1 /* $Id: mandoc.h,v 1.176 2014/12/01 08:05:52 schwarze Exp $ */
3 * Copyright (c) 2010, 2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2010-2014 Ingo Schwarze <schwarze@openbsd.org>
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.
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.
19 #define ASCII_NBRSP 31 /* non-breaking space */
20 #define ASCII_HYPH 30 /* breakable hyphen */
21 #define ASCII_BREAK 29 /* breakable zero-width space */
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
32 MANDOCLEVEL_WARNING, /* warnings: syntax, whitespace, etc. */
33 MANDOCLEVEL_ERROR, /* input has been thrown away */
34 MANDOCLEVEL_FATAL, /* input is borked */
35 MANDOCLEVEL_BADARG, /* bad argument in invocation */
36 MANDOCLEVEL_SYSERR, /* system error */
41 * All possible things that can go wrong within a parse, be it libroff,
47 MANDOCERR_WARNING, /* ===== start of warnings ===== */
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 */
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_SEC_ORDER, /* sections out of conventional order: Sh title */
70 MANDOCERR_SEC_REP, /* duplicate section title: Sh title */
71 MANDOCERR_SEC_MSEC, /* unexpected section: Sh title for ... only */
72 MANDOCERR_XR_ORDER, /* unusual Xr order: ... after ... */
73 MANDOCERR_XR_PUNCT, /* unusual Xr punctuation: ... after ... */
74 MANDOCERR_AN_MISSING, /* AUTHORS section without An macro */
76 /* related to macros and nesting */
77 MANDOCERR_MACRO_OBS, /* obsolete macro: macro */
78 MANDOCERR_MACRO_CALL, /* macro neither callable nor escaped: macro */
79 MANDOCERR_PAR_SKIP, /* skipping paragraph macro: macro ... */
80 MANDOCERR_PAR_MOVE, /* moving paragraph macro out of list: macro */
81 MANDOCERR_NS_SKIP, /* skipping no-space macro */
82 MANDOCERR_BLK_NEST, /* blocks badly nested: macro ... */
83 MANDOCERR_BD_NEST, /* nested displays are not portable: macro ... */
84 MANDOCERR_BL_MOVE, /* moving content out of list: macro */
85 MANDOCERR_VT_CHILD, /* .Vt block has child macro: macro */
86 MANDOCERR_FI_SKIP, /* fill mode already enabled, skipping: fi */
87 MANDOCERR_NF_SKIP, /* fill mode already disabled, skipping: nf */
88 MANDOCERR_BLK_LINE, /* line scope broken: macro breaks macro */
90 /* related to missing arguments */
91 MANDOCERR_REQ_EMPTY, /* skipping empty request: request */
92 MANDOCERR_COND_EMPTY, /* conditional request controls empty scope */
93 MANDOCERR_MACRO_EMPTY, /* skipping empty macro: macro */
94 MANDOCERR_ARG_EMPTY, /* empty argument, using 0n: macro arg */
95 MANDOCERR_ARGCWARN, /* argument count wrong */
96 MANDOCERR_BD_NOTYPE, /* missing display type, using -ragged: Bd */
97 MANDOCERR_BL_LATETYPE, /* list type is not the first argument: Bl arg */
98 MANDOCERR_BL_NOWIDTH, /* missing -width in -tag list, using 8n */
99 MANDOCERR_EX_NONAME, /* missing utility name, using "": Ex */
100 MANDOCERR_IT_NOHEAD, /* empty head in list item: Bl -type It */
101 MANDOCERR_IT_NOBODY, /* empty list item: Bl -type It */
102 MANDOCERR_BF_NOFONT, /* missing font type, using \fR: Bf */
103 MANDOCERR_BF_BADFONT, /* unknown font type, using \fR: Bf font */
104 MANDOCERR_PF_SKIP, /* nothing follows prefix: Pf arg */
105 MANDOCERR_ARG_STD, /* missing -std argument, adding it: macro */
