Annotation of src/usr.bin/mandoc/roff.c, Revision 1.204
1.204 ! schwarze 1: /* $OpenBSD: roff.c,v 1.203 2018/08/16 13:49:40 schwarze Exp $ */
1.1 schwarze 2: /*
1.139 schwarze 3: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.197 schwarze 4: * Copyright (c) 2010-2015, 2017, 2018 Ingo Schwarze <schwarze@openbsd.org>
1.1 schwarze 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: *
1.16 schwarze 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1 schwarze 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.16 schwarze 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1 schwarze 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: */
1.99 schwarze 18: #include <sys/types.h>
19:
1.1 schwarze 20: #include <assert.h>
1.3 schwarze 21: #include <ctype.h>
1.117 schwarze 22: #include <limits.h>
1.167 schwarze 23: #include <stddef.h>
24: #include <stdint.h>
1.51 schwarze 25: #include <stdio.h>
1.1 schwarze 26: #include <stdlib.h>
27: #include <string.h>
28:
29: #include "mandoc.h"
1.74 schwarze 30: #include "mandoc_aux.h"
1.167 schwarze 31: #include "mandoc_ohash.h"
1.137 schwarze 32: #include "roff.h"
1.98 daniel 33: #include "libmandoc.h"
1.138 schwarze 34: #include "roff_int.h"
1.27 schwarze 35: #include "libroff.h"
1.2 schwarze 36:
1.43 schwarze 37: /* Maximum number of string expansions per line, to break infinite loops. */
38: #define EXPAND_LIMIT 1000
39:
1.187 schwarze 40: /* Types of definitions of macros and strings. */
41: #define ROFFDEF_USER (1 << 1) /* User-defined. */
42: #define ROFFDEF_PRE (1 << 2) /* Predefined. */
43: #define ROFFDEF_REN (1 << 3) /* Renamed standard macro. */
44: #define ROFFDEF_STD (1 << 4) /* mdoc(7) or man(7) macro. */
45: #define ROFFDEF_ANY (ROFFDEF_USER | ROFFDEF_PRE | \
46: ROFFDEF_REN | ROFFDEF_STD)
1.197 schwarze 47: #define ROFFDEF_UNDEF (1 << 5) /* Completely undefined. */
1.187 schwarze 48:
1.138 schwarze 49: /* --- data types --------------------------------------------------------- */
50:
1.41 schwarze 51: /*
1.42 schwarze 52: * An incredibly-simple string buffer.
53: */
1.8 schwarze 54: struct roffstr {
1.42 schwarze 55: char *p; /* nil-terminated buffer */
56: size_t sz; /* saved strlen(p) */
57: };
58:
59: /*
60: * A key-value roffstr pair as part of a singly-linked list.
61: */
62: struct roffkv {
63: struct roffstr key;
64: struct roffstr val;
65: struct roffkv *next; /* next in list */
1.8 schwarze 66: };
67:
1.52 schwarze 68: /*
69: * A single number register as part of a singly-linked list.
70: */
71: struct roffreg {
72: struct roffstr key;
1.53 schwarze 73: int val;
1.199 schwarze 74: int step;
1.52 schwarze 75: struct roffreg *next;
76: };
77:
1.167 schwarze 78: /*
79: * Association of request and macro names with token IDs.
80: */
81: struct roffreq {
82: enum roff_tok tok;
83: char name[];
84: };
85:
1.1 schwarze 86: struct roff {
1.35 schwarze 87: struct mparse *parse; /* parse point */
1.168 schwarze 88: struct roff_man *man; /* mdoc or man parser */
1.1 schwarze 89: struct roffnode *last; /* leaf of stack */
1.96 schwarze 90: int *rstack; /* stack of inverted `ie' values */
1.167 schwarze 91: struct ohash *reqtab; /* request lookup table */
1.52 schwarze 92: struct roffreg *regtab; /* number registers */
1.42 schwarze 93: struct roffkv *strtab; /* user-defined strings & macros */
1.178 schwarze 94: struct roffkv *rentab; /* renamed strings & macros */
1.42 schwarze 95: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
96: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.16 schwarze 97: const char *current_string; /* value of last called user macro */
1.27 schwarze 98: struct tbl_node *first_tbl; /* first table parsed */
99: struct tbl_node *last_tbl; /* last table parsed */
100: struct tbl_node *tbl; /* current table being parsed */
1.191 schwarze 101: struct eqn_node *last_eqn; /* equation parser */
102: struct eqn_node *eqn; /* active equation parser */
1.102 schwarze 103: int eqn_inline; /* current equation is inline */
1.96 schwarze 104: int options; /* parse options */
105: int rstacksz; /* current size limit of rstack */
106: int rstackpos; /* position in rstack */
1.99 schwarze 107: int format; /* current file in mdoc or man format */
1.145 schwarze 108: int argc; /* number of args of the last macro */
1.96 schwarze 109: char control; /* control character */
1.175 schwarze 110: char escape; /* escape character */
1.1 schwarze 111: };
112:
113: struct roffnode {
1.166 schwarze 114: enum roff_tok tok; /* type of node */
1.1 schwarze 115: struct roffnode *parent; /* up one in stack */
116: int line; /* parse line */
117: int col; /* parse col */
1.16 schwarze 118: char *name; /* node name, e.g. macro name */
1.2 schwarze 119: char *end; /* end-rules: custom token */
120: int endspan; /* end-rules: next-line or infty */
1.71 schwarze 121: int rule; /* current evaluation rule */
1.1 schwarze 122: };
123:
124: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.166 schwarze 125: enum roff_tok tok, /* tok of macro */ \
1.110 schwarze 126: struct buf *buf, /* input buffer */ \
1.1 schwarze 127: int ln, /* parse line */ \
1.2 schwarze 128: int ppos, /* original pos in buffer */ \
129: int pos, /* current pos in buffer */ \
130: int *offs /* reset offset of buffer data */
1.1 schwarze 131:
132: typedef enum rofferr (*roffproc)(ROFF_ARGS);
133:
134: struct roffmac {
1.2 schwarze 135: roffproc proc; /* process new macro */
136: roffproc text; /* process as child text of macro */
137: roffproc sub; /* process as child of macro */
138: int flags;
139: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.1 schwarze 140: };
141:
1.37 schwarze 142: struct predef {
143: const char *name; /* predefined input name */
144: const char *str; /* replacement symbol */
145: };
146:
147: #define PREDEF(__name, __str) \
148: { (__name), (__str) },
149:
1.138 schwarze 150: /* --- function prototypes ------------------------------------------------ */
151:
1.42 schwarze 152: static void roffnode_cleanscope(struct roff *);
153: static void roffnode_pop(struct roff *);
1.166 schwarze 154: static void roffnode_push(struct roff *, enum roff_tok,
1.42 schwarze 155: const char *, int, int);
1.193 schwarze 156: static void roff_addtbl(struct roff_man *, struct tbl_node *);
1.183 schwarze 157: static enum rofferr roff_als(ROFF_ARGS);
1.2 schwarze 158: static enum rofferr roff_block(ROFF_ARGS);
159: static enum rofferr roff_block_text(ROFF_ARGS);
160: static enum rofferr roff_block_sub(ROFF_ARGS);
1.168 schwarze 161: static enum rofferr roff_br(ROFF_ARGS);
1.2 schwarze 162: static enum rofferr roff_cblock(ROFF_ARGS);
1.48 schwarze 163: static enum rofferr roff_cc(ROFF_ARGS);
1.68 schwarze 164: static void roff_ccond(struct roff *, int, int);
1.2 schwarze 165: static enum rofferr roff_cond(ROFF_ARGS);
166: static enum rofferr roff_cond_text(ROFF_ARGS);
167: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.7 schwarze 168: static enum rofferr roff_ds(ROFF_ARGS);
1.175 schwarze 169: static enum rofferr roff_ec(ROFF_ARGS);
170: static enum rofferr roff_eo(ROFF_ARGS);
1.110 schwarze 171: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.143 schwarze 172: static int roff_evalcond(struct roff *r, int, char *, int *);
1.106 schwarze 173: static int roff_evalnum(struct roff *, int,
174: const char *, int *, int *, int);
175: static int roff_evalpar(struct roff *, int,
1.133 schwarze 176: const char *, int *, int *, int);
1.71 schwarze 177: static int roff_evalstrcond(const char *, int *);
1.42 schwarze 178: static void roff_free1(struct roff *);
1.52 schwarze 179: static void roff_freereg(struct roffreg *);
1.42 schwarze 180: static void roff_freestr(struct roffkv *);
1.85 schwarze 181: static size_t roff_getname(struct roff *, char **, int, int);
1.133 schwarze 182: static int roff_getnum(const char *, int *, int *, int);
1.56 schwarze 183: static int roff_getop(const char *, int *, char *);
1.199 schwarze 184: static int roff_getregn(struct roff *,
185: const char *, size_t, char);
1.145 schwarze 186: static int roff_getregro(const struct roff *,
187: const char *name);
1.197 schwarze 188: static const char *roff_getstrn(struct roff *,
1.187 schwarze 189: const char *, size_t, int *);
1.143 schwarze 190: static int roff_hasregn(const struct roff *,
191: const char *, size_t);
1.123 schwarze 192: static enum rofferr roff_insec(ROFF_ARGS);
1.51 schwarze 193: static enum rofferr roff_it(ROFF_ARGS);
1.21 schwarze 194: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.137 schwarze 195: static void roff_man_alloc1(struct roff_man *);
196: static void roff_man_free1(struct roff_man *);
1.172 schwarze 197: static enum rofferr roff_manyarg(ROFF_ARGS);
1.202 schwarze 198: static enum rofferr roff_nop(ROFF_ARGS);
1.6 schwarze 199: static enum rofferr roff_nr(ROFF_ARGS);
1.169 schwarze 200: static enum rofferr roff_onearg(ROFF_ARGS);
1.166 schwarze 201: static enum roff_tok roff_parse(struct roff *, char *, int *,
1.87 schwarze 202: int, int);
1.177 schwarze 203: static enum rofferr roff_parsetext(struct roff *, struct buf *,
204: int, int *);
1.178 schwarze 205: static enum rofferr roff_renamed(ROFF_ARGS);
1.110 schwarze 206: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.29 schwarze 207: static enum rofferr roff_rm(ROFF_ARGS);
1.178 schwarze 208: static enum rofferr roff_rn(ROFF_ARGS);
1.76 schwarze 209: static enum rofferr roff_rr(ROFF_ARGS);
1.198 schwarze 210: static void roff_setregn(struct roff *, const char *,
1.199 schwarze 211: size_t, int, char, int);
1.8 schwarze 212: static void roff_setstr(struct roff *,
1.16 schwarze 213: const char *, const char *, int);
1.80 schwarze 214: static void roff_setstrn(struct roffkv **, const char *,
1.42 schwarze 215: size_t, const char *, size_t, int);
1.14 schwarze 216: static enum rofferr roff_so(ROFF_ARGS);
1.42 schwarze 217: static enum rofferr roff_tr(ROFF_ARGS);
1.47 schwarze 218: static enum rofferr roff_Dd(ROFF_ARGS);
1.27 schwarze 219: static enum rofferr roff_TE(ROFF_ARGS);
220: static enum rofferr roff_TS(ROFF_ARGS);
1.32 schwarze 221: static enum rofferr roff_EQ(ROFF_ARGS);
222: static enum rofferr roff_EN(ROFF_ARGS);
1.27 schwarze 223: static enum rofferr roff_T_(ROFF_ARGS);
1.123 schwarze 224: static enum rofferr roff_unsupp(ROFF_ARGS);
1.16 schwarze 225: static enum rofferr roff_userdef(ROFF_ARGS);
1.1 schwarze 226:
1.138 schwarze 227: /* --- constant data ------------------------------------------------------ */
228:
1.133 schwarze 229: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
230: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
231:
1.166 schwarze 232: const char *__roff_name[MAN_MAX + 1] = {
1.177 schwarze 233: "br", "ce", "ft", "ll",
1.184 schwarze 234: "mc", "po", "rj", "sp",
235: "ta", "ti", NULL,
1.166 schwarze 236: "ab", "ad", "af", "aln",
237: "als", "am", "am1", "ami",
238: "ami1", "as", "as1", "asciify",
239: "backtrace", "bd", "bleedat", "blm",
240: "box", "boxa", "bp", "BP",
241: "break", "breakchar", "brnl", "brp",
1.177 schwarze 242: "brpnl", "c2", "cc",
1.166 schwarze 243: "cf", "cflags", "ch", "char",
244: "chop", "class", "close", "CL",
245: "color", "composite", "continue", "cp",
246: "cropat", "cs", "cu", "da",
247: "dch", "Dd", "de", "de1",
248: "defcolor", "dei", "dei1", "device",
249: "devicem", "di", "do", "ds",
250: "ds1", "dwh", "dt", "ec",
251: "ecr", "ecs", "el", "em",
252: "EN", "eo", "EP", "EQ",
253: "errprint", "ev", "evc", "ex",
254: "fallback", "fam", "fc", "fchar",
255: "fcolor", "fdeferlig", "feature", "fkern",
256: "fl", "flig", "fp", "fps",
257: "fschar", "fspacewidth", "fspecial", "ftr",
258: "fzoom", "gcolor", "hc", "hcode",
259: "hidechar", "hla", "hlm", "hpf",
260: "hpfa", "hpfcode", "hw", "hy",
261: "hylang", "hylen", "hym", "hypp",
262: "hys", "ie", "if", "ig",
263: "index", "it", "itc", "IX",
264: "kern", "kernafter", "kernbefore", "kernpair",
265: "lc", "lc_ctype", "lds", "length",
266: "letadj", "lf", "lg", "lhang",
267: "linetabs", "lnr", "lnrf", "lpfx",
1.176 schwarze 268: "ls", "lsm", "lt",
1.166 schwarze 269: "mediasize", "minss", "mk", "mso",
270: "na", "ne", "nh", "nhychar",
271: "nm", "nn", "nop", "nr",
272: "nrf", "nroff", "ns", "nx",
273: "open", "opena", "os", "output",
274: "padj", "papersize", "pc", "pev",
275: "pi", "PI", "pl", "pm",
1.184 schwarze 276: "pn", "pnr", "ps",
1.166 schwarze 277: "psbb", "pshape", "pso", "ptr",
278: "pvs", "rchar", "rd", "recursionlimit",
1.181 schwarze 279: "return", "rfschar", "rhang",
1.166 schwarze 280: "rm", "rn", "rnn", "rr",
281: "rs", "rt", "schar", "sentchar",
282: "shc", "shift", "sizes", "so",
283: "spacewidth", "special", "spreadwarn", "ss",
284: "sty", "substring", "sv", "sy",
1.172 schwarze 285: "T&", "tc", "TE",
1.173 schwarze 286: "TH", "tkf", "tl",
1.166 schwarze 287: "tm", "tm1", "tmc", "tr",
288: "track", "transchar", "trf", "trimat",
289: "trin", "trnt", "troff", "TS",
290: "uf", "ul", "unformat", "unwatch",
291: "unwatchn", "vpt", "vs", "warn",
292: "warnscale", "watch", "watchlength", "watchn",
293: "wh", "while", "write", "writec",
294: "writem", "xflag", ".", NULL,
1.178 schwarze 295: NULL, "text",
1.166 schwarze 296: "Dd", "Dt", "Os", "Sh",
297: "Ss", "Pp", "D1", "Dl",
298: "Bd", "Ed", "Bl", "El",
299: "It", "Ad", "An", "Ap",
300: "Ar", "Cd", "Cm", "Dv",
301: "Er", "Ev", "Ex", "Fa",
302: "Fd", "Fl", "Fn", "Ft",
303: "Ic", "In", "Li", "Nd",
304: "Nm", "Op", "Ot", "Pa",
305: "Rv", "St", "Va", "Vt",
306: "Xr", "%A", "%B", "%D",
307: "%I", "%J", "%N", "%O",
308: "%P", "%R", "%T", "%V",
309: "Ac", "Ao", "Aq", "At",
310: "Bc", "Bf", "Bo", "Bq",
311: "Bsx", "Bx", "Db", "Dc",
312: "Do", "Dq", "Ec", "Ef",
313: "Em", "Eo", "Fx", "Ms",
314: "No", "Ns", "Nx", "Ox",
315: "Pc", "Pf", "Po", "Pq",
316: "Qc", "Ql", "Qo", "Qq",
317: "Re", "Rs", "Sc", "So",
318: "Sq", "Sm", "Sx", "Sy",
319: "Tn", "Ux", "Xc", "Xo",
320: "Fo", "Fc", "Oo", "Oc",
321: "Bk", "Ek", "Bt", "Hf",
322: "Fr", "Ud", "Lb", "Lp",
323: "Lk", "Mt", "Brq", "Bro",
324: "Brc", "%C", "Es", "En",
1.171 schwarze 325: "Dx", "%Q", "%U", "Ta",
326: NULL,
1.166 schwarze 327: "TH", "SH", "SS", "TP",
1.204 ! schwarze 328: "TQ",
1.166 schwarze 329: "LP", "PP", "P", "IP",
330: "HP", "SM", "SB", "BI",
331: "IB", "BR", "RB", "R",
332: "B", "I", "IR", "RI",
1.171 schwarze 333: "nf", "fi",
1.166 schwarze 334: "RE", "RS", "DT", "UC",
1.169 schwarze 335: "PD", "AT", "in",
1.166 schwarze 336: "OP", "EX", "EE", "UR",
1.189 bentley 337: "UE", "MT", "ME", NULL
1.166 schwarze 338: };
339: const char *const *roff_name = __roff_name;
340:
341: static struct roffmac roffs[TOKEN_NONE] = {
1.168 schwarze 342: { roff_br, NULL, NULL, 0 }, /* br */
1.177 schwarze 343: { roff_onearg, NULL, NULL, 0 }, /* ce */
1.169 schwarze 344: { roff_onearg, NULL, NULL, 0 }, /* ft */
1.170 schwarze 345: { roff_onearg, NULL, NULL, 0 }, /* ll */
1.176 schwarze 346: { roff_onearg, NULL, NULL, 0 }, /* mc */
1.184 schwarze 347: { roff_onearg, NULL, NULL, 0 }, /* po */
1.181 schwarze 348: { roff_onearg, NULL, NULL, 0 }, /* rj */
1.171 schwarze 349: { roff_onearg, NULL, NULL, 0 }, /* sp */
