Delete the mdoc_node.pending pointer and the function calculating
[mandoc.git] / mandoc.h
1 /* $Id: mandoc.h,v 1.200 2015/02/06 16:06:25 schwarze Exp $ */
2 /*
3 * Copyright (c) 2010, 2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2010-2015 Ingo Schwarze <schwarze@openbsd.org>
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
19 #define ASCII_NBRSP 31 /* non-breaking space */
20 #define ASCII_HYPH 30 /* breakable hyphen */
21 #define ASCII_BREAK 29 /* breakable zero-width space */
22
23 /*
24 * Status level. This refers to both internal status (i.e., whilst
25 * running, when warnings/errors are reported) and an indicator of a
26 * threshold of when to halt (when said internal state exceeds the
27 * threshold).
28 */
29 enum mandoclevel {
30 MANDOCLEVEL_OK = 0,
31 MANDOCLEVEL_RESERVED,
32 MANDOCLEVEL_WARNING, /* warnings: syntax, whitespace, etc. */
33 MANDOCLEVEL_ERROR, /* input has been thrown away */
34 MANDOCLEVEL_UNSUPP, /* input needs unimplemented features */
35 MANDOCLEVEL_BADARG, /* bad argument in invocation */
36 MANDOCLEVEL_SYSERR, /* system error */
37 MANDOCLEVEL_MAX
38 };
39
40 /*
41 * All possible things that can go wrong within a parse, be it libroff,
42 * libmdoc, or libman.
43 */
44 enum mandocerr {
45 MANDOCERR_OK,
46
47 MANDOCERR_WARNING, /* ===== start of warnings ===== */
48
49 /* related to the prologue */
50 MANDOCERR_DT_NOTITLE, /* missing manual title, using UNTITLED: line */
51 MANDOCERR_TH_NOTITLE, /* missing manual title, using "": [macro] */
52 MANDOCERR_TITLE_CASE, /* lower case character in document title */
53 MANDOCERR_MSEC_MISSING, /* missing manual section, using "": macro */
54 MANDOCERR_MSEC_BAD, /* unknown manual section: Dt ... section */
55 MANDOCERR_DATE_MISSING, /* missing date, using today's date */
56 MANDOCERR_DATE_BAD, /* cannot parse date, using it verbatim: date */
57 MANDOCERR_OS_MISSING, /* missing Os macro, using "" */
58 MANDOCERR_PROLOG_REP, /* duplicate prologue macro: macro */
59 MANDOCERR_PROLOG_LATE, /* late prologue macro: macro */
60 MANDOCERR_DT_LATE, /* skipping late title macro: Dt args */
61 MANDOCERR_PROLOG_ORDER, /* prologue macros out of order: macros */
62
63 /* related to document structure */
64 MANDOCERR_SO, /* .so is fragile, better use ln(1): so path */
65 MANDOCERR_DOC_EMPTY, /* no document body */
66 MANDOCERR_SEC_BEFORE, /* content before first section header: macro */
67 MANDOCERR_NAMESEC_FIRST, /* first section is not NAME: Sh title */
68 MANDOCERR_NAMESEC_BAD, /* bad NAME section contents: macro */
69 MANDOCERR_ND_EMPTY, /* missing description line, using "" */
70 MANDOCERR_SEC_ORDER, /* sections out of conventional order: Sh title */
71 MANDOCERR_SEC_REP, /* duplicate section title: Sh title */
72 MANDOCERR_SEC_MSEC, /* unexpected section: Sh title for ... only */
73 MANDOCERR_XR_ORDER, /* unusual Xr order: ... after ... */
74 MANDOCERR_XR_PUNCT, /* unusual Xr punctuation: ... after ... */
75 MANDOCERR_AN_MISSING, /* AUTHORS section without An macro */
76
77 /* related to macros and nesting */
78 MANDOCERR_MACRO_OBS, /* obsolete macro: macro */
79 MANDOCERR_MACRO_CALL, /* macro neither callable nor escaped: macro */
80 MANDOCERR_PAR_SKIP, /* skipping paragraph macro: macro ... */
81 MANDOCERR_PAR_MOVE, /* moving paragraph macro out of list: macro */
82 MANDOCERR_NS_SKIP, /* skipping no-space macro */
83 MANDOCERR_BLK_NEST, /* blocks badly nested: macro ... */
84 MANDOCERR_BD_NEST, /* nested displays are not portable: macro ... */
