Annotation of mandoc/mandoc.h, Revision 1.123
1.123 ! schwarze 1: /* $Id: mandoc.h,v 1.122 2014/06/20 23:02:31 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 */
1.121 schwarze 52: MANDOCERR_TH_MISSING, /* no TH macro in document */
53: MANDOCERR_TITLE_CASE, /* document title should be all caps */
54: MANDOCERR_MSEC_BAD, /* unknown manual section */
55: MANDOCERR_ARCH_BAD, /* unknown manual volume or arch */
56: MANDOCERR_DATE_MISSING, /* date missing, using today's date */
57: MANDOCERR_DATE_BAD, /* cannot parse date, using it verbatim */
58: MANDOCERR_PROLOG_ORDER, /* prologue macros out of order */
59: MANDOCERR_PROLOG_REP, /* duplicate prologue macro */
60: MANDOCERR_PROLOG_BAD, /* macro not allowed in prologue */
61: MANDOCERR_PROLOG_ONLY, /* macro not allowed in body */
1.32 kristaps 62:
63: /* related to document structure */
1.26 kristaps 64: MANDOCERR_SO, /* .so is fragile, better use ln(1) */
1.122 schwarze 65: MANDOCERR_DOC_EMPTY, /* no document body */
66: MANDOCERR_SEC_BEFORE, /* content before the first section header */
1.32 kristaps 67: MANDOCERR_NAMESECFIRST, /* NAME section must come first */
68: MANDOCERR_BADNAMESEC, /* bad NAME section contents */
1.9 kristaps 69: MANDOCERR_SECOOO, /* sections out of conventional order */
1.32 kristaps 70: MANDOCERR_SECREP, /* duplicate section name */
1.110 schwarze 71: MANDOCERR_SECMSEC, /* section header suited to sections ... */
1.32 kristaps 72:
73: /* related to macros and nesting */
74: MANDOCERR_MACROOBS, /* skipping obsolete macro */
1.35 kristaps 75: MANDOCERR_IGNPAR, /* skipping paragraph macro */
1.104 schwarze 76: MANDOCERR_MOVEPAR, /* moving paragraph macro out of list */
1.54 kristaps 77: MANDOCERR_IGNNS, /* skipping no-space macro */
1.14 schwarze 78: MANDOCERR_SCOPENEST, /* blocks badly nested */
1.32 kristaps 79: MANDOCERR_CHILD, /* child violates parent syntax */
80: MANDOCERR_NESTEDDISP, /* nested displays are not portable */
81: MANDOCERR_SCOPEREP, /* already in literal mode */
1.68 kristaps 82: MANDOCERR_LINESCOPE, /* line scope broken */
1.32 kristaps 83:
84: /* related to missing macro arguments */
85: MANDOCERR_MACROEMPTY, /* skipping empty macro */
1.46 schwarze 86: MANDOCERR_ARGCWARN, /* argument count wrong */
1.32 kristaps 87: MANDOCERR_DISPTYPE, /* missing display type */
88: MANDOCERR_LISTFIRST, /* list type must come first */
89: MANDOCERR_NOWIDTHARG, /* tag lists require a width argument */
90: MANDOCERR_FONTTYPE, /* missing font type */
1.53 kristaps 91: MANDOCERR_WNOSCOPE, /* skipping end of block that is not open */
1.13 schwarze 92:
1.32 kristaps 93: /* related to bad macro arguments */
94: MANDOCERR_IGNARGV, /* skipping argument */
95: MANDOCERR_ARGVREP, /* duplicate argument */
96: MANDOCERR_DISPREP, /* duplicate display type */
97: MANDOCERR_LISTREP, /* duplicate list type */
98: MANDOCERR_BADATT, /* unknown AT&T UNIX version */
1.5 kristaps 99: MANDOCERR_BADBOOL, /* bad Boolean value */
1.31 kristaps 100: MANDOCERR_BADFONT, /* unknown font */
1.32 kristaps 101: MANDOCERR_BADSTANDARD, /* unknown standard specifier */
102: MANDOCERR_BADWIDTH, /* bad width argument */
103:
104: /* related to plain text */
105: MANDOCERR_NOBLANKLN, /* blank line in non-literal context */
106: MANDOCERR_BADTAB, /* tab in non-literal context */
107: MANDOCERR_EOLNSPACE, /* end of line whitespace */
