGSLOST, /* line argument(s) will be lost */
151: MANDOCERR_BODYLOST, /* body argument(s) will be lost */
1.13 schwarze 152:
1.21 kristaps 153: MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
1.32 kristaps 154:
1.66 kristaps 155: MANDOCERR_NOTMANUAL, /* manual isn't really a manual */
1.8 kristaps 156: MANDOCERR_COLUMNS, /* column syntax is inconsistent */
1.32 kristaps 157: MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
1.5 kristaps 158: MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */
159: MANDOCERR_SYNTCHILD, /* child violates parent syntax */
160: MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
1.26 kristaps 161: MANDOCERR_SOPATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
1.5 kristaps 162: MANDOCERR_NODOCBODY, /* no document body */
163: MANDOCERR_NODOCPROLOG, /* no document prologue */
1.18 schwarze 164: MANDOCERR_MEM, /* static buffer exhausted */
1.1 kristaps 165: MANDOCERR_MAX
1.43 kristaps 166: };
167:
1.107 schwarze 168: struct tbl_opts {
1.43 kristaps 169: char tab; /* cell-separator */
170: char decimal; /* decimal point */
171: int linesize;
172: int opts;
173: #define TBL_OPT_CENTRE (1 << 0)
174: #define TBL_OPT_EXPAND (1 << 1)
175: #define TBL_OPT_BOX (1 << 2)
176: #define TBL_OPT_DBOX (1 << 3)
177: #define TBL_OPT_ALLBOX (1 << 4)
178: #define TBL_OPT_NOKEEP (1 << 5)
179: #define TBL_OPT_NOSPACE (1 << 6)
1.45 kristaps 180: int cols; /* number of columns */
1.38 kristaps 181: };
182:
1.41 kristaps 183: /*
184: * The head of a table specifies all of its columns. When formatting a
185: * tbl_span, iterate over these and plug in data from the tbl_span when
186: * appropriate, using tbl_cell as a guide to placement.
187: */
188: struct tbl_head {
1.45 kristaps 189: int ident; /* 0 <= unique id < cols */
1.101 schwarze 190: int vert; /* width of preceding vertical line */
1.41 kristaps 191: struct tbl_head *next;
192: struct tbl_head *prev;
193: };
194:
1.38 kristaps 195: enum tbl_cellt {
196: TBL_CELL_CENTRE, /* c, C */
197: TBL_CELL_RIGHT, /* r, R */
198: TBL_CELL_LEFT, /* l, L */
199: TBL_CELL_NUMBER, /* n, N */
200: TBL_CELL_SPAN, /* s, S */
201: TBL_CELL_LONG, /* a, A */
202: TBL_CELL_DOWN, /* ^ */
203: TBL_CELL_HORIZ, /* _, - */
204: TBL_CELL_DHORIZ, /* = */
205: TBL_CELL_MAX
206: };
207:
208: /*
209: * A cell in a layout row.
210: */
211: struct tbl_cell {
212: struct tbl_cell *next;
1.101 schwarze 213: int vert; /* width of preceding vertical line */
1.38 kristaps 214: enum tbl_cellt pos;
1.51 kristaps 215: size_t spacing;
1.38 kristaps 216: int flags;
217: #define TBL_CELL_TALIGN (1 << 0) /* t, T */
218: #define TBL_CELL_BALIGN (1 << 1) /* d, D */
219: #define TBL_CELL_BOLD (1 << 2) /* fB, B, b */
220: #define TBL_CELL_ITALIC (1 << 3) /* fI, I, i */
221: #define TBL_CELL_EQUAL (1 << 4) /* e, E */
222: #define TBL_CELL_UP (1 << 5) /* u, U */
223: #define TBL_CELL_WIGN (1 << 6) /* z, Z */
1.41 kristaps 224: struct tbl_head *head;
1.38 kristaps 225: };
226:
227: /*
228: * A layout row.
229: */
230: struct tbl_row {
231: struct tbl_row *next;
232: struct tbl_cell *first;
233: struct tbl_cell *last;
234: };
235:
1.40 kristaps 236: enum tbl_datt {
1.51 kristaps 237: TBL_DATA_NONE, /* has no data */
238: TBL_DATA_DATA, /* consists of data/string */
239: TBL_DATA_HORIZ, /* horizontal line */
240: TBL_DATA_DHORIZ, /* double-horizontal line */
241: TBL_DATA_NHORIZ, /* squeezed horizontal line */
242: TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
1.40 kristaps 243: };
244:
1.38 kristaps 245: /*
246: * A cell within a row of data. The "string" field contains the actual
247: * string value that's in the cell. The rest is layout.