85 MANDOCERR_BL_MOVE, /* moving content out of list: macro */
86 MANDOCERR_VT_CHILD, /* .Vt block has child macro: macro */
87 MANDOCERR_FI_SKIP, /* fill mode already enabled, skipping: fi */
88 MANDOCERR_NF_SKIP, /* fill mode already disabled, skipping: nf */
89 MANDOCERR_BLK_LINE, /* line scope broken: macro breaks macro */
90
91 /* related to missing arguments */
92 MANDOCERR_REQ_EMPTY, /* skipping empty request: request */
93 MANDOCERR_COND_EMPTY, /* conditional request controls empty scope */
94 MANDOCERR_MACRO_EMPTY, /* skipping empty macro: macro */
95 MANDOCERR_BLK_EMPTY, /* empty block: macro */
96 MANDOCERR_ARG_EMPTY, /* empty argument, using 0n: macro arg */
97 MANDOCERR_BD_NOTYPE, /* missing display type, using -ragged: Bd */
98 MANDOCERR_BL_LATETYPE, /* list type is not the first argument: Bl arg */
99 MANDOCERR_BL_NOWIDTH, /* missing -width in -tag list, using 8n */
100 MANDOCERR_EX_NONAME, /* missing utility name, using "": Ex */
101 MANDOCERR_FO_NOHEAD, /* missing function name, using "": Fo */
102 MANDOCERR_IT_NOHEAD, /* empty head in list item: Bl -type It */
103 MANDOCERR_IT_NOBODY, /* empty list item: Bl -type It */
104 MANDOCERR_BF_NOFONT, /* missing font type, using \fR: Bf */
105 MANDOCERR_BF_BADFONT, /* unknown font type, using \fR: Bf font */
106 MANDOCERR_PF_SKIP, /* nothing follows prefix: Pf arg */
107 MANDOCERR_RS_EMPTY, /* empty reference block: Rs */
108 MANDOCERR_ARG_STD, /* missing -std argument, adding it: macro */
109 MANDOCERR_OP_EMPTY, /* missing option string, using "": OP */
110 MANDOCERR_UR_NOHEAD, /* missing resource identifier, using "": UR */
111 MANDOCERR_EQN_NOBOX, /* missing eqn box, using "": op */
112
113 /* related to bad arguments */
114 MANDOCERR_ARG_QUOTE, /* unterminated quoted argument */
115 MANDOCERR_ARG_REP, /* duplicate argument: macro arg */
116 MANDOCERR_AN_REP, /* skipping duplicate argument: An -arg */
117 MANDOCERR_BD_REP, /* skipping duplicate display type: Bd -type */
118 MANDOCERR_BL_REP, /* skipping duplicate list type: Bl -type */
119 MANDOCERR_BL_SKIPW, /* skipping -width argument: Bl -type */
120 MANDOCERR_BL_COL, /* wrong number of cells */
121 MANDOCERR_AT_BAD, /* unknown AT&T UNIX version: At version */
122 MANDOCERR_FA_COMMA, /* comma in function argument: arg */
123 MANDOCERR_FN_PAREN, /* parenthesis in function name: arg */
124 MANDOCERR_RS_BAD, /* invalid content in Rs block: macro */
125 MANDOCERR_SM_BAD, /* invalid Boolean argument: macro arg */
126 MANDOCERR_FT_BAD, /* unknown font, skipping request: ft font */
127 MANDOCERR_TR_ODD, /* odd number of characters in request: tr char */
128
129 /* related to plain text */
130 MANDOCERR_FI_BLANK, /* blank line in fill mode, using .sp */
131 MANDOCERR_FI_TAB, /* tab in filled text */
132 MANDOCERR_SPACE_EOL, /* whitespace at end of input line */
133 MANDOCERR_COMMENT_BAD, /* bad comment style */
134 MANDOCERR_ESC_BAD, /* invalid escape sequence: esc */
135 MANDOCERR_STR_UNDEF, /* undefined string, using "": name */
136
137 /* related to tables */
138 MANDOCERR_TBLLAYOUT_SPAN, /* tbl line starts with span */
139 MANDOCERR_TBLLAYOUT_DOWN, /* tbl column starts with span */
140 MANDOCERR_TBLLAYOUT_VERT, /* skipping vertical bar in tbl layout */
141
142 MANDOCERR_ERROR, /* ===== start of errors ===== */
143
144 /* related to tables */
145 MANDOCERR_TBLOPT_ALPHA, /* non-alphabetic character in tbl options */
146 MANDOCERR_TBLOPT_BAD, /* skipping unknown tbl option: option */
147 MANDOCERR_TBLOPT_NOARG, /* missing tbl option argument: option */
