Annotation of mandoc/roff.c, Revision 1.337
1.337 ! schwarze 1: /* $Id: roff.c,v 1.336 2018/08/19 17:46:14 schwarze Exp $ */
1.1 kristaps 2: /*
1.267 schwarze 3: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.325 schwarze 4: * Copyright (c) 2010-2015, 2017, 2018 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.66 kristaps 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.106 kristaps 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.66 kristaps 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106 kristaps 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.66 kristaps 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.
1.1 kristaps 17: */
1.66 kristaps 18: #include "config.h"
1.225 schwarze 19:
20: #include <sys/types.h>
1.30 kristaps 21:
1.67 kristaps 22: #include <assert.h>
1.85 kristaps 23: #include <ctype.h>
1.245 schwarze 24: #include <limits.h>
1.295 schwarze 25: #include <stddef.h>
26: #include <stdint.h>
1.178 schwarze 27: #include <stdio.h>
1.1 kristaps 28: #include <stdlib.h>
1.67 kristaps 29: #include <string.h>
1.1 kristaps 30:
1.67 kristaps 31: #include "mandoc.h"
1.201 schwarze 32: #include "mandoc_aux.h"
1.295 schwarze 33: #include "mandoc_ohash.h"
1.265 schwarze 34: #include "roff.h"
1.226 schwarze 35: #include "libmandoc.h"
1.266 schwarze 36: #include "roff_int.h"
1.109 kristaps 37: #include "libroff.h"
1.82 kristaps 38:
1.170 schwarze 39: /* Maximum number of string expansions per line, to break infinite loops. */
40: #define EXPAND_LIMIT 1000
41:
1.315 schwarze 42: /* Types of definitions of macros and strings. */
43: #define ROFFDEF_USER (1 << 1) /* User-defined. */
44: #define ROFFDEF_PRE (1 << 2) /* Predefined. */
45: #define ROFFDEF_REN (1 << 3) /* Renamed standard macro. */
46: #define ROFFDEF_STD (1 << 4) /* mdoc(7) or man(7) macro. */
47: #define ROFFDEF_ANY (ROFFDEF_USER | ROFFDEF_PRE | \
48: ROFFDEF_REN | ROFFDEF_STD)
1.325 schwarze 49: #define ROFFDEF_UNDEF (1 << 5) /* Completely undefined. */
1.315 schwarze 50:
1.266 schwarze 51: /* --- data types --------------------------------------------------------- */
52:
1.147 kristaps 53: /*
1.167 kristaps 54: * An incredibly-simple string buffer.
55: */
1.94 kristaps 56: struct roffstr {
1.167 kristaps 57: char *p; /* nil-terminated buffer */
58: size_t sz; /* saved strlen(p) */
1.166 kristaps 59: };
60:
61: /*
1.167 kristaps 62: * A key-value roffstr pair as part of a singly-linked list.
1.166 kristaps 63: */
64: struct roffkv {
65: struct roffstr key;
66: struct roffstr val;
67: struct roffkv *next; /* next in list */
1.94 kristaps 68: };
69:
1.180 schwarze 70: /*
71: * A single number register as part of a singly-linked list.
72: */
73: struct roffreg {
74: struct roffstr key;
1.181 schwarze 75: int val;
1.327 schwarze 76: int step;
1.180 schwarze 77: struct roffreg *next;
78: };
79:
1.295 schwarze 80: /*
81: * Association of request and macro names with token IDs.