106 MANDOCERR_EQN_NOBOX, /* missing eqn box, using "": op */
108 /* related to bad arguments */
109 MANDOCERR_ARG_QUOTE, /* unterminated quoted argument */
110 MANDOCERR_ARG_REP, /* duplicate argument: macro arg */
111 MANDOCERR_AN_REP, /* skipping duplicate argument: An -arg */
112 MANDOCERR_BD_REP, /* skipping duplicate display type: Bd -type */
113 MANDOCERR_BL_REP, /* skipping duplicate list type: Bl -type */
114 MANDOCERR_BL_SKIPW, /* skipping -width argument: Bl -type */
115 MANDOCERR_AT_BAD, /* unknown AT&T UNIX version: At version */
116 MANDOCERR_FA_COMMA, /* comma in function argument: arg */
117 MANDOCERR_FN_PAREN, /* parenthesis in function name: arg */
118 MANDOCERR_RS_BAD, /* invalid content in Rs block: macro */
119 MANDOCERR_SM_BAD, /* invalid Boolean argument: macro arg */
120 MANDOCERR_FT_BAD, /* unknown font, skipping request: ft font */
122 /* related to plain text */
123 MANDOCERR_FI_BLANK, /* blank line in fill mode, using .sp */
124 MANDOCERR_FI_TAB, /* tab in filled text */
125 MANDOCERR_SPACE_EOL, /* whitespace at end of input line */
126 MANDOCERR_COMMENT_BAD, /* bad comment style */
127 MANDOCERR_ESC_BAD, /* invalid escape sequence: esc */
128 MANDOCERR_STR_UNDEF, /* undefined string, using "": name */
130 MANDOCERR_ERROR, /* ===== start of errors ===== */
132 /* related to equations */
133 MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
134 MANDOCERR_EQNSCOPE, /* equation scope open on exit */
135 MANDOCERR_EQNBADSCOPE, /* overlapping equation scopes */
136 MANDOCERR_EQNEOF, /* unexpected end of equation */
138 /* related to tables */
139 MANDOCERR_TBL, /* bad table syntax */
140 MANDOCERR_TBLOPT, /* bad table option */
141 MANDOCERR_TBLLAYOUT, /* bad table layout */
142 MANDOCERR_TBLNOLAYOUT, /* no table layout cells specified */
143 MANDOCERR_TBLNODATA, /* no table data cells specified */
144 MANDOCERR_TBLIGNDATA, /* ignore data in cell */
145 MANDOCERR_TBLBLOCK, /* data block still open */
146 MANDOCERR_TBLEXTRADAT, /* ignoring extra data cells */
148 /* related to document structure and macros */
149 MANDOCERR_ROFFLOOP, /* input stack limit exceeded, infinite loop? */
150 MANDOCERR_BADCHAR, /* skipping bad character: number */
151 MANDOCERR_MACRO, /* skipping unknown macro: macro */
152 MANDOCERR_IT_STRAY, /* skipping item outside list: It ... */
153 MANDOCERR_TA_STRAY, /* skipping column outside column list: Ta */
154 MANDOCERR_BLK_NOTOPEN, /* skipping end of block that is not open */
155 MANDOCERR_BLK_BROKEN, /* inserting missing end of block: macro ... */
156 MANDOCERR_BLK_NOEND, /* appending missing end of block: macro */
158 /* related to request and macro arguments */
159 MANDOCERR_NAMESC, /* escaped character not allowed in a name: name */
160 MANDOCERR_ARGCOUNT, /* argument count wrong */
161 MANDOCERR_BD_FILE, /* NOT IMPLEMENTED: Bd -file */
162 MANDOCERR_BL_NOTYPE, /* missing list type, using -item: Bl */
163 MANDOCERR_NM_NONAME, /* missing manual name, using "": Nm */
164 MANDOCERR_OS_UNAME, /* uname(3) system call failed, using UNKNOWN */
165 MANDOCERR_ST_BAD, /* unknown standard specifier: St standard */
166 MANDOCERR_IT_NONUM, /* skipping request without numeric argument */
167 MANDOCERR_ARG_SKIP, /* skipping all arguments: macro args */