1.172 schwarze 350: { roff_manyarg, NULL, NULL, 0 }, /* ta */
1.173 schwarze 351: { roff_onearg, NULL, NULL, 0 }, /* ti */
1.168 schwarze 352: { NULL, NULL, NULL, 0 }, /* ROFF_MAX */
1.167 schwarze 353: { roff_unsupp, NULL, NULL, 0 }, /* ab */
354: { roff_line_ignore, NULL, NULL, 0 }, /* ad */
355: { roff_line_ignore, NULL, NULL, 0 }, /* af */
356: { roff_unsupp, NULL, NULL, 0 }, /* aln */
1.183 schwarze 357: { roff_als, NULL, NULL, 0 }, /* als */
1.167 schwarze 358: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am */
359: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am1 */
360: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami */
361: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami1 */
362: { roff_ds, NULL, NULL, 0 }, /* as */
363: { roff_ds, NULL, NULL, 0 }, /* as1 */
364: { roff_unsupp, NULL, NULL, 0 }, /* asciify */
365: { roff_line_ignore, NULL, NULL, 0 }, /* backtrace */
366: { roff_line_ignore, NULL, NULL, 0 }, /* bd */
367: { roff_line_ignore, NULL, NULL, 0 }, /* bleedat */
368: { roff_unsupp, NULL, NULL, 0 }, /* blm */
369: { roff_unsupp, NULL, NULL, 0 }, /* box */
370: { roff_unsupp, NULL, NULL, 0 }, /* boxa */
371: { roff_line_ignore, NULL, NULL, 0 }, /* bp */
372: { roff_unsupp, NULL, NULL, 0 }, /* BP */
373: { roff_unsupp, NULL, NULL, 0 }, /* break */
374: { roff_line_ignore, NULL, NULL, 0 }, /* breakchar */
375: { roff_line_ignore, NULL, NULL, 0 }, /* brnl */
1.168 schwarze 376: { roff_br, NULL, NULL, 0 }, /* brp */
1.167 schwarze 377: { roff_line_ignore, NULL, NULL, 0 }, /* brpnl */
378: { roff_unsupp, NULL, NULL, 0 }, /* c2 */
379: { roff_cc, NULL, NULL, 0 }, /* cc */
380: { roff_insec, NULL, NULL, 0 }, /* cf */
381: { roff_line_ignore, NULL, NULL, 0 }, /* cflags */
382: { roff_line_ignore, NULL, NULL, 0 }, /* ch */
383: { roff_unsupp, NULL, NULL, 0 }, /* char */
384: { roff_unsupp, NULL, NULL, 0 }, /* chop */
385: { roff_line_ignore, NULL, NULL, 0 }, /* class */
386: { roff_insec, NULL, NULL, 0 }, /* close */
387: { roff_unsupp, NULL, NULL, 0 }, /* CL */
388: { roff_line_ignore, NULL, NULL, 0 }, /* color */
389: { roff_unsupp, NULL, NULL, 0 }, /* composite */
390: { roff_unsupp, NULL, NULL, 0 }, /* continue */
391: { roff_line_ignore, NULL, NULL, 0 }, /* cp */
392: { roff_line_ignore, NULL, NULL, 0 }, /* cropat */
393: { roff_line_ignore, NULL, NULL, 0 }, /* cs */
394: { roff_line_ignore, NULL, NULL, 0 }, /* cu */
395: { roff_unsupp, NULL, NULL, 0 }, /* da */
396: { roff_unsupp, NULL, NULL, 0 }, /* dch */
397: { roff_Dd, NULL, NULL, 0 }, /* Dd */
398: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de */
399: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de1 */
400: { roff_line_ignore, NULL, NULL, 0 }, /* defcolor */
401: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei */
402: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei1 */
403: { roff_unsupp, NULL, NULL, 0 }, /* device */
404: { roff_unsupp, NULL, NULL, 0 }, /* devicem */
405: { roff_unsupp, NULL, NULL, 0 }, /* di */
406: { roff_unsupp, NULL, NULL, 0 }, /* do */
407: { roff_ds, NULL, NULL, 0 }, /* ds */
408: { roff_ds, NULL, NULL, 0 }, /* ds1 */
409: { roff_unsupp, NULL, NULL, 0 }, /* dwh */
410: { roff_unsupp, NULL, NULL, 0 }, /* dt */
1.175 schwarze 411: { roff_ec, NULL, NULL, 0 }, /* ec */
1.167 schwarze 412: { roff_unsupp, NULL, NULL, 0 }, /* ecr */
413: { roff_unsupp, NULL, NULL, 0 }, /* ecs */
414: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* el */
415: { roff_unsupp, NULL, NULL, 0 }, /* em */
416: { roff_EN, NULL, NULL, 0 }, /* EN */
1.175 schwarze 417: { roff_eo, NULL, NULL, 0 }, /* eo */
1.167 schwarze 418: { roff_unsupp, NULL, NULL, 0 }, /* EP */
419: { roff_EQ, NULL, NULL, 0 }, /* EQ */
420: { roff_line_ignore, NULL, NULL, 0 }, /* errprint */
421: { roff_unsupp, NULL, NULL, 0 }, /* ev */
422: { roff_unsupp, NULL, NULL, 0 }, /* evc */
423: { roff_unsupp, NULL, NULL, 0 }, /* ex */
424: { roff_line_ignore, NULL, NULL, 0 }, /* fallback */
425: { roff_line_ignore, NULL, NULL, 0 }, /* fam */
426: { roff_unsupp, NULL, NULL, 0 }, /* fc */
427: { roff_unsupp, NULL, NULL, 0 }, /* fchar */
428: { roff_line_ignore, NULL, NULL, 0 }, /* fcolor */
429: { roff_line_ignore, NULL, NULL, 0 }, /* fdeferlig */
430: { roff_line_ignore, NULL, NULL, 0 }, /* feature */
431: { roff_line_ignore, NULL, NULL, 0 }, /* fkern */
432: { roff_line_ignore, NULL, NULL, 0 }, /* fl */
433: { roff_line_ignore, NULL, NULL, 0 }, /* flig */
434: { roff_line_ignore, NULL, NULL, 0 }, /* fp */
435: { roff_line_ignore, NULL, NULL, 0 }, /* fps */
436: { roff_unsupp, NULL, NULL, 0 }, /* fschar */
437: { roff_line_ignore, NULL, NULL, 0 }, /* fspacewidth */
438: { roff_line_ignore, NULL, NULL, 0 }, /* fspecial */
439: { roff_line_ignore, NULL, NULL, 0 }, /* ftr */
440: { roff_line_ignore, NULL, NULL, 0 }, /* fzoom */
441: { roff_line_ignore, NULL, NULL, 0 }, /* gcolor */
442: { roff_line_ignore, NULL, NULL, 0 }, /* hc */
443: { roff_line_ignore, NULL, NULL, 0 }, /* hcode */
444: { roff_line_ignore, NULL, NULL, 0 }, /* hidechar */
445: { roff_line_ignore, NULL, NULL, 0 }, /* hla */
446: { roff_line_ignore, NULL, NULL, 0 }, /* hlm */
447: { roff_line_ignore, NULL, NULL, 0 }, /* hpf */
448: { roff_line_ignore, NULL, NULL, 0 }, /* hpfa */
449: { roff_line_ignore, NULL, NULL, 0 }, /* hpfcode */
450: { roff_line_ignore, NULL, NULL, 0 }, /* hw */
451: { roff_line_ignore, NULL, NULL, 0 }, /* hy */
452: { roff_line_ignore, NULL, NULL, 0 }, /* hylang */
453: { roff_line_ignore, NULL, NULL, 0 }, /* hylen */
454: { roff_line_ignore, NULL, NULL, 0 }, /* hym */
455: { roff_line_ignore, NULL, NULL, 0 }, /* hypp */
456: { roff_line_ignore, NULL, NULL, 0 }, /* hys */
457: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* ie */
458: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* if */
459: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ig */
460: { roff_unsupp, NULL, NULL, 0 }, /* index */
461: { roff_it, NULL, NULL, 0 }, /* it */
462: { roff_unsupp, NULL, NULL, 0 }, /* itc */
463: { roff_line_ignore, NULL, NULL, 0 }, /* IX */
464: { roff_line_ignore, NULL, NULL, 0 }, /* kern */
465: { roff_line_ignore, NULL, NULL, 0 }, /* kernafter */
466: { roff_line_ignore, NULL, NULL, 0 }, /* kernbefore */
467: { roff_line_ignore, NULL, NULL, 0 }, /* kernpair */
468: { roff_unsupp, NULL, NULL, 0 }, /* lc */
469: { roff_unsupp, NULL, NULL, 0 }, /* lc_ctype */
470: { roff_unsupp, NULL, NULL, 0 }, /* lds */
471: { roff_unsupp, NULL, NULL, 0 }, /* length */
472: { roff_line_ignore, NULL, NULL, 0 }, /* letadj */
473: { roff_insec, NULL, NULL, 0 }, /* lf */
474: { roff_line_ignore, NULL, NULL, 0 }, /* lg */
475: { roff_line_ignore, NULL, NULL, 0 }, /* lhang */
476: { roff_unsupp, NULL, NULL, 0 }, /* linetabs */
477: { roff_unsupp, NULL, NULL, 0 }, /* lnr */
478: { roff_unsupp, NULL, NULL, 0 }, /* lnrf */
479: { roff_unsupp, NULL, NULL, 0 }, /* lpfx */
480: { roff_line_ignore, NULL, NULL, 0 }, /* ls */
481: { roff_unsupp, NULL, NULL, 0 }, /* lsm */
482: { roff_line_ignore, NULL, NULL, 0 }, /* lt */
483: { roff_line_ignore, NULL, NULL, 0 }, /* mediasize */
484: { roff_line_ignore, NULL, NULL, 0 }, /* minss */
485: { roff_line_ignore, NULL, NULL, 0 }, /* mk */
486: { roff_insec, NULL, NULL, 0 }, /* mso */
487: { roff_line_ignore, NULL, NULL, 0 }, /* na */
488: { roff_line_ignore, NULL, NULL, 0 }, /* ne */
489: { roff_line_ignore, NULL, NULL, 0 }, /* nh */
490: { roff_line_ignore, NULL, NULL, 0 }, /* nhychar */
491: { roff_unsupp, NULL, NULL, 0 }, /* nm */
492: { roff_unsupp, NULL, NULL, 0 }, /* nn */
1.202 schwarze 493: { roff_nop, NULL, NULL, 0 }, /* nop */
1.167 schwarze 494: { roff_nr, NULL, NULL, 0 }, /* nr */
495: { roff_unsupp, NULL, NULL, 0 }, /* nrf */
496: { roff_line_ignore, NULL, NULL, 0 }, /* nroff */
497: { roff_line_ignore, NULL, NULL, 0 }, /* ns */
498: { roff_insec, NULL, NULL, 0 }, /* nx */
499: { roff_insec, NULL, NULL, 0 }, /* open */
500: { roff_insec, NULL, NULL, 0 }, /* opena */
501: { roff_line_ignore, NULL, NULL, 0 }, /* os */
502: { roff_unsupp, NULL, NULL, 0 }, /* output */
503: { roff_line_ignore, NULL, NULL, 0 }, /* padj */
504: { roff_line_ignore, NULL, NULL, 0 }, /* papersize */
505: { roff_line_ignore, NULL, NULL, 0 }, /* pc */
506: { roff_line_ignore, NULL, NULL, 0 }, /* pev */
507: { roff_insec, NULL, NULL, 0 }, /* pi */
508: { roff_unsupp, NULL, NULL, 0 }, /* PI */
509: { roff_line_ignore, NULL, NULL, 0 }, /* pl */
510: { roff_line_ignore, NULL, NULL, 0 }, /* pm */
511: { roff_line_ignore, NULL, NULL, 0 }, /* pn */
512: { roff_line_ignore, NULL, NULL, 0 }, /* pnr */
513: { roff_line_ignore, NULL, NULL, 0 }, /* ps */
514: { roff_unsupp, NULL, NULL, 0 }, /* psbb */
515: { roff_unsupp, NULL, NULL, 0 }, /* pshape */
516: { roff_insec, NULL, NULL, 0 }, /* pso */
517: { roff_line_ignore, NULL, NULL, 0 }, /* ptr */
518: { roff_line_ignore, NULL, NULL, 0 }, /* pvs */
519: { roff_unsupp, NULL, NULL, 0 }, /* rchar */
520: { roff_line_ignore, NULL, NULL, 0 }, /* rd */
521: { roff_line_ignore, NULL, NULL, 0 }, /* recursionlimit */
522: { roff_unsupp, NULL, NULL, 0 }, /* return */
523: { roff_unsupp, NULL, NULL, 0 }, /* rfschar */
524: { roff_line_ignore, NULL, NULL, 0 }, /* rhang */
525: { roff_rm, NULL, NULL, 0 }, /* rm */
1.178 schwarze 526: { roff_rn, NULL, NULL, 0 }, /* rn */
1.167 schwarze 527: { roff_unsupp, NULL, NULL, 0 }, /* rnn */
528: { roff_rr, NULL, NULL, 0 }, /* rr */
529: { roff_line_ignore, NULL, NULL, 0 }, /* rs */
530: { roff_line_ignore, NULL, NULL, 0 }, /* rt */
531: { roff_unsupp, NULL, NULL, 0 }, /* schar */
532: { roff_line_ignore, NULL, NULL, 0 }, /* sentchar */
533: { roff_line_ignore, NULL, NULL, 0 }, /* shc */
534: { roff_unsupp, NULL, NULL, 0 }, /* shift */
535: { roff_line_ignore, NULL, NULL, 0 }, /* sizes */
536: { roff_so, NULL, NULL, 0 }, /* so */
537: { roff_line_ignore, NULL, NULL, 0 }, /* spacewidth */
538: { roff_line_ignore, NULL, NULL, 0 }, /* special */
539: { roff_line_ignore, NULL, NULL, 0 }, /* spreadwarn */
540: { roff_line_ignore, NULL, NULL, 0 }, /* ss */
541: { roff_line_ignore, NULL, NULL, 0 }, /* sty */
542: { roff_unsupp, NULL, NULL, 0 }, /* substring */
543: { roff_line_ignore, NULL, NULL, 0 }, /* sv */
544: { roff_insec, NULL, NULL, 0 }, /* sy */
545: { roff_T_, NULL, NULL, 0 }, /* T& */
546: { roff_unsupp, NULL, NULL, 0 }, /* tc */
547: { roff_TE, NULL, NULL, 0 }, /* TE */
1.187 schwarze 548: { roff_Dd, NULL, NULL, 0 }, /* TH */
1.167 schwarze 549: { roff_line_ignore, NULL, NULL, 0 }, /* tkf */
550: { roff_unsupp, NULL, NULL, 0 }, /* tl */
551: { roff_line_ignore, NULL, NULL, 0 }, /* tm */
552: { roff_line_ignore, NULL, NULL, 0 }, /* tm1 */
553: { roff_line_ignore, NULL, NULL, 0 }, /* tmc */
554: { roff_tr, NULL, NULL, 0 }, /* tr */
555: { roff_line_ignore, NULL, NULL, 0 }, /* track */
556: { roff_line_ignore, NULL, NULL, 0 }, /* transchar */
557: { roff_insec, NULL, NULL, 0 }, /* trf */
558: { roff_line_ignore, NULL, NULL, 0 }, /* trimat */
559: { roff_unsupp, NULL, NULL, 0 }, /* trin */
560: { roff_unsupp, NULL, NULL, 0 }, /* trnt */
561: { roff_line_ignore, NULL, NULL, 0 }, /* troff */
562: { roff_TS, NULL, NULL, 0 }, /* TS */
563: { roff_line_ignore, NULL, NULL, 0 }, /* uf */
564: { roff_line_ignore, NULL, NULL, 0 }, /* ul */
565: { roff_unsupp, NULL, NULL, 0 }, /* unformat */
566: { roff_line_ignore, NULL, NULL, 0 }, /* unwatch */
567: { roff_line_ignore, NULL, NULL, 0 }, /* unwatchn */
568: { roff_line_ignore, NULL, NULL, 0 }, /* vpt */
569: { roff_line_ignore, NULL, NULL, 0 }, /* vs */
570: { roff_line_ignore, NULL, NULL, 0 }, /* warn */
571: { roff_line_ignore, NULL, NULL, 0 }, /* warnscale */
572: { roff_line_ignore, NULL, NULL, 0 }, /* watch */
573: { roff_line_ignore, NULL, NULL, 0 }, /* watchlength */
574: { roff_line_ignore, NULL, NULL, 0 }, /* watchn */
575: { roff_unsupp, NULL, NULL, 0 }, /* wh */
576: { roff_unsupp, NULL, NULL, 0 }, /* while */
577: { roff_insec, NULL, NULL, 0 }, /* write */
578: { roff_insec, NULL, NULL, 0 }, /* writec */
579: { roff_insec, NULL, NULL, 0 }, /* writem */
580: { roff_line_ignore, NULL, NULL, 0 }, /* xflag */
581: { roff_cblock, NULL, NULL, 0 }, /* . */
1.178 schwarze 582: { roff_renamed, NULL, NULL, 0 },
1.167 schwarze 583: { roff_userdef, NULL, NULL, 0 }
1.1 schwarze 584: };
585:
1.37 schwarze 586: /* Array of injected predefined strings. */
587: #define PREDEFS_MAX 38
588: static const struct predef predefs[PREDEFS_MAX] = {
589: #include "predefs.in"
590: };
591:
1.177 schwarze 592: static int roffce_lines; /* number of input lines to center */
593: static struct roff_node *roffce_node; /* active request */
1.51 schwarze 594: static int roffit_lines; /* number of lines to delay */
595: static char *roffit_macro; /* nil-terminated macro line */
596:
1.80 schwarze 597:
1.138 schwarze 598: /* --- request table ------------------------------------------------------ */
599:
1.167 schwarze 600: struct ohash *
601: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.3 schwarze 602: {
1.167 schwarze 603: struct ohash *htab;
604: struct roffreq *req;
605: enum roff_tok tok;
606: size_t sz;
607: unsigned int slot;
1.3 schwarze 608:
1.167 schwarze 609: htab = mandoc_malloc(sizeof(*htab));
610: mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.3 schwarze 611:
1.167 schwarze 612: for (tok = mintok; tok < maxtok; tok++) {
1.168 schwarze 613: if (roff_name[tok] == NULL)
614: continue;
1.167 schwarze 615: sz = strlen(roff_name[tok]);
616: req = mandoc_malloc(sizeof(*req) + sz + 1);
617: req->tok = tok;
618: memcpy(req->name, roff_name[tok], sz + 1);
619: slot = ohash_qlookup(htab, req->name);
620: ohash_insert(htab, slot, req);
1.3 schwarze 621: }
1.167 schwarze 622: return htab;
1.3 schwarze 623: }
624:
1.167 schwarze 625: void
626: roffhash_free(struct ohash *htab)
1.1 schwarze 627: {
1.167 schwarze 628: struct roffreq *req;
629: unsigned int slot;
1.1 schwarze 630:
1.167 schwarze 631: if (htab == NULL)
632: return;
633: for (req = ohash_first(htab, &slot); req != NULL;
634: req = ohash_next(htab, &slot))
635: free(req);
636: ohash_delete(htab);
637: free(htab);
638: }
639:
640: enum roff_tok
641: roffhash_find(struct ohash *htab, const char *name, size_t sz)
642: {
643: struct roffreq *req;
644: const char *end;
645:
646: if (sz) {
647: end = name + sz;
648: req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
649: } else
650: req = ohash_find(htab, ohash_qlookup(htab, name));
651: return req == NULL ? TOKEN_NONE : req->tok;
1.1 schwarze 652: }
653:
1.138 schwarze 654: /* --- stack of request blocks -------------------------------------------- */
655:
1.1 schwarze 656: /*
657: * Pop the current node off of the stack of roff instructions currently
658: * pending.
659: */
660: static void
661: roffnode_pop(struct roff *r)
662: {
663: struct roffnode *p;
664:
1.2 schwarze 665: assert(r->last);
1.80 schwarze 666: p = r->last;
1.2 schwarze 667:
668: r->last = r->last->parent;
1.16 schwarze 669: free(p->name);
670: free(p->end);
1.1 schwarze 671: free(p);
672: }
673:
674: /*
675: * Push a roff node onto the instruction stack. This must later be
676: * removed with roffnode_pop().