82: */
83: struct roffreq {
84: enum roff_tok tok;
85: char name[];
86: };
87:
1.67 kristaps 88: struct roff {
1.128 kristaps 89: struct mparse *parse; /* parse point */
1.296 schwarze 90: struct roff_man *man; /* mdoc or man parser */
1.67 kristaps 91: struct roffnode *last; /* leaf of stack */
1.223 schwarze 92: int *rstack; /* stack of inverted `ie' values */
1.295 schwarze 93: struct ohash *reqtab; /* request lookup table */
1.180 schwarze 94: struct roffreg *regtab; /* number registers */
1.166 kristaps 95: struct roffkv *strtab; /* user-defined strings & macros */
1.306 schwarze 96: struct roffkv *rentab; /* renamed strings & macros */
1.167 kristaps 97: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
98: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.106 kristaps 99: const char *current_string; /* value of last called user macro */
1.118 kristaps 100: struct tbl_node *first_tbl; /* first table parsed */
101: struct tbl_node *last_tbl; /* last table parsed */
102: struct tbl_node *tbl; /* current table being parsed */
1.319 schwarze 103: struct eqn_node *last_eqn; /* equation parser */
104: struct eqn_node *eqn; /* active equation parser */
1.230 schwarze 105: int eqn_inline; /* current equation is inline */
1.223 schwarze 106: int options; /* parse options */
107: int rstacksz; /* current size limit of rstack */
108: int rstackpos; /* position in rstack */
1.227 schwarze 109: int format; /* current file in mdoc or man format */
1.223 schwarze 110: char control; /* control character */
1.303 schwarze 111: char escape; /* escape character */
1.79 kristaps 112: };
113:
1.67 kristaps 114: struct roffnode {
1.294 schwarze 115: enum roff_tok tok; /* type of node */
1.67 kristaps 116: struct roffnode *parent; /* up one in stack */
117: int line; /* parse line */
118: int col; /* parse col */
1.106 kristaps 119: char *name; /* node name, e.g. macro name */
1.79 kristaps 120: char *end; /* end-rules: custom token */
121: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 122: int rule; /* current evaluation rule */
1.67 kristaps 123: };
124:
125: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.294 schwarze 126: enum roff_tok tok, /* tok of macro */ \
1.238 schwarze 127: struct buf *buf, /* input buffer */ \
1.67 kristaps 128: int ln, /* parse line */ \
1.75 kristaps 129: int ppos, /* original pos in buffer */ \
130: int pos, /* current pos in buffer */ \
1.74 kristaps 131: int *offs /* reset offset of buffer data */
1.67 kristaps 132:
133: typedef enum rofferr (*roffproc)(ROFF_ARGS);
134:
135: struct roffmac {
1.79 kristaps 136: roffproc proc; /* process new macro */
137: roffproc text; /* process as child text of macro */
138: roffproc sub; /* process as child of macro */
139: int flags;
140: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.67 kristaps 141: };
142:
1.141 kristaps 143: struct predef {
144: const char *name; /* predefined input name */
145: const char *str; /* replacement symbol */
146: };
147:
148: #define PREDEF(__name, __str) \
149: { (__name), (__str) },
150:
1.266 schwarze 151: /* --- function prototypes ------------------------------------------------ */
152:
1.155 kristaps 153: static void roffnode_cleanscope(struct roff *);
154: static void roffnode_pop(struct roff *);
1.294 schwarze 155: static void roffnode_push(struct roff *, enum roff_tok,