1.5 kristaps 108: MANDOCERR_BADCOMMENT, /* bad comment style */
1.32 kristaps 109: MANDOCERR_BADESCAPE, /* unknown escape sequence */
110: MANDOCERR_BADQUOTE, /* unterminated quoted string */
1.79 kristaps 111:
112: /* related to equations */
113: MANDOCERR_EQNQUOTE, /* unexpected literal in equation */
1.32 kristaps 114:
115: MANDOCERR_ERROR, /* ===== start of errors ===== */
1.80 kristaps 116:
117: /* related to equations */
1.84 kristaps 118: MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
119: MANDOCERR_EQNSCOPE, /* equation scope open on exit */
1.89 kristaps 120: MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
121: MANDOCERR_EQNEOF, /* unexpected end of equation */
122: MANDOCERR_EQNSYNT, /* equation syntax error */
1.32 kristaps 123:
1.42 kristaps 124: /* related to tables */
1.36 kristaps 125: MANDOCERR_TBL, /* bad table syntax */
126: MANDOCERR_TBLOPT, /* bad table option */
1.37 kristaps 127: MANDOCERR_TBLLAYOUT, /* bad table layout */
128: MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
1.39 kristaps 129: MANDOCERR_TBLNODATA, /* no table data cells specified */
1.47 kristaps 130: MANDOCERR_TBLIGNDATA, /* ignore data in cell */
1.48 kristaps 131: MANDOCERR_TBLBLOCK, /* data block still open */
1.50 kristaps 132: MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
1.42 kristaps 133:
1.33 schwarze 134: MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
1.32 kristaps 135: MANDOCERR_BADCHAR, /* skipping bad character */
1.52 schwarze 136: MANDOCERR_NAMESC, /* escaped character not allowed in a name */
1.103 schwarze 137: MANDOCERR_NONAME, /* manual name not yet set */
1.32 kristaps 138: MANDOCERR_NOTEXT, /* skipping text before the first section header */
139: MANDOCERR_MACRO, /* skipping unknown macro */
1.25 kristaps 140: MANDOCERR_REQUEST, /* NOT IMPLEMENTED: skipping request */
1.5 kristaps 141: MANDOCERR_ARGCOUNT, /* argument count wrong */
1.106 schwarze 142: MANDOCERR_STRAYTA, /* skipping column outside column list */
1.32 kristaps 143: MANDOCERR_NOSCOPE, /* skipping end of block that is not open */
1.30 kristaps 144: MANDOCERR_SCOPEBROKEN, /* missing end of block */
1.19 schwarze 145: MANDOCERR_SCOPEEXIT, /* scope open on exit */
1.28 kristaps 146: MANDOCERR_UNAME, /* uname(3) system call failed */
1.5 kristaps 147: /* FIXME: merge following with MANDOCERR_ARGCOUNT */
148: MANDOCERR_NOARGS, /* macro requires line argument(s) */
149: MANDOCERR_NOBODY, /* macro requires body argument(s) */
150: MANDOCERR_NOARGV, /* macro requires argument(s) */
1.108 schwarze 151: MANDOCERR_NUMERIC, /* request requires a numeric argument */
1.10 kristaps 152: MANDOCERR_LISTTYPE, /* missing list type */
1.5 kristaps 153: MANDOCERR_ARGSLOST, /* line argument(s) will be lost */
154: MANDOCERR_BODYLOST, /* body argument(s) will be lost */
1.13 schwarze 155:
1.21 kristaps 156: MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
1.32 kristaps 157:
1.113 schwarze 158: MANDOCERR_TOOLARGE, /* input too large */
1.66 kristaps 159: MANDOCERR_NOTMANUAL, /* manual isn't really a manual */
1.8 kristaps 160: MANDOCERR_COLUMNS, /* column syntax is inconsistent */
1.32 kristaps 161: MANDOCERR_BADDISP, /* NOT IMPLEMENTED: .Bd -file */
1.5 kristaps 162: MANDOCERR_SYNTARGVCOUNT, /* argument count wrong, violates syntax */
163: MANDOCERR_SYNTCHILD, /* child violates parent syntax */