677: */
1.11 schwarze 678: static void
1.166 schwarze 679: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.16 schwarze 680: int line, int col)
1.1 schwarze 681: {
682: struct roffnode *p;
683:
1.11 schwarze 684: p = mandoc_calloc(1, sizeof(struct roffnode));
1.1 schwarze 685: p->tok = tok;
1.16 schwarze 686: if (name)
687: p->name = mandoc_strdup(name);
1.1 schwarze 688: p->parent = r->last;
689: p->line = line;
690: p->col = col;
1.71 schwarze 691: p->rule = p->parent ? p->parent->rule : 0;
1.1 schwarze 692:
693: r->last = p;
694: }
695:
1.138 schwarze 696: /* --- roff parser state data management ---------------------------------- */
697:
1.1 schwarze 698: static void
699: roff_free1(struct roff *r)
700: {
1.49 schwarze 701: struct tbl_node *tbl;
1.42 schwarze 702: int i;
1.27 schwarze 703:
1.49 schwarze 704: while (NULL != (tbl = r->first_tbl)) {
705: r->first_tbl = tbl->next;
706: tbl_free(tbl);
1.27 schwarze 707: }
708: r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1 schwarze 709:
1.191 schwarze 710: if (r->last_eqn != NULL)
711: eqn_free(r->last_eqn);
712: r->last_eqn = r->eqn = NULL;
1.32 schwarze 713:
1.1 schwarze 714: while (r->last)
715: roffnode_pop(r);
1.27 schwarze 716:
1.96 schwarze 717: free (r->rstack);
718: r->rstack = NULL;
719: r->rstacksz = 0;
720: r->rstackpos = -1;
721:
722: roff_freereg(r->regtab);
723: r->regtab = NULL;
724:
1.42 schwarze 725: roff_freestr(r->strtab);
1.178 schwarze 726: roff_freestr(r->rentab);
1.42 schwarze 727: roff_freestr(r->xmbtab);
1.178 schwarze 728: r->strtab = r->rentab = r->xmbtab = NULL;
1.42 schwarze 729:
730: if (r->xtab)
731: for (i = 0; i < 128; i++)
732: free(r->xtab[i].p);
733: free(r->xtab);
734: r->xtab = NULL;
1.1 schwarze 735: }
736:
737: void
738: roff_reset(struct roff *r)
739: {
740: roff_free1(r);
1.99 schwarze 741: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.175 schwarze 742: r->control = '\0';
743: r->escape = '\\';
1.179 schwarze 744: roffce_lines = 0;
745: roffce_node = NULL;
746: roffit_lines = 0;
747: roffit_macro = NULL;
1.1 schwarze 748: }
749:
750: void
751: roff_free(struct roff *r)
752: {
753: roff_free1(r);
1.167 schwarze 754: roffhash_free(r->reqtab);
1.1 schwarze 755: free(r);
756: }
757:
758: struct roff *
1.151 schwarze 759: roff_alloc(struct mparse *parse, int options)
1.1 schwarze 760: {
761: struct roff *r;
762:
1.11 schwarze 763: r = mandoc_calloc(1, sizeof(struct roff));
1.35 schwarze 764: r->parse = parse;
1.200 schwarze 765: r->reqtab = roffhash_alloc(0, ROFF_RENAMED);
1.72 schwarze 766: r->options = options;
1.99 schwarze 767: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2 schwarze 768: r->rstackpos = -1;
1.175 schwarze 769: r->escape = '\\';
1.149 schwarze 770: return r;
1.137 schwarze 771: }
772:
1.138 schwarze 773: /* --- syntax tree state data management ---------------------------------- */
774:
1.137 schwarze 775: static void
776: roff_man_free1(struct roff_man *man)
777: {
778:
1.138 schwarze 779: if (man->first != NULL)
780: roff_node_delete(man, man->first);
1.137 schwarze 781: free(man->meta.msec);
782: free(man->meta.vol);
783: free(man->meta.os);
784: free(man->meta.arch);
785: free(man->meta.title);
786: free(man->meta.name);
787: free(man->meta.date);
788: }
789:
790: static void
791: roff_man_alloc1(struct roff_man *man)
792: {
793:
794: memset(&man->meta, 0, sizeof(man->meta));
795: man->first = mandoc_calloc(1, sizeof(*man->first));
796: man->first->type = ROFFT_ROOT;
797: man->last = man->first;
798: man->last_es = NULL;
799: man->flags = 0;
800: man->macroset = MACROSET_NONE;
801: man->lastsec = man->lastnamed = SEC_NONE;
802: man->next = ROFF_NEXT_CHILD;
803: }
804:
805: void
806: roff_man_reset(struct roff_man *man)
807: {
808:
809: roff_man_free1(man);
810: roff_man_alloc1(man);
811: }
812:
813: void
814: roff_man_free(struct roff_man *man)
815: {
816:
817: roff_man_free1(man);
818: free(man);
819: }
820:
821: struct roff_man *
822: roff_man_alloc(struct roff *roff, struct mparse *parse,
1.188 schwarze 823: const char *os_s, int quick)
1.137 schwarze 824: {
825: struct roff_man *man;
826:
827: man = mandoc_calloc(1, sizeof(*man));
828: man->parse = parse;
829: man->roff = roff;
1.188 schwarze 830: man->os_s = os_s;
1.137 schwarze 831: man->quick = quick;
832: roff_man_alloc1(man);
1.168 schwarze 833: roff->man = man;
1.149 schwarze 834: return man;
1.1 schwarze 835: }
836:
1.138 schwarze 837: /* --- syntax tree handling ----------------------------------------------- */
838:
839: struct roff_node *
840: roff_node_alloc(struct roff_man *man, int line, int pos,
841: enum roff_type type, int tok)
842: {
843: struct roff_node *n;
844:
845: n = mandoc_calloc(1, sizeof(*n));
846: n->line = line;
847: n->pos = pos;
848: n->tok = tok;
849: n->type = type;
850: n->sec = man->lastsec;
851:
852: if (man->flags & MDOC_SYNOPSIS)
1.157 schwarze 853: n->flags |= NODE_SYNPRETTY;
1.138 schwarze 854: else
1.157 schwarze 855: n->flags &= ~NODE_SYNPRETTY;
1.138 schwarze 856: if (man->flags & MDOC_NEWLINE)
1.157 schwarze 857: n->flags |= NODE_LINE;
1.138 schwarze 858: man->flags &= ~MDOC_NEWLINE;
859:
1.149 schwarze 860: return n;
1.138 schwarze 861: }
862:
863: void
864: roff_node_append(struct roff_man *man, struct roff_node *n)
865: {
866:
867: switch (man->next) {
868: case ROFF_NEXT_SIBLING:
1.153 schwarze 869: if (man->last->next != NULL) {
870: n->next = man->last->next;
871: man->last->next->prev = n;
872: } else
873: man->last->parent->last = n;
1.138 schwarze 874: man->last->next = n;
875: n->prev = man->last;
876: n->parent = man->last->parent;
877: break;
878: case ROFF_NEXT_CHILD:
1.159 schwarze 879: if (man->last->child != NULL) {
880: n->next = man->last->child;
881: man->last->child->prev = n;
882: } else
883: man->last->last = n;
1.138 schwarze 884: man->last->child = n;
885: n->parent = man->last;
886: break;
887: default:
888: abort();
889: }
1.154 schwarze 890: man->last = n;
1.138 schwarze 891:
892: switch (n->type) {
893: case ROFFT_HEAD:
894: n->parent->head = n;
895: break;
896: case ROFFT_BODY:
1.154 schwarze 897: if (n->end != ENDBODY_NOT)
898: return;
1.138 schwarze 899: n->parent->body = n;
900: break;
901: case ROFFT_TAIL:
902: n->parent->tail = n;
903: break;
904: default:
1.154 schwarze 905: return;
1.138 schwarze 906: }
1.154 schwarze 907:
908: /*
909: * Copy over the normalised-data pointer of our parent. Not
910: * everybody has one, but copying a null pointer is fine.
911: */
912:
913: n->norm = n->parent->norm;
914: assert(n->parent->type == ROFFT_BLOCK);
1.138 schwarze 915: }
916:
1.139 schwarze 917: void
918: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
919: {
920: struct roff_node *n;
921:
922: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
923: n->string = roff_strdup(man->roff, word);
924: roff_node_append(man, n);
1.158 schwarze 925: n->flags |= NODE_VALID | NODE_ENDED;
1.139 schwarze 926: man->next = ROFF_NEXT_SIBLING;
927: }
928:
929: void
930: roff_word_append(struct roff_man *man, const char *word)
931: {
932: struct roff_node *n;
933: char *addstr, *newstr;
934:
935: n = man->last;
936: addstr = roff_strdup(man->roff, word);
937: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
938: free(addstr);
939: free(n->string);
940: n->string = newstr;
941: man->next = ROFF_NEXT_SIBLING;
1.140 schwarze 942: }
943:
944: void
945: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
946: {
947: struct roff_node *n;
948:
949: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
950: roff_node_append(man, n);
951: man->next = ROFF_NEXT_CHILD;
952: }
953:
954: struct roff_node *
955: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
956: {
957: struct roff_node *n;
958:
959: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
960: roff_node_append(man, n);
961: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 962: return n;
1.139 schwarze 963: }
964:
1.138 schwarze 965: struct roff_node *
966: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
967: {
968: struct roff_node *n;
969:
970: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
971: roff_node_append(man, n);
972: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 973: return n;
1.138 schwarze 974: }
975:
976: struct roff_node *
977: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
978: {
979: struct roff_node *n;
980:
981: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
982: roff_node_append(man, n);
983: man->next = ROFF_NEXT_CHILD;
1.149 schwarze 984: return n;
1.139 schwarze 985: }
986:
1.193 schwarze 987: static void
988: roff_addtbl(struct roff_man *man, struct tbl_node *tbl)
1.139 schwarze 989: {
990: struct roff_node *n;
1.193 schwarze 991: const struct tbl_span *span;
1.139 schwarze 992:
993: if (man->macroset == MACROSET_MAN)
1.174 schwarze 994: man_breakscope(man, ROFF_TS);
1.193 schwarze 995: while ((span = tbl_span(tbl)) != NULL) {
996: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
997: n->span = span;
998: roff_node_append(man, n);
999: n->flags |= NODE_VALID | NODE_ENDED;
1000: man->next = ROFF_NEXT_SIBLING;
1001: }
1.138 schwarze 1002: }
1003:
1004: void
1005: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1006: {
1007:
1008: /* Adjust siblings. */
1009:
1010: if (n->prev)
1011: n->prev->next = n->next;
1012: if (n->next)
1013: n->next->prev = n->prev;
1014:
1015: /* Adjust parent. */
1016:
1017: if (n->parent != NULL) {
1018: if (n->parent->child == n)
1019: n->parent->child = n->next;
1020: if (n->parent->last == n)
1021: n->parent->last = n->prev;
1022: }
1023:
1024: /* Adjust parse point. */
1025:
1026: if (man == NULL)
1027: return;
1028: if (man->last == n) {
1029: if (n->prev == NULL) {
1030: man->last = n->parent;
1031: man->next = ROFF_NEXT_CHILD;
1032: } else {
1033: man->last = n->prev;
1034: man->next = ROFF_NEXT_SIBLING;
1035: }
1036: }
1037: if (man->first == n)
1038: man->first = NULL;
1039: }
1040:
1041: void
1042: roff_node_free(struct roff_node *n)
1043: {
1044:
1045: if (n->args != NULL)
1046: mdoc_argv_free(n->args);
1047: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1048: free(n->norm);
1.191 schwarze 1049: if (n->eqn != NULL)
1050: eqn_box_free(n->eqn);
1.138 schwarze 1051: free(n->string);
1052: free(n);
1053: }
1054:
1055: void
1056: roff_node_delete(struct roff_man *man, struct roff_node *n)
1057: {
1058:
1059: while (n->child != NULL)
1060: roff_node_delete(man, n->child);
1061: roff_node_unlink(man, n);
1062: roff_node_free(n);
1.141 schwarze 1063: }
1064:
1065: void
1066: deroff(char **dest, const struct roff_node *n)
1067: {
1068: char *cp;
1069: size_t sz;
1070:
1071: if (n->type != ROFFT_TEXT) {
1072: for (n = n->child; n != NULL; n = n->next)
1073: deroff(dest, n);
1074: return;
1075: }
1076:
1.160 schwarze 1077: /* Skip leading whitespace. */
1.141 schwarze 1078:
1.160 schwarze 1079: for (cp = n->string; *cp != '\0'; cp++) {
1.161 schwarze 1080: if (cp[0] == '\\' && cp[1] != '\0' &&
1081: strchr(" %&0^|~", cp[1]) != NULL)
1.141 schwarze 1082: cp++;
1.160 schwarze 1083: else if ( ! isspace((unsigned char)*cp))
1.141 schwarze 1084: break;
1085: }
1086:
1.161 schwarze 1087: /* Skip trailing backslash. */
1088:
1089: sz = strlen(cp);
1.162 schwarze 1090: if (sz > 0 && cp[sz - 1] == '\\')
1.161 schwarze 1091: sz--;
1092:
1.141 schwarze 1093: /* Skip trailing whitespace. */
1094:
1.161 schwarze 1095: for (; sz; sz--)
1.141 schwarze 1096: if ( ! isspace((unsigned char)cp[sz-1]))
1097: break;
1098:
1099: /* Skip empty strings. */
1100:
1101: if (sz == 0)
1102: return;
1103:
1104: if (*dest == NULL) {
1105: *dest = mandoc_strndup(cp, sz);
1106: return;
1107: }
1108:
1109: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1110: free(*dest);
1111: *dest = cp;
1.138 schwarze 1112: }
1113:
1114: /* --- main functions of the roff parser ---------------------------------- */
1115:
1.8 schwarze 1116: /*
1.79 schwarze 1117: * In the current line, expand escape sequences that tend to get
1118: * used in numerical expressions and conditional requests.
1119: * Also check the syntax of the remaining escape sequences.
1.8 schwarze 1120: */
1.45 schwarze 1121: static enum rofferr
1.110 schwarze 1122: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8 schwarze 1123: {
1.81 schwarze 1124: char ubuf[24]; /* buffer to print the number */
1.200 schwarze 1125: struct roff_node *n; /* used for header comments */
1.78 schwarze 1126: const char *start; /* start of the string to process */
1.82 schwarze 1127: char *stesc; /* start of an escape sequence ('\\') */
1.200 schwarze 1128: char *ep; /* end of comment string */
1.23 schwarze 1129: const char *stnam; /* start of the name, after "[(*" */
1130: const char *cp; /* end of the name, e.g. before ']' */
1131: const char *res; /* the string to be substituted */
1.110 schwarze 1132: char *nbuf; /* new buffer to copy buf->buf to */
1.53 schwarze 1133: size_t maxl; /* expected length of the escape name */
1134: size_t naml; /* actual length of the escape name */
1.109 schwarze 1135: enum mandoc_esc esc; /* type of the escape sequence */
1136: int inaml; /* length returned from mandoc_escape() */
1.53 schwarze 1137: int expand_count; /* to avoid infinite loops */
1.79 schwarze 1138: int npos; /* position in numeric expression */
1.91 schwarze 1139: int arg_complete; /* argument not interrupted by eol */
1.175 schwarze 1140: int done; /* no more input available */
1.187 schwarze 1141: int deftype; /* type of definition to paste */
1.195 schwarze 1142: int rcsid; /* kind of RCS id seen */
1.199 schwarze 1143: char sign; /* increment number register */
1.79 schwarze 1144: char term; /* character terminating the escape */
1.8 schwarze 1145:
1.175 schwarze 1146: /* Search forward for comments. */
1147:
1148: done = 0;
1149: start = buf->buf + pos;
1150: for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
1151: if (stesc[0] != r->escape || stesc[1] == '\0')
1152: continue;
1153: stesc++;
1154: if (*stesc != '"' && *stesc != '#')
1155: continue;
1.185 schwarze 1156:
1157: /* Comment found, look for RCS id. */
1158:
1.195 schwarze 1159: rcsid = 0;
1.185 schwarze 1160: if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.195 schwarze 1161: rcsid = 1 << MANDOC_OS_OPENBSD;
1.185 schwarze 1162: cp += 8;
1163: } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.195 schwarze 1164: rcsid = 1 << MANDOC_OS_NETBSD;
1.185 schwarze 1165: cp += 7;
1166: }
1167: if (cp != NULL &&
1168: isalnum((unsigned char)*cp) == 0 &&
1.186 schwarze 1169: strchr(cp, '$') != NULL) {
1.195 schwarze 1170: if (r->man->meta.rcsids & rcsid)
1.186 schwarze 1171: mandoc_msg(MANDOCERR_RCS_REP, r->parse,
1172: ln, stesc + 1 - buf->buf, stesc + 1);
1.195 schwarze 1173: r->man->meta.rcsids |= rcsid;
1.186 schwarze 1174: }
1.185 schwarze 1175:
1176: /* Handle trailing whitespace. */
1177:
1.200 schwarze 1178: ep = strchr(stesc--, '\0') - 1;
1179: if (*ep == '\n') {
1.175 schwarze 1180: done = 1;
1.200 schwarze 1181: ep--;
1.175 schwarze 1182: }
1.200 schwarze 1183: if (*ep == ' ' || *ep == '\t')
1.175 schwarze 1184: mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
1.200 schwarze 1185: ln, ep - buf->buf, NULL);
1186:
1187: /*
1188: * Save comments preceding the title macro
1189: * in the syntax tree.
1190: */
1191:
1192: if (r->format == 0) {
1193: while (*ep == ' ' || *ep == '\t')
1194: ep--;
1195: ep[1] = '\0';
1196: n = roff_node_alloc(r->man,
1197: ln, stesc + 1 - buf->buf,
1198: ROFFT_COMMENT, TOKEN_NONE);
1199: n->string = mandoc_strdup(stesc + 2);
1200: roff_node_append(r->man, n);
1201: n->flags |= NODE_VALID | NODE_ENDED;
1202: r->man->next = ROFF_NEXT_SIBLING;
1203: }
1204:
1205: /* Discard comments. */
1206:
1.175 schwarze 1207: while (stesc > start && stesc[-1] == ' ')
1208: stesc--;
1209: *stesc = '\0';
1210: break;
1211: }
1212: if (stesc == start)
1213: return ROFF_CONT;
1214: stesc--;
1215:
1216: /* Notice the end of the input. */
1217:
1218: if (*stesc == '\n') {
1219: *stesc-- = '\0';
1220: done = 1;
1221: }
1222:
1.43 schwarze 1223: expand_count = 0;
1.175 schwarze 1224: while (stesc >= start) {
1.43 schwarze 1225:
1.78 schwarze 1226: /* Search backwards for the next backslash. */
1227:
1.175 schwarze 1228: if (*stesc != r->escape) {
1229: if (*stesc == '\\') {
1230: *stesc = '\0';
1231: buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
1232: buf->buf, stesc + 1) + 1;
1233: start = nbuf + pos;
1234: stesc = nbuf + (stesc - buf->buf);
1235: free(buf->buf);
1236: buf->buf = nbuf;
1237: }
1238: stesc--;
1.78 schwarze 1239: continue;
1.175 schwarze 1240: }
1.78 schwarze 1241:
1242: /* If it is escaped, skip it. */
1243:
1244: for (cp = stesc - 1; cp >= start; cp--)
1.175 schwarze 1245: if (*cp != r->escape)
1.78 schwarze 1246: break;
1247:
1.110 schwarze 1248: if ((stesc - cp) % 2 == 0) {
1.175 schwarze 1249: while (stesc > cp)
1250: *stesc-- = '\\';
1.78 schwarze 1251: continue;
1.175 schwarze 1252: } else if (stesc[1] != '\0') {
1253: *stesc = '\\';
1254: } else {
1255: *stesc-- = '\0';
1256: if (done)
1257: continue;
1258: else
1259: return ROFF_APPEND;
1.78 schwarze 1260: }
1.23 schwarze 1261:
1.79 schwarze 1262: /* Decide whether to expand or to check only. */
1.23 schwarze 1263:
1.79 schwarze 1264: term = '\0';
1.78 schwarze 1265: cp = stesc + 1;
1.53 schwarze 1266: switch (*cp) {
1.80 schwarze 1267: case '*':
1.53 schwarze 1268: res = NULL;
1269: break;
1.80 schwarze 1270: case 'B':
1271: case 'w':
1.79 schwarze 1272: term = cp[1];
1273: /* FALLTHROUGH */
1.80 schwarze 1274: case 'n':
1.199 schwarze 1275: sign = cp[1];
1276: if (sign == '+' || sign == '-')
1277: cp++;
1.53 schwarze 1278: res = ubuf;
1279: break;
1280: default:
1.109 schwarze 1281: esc = mandoc_escape(&cp, &stnam, &inaml);
1282: if (esc == ESCAPE_ERROR ||
1283: (esc == ESCAPE_SPECIAL &&
1.151 schwarze 1284: mchars_spec2cp(stnam, inaml) < 0))
1.92 schwarze 1285: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110 schwarze 1286: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1287: "%.*s", (int)(cp - stesc), stesc);
1.175 schwarze 1288: stesc--;
1.78 schwarze 1289: continue;
1.42 schwarze 1290: }
1291:
1.78 schwarze 1292: if (EXPAND_LIMIT < ++expand_count) {
1293: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110 schwarze 1294: ln, (int)(stesc - buf->buf), NULL);
1.149 schwarze 1295: return ROFF_IGN;
1.78 schwarze 1296: }
1.23 schwarze 1297:
1298: /*
1299: * The third character decides the length
1.53 schwarze 1300: * of the name of the string or register.
