mandoc.h revision 274880
1/*	$Id: mandoc.h,v 1.152 2014/08/06 15:09:05 schwarze Exp $ */
2/*
3 * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2010-2014 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#ifndef MANDOC_H
19#define MANDOC_H
20
21#define ASCII_NBRSP	 31  /* non-breaking space */
22#define	ASCII_HYPH	 30  /* breakable hyphen */
23#define	ASCII_BREAK	 29  /* breakable zero-width space */
24
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 */
31enum	mandoclevel {
32	MANDOCLEVEL_OK = 0,
33	MANDOCLEVEL_RESERVED,
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 */
39	MANDOCLEVEL_MAX
40};
41
42/*
43 * All possible things that can go wrong within a parse, be it libroff,
44 * libmdoc, or libman.
45 */
46enum	mandocerr {
47	MANDOCERR_OK,
48
49	MANDOCERR_WARNING, /* ===== start of warnings ===== */
50
51	/* related to the prologue */
52	MANDOCERR_DT_NOTITLE, /* missing manual title, using UNTITLED: line */
53	MANDOCERR_TH_NOTITLE, /* missing manual title, using "": [macro] */
54	MANDOCERR_TITLE_CASE, /* lower case character in document title */
55	MANDOCERR_MSEC_MISSING, /* missing manual section, using "": macro */
56	MANDOCERR_MSEC_BAD, /* unknown manual section: Dt ... section */
57	MANDOCERR_ARCH_BAD, /* unknown manual volume or arch: Dt ... volume */
58	MANDOCERR_DATE_MISSING, /* missing date, using today's date */
59	MANDOCERR_DATE_BAD, /* cannot parse date, using it verbatim: date */
60	MANDOCERR_OS_MISSING, /* missing Os macro, using "" */
61	MANDOCERR_PROLOG_REP, /* duplicate prologue macro: macro */
62	MANDOCERR_PROLOG_LATE, /* late prologue macro: macro */
63	MANDOCERR_DT_LATE, /* skipping late title macro: Dt args */
64	MANDOCERR_PROLOG_ORDER, /* prologue macros out of order: macros */
65
66	/* related to document structure */
67	MANDOCERR_SO, /* .so is fragile, better use ln(1): so path */
68	MANDOCERR_DOC_EMPTY, /* no document body */
69	MANDOCERR_SEC_BEFORE, /* content before first section header: macro */
70	MANDOCERR_NAMESEC_FIRST, /* first section is not NAME: Sh title */
71	MANDOCERR_NAMESEC_BAD, /* bad NAME section contents: macro */
72	MANDOCERR_SEC_ORDER, /* sections out of conventional order: Sh title */
73	MANDOCERR_SEC_REP, /* duplicate section title: Sh title */
74	MANDOCERR_SEC_MSEC, /* unexpected section: Sh title for ... only */
75
76	/* related to macros and nesting */
77	MANDOCERR_MACRO_OBS, /* obsolete macro: macro */
78	MANDOCERR_PAR_SKIP, /* skipping paragraph macro: macro ... */
79	MANDOCERR_PAR_MOVE, /* moving paragraph macro out of list: macro */
80	MANDOCERR_NS_SKIP, /* skipping no-space macro */
81	MANDOCERR_BLK_NEST, /* blocks badly nested: macro ... */
82	MANDOCERR_BD_NEST, /* nested displays are not portable: macro ... */
83	MANDOCERR_BL_MOVE, /* moving content out of list: macro */
84	MANDOCERR_VT_CHILD, /* .Vt block has child macro: macro */
85	MANDOCERR_FI_SKIP, /* fill mode already enabled, skipping: fi */
86	MANDOCERR_NF_SKIP, /* fill mode already disabled, skipping: nf */
87	MANDOCERR_BLK_LINE, /* line scope broken: macro breaks macro */
88
89	/* related to missing arguments */
90	MANDOCERR_REQ_EMPTY, /* skipping empty request: request */