164: MANDOCERR_SYNTARGCOUNT, /* argument count wrong, violates syntax */
1.123 ! schwarze 165: MANDOCERR_SO_PATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
! 166: MANDOCERR_SO_FAIL, /* .so request failed */
1.5 kristaps 167: MANDOCERR_NODOCPROLOG, /* no document prologue */
1.18 schwarze 168: MANDOCERR_MEM, /* static buffer exhausted */
1.113 schwarze 169:
170: /* ===== system errors ===== */
171:
172: MANDOCERR_SYSOPEN, /* cannot open file */
173: MANDOCERR_SYSSTAT, /* cannot stat file */
174: MANDOCERR_SYSREAD, /* cannot read file */
175:
1.1 kristaps 176: MANDOCERR_MAX
1.43 kristaps 177: };
178:
1.107 schwarze 179: struct tbl_opts {
1.43 kristaps 180: char tab; /* cell-separator */
181: char decimal; /* decimal point */
182: int linesize;
183: int opts;
184: #define TBL_OPT_CENTRE (1 << 0)
185: #define TBL_OPT_EXPAND (1 << 1)
186: #define TBL_OPT_BOX (1 << 2)
187: #define TBL_OPT_DBOX (1 << 3)
188: #define TBL_OPT_ALLBOX (1 << 4)
189: #define TBL_OPT_NOKEEP (1 << 5)
190: #define TBL_OPT_NOSPACE (1 << 6)
1.45 kristaps 191: int cols; /* number of columns */
1.38 kristaps 192: };
193:
1.41 kristaps 194: /*
195: * The head of a table specifies all of its columns. When formatting a
196: * tbl_span, iterate over these and plug in data from the tbl_span when
197: * appropriate, using tbl_cell as a guide to placement.
198: */
199: struct tbl_head {
1.45 kristaps 200: int ident; /* 0 <= unique id < cols */
1.101 schwarze 201: int vert; /* width of preceding vertical line */
1.41 kristaps 202: struct tbl_head *next;
203: struct tbl_head *prev;
204: };
205:
1.38 kristaps 206: enum tbl_cellt {
207: TBL_CELL_CENTRE, /* c, C */
208: TBL_CELL_RIGHT, /* r, R */
209: TBL_CELL_LEFT, /* l, L */
210: TBL_CELL_NUMBER, /* n, N */
211: TBL_CELL_SPAN, /* s, S */
212: TBL_CELL_LONG, /* a, A */
213: TBL_CELL_DOWN, /* ^ */
214: TBL_CELL_HORIZ, /* _, - */
215: TBL_CELL_DHORIZ, /* = */
216: TBL_CELL_MAX
217: };
218:
219: /*
220: * A cell in a layout row.
221: */
222: struct tbl_cell {
223: struct tbl_cell *next;
1.101 schwarze 224: int vert; /* width of preceding vertical line */
1.38 kristaps 225: enum tbl_cellt pos;
1.51 kristaps 226: size_t spacing;
1.38 kristaps 227: int flags;
228: #define TBL_CELL_TALIGN (1 << 0) /* t, T */
229: #define TBL_CELL_BALIGN (1 << 1) /* d, D */
230: #define TBL_CELL_BOLD (1 << 2) /* fB, B, b */
231: #define TBL_CELL_ITALIC (1 << 3) /* fI, I, i */
232: #define TBL_CELL_EQUAL (1 << 4) /* e, E */
233: #define TBL_CELL_UP (1 << 5) /* u, U */
234: #define TBL_CELL_WIGN (1 << 6) /* z, Z */
1.41 kristaps 235: struct tbl_head *head;
1.38 kristaps 236: };
237:
238: /*
239: * A layout row.
240: */
241: struct tbl_row {
242: struct tbl_row *next;
243: struct tbl_cell *first;
244: struct tbl_cell *last;
1.119 schwarze 245: int vert; /* trailing vertical line */
1.38 kristaps 246: };
247:
1.40 kristaps 248: enum tbl_datt {
1.51 kristaps 249: TBL_DATA_NONE, /* has no data */
250: TBL_DATA_DATA, /* consists of data/string */
251: TBL_DATA_HORIZ, /* horizontal line */
252: TBL_DATA_DHORIZ, /* double-horizontal line */
253: TBL_DATA_NHORIZ, /* squeezed horizontal line */
254: TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
1.40 kristaps 255: };
256:
1.38 kristaps 257: /*
258: * A cell within a row of data. The "string" field contains the actual
259: * string value that's in the cell. The rest is layout.