1.23 schwarze 1301: * Save a pointer to the name.
1302: */
1303:
1.110 schwarze 1304: if (term == '\0') {
1.79 schwarze 1305: switch (*++cp) {
1.80 schwarze 1306: case '\0':
1.79 schwarze 1307: maxl = 0;
1308: break;
1.80 schwarze 1309: case '(':
1.79 schwarze 1310: cp++;
1311: maxl = 2;
1312: break;
1.80 schwarze 1313: case '[':
1.79 schwarze 1314: cp++;
1315: term = ']';
1316: maxl = 0;
1317: break;
1318: default:
1319: maxl = 1;
1320: break;
1321: }
1322: } else {
1323: cp += 2;
1.8 schwarze 1324: maxl = 0;
1325: }
1.23 schwarze 1326: stnam = cp;
1.8 schwarze 1327:
1.23 schwarze 1328: /* Advance to the end of the name. */
1.8 schwarze 1329:
1.125 schwarze 1330: naml = 0;
1.91 schwarze 1331: arg_complete = 1;
1.125 schwarze 1332: while (maxl == 0 || naml < maxl) {
1.110 schwarze 1333: if (*cp == '\0') {
1.92 schwarze 1334: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 1335: ln, (int)(stesc - buf->buf), stesc);
1.91 schwarze 1336: arg_complete = 0;
1.79 schwarze 1337: break;
1.42 schwarze 1338: }
1.110 schwarze 1339: if (maxl == 0 && *cp == term) {
1.79 schwarze 1340: cp++;
1.125 schwarze 1341: break;
1342: }
1343: if (*cp++ != '\\' || stesc[1] != 'w') {
1344: naml++;
1345: continue;
1346: }
1347: switch (mandoc_escape(&cp, NULL, NULL)) {
1348: case ESCAPE_SPECIAL:
1349: case ESCAPE_UNICODE:
1350: case ESCAPE_NUMBERED:
1351: case ESCAPE_OVERSTRIKE:
1352: naml++;
1353: break;
1354: default:
1.8 schwarze 1355: break;
1.79 schwarze 1356: }
1.8 schwarze 1357: }
1358:
1.23 schwarze 1359: /*
1360: * Retrieve the replacement string; if it is
1361: * undefined, resume searching for escapes.
1362: */
1363:
1.79 schwarze 1364: switch (stesc[1]) {
1.80 schwarze 1365: case '*':
1.187 schwarze 1366: if (arg_complete) {
1367: deftype = ROFFDEF_USER | ROFFDEF_PRE;
1368: res = roff_getstrn(r, stnam, naml, &deftype);
1.203 schwarze 1369:
1370: /*
1371: * If not overriden, let \*(.T
1372: * through to the formatters.
1373: */
1374:
1375: if (res == NULL && naml == 2 &&
1376: stnam[0] == '.' && stnam[1] == 'T') {
1377: roff_setstrn(&r->strtab,
1378: ".T", 2, NULL, 0, 0);
1379: stesc--;
1380: continue;
1381: }
1.187 schwarze 1382: }
1.79 schwarze 1383: break;
1.80 schwarze 1384: case 'B':
1.79 schwarze 1385: npos = 0;
1.91 schwarze 1386: ubuf[0] = arg_complete &&
1.133 schwarze 1387: roff_evalnum(r, ln, stnam, &npos,
1388: NULL, ROFFNUM_SCALE) &&
1.91 schwarze 1389: stnam + npos + 1 == cp ? '1' : '0';
1.79 schwarze 1390: ubuf[1] = '\0';
1391: break;
1.80 schwarze 1392: case 'n':
1.91 schwarze 1393: if (arg_complete)
1394: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.199 schwarze 1395: roff_getregn(r, stnam, naml, sign));
1.91 schwarze 1396: else
1397: ubuf[0] = '\0';
1.79 schwarze 1398: break;
1.80 schwarze 1399: case 'w':
1.91 schwarze 1400: /* use even incomplete args */
1.81 schwarze 1401: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79 schwarze 1402: 24 * (int)naml);
1403: break;
1404: }
1.8 schwarze 1405:
1.110 schwarze 1406: if (res == NULL) {
1.92 schwarze 1407: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1408: r->parse, ln, (int)(stesc - buf->buf),
1.92 schwarze 1409: "%.*s", (int)naml, stnam);
1.37 schwarze 1410: res = "";
1.118 schwarze 1411: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1412: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1413: ln, (int)(stesc - buf->buf), NULL);
1.149 schwarze 1414: return ROFF_IGN;
1.8 schwarze 1415: }
1416:
1.23 schwarze 1417: /* Replace the escape sequence by the string. */
1418:
1.82 schwarze 1419: *stesc = '\0';
1.110 schwarze 1420: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1421: buf->buf, res, cp) + 1;
1.8 schwarze 1422:
1.78 schwarze 1423: /* Prepare for the next replacement. */
1.8 schwarze 1424:
1.78 schwarze 1425: start = nbuf + pos;
1.110 schwarze 1426: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1427: free(buf->buf);
1428: buf->buf = nbuf;
1.42 schwarze 1429: }
1.149 schwarze 1430: return ROFF_CONT;
1.42 schwarze 1431: }
1432:
1433: /*
1.147 schwarze 1434: * Process text streams.
1.42 schwarze 1435: */
1436: static enum rofferr
1.177 schwarze 1437: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.42 schwarze 1438: {
1439: size_t sz;
1440: const char *start;
1.51 schwarze 1441: char *p;
1442: int isz;
1.42 schwarze 1443: enum mandoc_esc esc;
1444:
1.147 schwarze 1445: /* Spring the input line trap. */
1446:
1447: if (roffit_lines == 1) {
1448: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1449: free(buf->buf);
1450: buf->buf = p;
1451: buf->sz = isz + 1;
1452: *offs = 0;
1453: free(roffit_macro);
1454: roffit_lines = 0;
1.149 schwarze 1455: return ROFF_REPARSE;
1.147 schwarze 1456: } else if (roffit_lines > 1)
1457: --roffit_lines;
1458:
1.177 schwarze 1459: if (roffce_node != NULL && buf->buf[pos] != '\0') {
1460: if (roffce_lines < 1) {
1461: r->man->last = roffce_node;
1462: r->man->next = ROFF_NEXT_SIBLING;
1463: roffce_lines = 0;
1464: roffce_node = NULL;
1465: } else
1466: roffce_lines--;
1467: }
1468:
1.147 schwarze 1469: /* Convert all breakable hyphens into ASCII_HYPH. */
1470:
1.110 schwarze 1471: start = p = buf->buf + pos;
1.42 schwarze 1472:
1.110 schwarze 1473: while (*p != '\0') {
1.42 schwarze 1474: sz = strcspn(p, "-\\");
1475: p += sz;
1476:
1.110 schwarze 1477: if (*p == '\0')
1.42 schwarze 1478: break;
1479:
1.110 schwarze 1480: if (*p == '\\') {
1.42 schwarze 1481: /* Skip over escapes. */
1482: p++;
1.62 schwarze 1483: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110 schwarze 1484: if (esc == ESCAPE_ERROR)
1.42 schwarze 1485: break;
1.136 schwarze 1486: while (*p == '-')
1487: p++;
1.42 schwarze 1488: continue;
1489: } else if (p == start) {
1490: p++;
1491: continue;
1492: }
1493:
1.44 schwarze 1494: if (isalpha((unsigned char)p[-1]) &&
1495: isalpha((unsigned char)p[1]))
1.42 schwarze 1496: *p = ASCII_HYPH;
1497: p++;
1.8 schwarze 1498: }
1.149 schwarze 1499: return ROFF_CONT;
1.8 schwarze 1500: }
1501:
1.1 schwarze 1502: enum rofferr
1.110 schwarze 1503: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1 schwarze 1504: {
1.166 schwarze 1505: enum roff_tok t;
1.27 schwarze 1506: enum rofferr e;
1.110 schwarze 1507: int pos; /* parse point */
1.115 schwarze 1508: int spos; /* saved parse point for messages */
1.110 schwarze 1509: int ppos; /* original offset in buf->buf */
1510: int ctl; /* macro line (boolean) */
1511:
1512: ppos = pos = *offs;
1.1 schwarze 1513:
1.102 schwarze 1514: /* Handle in-line equation delimiters. */
1515:
1.108 schwarze 1516: if (r->tbl == NULL &&
1517: r->last_eqn != NULL && r->last_eqn->delim &&
1.102 schwarze 1518: (r->eqn == NULL || r->eqn_inline)) {
1.110 schwarze 1519: e = roff_eqndelim(r, buf, pos);
1.102 schwarze 1520: if (e == ROFF_REPARSE)
1.149 schwarze 1521: return e;
1.102 schwarze 1522: assert(e == ROFF_CONT);
1523: }
1524:
1525: /* Expand some escape sequences. */
1.8 schwarze 1526:
1.110 schwarze 1527: e = roff_res(r, buf, ln, pos);
1.175 schwarze 1528: if (e == ROFF_IGN || e == ROFF_APPEND)
1.149 schwarze 1529: return e;
1.110 schwarze 1530: assert(e == ROFF_CONT);
1.8 schwarze 1531:
1.110 schwarze 1532: ctl = roff_getcontrol(r, buf->buf, &pos);
1.35 schwarze 1533:
1.8 schwarze 1534: /*
1.2 schwarze 1535: * First, if a scope is open and we're not a macro, pass the
1.124 schwarze 1536: * text through the macro's filter.
1537: * Equations process all content themselves.
1538: * Tables process almost all content themselves, but we want
1539: * to warn about macros before passing it there.
1.2 schwarze 1540: */
1541:
1.124 schwarze 1542: if (r->last != NULL && ! ctl) {
1.2 schwarze 1543: t = r->last->tok;
1.110 schwarze 1544: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.166 schwarze 1545: if (e == ROFF_IGN)
1.149 schwarze 1546: return e;
1.166 schwarze 1547: assert(e == ROFF_CONT);
1.54 schwarze 1548: }
1.191 schwarze 1549: if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
1550: eqn_read(r->eqn, buf->buf + ppos);
1551: return ROFF_IGN;
1552: }
1.193 schwarze 1553: if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
1554: tbl_read(r->tbl, ln, buf->buf, ppos);
1555: roff_addtbl(r->man, r->tbl);
1556: return ROFF_IGN;
1557: }
1.124 schwarze 1558: if ( ! ctl)
1.177 schwarze 1559: return roff_parsetext(r, buf, pos, offs);
1.100 schwarze 1560:
1561: /* Skip empty request lines. */
1562:
1.110 schwarze 1563: if (buf->buf[pos] == '"') {
1.100 schwarze 1564: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1565: ln, pos, NULL);
1.149 schwarze 1566: return ROFF_IGN;
1.110 schwarze 1567: } else if (buf->buf[pos] == '\0')
1.149 schwarze 1568: return ROFF_IGN;
1.2 schwarze 1569:
1570: /*
1571: * If a scope is open, go to the child handler for that macro,
1572: * as it may want to preprocess before doing anything with it.
1.32 schwarze 1573: * Don't do so if an equation is open.
1.2 schwarze 1574: */
1575:
1576: if (r->last) {
1.1 schwarze 1577: t = r->last->tok;
1.149 schwarze 1578: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.2 schwarze 1579: }
1580:
1.115 schwarze 1581: /* No scope is open. This is a new request or macro. */
1582:
1583: spos = pos;
1584: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1585:
1586: /* Tables ignore most macros. */
1587:
1.180 schwarze 1588: if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.181 schwarze 1589: t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.115 schwarze 1590: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1591: ln, pos, buf->buf + spos);
1.180 schwarze 1592: if (t != TOKEN_NONE)
1.149 schwarze 1593: return ROFF_IGN;
1.129 schwarze 1594: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1595: pos++;
1.163 schwarze 1596: while (buf->buf[pos] == ' ')
1.129 schwarze 1597: pos++;
1.193 schwarze 1598: tbl_read(r->tbl, ln, buf->buf, pos);
1599: roff_addtbl(r->man, r->tbl);
1600: return ROFF_IGN;
1.115 schwarze 1601: }
1602:
1.177 schwarze 1603: /* For now, let high level macros abort .ce mode. */
1604:
1605: if (ctl && roffce_node != NULL &&
1.196 schwarze 1606: (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
1607: t == ROFF_TH || t == ROFF_TS)) {
1.177 schwarze 1608: r->man->last = roffce_node;
1609: r->man->next = ROFF_NEXT_SIBLING;
1610: roffce_lines = 0;
1611: roffce_node = NULL;
1612: }
1613:
1.2 schwarze 1614: /*
1.115 schwarze 1615: * This is neither a roff request nor a user-defined macro.
1616: * Let the standard macro set parsers handle it.
1.2 schwarze 1617: */
1618:
1.166 schwarze 1619: if (t == TOKEN_NONE)
1.149 schwarze 1620: return ROFF_CONT;
1.115 schwarze 1621:
1622: /* Execute a roff request or a user defined macro. */
1.1 schwarze 1623:
1.168 schwarze 1624: return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.2 schwarze 1625: }
1626:
1.27 schwarze 1627: void
1.2 schwarze 1628: roff_endparse(struct roff *r)
1629: {
1.193 schwarze 1630: if (r->last != NULL)
1.94 schwarze 1631: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1632: r->last->line, r->last->col,
1.167 schwarze 1633: roff_name[r->last->tok]);
1.27 schwarze 1634:
1.193 schwarze 1635: if (r->eqn != NULL) {
1.94 schwarze 1636: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1.191 schwarze 1637: r->eqn->node->line, r->eqn->node->pos, "EQ");
1638: eqn_parse(r->eqn);
1639: r->eqn = NULL;
1.32 schwarze 1640: }
1641:
1.193 schwarze 1642: if (r->tbl != NULL) {
1.94 schwarze 1643: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1644: r->tbl->line, r->tbl->pos, "TS");
1.193 schwarze 1645: tbl_end(r->tbl);
1646: r->tbl = NULL;
1.27 schwarze 1647: }
1.1 schwarze 1648: }
1649:
1650: /*
1651: * Parse a roff node's type from the input buffer. This must be in the
1652: * form of ".foo xxx" in the usual way.
1653: */
1.166 schwarze 1654: static enum roff_tok
1.87 schwarze 1655: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1 schwarze 1656: {
1.87 schwarze 1657: char *cp;
1.16 schwarze 1658: const char *mac;
1659: size_t maclen;
1.187 schwarze 1660: int deftype;
1.166 schwarze 1661: enum roff_tok t;
1.1 schwarze 1662:
1.87 schwarze 1663: cp = buf + *pos;
1664:
1665: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.166 schwarze 1666: return TOKEN_NONE;
1.1 schwarze 1667:
1.87 schwarze 1668: mac = cp;
1669: maclen = roff_getname(r, &cp, ln, ppos);
1.1 schwarze 1670:
1.187 schwarze 1671: deftype = ROFFDEF_USER | ROFFDEF_REN;
1672: r->current_string = roff_getstrn(r, mac, maclen, &deftype);
1673: switch (deftype) {
1674: case ROFFDEF_USER:
1675: t = ROFF_USERDEF;
1676: break;
1677: case ROFFDEF_REN:
1678: t = ROFF_RENAMED;
1679: break;
1680: default:
1681: t = roffhash_find(r->reqtab, mac, maclen);
1682: break;
1683: }
1.166 schwarze 1684: if (t != TOKEN_NONE)
1.87 schwarze 1685: *pos = cp - buf;
1.197 schwarze 1686: else if (deftype == ROFFDEF_UNDEF) {
1687: /* Using an undefined macro defines it to be empty. */
1688: roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
1689: roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
1690: }
1.149 schwarze 1691: return t;
1.1 schwarze 1692: }
1693:
1.138 schwarze 1694: /* --- handling of request blocks ----------------------------------------- */
1695:
1.1 schwarze 1696: static enum rofferr
1.2 schwarze 1697: roff_cblock(ROFF_ARGS)
1.1 schwarze 1698: {
1699:
1.2 schwarze 1700: /*
1701: * A block-close `..' should only be invoked as a child of an
1702: * ignore macro, otherwise raise a warning and just ignore it.