1.155 kristaps 156: const char *, int, int);
1.321 schwarze 157: static void roff_addtbl(struct roff_man *, struct tbl_node *);
1.311 schwarze 158: static enum rofferr roff_als(ROFF_ARGS);
1.80 kristaps 159: static enum rofferr roff_block(ROFF_ARGS);
160: static enum rofferr roff_block_text(ROFF_ARGS);
161: static enum rofferr roff_block_sub(ROFF_ARGS);
1.296 schwarze 162: static enum rofferr roff_br(ROFF_ARGS);
1.80 kristaps 163: static enum rofferr roff_cblock(ROFF_ARGS);
1.174 kristaps 164: static enum rofferr roff_cc(ROFF_ARGS);
1.195 schwarze 165: static void roff_ccond(struct roff *, int, int);
1.82 kristaps 166: static enum rofferr roff_cond(ROFF_ARGS);
167: static enum rofferr roff_cond_text(ROFF_ARGS);
168: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.92 schwarze 169: static enum rofferr roff_ds(ROFF_ARGS);
1.303 schwarze 170: static enum rofferr roff_ec(ROFF_ARGS);
171: static enum rofferr roff_eo(ROFF_ARGS);
1.238 schwarze 172: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.271 schwarze 173: static int roff_evalcond(struct roff *r, int, char *, int *);
1.234 kristaps 174: static int roff_evalnum(struct roff *, int,
175: const char *, int *, int *, int);
176: static int roff_evalpar(struct roff *, int,
1.261 schwarze 177: const char *, int *, int *, int);
1.198 schwarze 178: static int roff_evalstrcond(const char *, int *);
1.155 kristaps 179: static void roff_free1(struct roff *);
1.180 schwarze 180: static void roff_freereg(struct roffreg *);
1.167 kristaps 181: static void roff_freestr(struct roffkv *);
1.212 schwarze 182: static size_t roff_getname(struct roff *, char **, int, int);
1.261 schwarze 183: static int roff_getnum(const char *, int *, int *, int);
1.184 schwarze 184: static int roff_getop(const char *, int *, char *);
1.327 schwarze 185: static int roff_getregn(struct roff *,
186: const char *, size_t, char);
1.273 schwarze 187: static int roff_getregro(const struct roff *,
188: const char *name);
1.325 schwarze 189: static const char *roff_getstrn(struct roff *,
1.315 schwarze 190: const char *, size_t, int *);
1.271 schwarze 191: static int roff_hasregn(const struct roff *,
192: const char *, size_t);
1.251 schwarze 193: static enum rofferr roff_insec(ROFF_ARGS);
1.178 schwarze 194: static enum rofferr roff_it(ROFF_ARGS);
1.103 kristaps 195: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.265 schwarze 196: static void roff_man_alloc1(struct roff_man *);
197: static void roff_man_free1(struct roff_man *);
1.300 schwarze 198: static enum rofferr roff_manyarg(ROFF_ARGS);
1.330 schwarze 199: static enum rofferr roff_nop(ROFF_ARGS);
1.89 kristaps 200: static enum rofferr roff_nr(ROFF_ARGS);
1.297 schwarze 201: static enum rofferr roff_onearg(ROFF_ARGS);
1.294 schwarze 202: static enum roff_tok roff_parse(struct roff *, char *, int *,
1.214 schwarze 203: int, int);
1.305 schwarze 204: static enum rofferr roff_parsetext(struct roff *, struct buf *,
205: int, int *);
1.306 schwarze 206: static enum rofferr roff_renamed(ROFF_ARGS);
1.238 schwarze 207: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.122 schwarze 208: static enum rofferr roff_rm(ROFF_ARGS);
1.306 schwarze 209: static enum rofferr roff_rn(ROFF_ARGS);
1.203 schwarze 210: static enum rofferr roff_rr(ROFF_ARGS);