248: */
249: struct tbl_dat {
1.51 kristaps 250: struct tbl_cell *layout; /* layout cell */
1.50 kristaps 251: int spans; /* how many spans follow */
1.38 kristaps 252: struct tbl_dat *next;
1.51 kristaps 253: char *string; /* data (NULL if not TBL_DATA_DATA) */
1.40 kristaps 254: enum tbl_datt pos;
255: };
256:
257: enum tbl_spant {
258: TBL_SPAN_DATA, /* span consists of data */
259: TBL_SPAN_HORIZ, /* span is horizontal line */
260: TBL_SPAN_DHORIZ /* span is double horizontal line */
1.38 kristaps 261: };
262:
263: /*
264: * A row of data in a table.
265: */
266: struct tbl_span {
1.107 schwarze 267: struct tbl_opts *opts;
1.44 kristaps 268: struct tbl_head *head;
1.51 kristaps 269: struct tbl_row *layout; /* layout row */
1.38 kristaps 270: struct tbl_dat *first;
271: struct tbl_dat *last;
1.56 kristaps 272: int line; /* parse line */
1.44 kristaps 273: int flags;
274: #define TBL_SPAN_FIRST (1 << 0)
275: #define TBL_SPAN_LAST (1 << 1)
1.40 kristaps 276: enum tbl_spant pos;
1.38 kristaps 277: struct tbl_span *next;
1.55 kristaps 278: };
279:
1.83 kristaps 280: enum eqn_boxt {
1.84 kristaps 281: EQN_ROOT, /* root of parse tree */
282: EQN_TEXT, /* text (number, variable, whatever) */
1.90 kristaps 283: EQN_SUBEXPR, /* nested `eqn' subexpression */
1.92 kristaps 284: EQN_LIST, /* subexpressions list */
285: EQN_MATRIX /* matrix subexpression */
1.83 kristaps 286: };
287:
1.85 kristaps 288: enum eqn_markt {
289: EQNMARK_NONE = 0,
290: EQNMARK_DOT,
291: EQNMARK_DOTDOT,
292: EQNMARK_HAT,
293: EQNMARK_TILDE,
294: EQNMARK_VEC,
295: EQNMARK_DYAD,
296: EQNMARK_BAR,
297: EQNMARK_UNDER,
298: EQNMARK__MAX
299: };
300:
1.86 kristaps 301: enum eqn_fontt {
302: EQNFONT_NONE = 0,
303: EQNFONT_ROMAN,
304: EQNFONT_BOLD,
1.93 kristaps 305: EQNFONT_FAT,
1.86 kristaps 306: EQNFONT_ITALIC,
307: EQNFONT__MAX
308: };
309:
1.87 kristaps 310: enum eqn_post {
311: EQNPOS_NONE = 0,
312: EQNPOS_OVER,
313: EQNPOS_SUP,
314: EQNPOS_SUB,
315: EQNPOS_TO,
316: EQNPOS_FROM,
317: EQNPOS__MAX
318: };
319:
1.89 kristaps 320: enum eqn_pilet {
321: EQNPILE_NONE = 0,
1.95 kristaps 322: EQNPILE_PILE,
1.89 kristaps 323: EQNPILE_CPILE,
324: EQNPILE_RPILE,
325: EQNPILE_LPILE,
1.95 kristaps 326: EQNPILE_COL,
1.92 kristaps 327: EQNPILE_CCOL,
328: EQNPILE_RCOL,
329: EQNPILE_LCOL,
1.89 kristaps 330: EQNPILE__MAX
331: };
332:
1.86 kristaps 333: /*
1.84 kristaps 334: * A "box" is a parsed mathematical expression as defined by the eqn.7
335: * grammar.