148 MANDOCERR_TBLOPT_ARGSZ, /* wrong tbl option argument size: option */
149 MANDOCERR_TBLLAYOUT_NONE, /* empty tbl layout */
150 MANDOCERR_TBLLAYOUT_CHAR, /* invalid character in tbl layout: char */
151 MANDOCERR_TBLLAYOUT_PAR, /* unmatched parenthesis in tbl layout */
152 MANDOCERR_TBLDATA_NONE, /* tbl without any data cells */
153 MANDOCERR_TBLDATA_SPAN, /* ignoring data in spanned tbl cell: data */
154 MANDOCERR_TBLDATA_EXTRA, /* ignoring extra tbl data cells: data */
155 MANDOCERR_TBLDATA_BLK, /* data block open at end of tbl: macro */
156
157 /* related to document structure and macros */
158 MANDOCERR_FILE, /* cannot open file */
159 MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
160 MANDOCERR_CHAR_BAD, /* skipping bad character: number */
161 MANDOCERR_MACRO, /* skipping unknown macro: macro */
162 MANDOCERR_REQ_INSEC, /* skipping insecure request: request */
163 MANDOCERR_IT_STRAY, /* skipping item outside list: It ... */
164 MANDOCERR_TA_STRAY, /* skipping column outside column list: Ta */
165 MANDOCERR_BLK_NOTOPEN, /* skipping end of block that is not open */
166 MANDOCERR_RE_NOTOPEN, /* fewer RS blocks open, skipping: RE arg */
167 MANDOCERR_BLK_BROKEN, /* inserting missing end of block: macro ... */
168 MANDOCERR_BLK_NOEND, /* appending missing end of block: macro */
169
170 /* related to request and macro arguments */
171 MANDOCERR_NAMESC, /* escaped character not allowed in a name: name */
172 MANDOCERR_BD_FILE, /* NOT IMPLEMENTED: Bd -file */
173 MANDOCERR_BL_NOTYPE, /* missing list type, using -item: Bl */
174 MANDOCERR_NM_NONAME, /* missing manual name, using "": Nm */
175 MANDOCERR_OS_UNAME, /* uname(3) system call failed, using UNKNOWN */
176 MANDOCERR_ST_BAD, /* unknown standard specifier: St standard */
177 MANDOCERR_IT_NONUM, /* skipping request without numeric argument */
178 MANDOCERR_SO_PATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
179 MANDOCERR_SO_FAIL, /* .so request failed */
180 MANDOCERR_ARG_SKIP, /* skipping all arguments: macro args */
181 MANDOCERR_ARG_EXCESS, /* skipping excess arguments: macro ... args */
182 MANDOCERR_DIVZERO, /* divide by zero */
183
184 MANDOCERR_UNSUPP, /* ===== start of unsupported features ===== */
185
186 MANDOCERR_TOOLARGE, /* input too large */
187 MANDOCERR_CHAR_UNSUPP, /* unsupported control character: number */
188 MANDOCERR_REQ_UNSUPP, /* unsupported roff request: request */
189 MANDOCERR_TBLOPT_EQN, /* eqn delim option in tbl: arg */
190 MANDOCERR_TBLLAYOUT_MOD, /* unsupported tbl layout modifier: m */
191 MANDOCERR_TBLMACRO, /* ignoring macro in table: macro */
192
193 MANDOCERR_MAX
194 };
195
196 struct tbl_opts {
197 char tab; /* cell-separator */
198 char decimal; /* decimal point */
199 int opts;
200 #define TBL_OPT_CENTRE (1 << 0)
201 #define TBL_OPT_EXPAND (1 << 1)
202 #define TBL_OPT_BOX (1 << 2)
203 #define TBL_OPT_DBOX (1 << 3)
204 #define TBL_OPT_ALLBOX (1 << 4)
205 #define TBL_OPT_NOKEEP (1 << 5)
206 #define TBL_OPT_NOSPACE (1 << 6)
207 #define TBL_OPT_NOWARN (1 << 7)
208 int cols; /* number of columns */
209 int lvert; /* width of left vertical line */
210 int rvert; /* width of right vertical line */
211 };
212
213 enum tbl_cellt {
214 TBL_CELL_CENTRE, /* c, C */
215 TBL_CELL_RIGHT, /* r, R */
216 TBL_CELL_LEFT, /* l, L */
217 TBL_CELL_NUMBER, /* n, N */
218 TBL_CELL_SPAN, /* s, S */
219 TBL_CELL_LONG, /* a, A */
220 TBL_CELL_DOWN, /* ^ */
221 TBL_CELL_HORIZ, /* _, - */
222 TBL_CELL_DHORIZ, /* = */
223 TBL_CELL_MAX
224 };
225
226 /*
227 * A cell in a layout row.