1703: */
1704:
1.110 schwarze 1705: if (r->last == NULL) {
1.94 schwarze 1706: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1707: ln, ppos, "..");
1.149 schwarze 1708: return ROFF_IGN;
1.2 schwarze 1709: }
1.1 schwarze 1710:
1.2 schwarze 1711: switch (r->last->tok) {
1.80 schwarze 1712: case ROFF_am:
1.93 schwarze 1713: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.80 schwarze 1714: case ROFF_ami:
1715: case ROFF_de:
1.23 schwarze 1716: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.80 schwarze 1717: case ROFF_dei:
1718: case ROFF_ig:
1.2 schwarze 1719: break;
1720: default:
1.94 schwarze 1721: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1722: ln, ppos, "..");
1.149 schwarze 1723: return ROFF_IGN;
1.2 schwarze 1724: }
1725:
1.110 schwarze 1726: if (buf->buf[pos] != '\0')
1.90 schwarze 1727: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 1728: ".. %s", buf->buf + pos);
1.2 schwarze 1729:
1730: roffnode_pop(r);
1731: roffnode_cleanscope(r);
1.149 schwarze 1732: return ROFF_IGN;
1.2 schwarze 1733:
1734: }
1.1 schwarze 1735:
1.2 schwarze 1736: static void
1737: roffnode_cleanscope(struct roff *r)
1738: {
1.1 schwarze 1739:
1.2 schwarze 1740: while (r->last) {
1.46 schwarze 1741: if (--r->last->endspan != 0)
1.2 schwarze 1742: break;
1743: roffnode_pop(r);
1744: }
1745: }
1.1 schwarze 1746:
1.68 schwarze 1747: static void
1748: roff_ccond(struct roff *r, int ln, int ppos)
1.2 schwarze 1749: {
1.1 schwarze 1750:
1.2 schwarze 1751: if (NULL == r->last) {
1.94 schwarze 1752: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1753: ln, ppos, "\\}");
1.68 schwarze 1754: return;
1.2 schwarze 1755: }
1.1 schwarze 1756:
1.2 schwarze 1757: switch (r->last->tok) {
1.80 schwarze 1758: case ROFF_el:
1759: case ROFF_ie:
1760: case ROFF_if:
1.2 schwarze 1761: break;
1762: default:
1.94 schwarze 1763: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1764: ln, ppos, "\\}");
1.68 schwarze 1765: return;
1.2 schwarze 1766: }
1.1 schwarze 1767:
1.2 schwarze 1768: if (r->last->endspan > -1) {
1.94 schwarze 1769: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1770: ln, ppos, "\\}");
1.68 schwarze 1771: return;
1.2 schwarze 1772: }
1773:
1774: roffnode_pop(r);
1775: roffnode_cleanscope(r);
1.68 schwarze 1776: return;
1.1 schwarze 1777: }
1778:
1779: static enum rofferr
1.2 schwarze 1780: roff_block(ROFF_ARGS)
1.1 schwarze 1781: {
1.187 schwarze 1782: const char *name, *value;
1783: char *call, *cp, *iname, *rname;
1784: size_t csz, namesz, rsz;
1785: int deftype;
1.16 schwarze 1786:
1.93 schwarze 1787: /* Ignore groff compatibility mode for now. */
1.2 schwarze 1788:
1.110 schwarze 1789: if (tok == ROFF_de1)
1.93 schwarze 1790: tok = ROFF_de;
1.123 schwarze 1791: else if (tok == ROFF_dei1)
1792: tok = ROFF_dei;
1.110 schwarze 1793: else if (tok == ROFF_am1)
1.93 schwarze 1794: tok = ROFF_am;
1.123 schwarze 1795: else if (tok == ROFF_ami1)
1796: tok = ROFF_ami;
1.93 schwarze 1797:
1798: /* Parse the macro name argument. */
1799:
1.110 schwarze 1800: cp = buf->buf + pos;
1801: if (tok == ROFF_ig) {
1.93 schwarze 1802: iname = NULL;
1803: namesz = 0;
1804: } else {
1805: iname = cp;
1806: namesz = roff_getname(r, &cp, ln, ppos);
1807: iname[namesz] = '\0';
1808: }
1.22 schwarze 1809:
1.93 schwarze 1810: /* Resolve the macro name argument if it is indirect. */
1.22 schwarze 1811:
1.110 schwarze 1812: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187 schwarze 1813: deftype = ROFFDEF_USER;
1814: name = roff_getstrn(r, iname, namesz, &deftype);
1815: if (name == NULL) {
1.93 schwarze 1816: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1817: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1818: "%.*s", (int)namesz, iname);
1819: namesz = 0;
1820: } else
1821: namesz = strlen(name);
1822: } else
1823: name = iname;
1.22 schwarze 1824:
1.110 schwarze 1825: if (namesz == 0 && tok != ROFF_ig) {
1.93 schwarze 1826: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.167 schwarze 1827: ln, ppos, roff_name[tok]);
1.149 schwarze 1828: return ROFF_IGN;
1.93 schwarze 1829: }
1.2 schwarze 1830:
1.16 schwarze 1831: roffnode_push(r, tok, name, ln, ppos);
1832:
1833: /*
1834: * At the beginning of a `de' macro, clear the existing string
1835: * with the same name, if there is one. New content will be
1.66 schwarze 1836: * appended from roff_block_text() in multiline mode.
1.16 schwarze 1837: */
1.22 schwarze 1838:
1.183 schwarze 1839: if (tok == ROFF_de || tok == ROFF_dei) {
1.86 schwarze 1840: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.183 schwarze 1841: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.187 schwarze 1842: } else if (tok == ROFF_am || tok == ROFF_ami) {
1843: deftype = ROFFDEF_ANY;
1844: value = roff_getstrn(r, iname, namesz, &deftype);
1845: switch (deftype) { /* Before appending, ... */
1846: case ROFFDEF_PRE: /* copy predefined to user-defined. */
1847: roff_setstrn(&r->strtab, name, namesz,
1848: value, strlen(value), 0);
1849: break;
1850: case ROFFDEF_REN: /* call original standard macro. */
1851: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1852: (int)strlen(value), value);
1853: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1854: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1855: free(call);
1856: break;
1857: case ROFFDEF_STD: /* rename and call standard macro. */
1858: rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
1859: roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
1860: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1861: (int)rsz, rname);
1862: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1863: free(call);
1864: free(rname);
1865: break;
1866: default:
1867: break;
1868: }
1.183 schwarze 1869: }
1.2 schwarze 1870:
1.110 schwarze 1871: if (*cp == '\0')
1.149 schwarze 1872: return ROFF_IGN;
1.1 schwarze 1873:
1.93 schwarze 1874: /* Get the custom end marker. */
1.22 schwarze 1875:
1.93 schwarze 1876: iname = cp;
1.86 schwarze 1877: namesz = roff_getname(r, &cp, ln, ppos);
1.93 schwarze 1878:
1879: /* Resolve the end marker if it is indirect. */
1880:
1.110 schwarze 1881: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.187 schwarze 1882: deftype = ROFFDEF_USER;
1883: name = roff_getstrn(r, iname, namesz, &deftype);
1884: if (name == NULL) {
1.93 schwarze 1885: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110 schwarze 1886: r->parse, ln, (int)(iname - buf->buf),
1.93 schwarze 1887: "%.*s", (int)namesz, iname);
1888: namesz = 0;
1889: } else
1890: namesz = strlen(name);
1891: } else
1892: name = iname;
1893:
1.86 schwarze 1894: if (namesz)
1895: r->last->end = mandoc_strndup(name, namesz);
1.2 schwarze 1896:
1.110 schwarze 1897: if (*cp != '\0')
1.90 schwarze 1898: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.167 schwarze 1899: ln, pos, ".%s ... %s", roff_name[tok], cp);
1.1 schwarze 1900:
1.149 schwarze 1901: return ROFF_IGN;
1.1 schwarze 1902: }
1903:
1904: static enum rofferr
1.2 schwarze 1905: roff_block_sub(ROFF_ARGS)
1.1 schwarze 1906: {
1.166 schwarze 1907: enum roff_tok t;
1.2 schwarze 1908: int i, j;
1909:
1910: /*
1911: * First check whether a custom macro exists at this level. If
1912: * it does, then check against it. This is some of groff's
1913: * stranger behaviours. If we encountered a custom end-scope
1914: * tag and that tag also happens to be a "real" macro, then we
1915: * need to try interpreting it again as a real macro. If it's
1916: * not, then return ignore. Else continue.
1917: */
1918:
1919: if (r->last->end) {
1.35 schwarze 1920: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110 schwarze 1921: if (buf->buf[i] != r->last->end[j])
1.2 schwarze 1922: break;
1.1 schwarze 1923:
1.110 schwarze 1924: if (r->last->end[j] == '\0' &&
1925: (buf->buf[i] == '\0' ||
1926: buf->buf[i] == ' ' ||
1927: buf->buf[i] == '\t')) {
1.2 schwarze 1928: roffnode_pop(r);
1929: roffnode_cleanscope(r);
1.1 schwarze 1930:
1.110 schwarze 1931: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35 schwarze 1932: i++;
1933:
1934: pos = i;
1.110 schwarze 1935: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.166 schwarze 1936: TOKEN_NONE)
1.149 schwarze 1937: return ROFF_RERUN;
1938: return ROFF_IGN;
1.2 schwarze 1939: }
1.1 schwarze 1940: }
1941:
1.2 schwarze 1942: /*
1943: * If we have no custom end-query or lookup failed, then try
1944: * pulling it out of the hashtable.
1945: */
1.1 schwarze 1946:
1.110 schwarze 1947: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1 schwarze 1948:
1.110 schwarze 1949: if (t != ROFF_cblock) {
1950: if (tok != ROFF_ig)
1951: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.149 schwarze 1952: return ROFF_IGN;
1.16 schwarze 1953: }
1.1 schwarze 1954:
1.149 schwarze 1955: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.2 schwarze 1956: }
1957:
1958: static enum rofferr
1959: roff_block_text(ROFF_ARGS)
1960: {
1961:
1.110 schwarze 1962: if (tok != ROFF_ig)
1963: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16 schwarze 1964:
1.149 schwarze 1965: return ROFF_IGN;
1.2 schwarze 1966: }
1967:
1968: static enum rofferr
1969: roff_cond_sub(ROFF_ARGS)
1970: {
1.166 schwarze 1971: enum roff_tok t;
1.37 schwarze 1972: char *ep;
1.71 schwarze 1973: int rr;
1.2 schwarze 1974:
1975: rr = r->last->rule;
1.37 schwarze 1976: roffnode_cleanscope(r);
1.39 schwarze 1977:
1.69 schwarze 1978: /*
1979: * If `\}' occurs on a macro line without a preceding macro,
1980: * drop the line completely.
1981: */
1982:
1.110 schwarze 1983: ep = buf->buf + pos;
1984: if (ep[0] == '\\' && ep[1] == '}')
1.71 schwarze 1985: rr = 0;
1.69 schwarze 1986:
1.50 schwarze 1987: /* Always check for the closing delimiter `\}'. */
1.39 schwarze 1988:
1.110 schwarze 1989: while ((ep = strchr(ep, '\\')) != NULL) {
1.190 schwarze 1990: switch (ep[1]) {
1991: case '}':
1992: memmove(ep, ep + 2, strlen(ep + 2) + 1);
1993: roff_ccond(r, ln, ep - buf->buf);
1994: break;
1995: case '\0':
1996: ++ep;
1997: break;
1998: default:
1999: ep += 2;
2000: break;
1.70 schwarze 2001: }
1.50 schwarze 2002: }
1.190 schwarze 2003:
2004: /*
2005: * Fully handle known macros when they are structurally
2006: * required or when the conditional evaluated to true.
2007: */
2008:
2009: t = roff_parse(r, buf->buf, &pos, ln, ppos);
2010: return t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT)
2011: ? (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs) : rr
2012: ? ROFF_CONT : ROFF_IGN;
1.2 schwarze 2013: }
2014:
2015: static enum rofferr
2016: roff_cond_text(ROFF_ARGS)
2017: {
1.37 schwarze 2018: char *ep;
1.71 schwarze 2019: int rr;
1.2 schwarze 2020:
2021: rr = r->last->rule;
1.37 schwarze 2022: roffnode_cleanscope(r);
1.1 schwarze 2023:
1.110 schwarze 2024: ep = buf->buf + pos;
2025: while ((ep = strchr(ep, '\\')) != NULL) {
2026: if (*(++ep) == '}') {
1.70 schwarze 2027: *ep = '&';
1.110 schwarze 2028: roff_ccond(r, ln, ep - buf->buf - 1);
1.70 schwarze 2029: }
1.119 schwarze 2030: if (*ep != '\0')
2031: ++ep;
1.2 schwarze 2032: }
1.149 schwarze 2033: return rr ? ROFF_CONT : ROFF_IGN;
1.2 schwarze 2034: }
2035:
1.138 schwarze 2036: /* --- handling of numeric and conditional expressions -------------------- */
2037:
1.77 schwarze 2038: /*
2039: * Parse a single signed integer number. Stop at the first non-digit.
2040: * If there is at least one digit, return success and advance the
2041: * parse point, else return failure and let the parse point unchanged.
2042: * Ignore overflows, treat them just like the C language.
2043: */
1.56 schwarze 2044: static int
1.133 schwarze 2045: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56 schwarze 2046: {
1.133 schwarze 2047: int myres, scaled, n, p;
1.79 schwarze 2048:
2049: if (NULL == res)
2050: res = &myres;
1.56 schwarze 2051:
2052: p = *pos;
2053: n = v[p] == '-';
1.133 schwarze 2054: if (n || v[p] == '+')
1.56 schwarze 2055: p++;
2056:
1.133 schwarze 2057: if (flags & ROFFNUM_WHITE)
2058: while (isspace((unsigned char)v[p]))
2059: p++;
2060:
1.56 schwarze 2061: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77 schwarze 2062: *res = 10 * *res + v[p] - '0';
1.56 schwarze 2063: if (p == *pos + n)
2064: return 0;
2065:
2066: if (n)
2067: *res = -*res;
2068:
1.126 schwarze 2069: /* Each number may be followed by one optional scaling unit. */
2070:
2071: switch (v[p]) {
2072: case 'f':
1.133 schwarze 2073: scaled = *res * 65536;
1.126 schwarze 2074: break;
2075: case 'i':
1.133 schwarze 2076: scaled = *res * 240;
1.126 schwarze 2077: break;
2078: case 'c':
1.133 schwarze 2079: scaled = *res * 240 / 2.54;
1.126 schwarze 2080: break;
2081: case 'v':
2082: case 'P':
1.133 schwarze 2083: scaled = *res * 40;
1.126 schwarze 2084: break;
2085: case 'm':
2086: case 'n':
1.133 schwarze 2087: scaled = *res * 24;
1.126 schwarze 2088: break;
2089: case 'p':
1.133 schwarze 2090: scaled = *res * 10 / 3;
1.126 schwarze 2091: break;
2092: case 'u':
1.133 schwarze 2093: scaled = *res;
1.126 schwarze 2094: break;
2095: case 'M':
1.133 schwarze 2096: scaled = *res * 6 / 25;
1.126 schwarze 2097: break;
2098: default:
1.133 schwarze 2099: scaled = *res;
1.126 schwarze 2100: p--;
2101: break;
2102: }
1.133 schwarze 2103: if (flags & ROFFNUM_SCALE)
2104: *res = scaled;
1.126 schwarze 2105:
2106: *pos = p + 1;
1.149 schwarze 2107: return 1;
1.56 schwarze 2108: }
2109:
1.71 schwarze 2110: /*
2111: * Evaluate a string comparison condition.
2112: * The first character is the delimiter.
2113: * Succeed if the string up to its second occurrence
2114: * matches the string up to its third occurence.
2115: * Advance the cursor after the third occurrence
2116: * or lacking that, to the end of the line.
2117: */
2118: static int
2119: roff_evalstrcond(const char *v, int *pos)
2120: {
2121: const char *s1, *s2, *s3;
2122: int match;
2123:
2124: match = 0;
2125: s1 = v + *pos; /* initial delimiter */
2126: s2 = s1 + 1; /* for scanning the first string */
2127: s3 = strchr(s2, *s1); /* for scanning the second string */
2128:
2129: if (NULL == s3) /* found no middle delimiter */
2130: goto out;
2131:
2132: while ('\0' != *++s3) {
2133: if (*s2 != *s3) { /* mismatch */
2134: s3 = strchr(s3, *s1);
2135: break;
2136: }
2137: if (*s3 == *s1) { /* found the final delimiter */
2138: match = 1;
2139: break;
2140: }
2141: s2++;
2142: }
2143:
2144: out:
2145: if (NULL == s3)
2146: s3 = strchr(s2, '\0');
1.114 schwarze 2147: else if (*s3 != '\0')
1.71 schwarze 2148: s3++;
2149: *pos = s3 - v;
1.149 schwarze 2150: return match;
1.71 schwarze 2151: }
2152:
1.77 schwarze 2153: /*
2154: * Evaluate an optionally negated single character, numerical,
2155: * or string condition.
2156: */
1.71 schwarze 2157: static int
1.143 schwarze 2158: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.5 schwarze 2159: {
1.143 schwarze 2160: char *cp, *name;
2161: size_t sz;
1.187 schwarze 2162: int deftype, number, savepos, istrue, wanttrue;
1.5 schwarze 2163:
1.71 schwarze 2164: if ('!' == v[*pos]) {
2165: wanttrue = 0;
2166: (*pos)++;
2167: } else
2168: wanttrue = 1;
2169:
1.5 schwarze 2170: switch (v[*pos]) {
1.112 schwarze 2171: case '\0':
1.149 schwarze 2172: return 0;
1.80 schwarze 2173: case 'n':
2174: case 'o':
1.5 schwarze 2175: (*pos)++;
1.149 schwarze 2176: return wanttrue;
1.80 schwarze 2177: case 'c':
2178: case 'e':
2179: case 't':
1.111 schwarze 2180: case 'v':
1.5 schwarze 2181: (*pos)++;
1.149 schwarze 2182: return !wanttrue;
1.182 schwarze 2183: case 'd':
1.143 schwarze 2184: case 'r':
1.182 schwarze 2185: cp = v + *pos + 1;
2186: while (*cp == ' ')
2187: cp++;
2188: name = cp;
2189: sz = roff_getname(r, &cp, ln, cp - v);
1.187 schwarze 2190: if (sz == 0)
2191: istrue = 0;
2192: else if (v[*pos] == 'r')
2193: istrue = roff_hasregn(r, name, sz);
2194: else {
2195: deftype = ROFFDEF_ANY;
2196: roff_getstrn(r, name, sz, &deftype);
2197: istrue = !!deftype;
2198: }
1.143 schwarze 2199: *pos = cp - v;
1.182 schwarze 2200: return istrue == wanttrue;
1.5 schwarze 2201: default:
2202: break;
2203: }
2204:
1.113 schwarze 2205: savepos = *pos;
1.133 schwarze 2206: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.149 schwarze 2207: return (number > 0) == wanttrue;
1.113 schwarze 2208: else if (*pos == savepos)
1.149 schwarze 2209: return roff_evalstrcond(v, pos) == wanttrue;
1.77 schwarze 2210: else
1.149 schwarze 2211: return 0;
1.5 schwarze 2212: }
2213:
1.2 schwarze 2214: static enum rofferr
1.21 schwarze 2215: roff_line_ignore(ROFF_ARGS)
1.6 schwarze 2216: {
1.30 schwarze 2217:
1.149 schwarze 2218: return ROFF_IGN;
1.21 schwarze 2219: }
2220:
2221: static enum rofferr
1.123 schwarze 2222: roff_insec(ROFF_ARGS)
2223: {
2224:
2225: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.167 schwarze 2226: ln, ppos, roff_name[tok]);
1.149 schwarze 2227: return ROFF_IGN;
1.123 schwarze 2228: }
2229:
2230: static enum rofferr
2231: roff_unsupp(ROFF_ARGS)
2232: {
2233:
2234: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.167 schwarze 2235: ln, ppos, roff_name[tok]);
1.149 schwarze 2236: return ROFF_IGN;
1.123 schwarze 2237: }
2238:
2239: static enum rofferr
1.2 schwarze 2240: roff_cond(ROFF_ARGS)
2241: {
1.46 schwarze 2242:
2243: roffnode_push(r, tok, NULL, ln, ppos);
1.2 schwarze 2244:
1.80 schwarze 2245: /*
1.35 schwarze 2246: * An `.el' has no conditional body: it will consume the value
2247: * of the current rstack entry set in prior `ie' calls or
1.80 schwarze 2248: * defaults to DENY.
1.35 schwarze 2249: *
2250: * If we're not an `el', however, then evaluate the conditional.
2251: */
1.1 schwarze 2252:
1.110 schwarze 2253: r->last->rule = tok == ROFF_el ?
1.80 schwarze 2254: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110 schwarze 2255: roff_evalcond(r, ln, buf->buf, &pos);
1.2 schwarze 2256:
1.35 schwarze 2257: /*
2258: * An if-else will put the NEGATION of the current evaluated
2259: * conditional into the stack of rules.