91	MANDOCERR_COND_EMPTY, /* conditional request controls empty scope */
92	MANDOCERR_MACRO_EMPTY, /* skipping empty macro: macro */
93	MANDOCERR_ARG_EMPTY, /* empty argument, using 0n: macro arg */
94	MANDOCERR_ARGCWARN, /* argument count wrong */
95	MANDOCERR_BD_NOTYPE, /* missing display type, using -ragged: Bd */
96	MANDOCERR_BL_LATETYPE, /* list type is not the first argument: Bl arg */
97	MANDOCERR_BL_NOWIDTH, /* missing -width in -tag list, using 8n */
98	MANDOCERR_EX_NONAME, /* missing utility name, using "": Ex */
99	MANDOCERR_IT_NOHEAD, /* empty head in list item: Bl -type It */
100	MANDOCERR_IT_NOBODY, /* empty list item: Bl -type It */
101	MANDOCERR_BF_NOFONT, /* missing font type, using \fR: Bf */
102	MANDOCERR_BF_BADFONT, /* unknown font type, using \fR: Bf font */
103	MANDOCERR_ARG_STD, /* missing -std argument, adding it: macro */
104
105	/* related to bad arguments */
106	MANDOCERR_ARG_QUOTE, /* unterminated quoted argument */
107	MANDOCERR_ARG_REP, /* duplicate argument: macro arg */
108	MANDOCERR_AN_REP, /* skipping duplicate argument: An -arg */
109	MANDOCERR_BD_REP, /* skipping duplicate display type: Bd -type */
110	MANDOCERR_BL_REP, /* skipping duplicate list type: Bl -type */
111	MANDOCERR_BL_SKIPW, /* skipping -width argument: Bl -type */
112	MANDOCERR_AT_BAD, /* unknown AT&T UNIX version: At version */
113	MANDOCERR_RS_BAD, /* invalid content in Rs block: macro */
114	MANDOCERR_SM_BAD, /* invalid Boolean argument: macro arg */
115	MANDOCERR_FT_BAD, /* unknown font, skipping request: ft font */
116
117	/* related to plain text */
118	MANDOCERR_FI_BLANK, /* blank line in fill mode, using .sp */
119	MANDOCERR_FI_TAB, /* tab in filled text */
120	MANDOCERR_SPACE_EOL, /* whitespace at end of input line */
121	MANDOCERR_COMMENT_BAD, /* bad comment style */
122	MANDOCERR_ESC_BAD, /* invalid escape sequence: esc */
123	MANDOCERR_STR_UNDEF, /* undefined string, using "": name */
124
125	MANDOCERR_ERROR, /* ===== start of errors ===== */
126
127	/* related to equations */
128	MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
129	MANDOCERR_EQNSCOPE, /* equation scope open on exit */
130	MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
131	MANDOCERR_EQNEOF, /* unexpected end of equation */
132	MANDOCERR_EQNSYNT, /* equation syntax error */
133
134	/* related to tables */
135	MANDOCERR_TBL, /* bad table syntax */
136	MANDOCERR_TBLOPT, /* bad table option */
137	MANDOCERR_TBLLAYOUT, /* bad table layout */
138	MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
139	MANDOCERR_TBLNODATA, /* no table data cells specified */
140	MANDOCERR_TBLIGNDATA, /* ignore data in cell */
141	MANDOCERR_TBLBLOCK, /* data block still open */
142	MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
143
144	/* related to document structure and macros */
145	MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
146	MANDOCERR_BADCHAR, /* skipping bad character: number */
147	MANDOCERR_MACRO, /* skipping unknown macro: macro */
148	MANDOCERR_IT_STRAY, /* skipping item outside list: It ... */
149	MANDOCERR_TA_STRAY, /* skipping column outside column list: Ta */