1.326 schwarze 211: static void roff_setregn(struct roff *, const char *,
1.327 schwarze 212: size_t, int, char, int);
1.94 kristaps 213: static void roff_setstr(struct roff *,
1.106 kristaps 214: const char *, const char *, int);
1.207 schwarze 215: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 216: size_t, const char *, size_t, int);
1.105 kristaps 217: static enum rofferr roff_so(ROFF_ARGS);
1.164 kristaps 218: static enum rofferr roff_tr(ROFF_ARGS);
1.175 schwarze 219: static enum rofferr roff_Dd(ROFF_ARGS);
1.109 kristaps 220: static enum rofferr roff_TE(ROFF_ARGS);
221: static enum rofferr roff_TS(ROFF_ARGS);
1.125 kristaps 222: static enum rofferr roff_EQ(ROFF_ARGS);
223: static enum rofferr roff_EN(ROFF_ARGS);
1.112 kristaps 224: static enum rofferr roff_T_(ROFF_ARGS);
1.251 schwarze 225: static enum rofferr roff_unsupp(ROFF_ARGS);
1.106 kristaps 226: static enum rofferr roff_userdef(ROFF_ARGS);
1.67 kristaps 227:
1.266 schwarze 228: /* --- constant data ------------------------------------------------------ */
229:
1.261 schwarze 230: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
231: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
232:
1.294 schwarze 233: const char *__roff_name[MAN_MAX + 1] = {
1.305 schwarze 234: "br", "ce", "ft", "ll",
1.312 schwarze 235: "mc", "po", "rj", "sp",
236: "ta", "ti", NULL,
1.294 schwarze 237: "ab", "ad", "af", "aln",
238: "als", "am", "am1", "ami",
239: "ami1", "as", "as1", "asciify",
240: "backtrace", "bd", "bleedat", "blm",
241: "box", "boxa", "bp", "BP",
242: "break", "breakchar", "brnl", "brp",
1.305 schwarze 243: "brpnl", "c2", "cc",
1.294 schwarze 244: "cf", "cflags", "ch", "char",
245: "chop", "class", "close", "CL",
246: "color", "composite", "continue", "cp",
247: "cropat", "cs", "cu", "da",
248: "dch", "Dd", "de", "de1",
249: "defcolor", "dei", "dei1", "device",
250: "devicem", "di", "do", "ds",
251: "ds1", "dwh", "dt", "ec",
252: "ecr", "ecs", "el", "em",
253: "EN", "eo", "EP", "EQ",
254: "errprint", "ev", "evc", "ex",
255: "fallback", "fam", "fc", "fchar",
256: "fcolor", "fdeferlig", "feature", "fkern",
257: "fl", "flig", "fp", "fps",
258: "fschar", "fspacewidth", "fspecial", "ftr",
259: "fzoom", "gcolor", "hc", "hcode",
260: "hidechar", "hla", "hlm", "hpf",
261: "hpfa", "hpfcode", "hw", "hy",
262: "hylang", "hylen", "hym", "hypp",
263: "hys", "ie", "if", "ig",
264: "index", "it", "itc", "IX",
265: "kern", "kernafter", "kernbefore", "kernpair",
266: "lc", "lc_ctype", "lds", "length",
267: "letadj", "lf", "lg", "lhang",
268: "linetabs", "lnr", "lnrf", "lpfx",
1.304 schwarze 269: "ls", "lsm", "lt",
1.294 schwarze 270: "mediasize", "minss", "mk", "mso",
271: "na", "ne", "nh", "nhychar",
272: "nm", "nn", "nop", "nr",
273: "nrf", "nroff", "ns", "nx",
274: "open", "opena", "os", "output",
275: "padj", "papersize", "pc", "pev",
276: "pi", "PI", "pl", "pm",
1.312 schwarze 277: "pn", "pnr", "ps",
1.294 schwarze 278: "psbb", "pshape", "pso", "ptr",
279: "pvs", "rchar", "rd", "recursionlimit",
1.309 schwarze 280: "return", "rfschar", "rhang",
1.294 schwarze 281: "rm", "rn", "rnn", "rr",
282: "rs", "rt", "schar", "sentchar",
283: "shc", "shift", "sizes", "so",
284: "spacewidth", "special", "spreadwarn", "ss",
285: "sty", "substring", "sv", "sy",
1.300 schwarze 286: "T&", "tc", "TE",