168 MANDOCERR_ARG_EXCESS, /* skipping excess arguments: macro ... args */
169 MANDOCERR_DIVZERO, /* divide by zero */
171 MANDOCERR_FATAL, /* ===== start of fatal errors ===== */
173 MANDOCERR_TOOLARGE, /* input too large */
174 MANDOCERR_SO_PATH, /* NOT IMPLEMENTED: .so with absolute path or ".." */
175 MANDOCERR_SO_FAIL, /* .so request failed */
177 /* ===== system errors ===== */
179 MANDOCERR_SYSDUP, /* cannot dup file descriptor */
180 MANDOCERR_SYSEXEC, /* cannot exec */
181 MANDOCERR_SYSEXIT, /* gunzip failed with code */
182 MANDOCERR_SYSFORK, /* cannot fork */
183 MANDOCERR_SYSOPEN, /* cannot open file */
184 MANDOCERR_SYSPIPE, /* cannot open pipe */
185 MANDOCERR_SYSREAD, /* cannot read file */
186 MANDOCERR_SYSSIG, /* gunzip died from signal */
187 MANDOCERR_SYSSTAT, /* cannot stat file */
188 MANDOCERR_SYSWAIT, /* wait failed */
194 char tab; /* cell-separator */
195 char decimal; /* decimal point */
198 #define TBL_OPT_CENTRE (1 << 0)
199 #define TBL_OPT_EXPAND (1 << 1)
200 #define TBL_OPT_BOX (1 << 2)
201 #define TBL_OPT_DBOX (1 << 3)
202 #define TBL_OPT_ALLBOX (1 << 4)
203 #define TBL_OPT_NOKEEP (1 << 5)
204 #define TBL_OPT_NOSPACE (1 << 6)
205 int cols; /* number of columns */
209 * The head of a table specifies all of its columns. When formatting a
210 * tbl_span, iterate over these and plug in data from the tbl_span when
211 * appropriate, using tbl_cell as a guide to placement.
214 int ident; /* 0 <= unique id < cols */
215 int vert; /* width of preceding vertical line */
216 struct tbl_head *next;
217 struct tbl_head *prev;
221 TBL_CELL_CENTRE, /* c, C */
222 TBL_CELL_RIGHT, /* r, R */
223 TBL_CELL_LEFT, /* l, L */
224 TBL_CELL_NUMBER, /* n, N */
225 TBL_CELL_SPAN, /* s, S */
226 TBL_CELL_LONG, /* a, A */
227 TBL_CELL_DOWN, /* ^ */
228 TBL_CELL_HORIZ, /* _, - */
229 TBL_CELL_DHORIZ, /* = */
234 * A cell in a layout row.
237 struct tbl_cell *next;
238 int vert; /* width of preceding vertical line */
242 #define TBL_CELL_TALIGN (1 << 0) /* t, T */
243 #define TBL_CELL_BALIGN (1 << 1) /* d, D */
244 #define TBL_CELL_BOLD (1 << 2) /* fB, B, b */
245 #define TBL_CELL_ITALIC (1 << 3) /* fI, I, i */
246 #define TBL_CELL_EQUAL (1 << 4) /* e, E */
247 #define TBL_CELL_UP (1 << 5) /* u, U */
248 #define TBL_CELL_WIGN (1 << 6) /* z, Z */
249 #define TBL_CELL_WMAX (1 << 7) /* x, X */
250 struct tbl_head *head;
257 struct tbl_row *next;
258 struct tbl_cell *first;
259 struct tbl_cell *last;
260 int vert; /* trailing vertical line */
264 TBL_DATA_NONE, /* has no data */
265 TBL_DATA_DATA, /* consists of data/string */
266 TBL_DATA_HORIZ, /* horizontal line */
267 TBL_DATA_DHORIZ, /* double-horizontal line */
268 TBL_DATA_NHORIZ, /* squeezed horizontal line */
269 TBL_DATA_NDHORIZ /* squeezed double-horizontal line */
273 * A cell within a row of data. The "string" field contains the actual
274 * string value that's in the cell. The rest is layout.
277 struct tbl_cell *layout; /* layout cell */
278 int spans; /* how many spans follow */
279 struct tbl_dat *next;
280 char *string; /* data (NULL if not TBL_DATA_DATA) */
285 TBL_SPAN_DATA, /* span consists of data */
286 TBL_SPAN_HORIZ, /* span is horizontal line */
287 TBL_SPAN_DHORIZ /* span is double horizontal line */
291 * A row of data in a table.