150	MANDOCERR_BLK_NOTOPEN, /* skipping end of block that is not open */
151	MANDOCERR_BLK_BROKEN, /* inserting missing end of block: macro ... */
152	MANDOCERR_BLK_NOEND, /* appending missing end of block: macro */
153
154	/* related to request and macro arguments */
155	MANDOCERR_NAMESC, /* escaped character not allowed in a name: name */
156	MANDOCERR_ARGCOUNT, /* argument count wrong */
157	MANDOCERR_BL_NOTYPE, /* missing list type, using -item: Bl */
158	MANDOCERR_NM_NONAME, /* missing manual name, using "": Nm */
159	MANDOCERR_OS_UNAME, /* uname(3) system call failed, using UNKNOWN */
160	MANDOCERR_ST_BAD, /* unknown standard specifier: St standard */
161	MANDOCERR_IT_NONUM, /* skipping request without numeric argument */
162	MANDOCERR_ARG_SKIP, /* skipping all arguments: macro args */
163	MANDOCERR_ARG_EXCESS, /* skipping excess arguments: macro ... args */
164
165	MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
166
167	MANDOCERR_TOOLARGE, /* input too large */
168	MANDOCERR_BD_FILE, /* NOT IMPLEMENTED: Bd -file */
169	MANDOCERR_SO_PATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
170	MANDOCERR_SO_FAIL, /* .so request failed */
171
172	/* ===== system errors ===== */
173
174	MANDOCERR_SYSOPEN, /* cannot open file */
175	MANDOCERR_SYSSTAT, /* cannot stat file */
176	MANDOCERR_SYSREAD, /* cannot read file */
177
178	MANDOCERR_MAX
179};
180
181struct	tbl_opts {
182	char		  tab; /* cell-separator */
183	char		  decimal; /* decimal point */
184	int		  linesize;
185	int		  opts;
186#define	TBL_OPT_CENTRE	 (1 << 0)
187#define	TBL_OPT_EXPAND	 (1 << 1)
188#define	TBL_OPT_BOX	 (1 << 2)
189#define	TBL_OPT_DBOX	 (1 << 3)
190#define	TBL_OPT_ALLBOX	 (1 << 4)
191#define	TBL_OPT_NOKEEP	 (1 << 5)
192#define	TBL_OPT_NOSPACE	 (1 << 6)
193	int		  cols; /* number of columns */
194};
195
196/*
197 * The head of a table specifies all of its columns.  When formatting a
198 * tbl_span, iterate over these and plug in data from the tbl_span when
199 * appropriate, using tbl_cell as a guide to placement.
200 */
201struct	tbl_head {
202	int		  ident; /* 0 <= unique id < cols */
203	int		  vert; /* width of preceding vertical line */
204	struct tbl_head	 *next;
205	struct tbl_head	 *prev;
206};
207
208enum	tbl_cellt {
209	TBL_CELL_CENTRE, /* c, C */
210	TBL_CELL_RIGHT, /* r, R */
211	TBL_CELL_LEFT, /* l, L */
212	TBL_CELL_NUMBER, /* n, N */
213	TBL_CELL_SPAN, /* s, S */
214	TBL_CELL_LONG, /* a, A */
215	TBL_CELL_DOWN, /* ^ */
216	TBL_CELL_HORIZ, /* _, - */
217	TBL_CELL_DHORIZ, /* = */
218	TBL_CELL_MAX
219};
220
221/*
222 * A cell in a layout row.
223 */
224struct	tbl_cell {
225	struct tbl_cell	 *next;
226	int		  vert; /* width of preceding vertical line */
227	enum tbl_cellt	  pos;
228	size_t		  spacing;
229	int		  flags;
230#define	TBL_CELL_TALIGN	 (1 << 0) /* t, T */
231#define	TBL_CELL_BALIGN	 (1 << 1) /* d, D */
232#define	TBL_CELL_BOLD	 (1 << 2) /* fB, B, b */
233#define	TBL_CELL_ITALIC	 (1 << 3) /* fI, I, i */
234#define	TBL_CELL_EQUAL	 (1 << 4) /* e, E */
235#define	TBL_CELL_UP	 (1 << 5) /* u, U */
236#define	TBL_CELL_WIGN	 (1 << 6) /* z, Z */
237	struct tbl_head	 *head;
238};
239
240/*
241 * A layout row.