1.301 schwarze 287: "TH", "tkf", "tl",
1.294 schwarze 288: "tm", "tm1", "tmc", "tr",
289: "track", "transchar", "trf", "trimat",
290: "trin", "trnt", "troff", "TS",
291: "uf", "ul", "unformat", "unwatch",
292: "unwatchn", "vpt", "vs", "warn",
293: "warnscale", "watch", "watchlength", "watchn",
294: "wh", "while", "write", "writec",
295: "writem", "xflag", ".", NULL,
1.306 schwarze 296: NULL, "text",
1.294 schwarze 297: "Dd", "Dt", "Os", "Sh",
298: "Ss", "Pp", "D1", "Dl",
299: "Bd", "Ed", "Bl", "El",
300: "It", "Ad", "An", "Ap",
301: "Ar", "Cd", "Cm", "Dv",
302: "Er", "Ev", "Ex", "Fa",
303: "Fd", "Fl", "Fn", "Ft",
304: "Ic", "In", "Li", "Nd",
305: "Nm", "Op", "Ot", "Pa",
306: "Rv", "St", "Va", "Vt",
307: "Xr", "%A", "%B", "%D",
308: "%I", "%J", "%N", "%O",
309: "%P", "%R", "%T", "%V",
310: "Ac", "Ao", "Aq", "At",
311: "Bc", "Bf", "Bo", "Bq",
312: "Bsx", "Bx", "Db", "Dc",
313: "Do", "Dq", "Ec", "Ef",
314: "Em", "Eo", "Fx", "Ms",
315: "No", "Ns", "Nx", "Ox",
316: "Pc", "Pf", "Po", "Pq",
317: "Qc", "Ql", "Qo", "Qq",
318: "Re", "Rs", "Sc", "So",
319: "Sq", "Sm", "Sx", "Sy",
320: "Tn", "Ux", "Xc", "Xo",
321: "Fo", "Fc", "Oo", "Oc",
322: "Bk", "Ek", "Bt", "Hf",
323: "Fr", "Ud", "Lb", "Lp",
324: "Lk", "Mt", "Brq", "Bro",
325: "Brc", "%C", "Es", "En",
1.299 schwarze 326: "Dx", "%Q", "%U", "Ta",
327: NULL,
1.294 schwarze 328: "TH", "SH", "SS", "TP",
1.332 schwarze 329: "TQ",
1.294 schwarze 330: "LP", "PP", "P", "IP",
331: "HP", "SM", "SB", "BI",
332: "IB", "BR", "RB", "R",
333: "B", "I", "IR", "RI",
1.299 schwarze 334: "nf", "fi",
1.294 schwarze 335: "RE", "RS", "DT", "UC",
1.297 schwarze 336: "PD", "AT", "in",
1.333 schwarze 337: "SY", "YS", "OP",
338: "EX", "EE", "UR",
1.317 schwarze 339: "UE", "MT", "ME", NULL
1.294 schwarze 340: };
341: const char *const *roff_name = __roff_name;
342:
343: static struct roffmac roffs[TOKEN_NONE] = {
1.296 schwarze 344: { roff_br, NULL, NULL, 0 }, /* br */
1.305 schwarze 345: { roff_onearg, NULL, NULL, 0 }, /* ce */
1.297 schwarze 346: { roff_onearg, NULL, NULL, 0 }, /* ft */
1.298 schwarze 347: { roff_onearg, NULL, NULL, 0 }, /* ll */
1.304 schwarze 348: { roff_onearg, NULL, NULL, 0 }, /* mc */
1.312 schwarze 349: { roff_onearg, NULL, NULL, 0 }, /* po */
1.309 schwarze 350: { roff_onearg, NULL, NULL, 0 }, /* rj */
1.299 schwarze 351: { roff_onearg, NULL, NULL, 0 }, /* sp */
1.300 schwarze 352: { roff_manyarg, NULL, NULL, 0 }, /* ta */
1.301 schwarze 353: { roff_onearg, NULL, NULL, 0 }, /* ti */
1.296 schwarze 354: { NULL, NULL, NULL, 0 }, /* ROFF_MAX */
1.295 schwarze 355: { roff_unsupp, NULL, NULL, 0 }, /* ab */
356: { roff_line_ignore, NULL, NULL, 0 }, /* ad */
357: { roff_line_ignore, NULL, NULL, 0 }, /* af */
358: { roff_unsupp, NULL, NULL, 0 }, /* aln */
1.311 schwarze 359: { roff_als, NULL, NULL, 0 }, /* als */
1.295 schwarze 360: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am */
361: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am1 */
362: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami */
363: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami1 */
364: { roff_ds, NULL, NULL, 0 }, /* as */
365: { roff_ds, NULL, NULL, 0 }, /* as1 */
366: { roff_unsupp, NULL, NULL, 0 }, /* asciify */
367: { roff_line_ignore, NULL, NULL, 0 }, /* backtrace */
368: { roff_line_ignore, NULL, NULL, 0 }, /* bd */