228 */
229 struct tbl_cell {
230 struct tbl_cell *next;
231 int vert; /* width of subsequent vertical line */
232 enum tbl_cellt pos;
233 size_t spacing;
234 int col; /* column number, starting from 0 */
235 int flags;
236 #define TBL_CELL_TALIGN (1 << 0) /* t, T */
237 #define TBL_CELL_BALIGN (1 << 1) /* d, D */
238 #define TBL_CELL_BOLD (1 << 2) /* fB, B, b */
239 #define TBL_CELL_ITALIC (1 << 3) /* fI, I, i */
240 #define TBL_CELL_EQUAL (1 << 4) /* e, E */
241 #define TBL_CELL_UP (1 << 5) /* u, U */
242 #define TBL_CELL_WIGN (1 << 6) /* z, Z */
243 #define TBL_CELL_WMAX (1 << 7) /* x, X */
244 };
245
246 /*
247 * A layout row.
248 */
249 struct tbl_row {
250 struct tbl_row *next;
251 struct tbl_cell *first;
252 struct tbl_cell *last;
253 int vert; /* width of left vertical line */
254 };
255
256 enum tbl_datt {
257 TBL_DATA_NONE, /* has no data */
258 TBL_DATA_DATA, /* consists of data/string */
259 TBL_DATA_HORIZ, /* horizontal line */
260 TBL_DATA_DHORIZ, /* double-horizontal line */
261 TBL_DATA_NHORIZ, /* squeezed horizontal line */
262 TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
263 };
264
265 /*
266 * A cell within a row of data. The "string" field contains the actual
267 * string value that's in the cell. The rest is layout.
268 */
269 struct tbl_dat {
270 struct tbl_cell *layout; /* layout cell */
271 int spans; /* how many spans follow */
272 struct tbl_dat *next;
273 char *string; /* data (NULL if not TBL_DATA_DATA) */
274 enum tbl_datt pos;
275 };
276
277 enum tbl_spant {
278 TBL_SPAN_DATA, /* span consists of data */
279 TBL_SPAN_HORIZ, /* span is horizontal line */
280 TBL_SPAN_DHORIZ /* span is double horizontal line */
281 };
282
283 /*
284 * A row of data in a table.
285 */
286 struct tbl_span {
287 struct tbl_opts *opts;
288 struct tbl_row *layout; /* layout row */
289 struct tbl_dat *first;
290 struct tbl_dat *last;
291 struct tbl_span *prev;
292 struct tbl_span *next;
293 int line; /* parse line */
294 enum tbl_spant pos;
295 };
296
297 enum eqn_boxt {
298 EQN_ROOT, /* root of parse tree */
299 EQN_TEXT, /* text (number, variable, whatever) */
300 EQN_SUBEXPR, /* nested `eqn' subexpression */
301 EQN_LIST, /* list (braces, etc.) */
302 EQN_LISTONE, /* singleton list */
303 EQN_PILE, /* vertical pile */
304 EQN_MATRIX /* pile of piles */
305 };
306
307 enum eqn_fontt {
308 EQNFONT_NONE = 0,
309 EQNFONT_ROMAN,
310 EQNFONT_BOLD,
311 EQNFONT_FAT,
312 EQNFONT_ITALIC,
313 EQNFONT__MAX
314 };
315
316 enum eqn_post {
317 EQNPOS_NONE = 0,
318 EQNPOS_SUP,
319 EQNPOS_SUBSUP,
320 EQNPOS_SUB,
321 EQNPOS_TO,
322 EQNPOS_FROM,
323 EQNPOS_FROMTO,
324 EQNPOS_OVER,
325 EQNPOS_SQRT,
326 EQNPOS__MAX
327 };
328
329 enum eqn_pilet {
330 EQNPILE_NONE = 0,
331 EQNPILE_PILE,
332 EQNPILE_CPILE,
333 EQNPILE_RPILE,
334 EQNPILE_LPILE,
335 EQNPILE_COL,
336 EQNPILE_CCOL,
337 EQNPILE_RCOL,
338 EQNPILE_LCOL,
339 EQNPILE__MAX
340 };
341
342 /*
343 * A "box" is a parsed mathematical expression as defined by the eqn.7
344 * grammar.