242 */
243struct	tbl_row {
244	struct tbl_row	 *next;
245	struct tbl_cell	 *first;
246	struct tbl_cell	 *last;
247	int		  vert; /* trailing vertical line */
248};
249
250enum	tbl_datt {
251	TBL_DATA_NONE, /* has no data */
252	TBL_DATA_DATA, /* consists of data/string */
253	TBL_DATA_HORIZ, /* horizontal line */
254	TBL_DATA_DHORIZ, /* double-horizontal line */
255	TBL_DATA_NHORIZ, /* squeezed horizontal line */
256	TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
257};
258
259/*
260 * A cell within a row of data.  The "string" field contains the actual
261 * string value that's in the cell.  The rest is layout.
262 */
263struct	tbl_dat {
264	struct tbl_cell	 *layout; /* layout cell */
265	int		  spans; /* how many spans follow */
266	struct tbl_dat	 *next;
267	char		 *string; /* data (NULL if not TBL_DATA_DATA) */
268	enum tbl_datt	  pos;
269};
270
271enum	tbl_spant {
272	TBL_SPAN_DATA, /* span consists of data */
273	TBL_SPAN_HORIZ, /* span is horizontal line */
274	TBL_SPAN_DHORIZ /* span is double horizontal line */
275};
276
277/*
278 * A row of data in a table.
279 */
280struct	tbl_span {
281	struct tbl_opts	 *opts;
282	struct tbl_head	 *head;
283	struct tbl_row	 *layout; /* layout row */
284	struct tbl_dat	 *first;
285	struct tbl_dat	 *last;
286	int		  line; /* parse line */
287	int		  flags;
288#define	TBL_SPAN_FIRST	 (1 << 0)
289#define	TBL_SPAN_LAST	 (1 << 1)
290	enum tbl_spant	  pos;
291	struct tbl_span	 *next;
292};
293
294enum	eqn_boxt {
295	EQN_ROOT, /* root of parse tree */
296	EQN_TEXT, /* text (number, variable, whatever) */
297	EQN_SUBEXPR, /* nested `eqn' subexpression */
298	EQN_LIST, /* subexpressions list */
299	EQN_MATRIX /* matrix subexpression */
300};
301
302enum	eqn_markt {
303	EQNMARK_NONE = 0,
304	EQNMARK_DOT,
305	EQNMARK_DOTDOT,
306	EQNMARK_HAT,
307	EQNMARK_TILDE,
308	EQNMARK_VEC,
309	EQNMARK_DYAD,
310	EQNMARK_BAR,
311	EQNMARK_UNDER,
312	EQNMARK__MAX
313};
314
315enum	eqn_fontt {
316	EQNFONT_NONE = 0,
317	EQNFONT_ROMAN,
318	EQNFONT_BOLD,
319	EQNFONT_FAT,
320	EQNFONT_ITALIC,
321	EQNFONT__MAX
322};
323
324enum	eqn_post {
325	EQNPOS_NONE = 0,
326	EQNPOS_OVER,
327	EQNPOS_SUP,
328	EQNPOS_SUB,
329	EQNPOS_TO,
330	EQNPOS_FROM,
331	EQNPOS__MAX
332};
333
334enum	eqn_pilet {
335	EQNPILE_NONE = 0,
336	EQNPILE_PILE,
337	EQNPILE_CPILE,
338	EQNPILE_RPILE,
339	EQNPILE_LPILE,
340	EQNPILE_COL,
341	EQNPILE_CCOL,
342	EQNPILE_RCOL,
343	EQNPILE_LCOL,
344	EQNPILE__MAX
345};
346
347 /*
348 * A "box" is a parsed mathematical expression as defined by the eqn.7
349 * grammar.