260: */
261: struct tbl_dat {
1.51 kristaps 262: struct tbl_cell *layout; /* layout cell */
1.50 kristaps 263: int spans; /* how many spans follow */
1.38 kristaps 264: struct tbl_dat *next;
1.51 kristaps 265: char *string; /* data (NULL if not TBL_DATA_DATA) */
1.40 kristaps 266: enum tbl_datt pos;
267: };
268:
269: enum tbl_spant {
270: TBL_SPAN_DATA, /* span consists of data */
271: TBL_SPAN_HORIZ, /* span is horizontal line */
272: TBL_SPAN_DHORIZ /* span is double horizontal line */
1.38 kristaps 273: };
274:
275: /*
276: * A row of data in a table.
277: */
278: struct tbl_span {
1.107 schwarze 279: struct tbl_opts *opts;
1.44 kristaps 280: struct tbl_head *head;
1.51 kristaps 281: struct tbl_row *layout; /* layout row */
1.38 kristaps 282: struct tbl_dat *first;
283: struct tbl_dat *last;
1.56 kristaps 284: int line; /* parse line */
1.44 kristaps 285: int flags;
286: #define TBL_SPAN_FIRST (1 << 0)
287: #define TBL_SPAN_LAST (1 << 1)
1.40 kristaps 288: enum tbl_spant pos;
1.38 kristaps 289: struct tbl_span *next;
1.55 kristaps 290: };
291:
1.83 kristaps 292: enum eqn_boxt {
1.84 kristaps 293: EQN_ROOT, /* root of parse tree */
294: EQN_TEXT, /* text (number, variable, whatever) */
1.90 kristaps 295: EQN_SUBEXPR, /* nested `eqn' subexpression */
1.92 kristaps 296: EQN_LIST, /* subexpressions list */
297: EQN_MATRIX /* matrix subexpression */
1.83 kristaps 298: };
299:
1.85 kristaps 300: enum eqn_markt {
301: EQNMARK_NONE = 0,
302: EQNMARK_DOT,
303: EQNMARK_DOTDOT,
304: EQNMARK_HAT,
305: EQNMARK_TILDE,
306: EQNMARK_VEC,
307: EQNMARK_DYAD,
308: EQNMARK_BAR,
309: EQNMARK_UNDER,
310: EQNMARK__MAX
311: };
312:
1.86 kristaps 313: enum eqn_fontt {
314: EQNFONT_NONE = 0,
315: EQNFONT_ROMAN,
316: EQNFONT_BOLD,
1.93 kristaps 317: EQNFONT_FAT,
1.86 kristaps 318: EQNFONT_ITALIC,
319: EQNFONT__MAX
320: };
321:
1.87 kristaps 322: enum eqn_post {
323: EQNPOS_NONE = 0,
324: EQNPOS_OVER,
325: EQNPOS_SUP,
326: EQNPOS_SUB,
327: EQNPOS_TO,
328: EQNPOS_FROM,
329: EQNPOS__MAX
330: };
331:
1.89 kristaps 332: enum eqn_pilet {
333: EQNPILE_NONE = 0,
1.95 kristaps 334: EQNPILE_PILE,
1.89 kristaps 335: EQNPILE_CPILE,
336: EQNPILE_RPILE,
337: EQNPILE_LPILE,
1.95 kristaps 338: EQNPILE_COL,
1.92 kristaps 339: EQNPILE_CCOL,
340: EQNPILE_RCOL,
341: EQNPILE_LCOL,
1.89 kristaps 342: EQNPILE__MAX
343: };
344:
1.86 kristaps 345: /*
1.84 kristaps 346: * A "box" is a parsed mathematical expression as defined by the eqn.7
347: * grammar.
348: */
1.83 kristaps 349: struct eqn_box {
1.88 kristaps 350: int size; /* font size of expression */
351: #define EQN_DEFSIZE INT_MIN
1.84 kristaps 352: enum eqn_boxt type; /* type of node */
1.89 kristaps 353: struct eqn_box *first; /* first child node */
354: struct eqn_box *last; /* last child node */
355: struct eqn_box *next; /* node sibling */
356: struct eqn_box *parent; /* node sibling */
357: char *text; /* text (or NULL) */
358: char *left;
359: char *right;
1.87 kristaps 360: enum eqn_post pos; /* position of next box */
1.86 kristaps 361: enum eqn_markt mark; /* a mark about the box */
362: enum eqn_fontt font; /* font of box */
1.89 kristaps 363: enum eqn_pilet pile; /* equation piling */
1.83 kristaps 364: };
365:
1.87 kristaps 366: /*
367: * An equation consists of a tree of expressions starting at a given
1.120 schwarze 368: * line and position.