369: { roff_line_ignore, NULL, NULL, 0 }, /* bleedat */
370: { roff_unsupp, NULL, NULL, 0 }, /* blm */
371: { roff_unsupp, NULL, NULL, 0 }, /* box */
372: { roff_unsupp, NULL, NULL, 0 }, /* boxa */
373: { roff_line_ignore, NULL, NULL, 0 }, /* bp */
374: { roff_unsupp, NULL, NULL, 0 }, /* BP */
375: { roff_unsupp, NULL, NULL, 0 }, /* break */
376: { roff_line_ignore, NULL, NULL, 0 }, /* breakchar */
377: { roff_line_ignore, NULL, NULL, 0 }, /* brnl */
1.296 schwarze 378: { roff_br, NULL, NULL, 0 }, /* brp */
1.295 schwarze 379: { roff_line_ignore, NULL, NULL, 0 }, /* brpnl */
380: { roff_unsupp, NULL, NULL, 0 }, /* c2 */
381: { roff_cc, NULL, NULL, 0 }, /* cc */
382: { roff_insec, NULL, NULL, 0 }, /* cf */
383: { roff_line_ignore, NULL, NULL, 0 }, /* cflags */
384: { roff_line_ignore, NULL, NULL, 0 }, /* ch */
385: { roff_unsupp, NULL, NULL, 0 }, /* char */
386: { roff_unsupp, NULL, NULL, 0 }, /* chop */
387: { roff_line_ignore, NULL, NULL, 0 }, /* class */
388: { roff_insec, NULL, NULL, 0 }, /* close */
389: { roff_unsupp, NULL, NULL, 0 }, /* CL */
390: { roff_line_ignore, NULL, NULL, 0 }, /* color */
391: { roff_unsupp, NULL, NULL, 0 }, /* composite */
392: { roff_unsupp, NULL, NULL, 0 }, /* continue */
393: { roff_line_ignore, NULL, NULL, 0 }, /* cp */
394: { roff_line_ignore, NULL, NULL, 0 }, /* cropat */
395: { roff_line_ignore, NULL, NULL, 0 }, /* cs */
396: { roff_line_ignore, NULL, NULL, 0 }, /* cu */
397: { roff_unsupp, NULL, NULL, 0 }, /* da */
398: { roff_unsupp, NULL, NULL, 0 }, /* dch */
399: { roff_Dd, NULL, NULL, 0 }, /* Dd */
400: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de */
401: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de1 */
402: { roff_line_ignore, NULL, NULL, 0 }, /* defcolor */
403: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei */
404: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei1 */
405: { roff_unsupp, NULL, NULL, 0 }, /* device */
406: { roff_unsupp, NULL, NULL, 0 }, /* devicem */
407: { roff_unsupp, NULL, NULL, 0 }, /* di */
408: { roff_unsupp, NULL, NULL, 0 }, /* do */
409: { roff_ds, NULL, NULL, 0 }, /* ds */
410: { roff_ds, NULL, NULL, 0 }, /* ds1 */
411: { roff_unsupp, NULL, NULL, 0 }, /* dwh */
412: { roff_unsupp, NULL, NULL, 0 }, /* dt */
1.303 schwarze 413: { roff_ec, NULL, NULL, 0 }, /* ec */
1.295 schwarze 414: { roff_unsupp, NULL, NULL, 0 }, /* ecr */
415: { roff_unsupp, NULL, NULL, 0 }, /* ecs */
416: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* el */
417: { roff_unsupp, NULL, NULL, 0 }, /* em */
418: { roff_EN, NULL, NULL, 0 }, /* EN */
1.303 schwarze 419: { roff_eo, NULL, NULL, 0 }, /* eo */
1.295 schwarze 420: { roff_unsupp, NULL, NULL, 0 }, /* EP */
421: { roff_EQ, NULL, NULL, 0 }, /* EQ */
422: { roff_line_ignore, NULL, NULL, 0 }, /* errprint */
423: { roff_unsupp, NULL, NULL, 0 }, /* ev */
424: { roff_unsupp, NULL, NULL, 0 }, /* evc */
425: { roff_unsupp, NULL, NULL, 0 }, /* ex */
426: { roff_line_ignore, NULL, NULL, 0 }, /* fallback */
427: { roff_line_ignore, NULL, NULL, 0 }, /* fam */
428: { roff_unsupp, NULL, NULL, 0 }, /* fc */
429: { roff_unsupp, NULL, NULL, 0 }, /* fchar */
430: { roff_line_ignore, NULL, NULL, 0 }, /* fcolor */
431: { roff_line_ignore, NULL, NULL, 0 }, /* fdeferlig */
432: { roff_line_ignore, NULL, NULL, 0 }, /* feature */