350 */
351struct	eqn_box {
352	int		  size; /* font size of expression */
353#define	EQN_DEFSIZE	  INT_MIN
354	enum eqn_boxt	  type; /* type of node */
355	struct eqn_box	 *first; /* first child node */
356	struct eqn_box	 *last; /* last child node */
357	struct eqn_box	 *next; /* node sibling */
358	struct eqn_box	 *parent; /* node sibling */
359	char		 *text; /* text (or NULL) */
360	char		 *left;
361	char		 *right;
362	enum eqn_post	  pos; /* position of next box */
363	enum eqn_markt	  mark; /* a mark about the box */
364	enum eqn_fontt	  font; /* font of box */
365	enum eqn_pilet	  pile; /* equation piling */
366};
367
368/*
369 * An equation consists of a tree of expressions starting at a given
370 * line and position.
371 */
372struct	eqn {
373	char		 *name; /* identifier (or NULL) */
374	struct eqn_box	 *root; /* root mathematical expression */
375	int		  ln; /* invocation line */
376	int		  pos; /* invocation position */
377};
378
379/*
380 * Parse options.
381 */
382#define	MPARSE_MDOC	1  /* assume -mdoc */
383#define	MPARSE_MAN	2  /* assume -man */
384#define	MPARSE_SO	4  /* honour .so requests */
385#define	MPARSE_QUICK	8  /* abort the parse early */
386
387enum	mandoc_esc {
388	ESCAPE_ERROR = 0, /* bail! unparsable escape */
389	ESCAPE_IGNORE, /* escape to be ignored */
390	ESCAPE_SPECIAL, /* a regular special character */
391	ESCAPE_FONT, /* a generic font mode */
392	ESCAPE_FONTBOLD, /* bold font mode */
393	ESCAPE_FONTITALIC, /* italic font mode */
394	ESCAPE_FONTBI, /* bold italic font mode */
395	ESCAPE_FONTROMAN, /* roman font mode */
396	ESCAPE_FONTPREV, /* previous font mode */
397	ESCAPE_NUMBERED, /* a numbered glyph */
398	ESCAPE_UNICODE, /* a unicode codepoint */
399	ESCAPE_NOSPACE, /* suppress space if the last on a line */
400	ESCAPE_SKIPCHAR /* skip the next character */
401};
402
403typedef	void	(*mandocmsg)(enum mandocerr, enum mandoclevel,
404			const char *, int, int, const char *);
405
406struct	mparse;
407struct	mchars;
408struct	mdoc;
409struct	man;
410
411__BEGIN_DECLS
412
413enum mandoc_esc	  mandoc_escape(const char **, const char **, int *);
414struct mchars	 *mchars_alloc(void);
415void		  mchars_free(struct mchars *);
416char		  mchars_num2char(const char *, size_t);
417int		  mchars_num2uc(const char *, size_t);
418int		  mchars_spec2cp(const struct mchars *,
419			const char *, size_t);
420const char	 *mchars_spec2str(const struct mchars *,
421			const char *, size_t, size_t *);
422struct mparse	 *mparse_alloc(int, enum mandoclevel, mandocmsg,
423			const char *);
424void		  mparse_free(struct mparse *);
425void		  mparse_keep(struct mparse *);
426enum mandoclevel  mparse_readfd(struct mparse *, int, const char *);
427enum mandoclevel  mparse_readmem(struct mparse *, const void *, size_t,
428			const char *);
429void		  mparse_reset(struct mparse *);
430void		  mparse_result(struct mparse *,
431			struct mdoc **, struct man **, char **);
432const char	 *mparse_getkeep(const struct mparse *);
433const char	 *mparse_strerror(enum mandocerr);
434const char	 *mparse_strlevel(enum mandoclevel);
435
436__END_DECLS
437
438#endif /*!MANDOC_H*/
439