2260: */
2261:
1.110 schwarze 2262: if (tok == ROFF_ie) {
1.96 schwarze 2263: if (r->rstackpos + 1 == r->rstacksz) {
2264: r->rstacksz += 16;
2265: r->rstack = mandoc_reallocarray(r->rstack,
2266: r->rstacksz, sizeof(int));
1.35 schwarze 2267: }
1.71 schwarze 2268: r->rstack[++r->rstackpos] = !r->last->rule;
1.2 schwarze 2269: }
1.5 schwarze 2270:
2271: /* If the parent has false as its rule, then so do we. */
2272:
1.71 schwarze 2273: if (r->last->parent && !r->last->parent->rule)
2274: r->last->rule = 0;
1.5 schwarze 2275:
2276: /*
1.46 schwarze 2277: * Determine scope.
2278: * If there is nothing on the line after the conditional,
2279: * not even whitespace, use next-line scope.
1.5 schwarze 2280: */
1.2 schwarze 2281:
1.110 schwarze 2282: if (buf->buf[pos] == '\0') {
1.46 schwarze 2283: r->last->endspan = 2;
2284: goto out;
2285: }
2286:
1.110 schwarze 2287: while (buf->buf[pos] == ' ')
1.46 schwarze 2288: pos++;
2289:
2290: /* An opening brace requests multiline scope. */
1.2 schwarze 2291:
1.110 schwarze 2292: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2 schwarze 2293: r->last->endspan = -1;
2294: pos += 2;
1.144 schwarze 2295: while (buf->buf[pos] == ' ')
2296: pos++;
1.46 schwarze 2297: goto out;
1.80 schwarze 2298: }
1.2 schwarze 2299:
2300: /*
1.46 schwarze 2301: * Anything else following the conditional causes
2302: * single-line scope. Warn if the scope contains
2303: * nothing but trailing whitespace.
1.2 schwarze 2304: */
2305:
1.110 schwarze 2306: if (buf->buf[pos] == '\0')
1.89 schwarze 2307: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.167 schwarze 2308: ln, ppos, roff_name[tok]);
1.2 schwarze 2309:
1.46 schwarze 2310: r->last->endspan = 1;
1.1 schwarze 2311:
1.46 schwarze 2312: out:
1.2 schwarze 2313: *offs = pos;
1.149 schwarze 2314: return ROFF_RERUN;
1.1 schwarze 2315: }
2316:
1.2 schwarze 2317: static enum rofferr
1.7 schwarze 2318: roff_ds(ROFF_ARGS)
2319: {
1.85 schwarze 2320: char *string;
2321: const char *name;
2322: size_t namesz;
1.10 schwarze 2323:
1.123 schwarze 2324: /* Ignore groff compatibility mode for now. */
2325:
2326: if (tok == ROFF_ds1)
2327: tok = ROFF_ds;
2328: else if (tok == ROFF_as1)
2329: tok = ROFF_as;
2330:
1.10 schwarze 2331: /*
1.85 schwarze 2332: * The first word is the name of the string.
2333: * If it is empty or terminated by an escape sequence,
2334: * abort the `ds' request without defining anything.
1.10 schwarze 2335: */
1.7 schwarze 2336:
1.110 schwarze 2337: name = string = buf->buf + pos;
2338: if (*name == '\0')
1.149 schwarze 2339: return ROFF_IGN;
1.7 schwarze 2340:
1.85 schwarze 2341: namesz = roff_getname(r, &string, ln, pos);
1.110 schwarze 2342: if (name[namesz] == '\\')
1.149 schwarze 2343: return ROFF_IGN;
1.85 schwarze 2344:
2345: /* Read past the initial double-quote, if any. */
1.110 schwarze 2346: if (*string == '"')
1.7 schwarze 2347: string++;
2348:
1.10 schwarze 2349: /* The rest is the value. */
1.85 schwarze 2350: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2351: ROFF_as == tok);
1.183 schwarze 2352: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.149 schwarze 2353: return ROFF_IGN;
1.7 schwarze 2354: }
2355:
1.77 schwarze 2356: /*
2357: * Parse a single operator, one or two characters long.
2358: * If the operator is recognized, return success and advance the
2359: * parse point, else return failure and let the parse point unchanged.
2360: */
2361: static int
2362: roff_getop(const char *v, int *pos, char *res)
2363: {
2364:
2365: *res = v[*pos];
2366:
2367: switch (*res) {
1.80 schwarze 2368: case '+':
2369: case '-':
2370: case '*':
2371: case '/':
2372: case '%':
2373: case '&':
2374: case ':':
1.77 schwarze 2375: break;
2376: case '<':
2377: switch (v[*pos + 1]) {
1.80 schwarze 2378: case '=':
1.77 schwarze 2379: *res = 'l';
2380: (*pos)++;
2381: break;
1.80 schwarze 2382: case '>':
1.77 schwarze 2383: *res = '!';
2384: (*pos)++;
2385: break;
1.80 schwarze 2386: case '?':
1.77 schwarze 2387: *res = 'i';
2388: (*pos)++;
2389: break;
2390: default:
2391: break;
2392: }
2393: break;
2394: case '>':
2395: switch (v[*pos + 1]) {
1.80 schwarze 2396: case '=':
1.77 schwarze 2397: *res = 'g';
2398: (*pos)++;
2399: break;
1.80 schwarze 2400: case '?':
1.77 schwarze 2401: *res = 'a';
2402: (*pos)++;
2403: break;
2404: default:
2405: break;
2406: }
2407: break;
2408: case '=':
2409: if ('=' == v[*pos + 1])
2410: (*pos)++;
2411: break;
2412: default:
1.149 schwarze 2413: return 0;
1.77 schwarze 2414: }
2415: (*pos)++;
2416:
1.149 schwarze 2417: return *res;
1.77 schwarze 2418: }
2419:
2420: /*
2421: * Evaluate either a parenthesized numeric expression
2422: * or a single signed integer number.
2423: */
2424: static int
1.106 schwarze 2425: roff_evalpar(struct roff *r, int ln,
1.133 schwarze 2426: const char *v, int *pos, int *res, int flags)
1.77 schwarze 2427: {
2428:
2429: if ('(' != v[*pos])
1.149 schwarze 2430: return roff_getnum(v, pos, res, flags);
1.77 schwarze 2431:
2432: (*pos)++;
1.133 schwarze 2433: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.149 schwarze 2434: return 0;
1.77 schwarze 2435:
1.79 schwarze 2436: /*
2437: * Omission of the closing parenthesis
2438: * is an error in validation mode,
2439: * but ignored in evaluation mode.
2440: */
2441:
1.77 schwarze 2442: if (')' == v[*pos])
2443: (*pos)++;
1.79 schwarze 2444: else if (NULL == res)
1.149 schwarze 2445: return 0;
1.77 schwarze 2446:
1.149 schwarze 2447: return 1;
1.77 schwarze 2448: }
2449:
2450: /*
2451: * Evaluate a complete numeric expression.
2452: * Proceed left to right, there is no concept of precedence.
2453: */
2454: static int
1.106 schwarze 2455: roff_evalnum(struct roff *r, int ln, const char *v,
1.133 schwarze 2456: int *pos, int *res, int flags)
1.77 schwarze 2457: {
2458: int mypos, operand2;
2459: char operator;
2460:
2461: if (NULL == pos) {
2462: mypos = 0;
2463: pos = &mypos;
2464: }
2465:
1.133 schwarze 2466: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2467: while (isspace((unsigned char)v[*pos]))
2468: (*pos)++;
2469:
1.133 schwarze 2470: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.149 schwarze 2471: return 0;
1.77 schwarze 2472:
2473: while (1) {
1.133 schwarze 2474: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2475: while (isspace((unsigned char)v[*pos]))
2476: (*pos)++;
2477:
2478: if ( ! roff_getop(v, pos, &operator))
2479: break;
2480:
1.133 schwarze 2481: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2482: while (isspace((unsigned char)v[*pos]))
2483: (*pos)++;
2484:
1.133 schwarze 2485: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.149 schwarze 2486: return 0;
1.77 schwarze 2487:
1.133 schwarze 2488: if (flags & ROFFNUM_WHITE)
1.77 schwarze 2489: while (isspace((unsigned char)v[*pos]))
2490: (*pos)++;
1.79 schwarze 2491:
2492: if (NULL == res)
2493: continue;
1.77 schwarze 2494:
2495: switch (operator) {
1.80 schwarze 2496: case '+':
1.77 schwarze 2497: *res += operand2;
2498: break;
1.80 schwarze 2499: case '-':
1.77 schwarze 2500: *res -= operand2;
2501: break;
1.80 schwarze 2502: case '*':
1.77 schwarze 2503: *res *= operand2;
2504: break;
1.80 schwarze 2505: case '/':
1.116 schwarze 2506: if (operand2 == 0) {
1.106 schwarze 2507: mandoc_msg(MANDOCERR_DIVZERO,
2508: r->parse, ln, *pos, v);
2509: *res = 0;
2510: break;
2511: }
1.77 schwarze 2512: *res /= operand2;
2513: break;
1.80 schwarze 2514: case '%':
1.116 schwarze 2515: if (operand2 == 0) {
2516: mandoc_msg(MANDOCERR_DIVZERO,
2517: r->parse, ln, *pos, v);
2518: *res = 0;
2519: break;
2520: }
1.77 schwarze 2521: *res %= operand2;
2522: break;
1.80 schwarze 2523: case '<':
1.77 schwarze 2524: *res = *res < operand2;
2525: break;
1.80 schwarze 2526: case '>':
1.77 schwarze 2527: *res = *res > operand2;
2528: break;
1.80 schwarze 2529: case 'l':
1.77 schwarze 2530: *res = *res <= operand2;
2531: break;
1.80 schwarze 2532: case 'g':
1.77 schwarze 2533: *res = *res >= operand2;
2534: break;
1.80 schwarze 2535: case '=':
1.77 schwarze 2536: *res = *res == operand2;
2537: break;
1.80 schwarze 2538: case '!':
1.77 schwarze 2539: *res = *res != operand2;
2540: break;
1.80 schwarze 2541: case '&':
1.77 schwarze 2542: *res = *res && operand2;
2543: break;
1.80 schwarze 2544: case ':':
1.77 schwarze 2545: *res = *res || operand2;
2546: break;
1.80 schwarze 2547: case 'i':
1.77 schwarze 2548: if (operand2 < *res)
2549: *res = operand2;
2550: break;
1.80 schwarze 2551: case 'a':
1.77 schwarze 2552: if (operand2 > *res)
2553: *res = operand2;
2554: break;
2555: default:
2556: abort();
2557: }
2558: }
1.149 schwarze 2559: return 1;
1.77 schwarze 2560: }
2561:
1.138 schwarze 2562: /* --- register management ------------------------------------------------ */
2563:
1.52 schwarze 2564: void
1.60 schwarze 2565: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41 schwarze 2566: {
1.199 schwarze 2567: roff_setregn(r, name, strlen(name), val, sign, INT_MIN);
1.198 schwarze 2568: }
2569:
2570: static void
2571: roff_setregn(struct roff *r, const char *name, size_t len,
1.199 schwarze 2572: int val, char sign, int step)
1.198 schwarze 2573: {
1.52 schwarze 2574: struct roffreg *reg;
2575:
2576: /* Search for an existing register with the same name. */
2577: reg = r->regtab;
2578:
1.198 schwarze 2579: while (reg != NULL && (reg->key.sz != len ||
2580: strncmp(reg->key.p, name, len) != 0))
1.52 schwarze 2581: reg = reg->next;
1.41 schwarze 2582:
1.52 schwarze 2583: if (NULL == reg) {
2584: /* Create a new register. */
2585: reg = mandoc_malloc(sizeof(struct roffreg));
1.198 schwarze 2586: reg->key.p = mandoc_strndup(name, len);
2587: reg->key.sz = len;
1.60 schwarze 2588: reg->val = 0;
1.199 schwarze 2589: reg->step = 0;
1.52 schwarze 2590: reg->next = r->regtab;
2591: r->regtab = reg;
2592: }
2593:
1.60 schwarze 2594: if ('+' == sign)
2595: reg->val += val;
2596: else if ('-' == sign)
2597: reg->val -= val;
2598: else
2599: reg->val = val;
1.199 schwarze 2600: if (step != INT_MIN)
2601: reg->step = step;
1.41 schwarze 2602: }
2603:
1.65 schwarze 2604: /*
2605: * Handle some predefined read-only number registers.
2606: * For now, return -1 if the requested register is not predefined;
2607: * in case a predefined read-only register having the value -1
2608: * were to turn up, another special value would have to be chosen.
2609: */
2610: static int
1.145 schwarze 2611: roff_getregro(const struct roff *r, const char *name)
1.65 schwarze 2612: {
2613:
2614: switch (*name) {
1.145 schwarze 2615: case '$': /* Number of arguments of the last macro evaluated. */
1.149 schwarze 2616: return r->argc;
1.80 schwarze 2617: case 'A': /* ASCII approximation mode is always off. */
1.149 schwarze 2618: return 0;
1.80 schwarze 2619: case 'g': /* Groff compatibility mode is always on. */
1.149 schwarze 2620: return 1;
1.80 schwarze 2621: case 'H': /* Fixed horizontal resolution. */
1.149 schwarze 2622: return 24;
1.80 schwarze 2623: case 'j': /* Always adjust left margin only. */
1.149 schwarze 2624: return 0;
1.80 schwarze 2625: case 'T': /* Some output device is always defined. */
1.149 schwarze 2626: return 1;
1.80 schwarze 2627: case 'V': /* Fixed vertical resolution. */
1.149 schwarze 2628: return 40;
1.65 schwarze 2629: default:
1.149 schwarze 2630: return -1;
1.65 schwarze 2631: }
2632: }
2633:
1.53 schwarze 2634: int
1.198 schwarze 2635: roff_getreg(struct roff *r, const char *name)
1.41 schwarze 2636: {
1.199 schwarze 2637: return roff_getregn(r, name, strlen(name), '\0');
1.53 schwarze 2638: }
2639:
2640: static int
1.199 schwarze 2641: roff_getregn(struct roff *r, const char *name, size_t len, char sign)
1.53 schwarze 2642: {
2643: struct roffreg *reg;
1.65 schwarze 2644: int val;
2645:
2646: if ('.' == name[0] && 2 == len) {
1.145 schwarze 2647: val = roff_getregro(r, name + 1);
1.65 schwarze 2648: if (-1 != val)
1.149 schwarze 2649: return val;
1.65 schwarze 2650: }
1.53 schwarze 2651:
1.199 schwarze 2652: for (reg = r->regtab; reg; reg = reg->next) {
1.53 schwarze 2653: if (len == reg->key.sz &&
1.199 schwarze 2654: 0 == strncmp(name, reg->key.p, len)) {
2655: switch (sign) {
2656: case '+':
2657: reg->val += reg->step;
2658: break;
2659: case '-':
2660: reg->val -= reg->step;
2661: break;
2662: default:
2663: break;
2664: }
1.149 schwarze 2665: return reg->val;
1.199 schwarze 2666: }
2667: }
1.143 schwarze 2668:
1.199 schwarze 2669: roff_setregn(r, name, len, 0, '\0', INT_MIN);
1.149 schwarze 2670: return 0;
1.143 schwarze 2671: }
2672:
2673: static int
2674: roff_hasregn(const struct roff *r, const char *name, size_t len)
2675: {
2676: struct roffreg *reg;
2677: int val;
2678:
2679: if ('.' == name[0] && 2 == len) {
1.145 schwarze 2680: val = roff_getregro(r, name + 1);
1.143 schwarze 2681: if (-1 != val)
1.149 schwarze 2682: return 1;
1.143 schwarze 2683: }
2684:
2685: for (reg = r->regtab; reg; reg = reg->next)
2686: if (len == reg->key.sz &&
2687: 0 == strncmp(name, reg->key.p, len))
1.149 schwarze 2688: return 1;
1.41 schwarze 2689:
1.149 schwarze 2690: return 0;
1.41 schwarze 2691: }
2692:
1.52 schwarze 2693: static void
2694: roff_freereg(struct roffreg *reg)
1.41 schwarze 2695: {
1.52 schwarze 2696: struct roffreg *old_reg;
1.41 schwarze 2697:
1.52 schwarze 2698: while (NULL != reg) {
2699: free(reg->key.p);
2700: old_reg = reg;
2701: reg = reg->next;
2702: free(old_reg);
2703: }
1.41 schwarze 2704: }
1.7 schwarze 2705:
2706: static enum rofferr
1.6 schwarze 2707: roff_nr(ROFF_ARGS)
1.1 schwarze 2708: {
1.199 schwarze 2709: char *key, *val, *step;
1.85 schwarze 2710: size_t keysz;
1.199 schwarze 2711: int iv, is, len;
1.60 schwarze 2712: char sign;
1.6 schwarze 2713:
1.110 schwarze 2714: key = val = buf->buf + pos;
2715: if (*key == '\0')
1.149 schwarze 2716: return ROFF_IGN;
1.85 schwarze 2717:
2718: keysz = roff_getname(r, &val, ln, pos);
1.110 schwarze 2719: if (key[keysz] == '\\')
1.149 schwarze 2720: return ROFF_IGN;
1.6 schwarze 2721:
1.60 schwarze 2722: sign = *val;
1.110 schwarze 2723: if (sign == '+' || sign == '-')
1.60 schwarze 2724: val++;
2725:
1.199 schwarze 2726: len = 0;
2727: if (roff_evalnum(r, ln, val, &len, &iv, ROFFNUM_SCALE) == 0)
2728: return ROFF_IGN;
2729:
2730: step = val + len;
2731: while (isspace((unsigned char)*step))
2732: step++;
2733: if (roff_evalnum(r, ln, step, NULL, &is, 0) == 0)
2734: is = INT_MIN;
1.1 schwarze 2735:
1.199 schwarze 2736: roff_setregn(r, key, keysz, iv, sign, is);
1.149 schwarze 2737: return ROFF_IGN;
1.76 schwarze 2738: }
2739:
2740: static enum rofferr
2741: roff_rr(ROFF_ARGS)
2742: {
2743: struct roffreg *reg, **prev;
1.85 schwarze 2744: char *name, *cp;
2745: size_t namesz;
1.76 schwarze 2746:
1.110 schwarze 2747: name = cp = buf->buf + pos;
2748: if (*name == '\0')
1.149 schwarze 2749: return ROFF_IGN;
1.85 schwarze 2750: namesz = roff_getname(r, &cp, ln, pos);
2751: name[namesz] = '\0';
1.76 schwarze 2752:
2753: prev = &r->regtab;
2754: while (1) {
2755: reg = *prev;
1.110 schwarze 2756: if (reg == NULL || !strcmp(name, reg->key.p))
1.76 schwarze 2757: break;
2758: prev = ®->next;
2759: }
1.110 schwarze 2760: if (reg != NULL) {
1.76 schwarze 2761: *prev = reg->next;
2762: free(reg->key.p);
2763: free(reg);
2764: }
1.149 schwarze 2765: return ROFF_IGN;
1.29 schwarze 2766: }
2767:
1.138 schwarze 2768: /* --- handler functions for roff requests -------------------------------- */
2769:
1.29 schwarze 2770: static enum rofferr
2771: roff_rm(ROFF_ARGS)
2772: {
2773: const char *name;
2774: char *cp;
1.85 schwarze 2775: size_t namesz;
1.29 schwarze 2776:
1.110 schwarze 2777: cp = buf->buf + pos;
2778: while (*cp != '\0') {
1.85 schwarze 2779: name = cp;
1.110 schwarze 2780: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85 schwarze 2781: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.183 schwarze 2782: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.110 schwarze 2783: if (name[namesz] == '\\')
1.85 schwarze 2784: break;
1.29 schwarze 2785: }
1.149 schwarze 2786: return ROFF_IGN;
1.51 schwarze 2787: }
2788:
2789: static enum rofferr
2790: roff_it(ROFF_ARGS)
2791: {
2792: int iv;
2793:
2794: /* Parse the number of lines. */
1.133 schwarze 2795:
2796: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95 schwarze 2797: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110 schwarze 2798: ln, ppos, buf->buf + 1);
1.149 schwarze 2799: return ROFF_IGN;
1.51 schwarze 2800: }
2801:
1.134 schwarze 2802: while (isspace((unsigned char)buf->buf[pos]))
2803: pos++;
2804:
2805: /*
2806: * Arm the input line trap.