433: { roff_line_ignore, NULL, NULL, 0 }, /* fkern */
434: { roff_line_ignore, NULL, NULL, 0 }, /* fl */
435: { roff_line_ignore, NULL, NULL, 0 }, /* flig */
436: { roff_line_ignore, NULL, NULL, 0 }, /* fp */
437: { roff_line_ignore, NULL, NULL, 0 }, /* fps */
438: { roff_unsupp, NULL, NULL, 0 }, /* fschar */
439: { roff_line_ignore, NULL, NULL, 0 }, /* fspacewidth */
440: { roff_line_ignore, NULL, NULL, 0 }, /* fspecial */
441: { roff_line_ignore, NULL, NULL, 0 }, /* ftr */
442: { roff_line_ignore, NULL, NULL, 0 }, /* fzoom */
443: { roff_line_ignore, NULL, NULL, 0 }, /* gcolor */
444: { roff_line_ignore, NULL, NULL, 0 }, /* hc */
445: { roff_line_ignore, NULL, NULL, 0 }, /* hcode */
446: { roff_line_ignore, NULL, NULL, 0 }, /* hidechar */
447: { roff_line_ignore, NULL, NULL, 0 }, /* hla */
448: { roff_line_ignore, NULL, NULL, 0 }, /* hlm */
449: { roff_line_ignore, NULL, NULL, 0 }, /* hpf */
450: { roff_line_ignore, NULL, NULL, 0 }, /* hpfa */
451: { roff_line_ignore, NULL, NULL, 0 }, /* hpfcode */
452: { roff_line_ignore, NULL, NULL, 0 }, /* hw */
453: { roff_line_ignore, NULL, NULL, 0 }, /* hy */
454: { roff_line_ignore, NULL, NULL, 0 }, /* hylang */
455: { roff_line_ignore, NULL, NULL, 0 }, /* hylen */
456: { roff_line_ignore, NULL, NULL, 0 }, /* hym */
457: { roff_line_ignore, NULL, NULL, 0 }, /* hypp */
458: { roff_line_ignore, NULL, NULL, 0 }, /* hys */
459: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* ie */
460: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* if */
461: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ig */
462: { roff_unsupp, NULL, NULL, 0 }, /* index */
463: { roff_it, NULL, NULL, 0 }, /* it */
464: { roff_unsupp, NULL, NULL, 0 }, /* itc */
465: { roff_line_ignore, NULL, NULL, 0 }, /* IX */
466: { roff_line_ignore, NULL, NULL, 0 }, /* kern */
467: { roff_line_ignore, NULL, NULL, 0 }, /* kernafter */
468: { roff_line_ignore, NULL, NULL, 0 }, /* kernbefore */
469: { roff_line_ignore, NULL, NULL, 0 }, /* kernpair */
470: { roff_unsupp, NULL, NULL, 0 }, /* lc */
471: { roff_unsupp, NULL, NULL, 0 }, /* lc_ctype */
472: { roff_unsupp, NULL, NULL, 0 }, /* lds */
473: { roff_unsupp, NULL, NULL, 0 }, /* length */
474: { roff_line_ignore, NULL, NULL, 0 }, /* letadj */
475: { roff_insec, NULL, NULL, 0 }, /* lf */
476: { roff_line_ignore, NULL, NULL, 0 }, /* lg */
477: { roff_line_ignore, NULL, NULL, 0 }, /* lhang */
478: { roff_unsupp, NULL, NULL, 0 }, /* linetabs */
479: { roff_unsupp, NULL, NULL, 0 }, /* lnr */
480: { roff_unsupp, NULL, NULL, 0 }, /* lnrf */
481: { roff_unsupp, NULL, NULL, 0 }, /* lpfx */
482: { roff_line_ignore, NULL, NULL, 0 }, /* ls */
483: { roff_unsupp, NULL, NULL, 0 }, /* lsm */
484: { roff_line_ignore, NULL, NULL, 0 }, /* lt */
485: { roff_line_ignore, NULL, NULL, 0 }, /* mediasize */
486: { roff_line_ignore, NULL, NULL, 0 }, /* minss */
487: { roff_line_ignore, NULL, NULL, 0 }, /* mk */
488: { roff_insec, NULL, NULL, 0 }, /* mso */
489: { roff_line_ignore, NULL, NULL, 0 }, /* na */
490: { roff_line_ignore, NULL, NULL, 0 }, /* ne */
491: { roff_line_ignore, NULL, NULL, 0 }, /* nh */
492: { roff_line_ignore, NULL, NULL, 0 }, /* nhychar */
493: { roff_unsupp, NULL, NULL, 0 }, /* nm */
494: { roff_unsupp, NULL, NULL, 0 }, /* nn */
1.330 schwarze 495: { roff_nop, NULL, NULL, 0 }, /* nop */
1.295 schwarze 496: { roff_nr, NULL, NULL, 0 }, /* nr */