294 struct tbl_opts *opts;
295 struct tbl_head *head;
296 struct tbl_row *layout; /* layout row */
297 struct tbl_dat *first;
298 struct tbl_dat *last;
299 int line; /* parse line */
301 #define TBL_SPAN_FIRST (1 << 0)
302 #define TBL_SPAN_LAST (1 << 1)
304 struct tbl_span *next;
308 EQN_ROOT, /* root of parse tree */
309 EQN_TEXT, /* text (number, variable, whatever) */
310 EQN_SUBEXPR, /* nested `eqn' subexpression */
311 EQN_LIST, /* list (braces, etc.) */
312 EQN_LISTONE, /* singleton list */
313 EQN_PILE, /* vertical pile */
314 EQN_MATRIX /* pile of piles */
353 * A "box" is a parsed mathematical expression as defined by the eqn.7
357 int size; /* font size of expression */
358 #define EQN_DEFSIZE INT_MIN
359 enum eqn_boxt type; /* type of node */
360 struct eqn_box *first; /* first child node */
361 struct eqn_box *last; /* last child node */
362 struct eqn_box *next; /* node sibling */
363 struct eqn_box *prev; /* node sibling */
364 struct eqn_box *parent; /* node sibling */
365 char *text; /* text (or NULL) */
366 char *left; /* fence left-hand */
367 char *right; /* fence right-hand */
368 char *top; /* expression over-symbol */
369 char *bottom; /* expression under-symbol */
370 size_t args; /* arguments in parent */
371 size_t expectargs; /* max arguments in parent */
372 enum eqn_post pos; /* position of next box */
373 enum eqn_fontt font; /* font of box */
374 enum eqn_pilet pile; /* equation piling */
378 * An equation consists of a tree of expressions starting at a given
382 char *name; /* identifier (or NULL) */
383 struct eqn_box *root; /* root mathematical expression */
384 int ln; /* invocation line */
385 int pos; /* invocation position */
391 #define MPARSE_MDOC 1 /* assume -mdoc */
392 #define MPARSE_MAN 2 /* assume -man */
393 #define MPARSE_SO 4 /* honour .so requests */
394 #define MPARSE_QUICK 8 /* abort the parse early */
395 #define MPARSE_UTF8 16 /* accept UTF-8 input */
396 #define MPARSE_LATIN1 32 /* accept ISO-LATIN-1 input */
399 ESCAPE_ERROR = 0, /* bail! unparsable escape */
400 ESCAPE_IGNORE, /* escape to be ignored */
401 ESCAPE_SPECIAL, /* a regular special character */
402 ESCAPE_FONT, /* a generic font mode */
403 ESCAPE_FONTBOLD, /* bold font mode */
404 ESCAPE_FONTITALIC, /* italic font mode */
405 ESCAPE_FONTBI, /* bold italic font mode */
406 ESCAPE_FONTROMAN, /* roman font mode */
407 ESCAPE_FONTPREV, /* previous font mode */
408 ESCAPE_NUMBERED, /* a numbered glyph */
409 ESCAPE_UNICODE, /* a unicode codepoint */
410 ESCAPE_NOSPACE, /* suppress space if the last on a line */
411 ESCAPE_SKIPCHAR /* skip the next character */
414 typedef void (*mandocmsg)(enum mandocerr, enum mandoclevel,
415 const char *, int, int, const char *);
424 enum mandoc_esc mandoc_escape(const char **, const char **, int *);
425 struct mchars *mchars_alloc(void);
426 void mchars_free(struct mchars *);
427 int mchars_num2char(const char *, size_t);
428 const char *mchars_uc2str(int);
429 int mchars_num2uc(const char *, size_t);
430 int mchars_spec2cp(const struct mchars *,
431 const char *, size_t);
432 const char *mchars_spec2str(const struct mchars *,
433 const char *, size_t, size_t *);
434 struct mparse *mparse_alloc(int, enum mandoclevel, mandocmsg,
435 const struct mchars *, const char *);
436 void mparse_free(struct mparse *);
437 void mparse_keep(struct mparse *);
438 enum mandoclevel mparse_open(struct mparse *, int *, const char *);
439 enum mandoclevel mparse_readfd(struct mparse *, int, const char *);
440 enum mandoclevel mparse_readmem(struct mparse *, void *, size_t,
442 void mparse_reset(struct mparse *);
443 void mparse_result(struct mparse *,
444 struct mdoc **, struct man **, char **);
445 const char *mparse_getkeep(const struct mparse *);
446 const char *mparse_strerror(enum mandocerr);
447 const char *mparse_strlevel(enum mandoclevel);
448 enum mandoclevel mparse_wait(struct mparse *);