336: */
1.83 kristaps 337: struct eqn_box {
1.88 kristaps 338: int size; /* font size of expression */
339: #define EQN_DEFSIZE INT_MIN
1.84 kristaps 340: enum eqn_boxt type; /* type of node */
1.89 kristaps 341: struct eqn_box *first; /* first child node */
342: struct eqn_box *last; /* last child node */
343: struct eqn_box *next; /* node sibling */
344: struct eqn_box *parent; /* node sibling */
345: char *text; /* text (or NULL) */
346: char *left;
347: char *right;
1.87 kristaps 348: enum eqn_post pos; /* position of next box */
1.86 kristaps 349: enum eqn_markt mark; /* a mark about the box */
350: enum eqn_fontt font; /* font of box */
1.89 kristaps 351: enum eqn_pilet pile; /* equation piling */
1.83 kristaps 352: };
353:
1.87 kristaps 354: /*
355: * An equation consists of a tree of expressions starting at a given
356: * line and position.
357: */
1.55 kristaps 358: struct eqn {
1.94 kristaps 359: char *name; /* identifier (or NULL) */
1.84 kristaps 360: struct eqn_box *root; /* root mathematical expression */
1.82 kristaps 361: int ln; /* invocation line */
1.57 kristaps 362: int pos; /* invocation position */
1.16 kristaps 363: };
1.22 kristaps 364:
1.60 kristaps 365: /*
1.62 kristaps 366: * The type of parse sequence. This value is usually passed via the
367: * mandoc(1) command line of -man and -mdoc. It's almost exclusively
368: * -mandoc but the others have been retained for compatibility.
369: */
370: enum mparset {
371: MPARSE_AUTO, /* magically determine the document type */
372: MPARSE_MDOC, /* assume -mdoc */
373: MPARSE_MAN /* assume -man */
374: };
375:
1.70 kristaps 376: enum mandoc_esc {
377: ESCAPE_ERROR = 0, /* bail! unparsable escape */
378: ESCAPE_IGNORE, /* escape to be ignored */
379: ESCAPE_SPECIAL, /* a regular special character */
380: ESCAPE_FONT, /* a generic font mode */
381: ESCAPE_FONTBOLD, /* bold font mode */
382: ESCAPE_FONTITALIC, /* italic font mode */
1.109 schwarze 383: ESCAPE_FONTBI, /* bold italic font mode */
1.70 kristaps 384: ESCAPE_FONTROMAN, /* roman font mode */
385: ESCAPE_FONTPREV, /* previous font mode */
386: ESCAPE_NUMBERED, /* a numbered glyph */
1.75 kristaps 387: ESCAPE_UNICODE, /* a unicode codepoint */
1.102 schwarze 388: ESCAPE_NOSPACE, /* suppress space if the last on a line */
389: ESCAPE_SKIPCHAR /* skip the next character */
1.70 kristaps 390: };
391:
1.63 kristaps 392: typedef void (*mandocmsg)(enum mandocerr, enum mandoclevel,
393: const char *, int, int, const char *);
1.62 kristaps 394:
395: struct mparse;
1.71 kristaps 396: struct mchars;
1.62 kristaps 397: struct mdoc;
398: struct man;
1.59 kristaps 399:
1.22 kristaps 400: __BEGIN_DECLS
1.62 kristaps 401:
1.59 kristaps 402: void *mandoc_calloc(size_t, size_t);
1.112 ! schwarze 403: enum mandoc_esc mandoc_escape(const char **, const char **, int *);
1.59 kristaps 404: void *mandoc_malloc(size_t);
405: void *mandoc_realloc(void *, size_t);
1.78 kristaps 406: char *mandoc_strdup(const char *);
1.89 kristaps 407: char *mandoc_strndup(const char *, size_t);
1.73 kristaps 408: struct mchars *mchars_alloc(void);
1.96 kristaps 409: void mchars_free(struct mchars *);
1.74 kristaps 410: char mchars_num2char(const char *, size_t);
1.76 kristaps 411: int mchars_num2uc(const char *, size_t);
1.98 kristaps 412: int mchars_spec2cp(const struct mchars *,
1.96 kristaps 413: const char *, size_t);
1.98 kristaps 414: const char *mchars_spec2str(const struct mchars *,
1.96 kristaps 415: const char *, size_t, size_t *);
1.100 schwarze 416: struct mparse *mparse_alloc(enum mparset, enum mandoclevel,
417: mandocmsg, void *, 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 *,
425: struct mdoc **, struct man **);
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