1.87 kristaps 369: */
1.55 kristaps 370: struct eqn {
1.94 kristaps 371: char *name; /* identifier (or NULL) */
1.84 kristaps 372: struct eqn_box *root; /* root mathematical expression */
1.82 kristaps 373: int ln; /* invocation line */
1.57 kristaps 374: int pos; /* invocation position */
1.16 kristaps 375: };
1.22 kristaps 376:
1.60 kristaps 377: /*
1.116 schwarze 378: * Parse options.
379: */
380: #define MPARSE_MDOC 1 /* assume -mdoc */
381: #define MPARSE_MAN 2 /* assume -man */
382: #define MPARSE_SO 4 /* honour .so requests */
383: #define MPARSE_QUICK 8 /* abort the parse early */
1.62 kristaps 384:
1.70 kristaps 385: enum mandoc_esc {
386: ESCAPE_ERROR = 0, /* bail! unparsable escape */
387: ESCAPE_IGNORE, /* escape to be ignored */
388: ESCAPE_SPECIAL, /* a regular special character */
389: ESCAPE_FONT, /* a generic font mode */
390: ESCAPE_FONTBOLD, /* bold font mode */
391: ESCAPE_FONTITALIC, /* italic font mode */
1.109 schwarze 392: ESCAPE_FONTBI, /* bold italic font mode */
1.70 kristaps 393: ESCAPE_FONTROMAN, /* roman font mode */
394: ESCAPE_FONTPREV, /* previous font mode */
395: ESCAPE_NUMBERED, /* a numbered glyph */
1.75 kristaps 396: ESCAPE_UNICODE, /* a unicode codepoint */
1.102 schwarze 397: ESCAPE_NOSPACE, /* suppress space if the last on a line */
398: ESCAPE_SKIPCHAR /* skip the next character */
1.70 kristaps 399: };
400:
1.63 kristaps 401: typedef void (*mandocmsg)(enum mandocerr, enum mandoclevel,
402: const char *, int, int, const char *);
1.62 kristaps 403:
404: struct mparse;
1.71 kristaps 405: struct mchars;
1.62 kristaps 406: struct mdoc;
407: struct man;
1.59 kristaps 408:
1.22 kristaps 409: __BEGIN_DECLS
1.62 kristaps 410:
1.112 schwarze 411: enum mandoc_esc mandoc_escape(const char **, const char **, int *);
1.73 kristaps 412: struct mchars *mchars_alloc(void);
1.96 kristaps 413: void mchars_free(struct mchars *);
1.120 schwarze 414: char mchars_num2char(const char *, size_t);
1.76 kristaps 415: int mchars_num2uc(const char *, size_t);
1.120 schwarze 416: int mchars_spec2cp(const struct mchars *,
1.96 kristaps 417: const char *, size_t);
1.120 schwarze 418: const char *mchars_spec2str(const struct mchars *,
1.96 kristaps 419: const char *, size_t, size_t *);
1.116 schwarze 420: struct mparse *mparse_alloc(int, enum mandoclevel, mandocmsg, char *);
1.96 kristaps 421: void mparse_free(struct mparse *);
422: void mparse_keep(struct mparse *);
423: enum mandoclevel mparse_readfd(struct mparse *, int, const char *);
1.99 joerg 424: enum mandoclevel mparse_readmem(struct mparse *, const void *, size_t,
425: const char *);
1.96 kristaps 426: void mparse_reset(struct mparse *);
1.120 schwarze 427: void mparse_result(struct mparse *,
1.117 schwarze 428: struct mdoc **, struct man **, char **);
1.96 kristaps 429: const char *mparse_getkeep(const struct mparse *);
430: const char *mparse_strerror(enum mandocerr);
431: const char *mparse_strlevel(enum mandoclevel);
1.1 kristaps 432:
433: __END_DECLS
434:
435: #endif /*!MANDOC_H*/
CVSweb