345 */
346 struct eqn_box {
347 int size; /* font size of expression */
348 #define EQN_DEFSIZE INT_MIN
349 enum eqn_boxt type; /* type of node */
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 *prev; /* node sibling */
354 struct eqn_box *parent; /* node sibling */
355 char *text; /* text (or NULL) */
356 char *left; /* fence left-hand */
357 char *right; /* fence right-hand */
358 char *top; /* expression over-symbol */
359 char *bottom; /* expression under-symbol */
360 size_t args; /* arguments in parent */
361 size_t expectargs; /* max arguments in parent */
362 enum eqn_post pos; /* position of next box */
363 enum eqn_fontt font; /* font of box */
364 enum eqn_pilet pile; /* equation piling */
365 };
366
367 /*
368 * An equation consists of a tree of expressions starting at a given
369 * line and position.
370 */
371 struct eqn {
372 char *name; /* identifier (or NULL) */
373 struct eqn_box *root; /* root mathematical expression */
374 int ln; /* invocation line */
375 int pos; /* invocation position */
376 };
377
378 /*
379 * Parse options.
380 */
381 #define MPARSE_MDOC 1 /* assume -mdoc */
382 #define MPARSE_MAN 2 /* assume -man */
383 #define MPARSE_SO 4 /* honour .so requests */
384 #define MPARSE_QUICK 8 /* abort the parse early */
385 #define MPARSE_UTF8 16 /* accept UTF-8 input */
386 #define MPARSE_LATIN1 32 /* accept ISO-LATIN-1 input */
387
388 enum mandoc_esc {
389 ESCAPE_ERROR = 0, /* bail! unparsable escape */
390 ESCAPE_IGNORE, /* escape to be ignored */
391 ESCAPE_SPECIAL, /* a regular special character */
392 ESCAPE_FONT, /* a generic font mode */
393 ESCAPE_FONTBOLD, /* bold font mode */
394 ESCAPE_FONTITALIC, /* italic font mode */
395 ESCAPE_FONTBI, /* bold italic font mode */
396 ESCAPE_FONTROMAN, /* roman font mode */
397 ESCAPE_FONTPREV, /* previous font mode */
398 ESCAPE_NUMBERED, /* a numbered glyph */
399 ESCAPE_UNICODE, /* a unicode codepoint */
400 ESCAPE_NOSPACE, /* suppress space if the last on a line */
401 ESCAPE_SKIPCHAR, /* skip the next character */
402 ESCAPE_OVERSTRIKE /* overstrike all chars in the argument */
403 };
404
405 typedef void (*mandocmsg)(enum mandocerr, enum mandoclevel,
406 const char *, int, int, const char *);
407
408 __BEGIN_DECLS
409
410 struct mparse;
411 struct mchars;
412 struct mdoc;
413 struct man;
414
415 enum mandoc_esc mandoc_escape(const char **, const char **, int *);
416 struct mchars *mchars_alloc(void);
417 void mchars_free(struct mchars *);
418 int mchars_num2char(const char *, size_t);
419 const char *mchars_uc2str(int);
420 int mchars_num2uc(const char *, size_t);
421 int mchars_spec2cp(const struct mchars *,
422 const char *, size_t);
423 const char *mchars_spec2str(const struct mchars *,
424 const char *, size_t, size_t *);
425 struct mparse *mparse_alloc(int, enum mandoclevel, mandocmsg,
426 const struct mchars *, const char *);
427 void mparse_free(struct mparse *);
428 void mparse_keep(struct mparse *);
429 enum mandoclevel mparse_open(struct mparse *, int *, const char *);
430 enum mandoclevel mparse_readfd(struct mparse *, int, const char *);
431 enum mandoclevel mparse_readmem(struct mparse *, void *, size_t,
432 const char *);
433 void mparse_reset(struct mparse *);
434 void mparse_result(struct mparse *,
435 struct mdoc **, struct man **, char **);
436 const char *mparse_getkeep(const struct mparse *);
437 const char *mparse_strerror(enum mandocerr);
438 const char *mparse_strlevel(enum mandoclevel);
439 enum mandoclevel mparse_wait(struct mparse *);
440
441 __END_DECLS