2807: * Special-casing "an-trap" is an ugly workaround to cope
2808: * with DocBook stupidly fiddling with man(7) internals.
2809: */
1.133 schwarze 2810:
1.51 schwarze 2811: roffit_lines = iv;
1.134 schwarze 2812: roffit_macro = mandoc_strdup(iv != 1 ||
2813: strcmp(buf->buf + pos, "an-trap") ?
2814: buf->buf + pos : "br");
1.149 schwarze 2815: return ROFF_IGN;
1.47 schwarze 2816: }
2817:
2818: static enum rofferr
2819: roff_Dd(ROFF_ARGS)
2820: {
1.187 schwarze 2821: int mask;
2822: enum roff_tok t, te;
1.99 schwarze 2823:
1.187 schwarze 2824: switch (tok) {
2825: case ROFF_Dd:
2826: tok = MDOC_Dd;
2827: te = MDOC_MAX;
2828: if (r->format == 0)
2829: r->format = MPARSE_MDOC;
2830: mask = MPARSE_MDOC | MPARSE_QUICK;
2831: break;
2832: case ROFF_TH:
2833: tok = MAN_TH;
2834: te = MAN_MAX;
2835: if (r->format == 0)
2836: r->format = MPARSE_MAN;
2837: mask = MPARSE_QUICK;
2838: break;
2839: default:
2840: abort();
2841: }
2842: if ((r->options & mask) == 0)
2843: for (t = tok; t < te; t++)
2844: roff_setstr(r, roff_name[t], NULL, 0);
1.149 schwarze 2845: return ROFF_CONT;
1.14 schwarze 2846: }
2847:
2848: static enum rofferr
1.27 schwarze 2849: roff_TE(ROFF_ARGS)
2850: {
1.193 schwarze 2851: if (r->tbl == NULL) {
1.94 schwarze 2852: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2853: ln, ppos, "TE");
1.193 schwarze 2854: return ROFF_IGN;
2855: }
2856: if (tbl_end(r->tbl) == 0) {
2857: r->tbl = NULL;
1.130 schwarze 2858: free(buf->buf);
2859: buf->buf = mandoc_strdup(".sp");
2860: buf->sz = 4;
1.201 schwarze 2861: *offs = 0;
1.149 schwarze 2862: return ROFF_REPARSE;
1.130 schwarze 2863: }
1.193 schwarze 2864: r->tbl = NULL;
1.149 schwarze 2865: return ROFF_IGN;
1.27 schwarze 2866: }
2867:
2868: static enum rofferr
2869: roff_T_(ROFF_ARGS)
2870: {
2871:
2872: if (NULL == r->tbl)
1.94 schwarze 2873: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2874: ln, ppos, "T&");
1.27 schwarze 2875: else
1.168 schwarze 2876: tbl_restart(ln, ppos, r->tbl);
1.27 schwarze 2877:
1.149 schwarze 2878: return ROFF_IGN;
1.27 schwarze 2879: }
2880:
1.102 schwarze 2881: /*
2882: * Handle in-line equation delimiters.
2883: */
2884: static enum rofferr
1.110 schwarze 2885: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41 schwarze 2886: {
1.105 schwarze 2887: char *cp1, *cp2;
2888: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41 schwarze 2889:
1.102 schwarze 2890: /*
2891: * Outside equations, look for an opening delimiter.
2892: * If we are inside an equation, we already know it is
2893: * in-line, or this function wouldn't have been called;
2894: * so look for a closing delimiter.
2895: */
2896:
1.110 schwarze 2897: cp1 = buf->buf + pos;
1.102 schwarze 2898: cp2 = strchr(cp1, r->eqn == NULL ?
2899: r->last_eqn->odelim : r->last_eqn->cdelim);
2900: if (cp2 == NULL)
1.149 schwarze 2901: return ROFF_CONT;
1.102 schwarze 2902:
1.105 schwarze 2903: *cp2++ = '\0';
2904: bef_pr = bef_nl = aft_nl = aft_pr = "";
2905:
2906: /* Handle preceding text, protecting whitespace. */
2907:
1.110 schwarze 2908: if (*buf->buf != '\0') {
1.105 schwarze 2909: if (r->eqn == NULL)
2910: bef_pr = "\\&";
2911: bef_nl = "\n";
2912: }
2913:
2914: /*
2915: * Prepare replacing the delimiter with an equation macro
2916: * and drop leading white space from the equation.
2917: */
1.102 schwarze 2918:
1.105 schwarze 2919: if (r->eqn == NULL) {
2920: while (*cp2 == ' ')
2921: cp2++;
2922: mac = ".EQ";
2923: } else
2924: mac = ".EN";
2925:
2926: /* Handle following text, protecting whitespace. */
2927:
2928: if (*cp2 != '\0') {
2929: aft_nl = "\n";
2930: if (r->eqn != NULL)
2931: aft_pr = "\\&";
2932: }
2933:
2934: /* Do the actual replacement. */
2935:
1.110 schwarze 2936: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105 schwarze 2937: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110 schwarze 2938: free(buf->buf);
2939: buf->buf = cp1;
1.102 schwarze 2940:
2941: /* Toggle the in-line state of the eqn subsystem. */
2942:
2943: r->eqn_inline = r->eqn == NULL;
1.149 schwarze 2944: return ROFF_REPARSE;
1.41 schwarze 2945: }
2946:
1.107 schwarze 2947: static enum rofferr
2948: roff_EQ(ROFF_ARGS)
1.32 schwarze 2949: {
1.191 schwarze 2950: struct roff_node *n;
2951:
1.194 schwarze 2952: if (r->man->macroset == MACROSET_MAN)
2953: man_breakscope(r->man, ROFF_EQ);
1.191 schwarze 2954: n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
2955: if (ln > r->man->last->line)
2956: n->flags |= NODE_LINE;
2957: n->eqn = mandoc_calloc(1, sizeof(*n->eqn));
2958: n->eqn->expectargs = UINT_MAX;
2959: roff_node_append(r->man, n);
2960: r->man->next = ROFF_NEXT_SIBLING;
1.32 schwarze 2961:
1.110 schwarze 2962: assert(r->eqn == NULL);
1.191 schwarze 2963: if (r->last_eqn == NULL)
2964: r->last_eqn = eqn_alloc(r->parse);
2965: else
2966: eqn_reset(r->last_eqn);
2967: r->eqn = r->last_eqn;
2968: r->eqn->node = n;
1.41 schwarze 2969:
1.110 schwarze 2970: if (buf->buf[pos] != '\0')
1.107 schwarze 2971: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110 schwarze 2972: ".EQ %s", buf->buf + pos);
1.41 schwarze 2973:
1.149 schwarze 2974: return ROFF_IGN;
1.32 schwarze 2975: }
2976:
2977: static enum rofferr
2978: roff_EN(ROFF_ARGS)
2979: {
1.191 schwarze 2980: if (r->eqn != NULL) {
2981: eqn_parse(r->eqn);
2982: r->eqn = NULL;
2983: } else
2984: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
2985: if (buf->buf[pos] != '\0')
2986: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
2987: "EN %s", buf->buf + pos);
1.149 schwarze 2988: return ROFF_IGN;
1.32 schwarze 2989: }
2990:
2991: static enum rofferr
1.27 schwarze 2992: roff_TS(ROFF_ARGS)
2993: {
1.193 schwarze 2994: if (r->tbl != NULL) {
1.94 schwarze 2995: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2996: ln, ppos, "TS breaks TS");
1.193 schwarze 2997: tbl_end(r->tbl);
1.27 schwarze 2998: }
1.193 schwarze 2999: r->tbl = tbl_alloc(ppos, ln, r->parse);
1.27 schwarze 3000: if (r->last_tbl)
1.193 schwarze 3001: r->last_tbl->next = r->tbl;
1.27 schwarze 3002: else
1.193 schwarze 3003: r->first_tbl = r->tbl;
3004: r->last_tbl = r->tbl;
1.169 schwarze 3005: return ROFF_IGN;
3006: }
3007:
3008: static enum rofferr
3009: roff_onearg(ROFF_ARGS)
3010: {
3011: struct roff_node *n;
3012: char *cp;
1.177 schwarze 3013: int npos;
1.169 schwarze 3014:
1.174 schwarze 3015: if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.192 schwarze 3016: (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
3017: tok == ROFF_ti))
1.174 schwarze 3018: man_breakscope(r->man, tok);
3019:
1.181 schwarze 3020: if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.177 schwarze 3021: r->man->last = roffce_node;
3022: r->man->next = ROFF_NEXT_SIBLING;
3023: }
3024:
1.169 schwarze 3025: roff_elem_alloc(r->man, ln, ppos, tok);
3026: n = r->man->last;
3027:
3028: cp = buf->buf + pos;
3029: if (*cp != '\0') {
3030: while (*cp != '\0' && *cp != ' ')
3031: cp++;
3032: while (*cp == ' ')
3033: *cp++ = '\0';
3034: if (*cp != '\0')
3035: mandoc_vmsg(MANDOCERR_ARG_EXCESS,
3036: r->parse, ln, cp - buf->buf,
3037: "%s ... %s", roff_name[tok], cp);
3038: roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.172 schwarze 3039: }
3040:
1.181 schwarze 3041: if (tok == ROFF_ce || tok == ROFF_rj) {
3042: if (r->man->last->type == ROFFT_ELEM) {
1.177 schwarze 3043: roff_word_alloc(r->man, ln, pos, "1");
3044: r->man->last->flags |= NODE_NOSRC;
3045: }
3046: npos = 0;
3047: if (roff_evalnum(r, ln, r->man->last->string, &npos,
3048: &roffce_lines, 0) == 0) {
3049: mandoc_vmsg(MANDOCERR_CE_NONUM,
3050: r->parse, ln, pos, "ce %s", buf->buf + pos);
3051: roffce_lines = 1;
3052: }
3053: if (roffce_lines < 1) {
3054: r->man->last = r->man->last->parent;
3055: roffce_node = NULL;
3056: roffce_lines = 0;
3057: } else
3058: roffce_node = r->man->last->parent;
3059: } else {
3060: n->flags |= NODE_VALID | NODE_ENDED;
3061: r->man->last = n;
3062: }
3063: n->flags |= NODE_LINE;
1.172 schwarze 3064: r->man->next = ROFF_NEXT_SIBLING;
3065: return ROFF_IGN;
3066: }
3067:
3068: static enum rofferr
3069: roff_manyarg(ROFF_ARGS)
3070: {
3071: struct roff_node *n;
3072: char *sp, *ep;
3073:
3074: roff_elem_alloc(r->man, ln, ppos, tok);
3075: n = r->man->last;
3076:
3077: for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
3078: while (*ep != '\0' && *ep != ' ')
3079: ep++;
3080: while (*ep == ' ')
3081: *ep++ = '\0';
3082: roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.169 schwarze 3083: }
3084:
3085: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3086: r->man->last = n;
3087: r->man->next = ROFF_NEXT_SIBLING;
1.149 schwarze 3088: return ROFF_IGN;
1.123 schwarze 3089: }
3090:
3091: static enum rofferr
1.183 schwarze 3092: roff_als(ROFF_ARGS)
3093: {
3094: char *oldn, *newn, *end, *value;
3095: size_t oldsz, newsz, valsz;
3096:
3097: newn = oldn = buf->buf + pos;
3098: if (*newn == '\0')
3099: return ROFF_IGN;
3100:
3101: newsz = roff_getname(r, &oldn, ln, pos);
3102: if (newn[newsz] == '\\' || *oldn == '\0')
3103: return ROFF_IGN;
3104:
3105: end = oldn;
3106: oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
3107: if (oldsz == 0)
3108: return ROFF_IGN;
3109:
1.187 schwarze 3110: valsz = mandoc_asprintf(&value, ".%.*s \\$*\\\"\n",
3111: (int)oldsz, oldn);
1.183 schwarze 3112: roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
3113: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3114: free(value);
3115: return ROFF_IGN;
3116: }
3117:
3118: static enum rofferr
1.168 schwarze 3119: roff_br(ROFF_ARGS)
1.123 schwarze 3120: {
1.174 schwarze 3121: if (r->man->flags & (MAN_BLINE | MAN_ELINE))
3122: man_breakscope(r->man, ROFF_br);
1.168 schwarze 3123: roff_elem_alloc(r->man, ln, ppos, ROFF_br);
3124: if (buf->buf[pos] != '\0')
3125: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
3126: "%s %s", roff_name[tok], buf->buf + pos);
3127: r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3128: r->man->next = ROFF_NEXT_SIBLING;
3129: return ROFF_IGN;
1.27 schwarze 3130: }
3131:
3132: static enum rofferr
1.48 schwarze 3133: roff_cc(ROFF_ARGS)
3134: {
3135: const char *p;
3136:
1.110 schwarze 3137: p = buf->buf + pos;
1.48 schwarze 3138:
1.110 schwarze 3139: if (*p == '\0' || (r->control = *p++) == '.')
1.175 schwarze 3140: r->control = '\0';
1.48 schwarze 3141:
1.110 schwarze 3142: if (*p != '\0')
1.132 schwarze 3143: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3144: ln, p - buf->buf, "cc ... %s", p);
1.48 schwarze 3145:
1.149 schwarze 3146: return ROFF_IGN;
1.48 schwarze 3147: }
3148:
3149: static enum rofferr
1.175 schwarze 3150: roff_ec(ROFF_ARGS)
3151: {
3152: const char *p;
3153:
3154: p = buf->buf + pos;
3155: if (*p == '\0')
3156: r->escape = '\\';
3157: else {
3158: r->escape = *p;
3159: if (*++p != '\0')
3160: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3161: ln, p - buf->buf, "ec ... %s", p);
3162: }
3163: return ROFF_IGN;
3164: }
3165:
3166: static enum rofferr
3167: roff_eo(ROFF_ARGS)
3168: {
3169: r->escape = '\0';
3170: if (buf->buf[pos] != '\0')
3171: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
3172: ln, pos, "eo %s", buf->buf + pos);
3173: return ROFF_IGN;
1.202 schwarze 3174: }
3175:
3176: static enum rofferr
3177: roff_nop(ROFF_ARGS)
3178: {
3179: while (buf->buf[pos] == ' ')
3180: pos++;
3181: *offs = pos;
3182: return ROFF_RERUN;
1.175 schwarze 3183: }
3184:
3185: static enum rofferr
1.42 schwarze 3186: roff_tr(ROFF_ARGS)
3187: {
3188: const char *p, *first, *second;
3189: size_t fsz, ssz;
3190: enum mandoc_esc esc;
3191:
1.110 schwarze 3192: p = buf->buf + pos;
1.42 schwarze 3193:
1.110 schwarze 3194: if (*p == '\0') {
1.132 schwarze 3195: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.149 schwarze 3196: return ROFF_IGN;
1.42 schwarze 3197: }
3198:
1.110 schwarze 3199: while (*p != '\0') {
1.42 schwarze 3200: fsz = ssz = 1;
3201:
3202: first = p++;
1.110 schwarze 3203: if (*first == '\\') {
1.42 schwarze 3204: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 3205: if (esc == ESCAPE_ERROR) {
1.92 schwarze 3206: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 3207: ln, (int)(p - buf->buf), first);
1.149 schwarze 3208: return ROFF_IGN;
1.42 schwarze 3209: }
3210: fsz = (size_t)(p - first);
3211: }
3212:
3213: second = p++;
1.110 schwarze 3214: if (*second == '\\') {
1.42 schwarze 3215: esc = mandoc_escape(&p, NULL, NULL);
1.110 schwarze 3216: if (esc == ESCAPE_ERROR) {
1.92 schwarze 3217: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110 schwarze 3218: ln, (int)(p - buf->buf), second);
1.149 schwarze 3219: return ROFF_IGN;
1.42 schwarze 3220: }
3221: ssz = (size_t)(p - second);
1.110 schwarze 3222: } else if (*second == '\0') {
1.132 schwarze 3223: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
3224: ln, first - buf->buf, "tr %s", first);
1.42 schwarze 3225: second = " ";
3226: p--;
3227: }
3228:
3229: if (fsz > 1) {
1.80 schwarze 3230: roff_setstrn(&r->xmbtab, first, fsz,
3231: second, ssz, 0);
1.42 schwarze 3232: continue;
3233: }
3234:
1.110 schwarze 3235: if (r->xtab == NULL)
1.80 schwarze 3236: r->xtab = mandoc_calloc(128,
3237: sizeof(struct roffstr));
1.42 schwarze 3238:
3239: free(r->xtab[(int)*first].p);
3240: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
3241: r->xtab[(int)*first].sz = ssz;
3242: }
3243:
1.149 schwarze 3244: return ROFF_IGN;
1.42 schwarze 3245: }
3246:
3247: static enum rofferr
1.178 schwarze 3248: roff_rn(ROFF_ARGS)
3249: {
3250: const char *value;
3251: char *oldn, *newn, *end;
3252: size_t oldsz, newsz;
1.187 schwarze 3253: int deftype;
1.178 schwarze 3254:
3255: oldn = newn = buf->buf + pos;
3256: if (*oldn == '\0')
3257: return ROFF_IGN;
3258:
3259: oldsz = roff_getname(r, &newn, ln, pos);
3260: if (oldn[oldsz] == '\\' || *newn == '\0')
3261: return ROFF_IGN;
3262:
3263: end = newn;
3264: newsz = roff_getname(r, &end, ln, newn - buf->buf);
3265: if (newsz == 0)
3266: return ROFF_IGN;
3267:
1.187 schwarze 3268: deftype = ROFFDEF_ANY;
3269: value = roff_getstrn(r, oldn, oldsz, &deftype);
3270: switch (deftype) {
3271: case ROFFDEF_USER:
1.178 schwarze 3272: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3273: roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
3274: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.187 schwarze 3275: break;
3276: case ROFFDEF_PRE:
3277: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3278: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3279: break;
3280: case ROFFDEF_REN:
1.178 schwarze 3281: roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
3282: roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.187 schwarze 3283: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3284: break;
3285: case ROFFDEF_STD:
1.178 schwarze 3286: roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.187 schwarze 3287: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3288: break;
3289: default:
3290: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3291: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3292: break;
3293: }
1.178 schwarze 3294: return ROFF_IGN;
3295: }
3296:
3297: static enum rofferr
1.14 schwarze 3298: roff_so(ROFF_ARGS)
3299: {
1.121 schwarze 3300: char *name, *cp;
1.15 schwarze 3301:
1.110 schwarze 3302: name = buf->buf + pos;
1.97 schwarze 3303: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14 schwarze 3304:
1.22 schwarze 3305: /*
3306: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3307: * opening anything that's not in our cwd or anything beneath
3308: * it. Thus, explicitly disallow traversing up the file-system
3309: * or using absolute paths.
3310: */
3311:
1.110 schwarze 3312: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83 schwarze 3313: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3314: ".so %s", name);
1.121 schwarze 3315: buf->sz = mandoc_asprintf(&cp,
3316: ".sp\nSee the file %s.\n.sp", name) + 1;
3317: free(buf->buf);
3318: buf->buf = cp;
3319: *offs = 0;
1.149 schwarze 3320: return ROFF_REPARSE;
1.14 schwarze 3321: }
3322:
3323: *offs = pos;
1.149 schwarze 3324: return ROFF_SO;
1.7 schwarze 3325: }
3326:
1.138 schwarze 3327: /* --- user defined strings and macros ------------------------------------ */
3328:
1.16 schwarze 3329: static enum rofferr
3330: roff_userdef(ROFF_ARGS)
1.12 schwarze 3331: {
1.183 schwarze 3332: const char *arg[16], *ap;
1.16 schwarze 3333: char *cp, *n1, *n2;
1.164 schwarze 3334: int expand_count, i, ib, ie;
1.135 schwarze 3335: size_t asz, rsz;
1.12 schwarze 3336:
1.16 schwarze 3337: /*
3338: * Collect pointers to macro argument strings
1.61 schwarze 3339: * and NUL-terminate them.
1.16 schwarze 3340: */
1.135 schwarze 3341:
1.145 schwarze 3342: r->argc = 0;
1.110 schwarze 3343: cp = buf->buf + pos;
1.183 schwarze 3344: for (i = 0; i < 16; i++) {
1.145 schwarze 3345: if (*cp == '\0')
3346: arg[i] = "";
3347: else {
3348: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
3349: r->argc = i + 1;
3350: }
3351: }
1.16 schwarze 3352:
3353: /*
3354: * Expand macro arguments.
1.12 schwarze 3355: */
1.135 schwarze 3356:
3357: buf->sz = strlen(r->current_string) + 1;
1.164 schwarze 3358: n1 = n2 = cp = mandoc_malloc(buf->sz);
1.135 schwarze 3359: memcpy(n1, r->current_string, buf->sz);
1.164 schwarze 3360: expand_count = 0;
1.135 schwarze 3361: while (*cp != '\0') {
3362:
3363: /* Scan ahead for the next argument invocation. */
3364:
3365: if (*cp++ != '\\')
3366: continue;
3367: if (*cp++ != '$')
3368: continue;
1.146 schwarze 3369: if (*cp == '*') { /* \\$* inserts all arguments */
3370: ib = 0;
3371: ie = r->argc - 1;
3372: } else { /* \\$1 .. \\$9 insert one argument */
3373: ib = ie = *cp - '1';
3374: if (ib < 0 || ib > 8)
3375: continue;
3376: }
1.135 schwarze 3377: cp -= 2;
1.164 schwarze 3378:
3379: /*
3380: * Prevent infinite recursion.
3381: */
3382:
3383: if (cp >= n2)
3384: expand_count = 1;
3385: else if (++expand_count > EXPAND_LIMIT) {
3386: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
3387: ln, (int)(cp - n1), NULL);
1.165 schwarze 3388: free(buf->buf);
3389: buf->buf = n1;
1.201 schwarze 3390: *offs = 0;
1.164 schwarze 3391: return ROFF_IGN;
3392: }
1.135 schwarze 3393:
3394: /*
3395: * Determine the size of the expanded argument,
3396: * taking escaping of quotes into account.
3397: */
3398:
1.146 schwarze 3399: asz = ie > ib ? ie - ib : 0; /* for blanks */
3400: for (i = ib; i <= ie; i++) {
3401: for (ap = arg[i]; *ap != '\0'; ap++) {
3402: asz++;
3403: if (*ap == '"')
3404: asz += 3;
3405: }
1.135 schwarze 3406: }
3407: if (asz != 3) {
3408:
3409: /*
3410: * Determine the size of the rest of the
3411: * unexpanded macro, including the NUL.
3412: */
3413:
3414: rsz = buf->sz - (cp - n1) - 3;
3415:
3416: /*
3417: * When shrinking, move before
3418: * releasing the storage.
3419: */
3420:
3421: if (asz < 3)
3422: memmove(cp + asz, cp + 3, rsz);
3423:
3424: /*
3425: * Resize the storage for the macro
3426: * and readjust the parse pointer.
3427: */
3428:
3429: buf->sz += asz - 3;
3430: n2 = mandoc_realloc(n1, buf->sz);
3431: cp = n2 + (cp - n1);
3432: n1 = n2;
3433:
3434: /*
3435: * When growing, make room
3436: * for the expanded argument.
3437: */
3438:
3439: if (asz > 3)
3440: memmove(cp + asz, cp + 3, rsz);
3441: }
3442:
3443: /* Copy the expanded argument, escaping quotes. */
3444:
3445: n2 = cp;
1.146 schwarze 3446: for (i = ib; i <= ie; i++) {
3447: for (ap = arg[i]; *ap != '\0'; ap++) {
3448: if (*ap == '"') {
3449: memcpy(n2, "\\(dq", 4);
3450: n2 += 4;
3451: } else
3452: *n2++ = *ap;
3453: }
3454: if (i < ie)
3455: *n2++ = ' ';
1.16 schwarze 3456: }
1.12 schwarze 3457: }
3458:
1.16 schwarze 3459: /*
3460: * Replace the macro invocation
3461: * by the expanded macro.
3462: */
1.135 schwarze 3463:
1.110 schwarze 3464: free(buf->buf);
3465: buf->buf = n1;
1.120 schwarze 3466: *offs = 0;
1.16 schwarze 3467:
1.149 schwarze 3468: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
3469: ROFF_REPARSE : ROFF_APPEND;
1.12 schwarze 3470: }
1.28 schwarze 3471:
1.178 schwarze 3472: /*
3473: * Calling a high-level macro that was renamed with .rn.
3474: * r->current_string has already been set up by roff_parse().
3475: */
3476: static enum rofferr
3477: roff_renamed(ROFF_ARGS)
3478: {
3479: char *nbuf;
3480:
1.187 schwarze 3481: buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
3482: buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.178 schwarze 3483: free(buf->buf);
3484: buf->buf = nbuf;
1.201 schwarze 3485: *offs = 0;
1.178 schwarze 3486: return ROFF_CONT;
3487: }
3488:
1.85 schwarze 3489: static size_t
1.28 schwarze 3490: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3491: {
3492: char *name, *cp;
1.85 schwarze 3493: size_t namesz;
1.28 schwarze 3494:
3495: name = *cpp;
3496: if ('\0' == *name)
1.149 schwarze 3497: return 0;
1.28 schwarze 3498:
1.85 schwarze 3499: /* Read until end of name and terminate it with NUL. */
3500: for (cp = name; 1; cp++) {
3501: if ('\0' == *cp || ' ' == *cp) {
3502: namesz = cp - name;
3503: break;
3504: }
1.28 schwarze 3505: if ('\\' != *cp)
3506: continue;
1.88 schwarze 3507: namesz = cp - name;
3508: if ('{' == cp[1] || '}' == cp[1])
3509: break;
1.28 schwarze 3510: cp++;
3511: if ('\\' == *cp)
3512: continue;
1.97 schwarze 3513: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3514: "%.*s", (int)(cp - name + 1), name);
1.85 schwarze 3515: mandoc_escape((const char **)&cp, NULL, NULL);
3516: break;
1.28 schwarze 3517: }
3518:
3519: /* Read past spaces. */
3520: while (' ' == *cp)
3521: cp++;
3522:
3523: *cpp = cp;
1.149 schwarze 3524: return namesz;
1.28 schwarze 3525: }
3526:
1.16 schwarze 3527: /*
3528: * Store *string into the user-defined string called *name.
3529: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66 schwarze 3530: * append == 0: replace mode
3531: * append == 1: single-line append mode
3532: * append == 2: multiline append mode, append '\n' after each call
1.16 schwarze 3533: */
1.8 schwarze 3534: static void
1.16 schwarze 3535: roff_setstr(struct roff *r, const char *name, const char *string,
1.66 schwarze 3536: int append)
1.7 schwarze 3537: {
1.187 schwarze 3538: size_t namesz;
1.42 schwarze 3539:
1.187 schwarze 3540: namesz = strlen(name);
3541: roff_setstrn(&r->strtab, name, namesz, string,
1.80 schwarze 3542: string ? strlen(string) : 0, append);
1.187 schwarze 3543: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.42 schwarze 3544: }
3545:
3546: static void
3547: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66 schwarze 3548: const char *string, size_t stringsz, int append)
1.42 schwarze 3549: {
3550: struct roffkv *n;
3551: char *c;
3552: int i;
3553: size_t oldch, newch;
1.7 schwarze 3554:
1.16 schwarze 3555: /* Search for an existing string with the same name. */
1.42 schwarze 3556: n = *r;
3557:
1.84 schwarze 3558: while (n && (namesz != n->key.sz ||
3559: strncmp(n->key.p, name, namesz)))
1.7 schwarze 3560: n = n->next;
1.8 schwarze 3561:
3562: if (NULL == n) {
1.16 schwarze 3563: /* Create a new string table entry. */
1.42 schwarze 3564: n = mandoc_malloc(sizeof(struct roffkv));
3565: n->key.p = mandoc_strndup(name, namesz);
3566: n->key.sz = namesz;
3567: n->val.p = NULL;
3568: n->val.sz = 0;
3569: n->next = *r;
3570: *r = n;
1.66 schwarze 3571: } else if (0 == append) {
1.42 schwarze 3572: free(n->val.p);
3573: n->val.p = NULL;
3574: n->val.sz = 0;
1.16 schwarze 3575: }
3576:
3577: if (NULL == string)
3578: return;
3579:
3580: /*
3581: * One additional byte for the '\n' in multiline mode,
3582: * and one for the terminating '\0'.
3583: */
1.66 schwarze 3584: newch = stringsz + (1 < append ? 2u : 1u);
1.42 schwarze 3585:
3586: if (NULL == n->val.p) {
3587: n->val.p = mandoc_malloc(newch);
3588: *n->val.p = '\0';
1.16 schwarze 3589: oldch = 0;
3590: } else {
1.42 schwarze 3591: oldch = n->val.sz;
3592: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16 schwarze 3593: }
3594:
3595: /* Skip existing content in the destination buffer. */
1.42 schwarze 3596: c = n->val.p + (int)oldch;
1.16 schwarze 3597:
3598: /* Append new content to the destination buffer. */
1.42 schwarze 3599: i = 0;
3600: while (i < (int)stringsz) {
1.16 schwarze 3601: /*
3602: * Rudimentary roff copy mode:
3603: * Handle escaped backslashes.
3604: */
1.42 schwarze 3605: if ('\\' == string[i] && '\\' == string[i + 1])
3606: i++;
3607: *c++ = string[i++];
1.16 schwarze 3608: }
1.8 schwarze 3609:
1.16 schwarze 3610: /* Append terminating bytes. */
1.66 schwarze 3611: if (1 < append)
1.16 schwarze 3612: *c++ = '\n';
1.42 schwarze 3613:
1.16 schwarze 3614: *c = '\0';
1.42 schwarze 3615: n->val.sz = (int)(c - n->val.p);
1.7 schwarze 3616: }
3617:
1.8 schwarze 3618: static const char *
1.197 schwarze 3619: roff_getstrn(struct roff *r, const char *name, size_t len,
1.187 schwarze 3620: int *deftype)
1.7 schwarze 3621: {
1.187 schwarze 3622: const struct roffkv *n;
1.197 schwarze 3623: int found, i;
1.187 schwarze 3624: enum roff_tok tok;
3625:
1.197 schwarze 3626: found = 0;
3627: for (n = r->strtab; n != NULL; n = n->next) {
3628: if (strncmp(name, n->key.p, len) != 0 ||
3629: n->key.p[len] != '\0' || n->val.p == NULL)
3630: continue;
3631: if (*deftype & ROFFDEF_USER) {
3632: *deftype = ROFFDEF_USER;
3633: return n->val.p;
3634: } else {
3635: found = 1;
3636: break;
3637: }
3638: }
3639: for (n = r->rentab; n != NULL; n = n->next) {
3640: if (strncmp(name, n->key.p, len) != 0 ||
3641: n->key.p[len] != '\0' || n->val.p == NULL)
3642: continue;
3643: if (*deftype & ROFFDEF_REN) {
3644: *deftype = ROFFDEF_REN;
3645: return n->val.p;
3646: } else {
3647: found = 1;
3648: break;
1.187 schwarze 3649: }
3650: }
1.197 schwarze 3651: for (i = 0; i < PREDEFS_MAX; i++) {
3652: if (strncmp(name, predefs[i].name, len) != 0 ||
3653: predefs[i].name[len] != '\0')
3654: continue;
3655: if (*deftype & ROFFDEF_PRE) {
3656: *deftype = ROFFDEF_PRE;
3657: return predefs[i].str;
3658: } else {
3659: found = 1;
3660: break;
1.187 schwarze 3661: }
3662: }
1.197 schwarze 3663: if (r->man->macroset != MACROSET_MAN) {
3664: for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
3665: if (strncmp(name, roff_name[tok], len) != 0 ||
3666: roff_name[tok][len] != '\0')
3667: continue;
3668: if (*deftype & ROFFDEF_STD) {
3669: *deftype = ROFFDEF_STD;
3670: return NULL;
3671: } else {
3672: found = 1;
3673: break;
1.187 schwarze 3674: }
3675: }
3676: }
1.197 schwarze 3677: if (r->man->macroset != MACROSET_MDOC) {
3678: for (tok = MAN_TH; tok < MAN_MAX; tok++) {
3679: if (strncmp(name, roff_name[tok], len) != 0 ||
3680: roff_name[tok][len] != '\0')
3681: continue;
3682: if (*deftype & ROFFDEF_STD) {
3683: *deftype = ROFFDEF_STD;
3684: return NULL;
3685: } else {
3686: found = 1;
3687: break;
1.187 schwarze 3688: }
3689: }
1.197 schwarze 3690: }
3691:
3692: if (found == 0 && *deftype != ROFFDEF_ANY) {
3693: if (*deftype & ROFFDEF_REN) {
3694: /*
3695: * This might still be a request,
3696: * so do not treat it as undefined yet.
3697: */
3698: *deftype = ROFFDEF_UNDEF;
3699: return NULL;
1.187 schwarze 3700: }
1.197 schwarze 3701:
3702: /* Using an undefined string defines it to be empty. */
3703:
3704: roff_setstrn(&r->strtab, name, len, "", 0, 0);
3705: roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.187 schwarze 3706: }
1.197 schwarze 3707:
1.187 schwarze 3708: *deftype = 0;
1.149 schwarze 3709: return NULL;
1.7 schwarze 3710: }
3711:
1.8 schwarze 3712: static void
1.42 schwarze 3713: roff_freestr(struct roffkv *r)
1.7 schwarze 3714: {
1.42 schwarze 3715: struct roffkv *n, *nn;
1.7 schwarze 3716:
1.42 schwarze 3717: for (n = r; n; n = nn) {
3718: free(n->key.p);
3719: free(n->val.p);
1.7 schwarze 3720: nn = n->next;
3721: free(n);
3722: }
1.27 schwarze 3723: }
1.138 schwarze 3724:
3725: /* --- accessors and utility functions ------------------------------------ */
1.42 schwarze 3726:
3727: /*
3728: * Duplicate an input string, making the appropriate character
3729: * conversations (as stipulated by `tr') along the way.
3730: * Returns a heap-allocated string with all the replacements made.
3731: */
3732: char *
3733: roff_strdup(const struct roff *r, const char *p)
3734: {
3735: const struct roffkv *cp;
3736: char *res;
3737: const char *pp;
3738: size_t ssz, sz;
3739: enum mandoc_esc esc;
3740:
3741: if (NULL == r->xmbtab && NULL == r->xtab)
1.149 schwarze 3742: return mandoc_strdup(p);
1.42 schwarze 3743: else if ('\0' == *p)
1.149 schwarze 3744: return mandoc_strdup("");
1.42 schwarze 3745:
3746: /*
3747: * Step through each character looking for term matches
3748: * (remember that a `tr' can be invoked with an escape, which is
3749: * a glyph but the escape is multi-character).
3750: * We only do this if the character hash has been initialised
3751: * and the string is >0 length.
3752: */
3753:
3754: res = NULL;
3755: ssz = 0;
3756:
3757: while ('\0' != *p) {
1.161 schwarze 3758: assert((unsigned int)*p < 128);
3759: if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.42 schwarze 3760: sz = r->xtab[(int)*p].sz;
3761: res = mandoc_realloc(res, ssz + sz + 1);
3762: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3763: ssz += sz;
3764: p++;
3765: continue;
3766: } else if ('\\' != *p) {
3767: res = mandoc_realloc(res, ssz + 2);
3768: res[ssz++] = *p++;
3769: continue;
3770: }
3771:
3772: /* Search for term matches. */
3773: for (cp = r->xmbtab; cp; cp = cp->next)
3774: if (0 == strncmp(p, cp->key.p, cp->key.sz))
3775: break;
3776:
3777: if (NULL != cp) {
3778: /*
3779: * A match has been found.
3780: * Append the match to the array and move
3781: * forward by its keysize.
3782: */
1.80 schwarze 3783: res = mandoc_realloc(res,
3784: ssz + cp->val.sz + 1);
1.42 schwarze 3785: memcpy(res + ssz, cp->val.p, cp->val.sz);
3786: ssz += cp->val.sz;
3787: p += (int)cp->key.sz;
3788: continue;
3789: }
3790:
3791: /*
3792: * Handle escapes carefully: we need to copy
3793: * over just the escape itself, or else we might
3794: * do replacements within the escape itself.
3795: * Make sure to pass along the bogus string.
3796: */
3797: pp = p++;
3798: esc = mandoc_escape(&p, NULL, NULL);
3799: if (ESCAPE_ERROR == esc) {
3800: sz = strlen(pp);
3801: res = mandoc_realloc(res, ssz + sz + 1);
3802: memcpy(res + ssz, pp, sz);
3803: break;
3804: }
1.80 schwarze 3805: /*
3806: * We bail out on bad escapes.
1.42 schwarze 3807: * No need to warn: we already did so when
3808: * roff_res() was called.
3809: */
3810: sz = (int)(p - pp);
3811: res = mandoc_realloc(res, ssz + sz + 1);
3812: memcpy(res + ssz, pp, sz);
3813: ssz += sz;
3814: }
3815:
3816: res[(int)ssz] = '\0';
1.149 schwarze 3817: return res;
1.99 schwarze 3818: }
3819:
3820: int
3821: roff_getformat(const struct roff *r)
3822: {
3823:
1.149 schwarze 3824: return r->format;
1.48 schwarze 3825: }
3826:
3827: /*
1.80 schwarze 3828: * Find out whether a line is a macro line or not.
1.48 schwarze 3829: * If it is, adjust the current position and return one; if it isn't,
3830: * return zero and don't change the current position.
3831: * If the control character has been set with `.cc', then let that grain
3832: * precedence.
3833: * This is slighly contrary to groff, where using the non-breaking
3834: * control character when `cc' has been invoked will cause the
3835: * non-breaking macro contents to be printed verbatim.
3836: */
3837: int
3838: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3839: {
3840: int pos;
3841:
3842: pos = *ppos;
3843:
1.175 schwarze 3844: if (r->control != '\0' && cp[pos] == r->control)
1.48 schwarze 3845: pos++;
1.175 schwarze 3846: else if (r->control != '\0')
1.149 schwarze 3847: return 0;
1.48 schwarze 3848: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3849: pos += 2;
3850: else if ('.' == cp[pos] || '\'' == cp[pos])
3851: pos++;
3852: else
1.149 schwarze 3853: return 0;
1.48 schwarze 3854:
3855: while (' ' == cp[pos] || '\t' == cp[pos])
3856: pos++;
3857:
3858: *ppos = pos;
1.149 schwarze 3859: return 1;
1.1 schwarze 3860: }
![[BACK]](/icons/back.gif){kind=icon}
Return to roff.c CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [local] / src / usr.bin / mandoc