497: { roff_unsupp, NULL, NULL, 0 }, /* nrf */
498: { roff_line_ignore, NULL, NULL, 0 }, /* nroff */
499: { roff_line_ignore, NULL, NULL, 0 }, /* ns */
500: { roff_insec, NULL, NULL, 0 }, /* nx */
501: { roff_insec, NULL, NULL, 0 }, /* open */
502: { roff_insec, NULL, NULL, 0 }, /* opena */
503: { roff_line_ignore, NULL, NULL, 0 }, /* os */
504: { roff_unsupp, NULL, NULL, 0 }, /* output */
505: { roff_line_ignore, NULL, NULL, 0 }, /* padj */
506: { roff_line_ignore, NULL, NULL, 0 }, /* papersize */
507: { roff_line_ignore, NULL, NULL, 0 }, /* pc */
508: { roff_line_ignore, NULL, NULL, 0 }, /* pev */
509: { roff_insec, NULL, NULL, 0 }, /* pi */
510: { roff_unsupp, NULL, NULL, 0 }, /* PI */
511: { roff_line_ignore, NULL, NULL, 0 }, /* pl */
512: { roff_line_ignore, NULL, NULL, 0 }, /* pm */
513: { roff_line_ignore, NULL, NULL, 0 }, /* pn */
514: { roff_line_ignore, NULL, NULL, 0 }, /* pnr */
515: { roff_line_ignore, NULL, NULL, 0 }, /* ps */
516: { roff_unsupp, NULL, NULL, 0 }, /* psbb */
517: { roff_unsupp, NULL, NULL, 0 }, /* pshape */
518: { roff_insec, NULL, NULL, 0 }, /* pso */
519: { roff_line_ignore, NULL, NULL, 0 }, /* ptr */
520: { roff_line_ignore, NULL, NULL, 0 }, /* pvs */
521: { roff_unsupp, NULL, NULL, 0 }, /* rchar */
522: { roff_line_ignore, NULL, NULL, 0 }, /* rd */
523: { roff_line_ignore, NULL, NULL, 0 }, /* recursionlimit */
524: { roff_unsupp, NULL, NULL, 0 }, /* return */
525: { roff_unsupp, NULL, NULL, 0 }, /* rfschar */
526: { roff_line_ignore, NULL, NULL, 0 }, /* rhang */
527: { roff_rm, NULL, NULL, 0 }, /* rm */
1.306 schwarze 528: { roff_rn, NULL, NULL, 0 }, /* rn */
1.295 schwarze 529: { roff_unsupp, NULL, NULL, 0 }, /* rnn */
530: { roff_rr, NULL, NULL, 0 }, /* rr */
531: { roff_line_ignore, NULL, NULL, 0 }, /* rs */
532: { roff_line_ignore, NULL, NULL, 0 }, /* rt */
533: { roff_unsupp, NULL, NULL, 0 }, /* schar */
534: { roff_line_ignore, NULL, NULL, 0 }, /* sentchar */
535: { roff_line_ignore, NULL, NULL, 0 }, /* shc */
536: { roff_unsupp, NULL, NULL, 0 }, /* shift */
537: { roff_line_ignore, NULL, NULL, 0 }, /* sizes */
538: { roff_so, NULL, NULL, 0 }, /* so */
539: { roff_line_ignore, NULL, NULL, 0 }, /* spacewidth */
540: { roff_line_ignore, NULL, NULL, 0 }, /* special */
541: { roff_line_ignore, NULL, NULL, 0 }, /* spreadwarn */
542: { roff_line_ignore, NULL, NULL, 0 }, /* ss */
543: { roff_line_ignore, NULL, NULL, 0 }, /* sty */
544: { roff_unsupp, NULL, NULL, 0 }, /* substring */
545: { roff_line_ignore, NULL, NULL, 0 }, /* sv */
546: { roff_insec, NULL, NULL, 0 }, /* sy */
547: { roff_T_, NULL, NULL, 0 }, /* T& */
548: { roff_unsupp, NULL, NULL, 0 }, /* tc */
549: { roff_TE, NULL, NULL, 0 }, /* TE */
1.315 schwarze 550: { roff_Dd, NULL, NULL, 0 }, /* TH */
1.295 schwarze 551: { roff_line_ignore, NULL, NULL, 0 }, /* tkf */
552: { roff_unsupp, NULL, NULL, 0 }, /* tl */
553: { roff_line_ignore, NULL, NULL, 0 }, /* tm */
554: { roff_line_ignore, NULL, NULL, 0 }, /* tm1 */
555: { roff_line_ignore, NULL, NULL, 0 }, /* tmc */
556: { roff_tr, NULL, NULL, 0 }, /* tr */
557: { roff_line_ignore, NULL, NULL, 0 }, /* track */
558: { roff_line_ignore, NULL, NULL, 0 }, /* transchar */
559: { roff_insec, NULL, NULL, 0 }, /* trf */
560: { roff_line_ignore, NULL, NULL, 0 }, /* trimat */
561: { roff_unsupp, NULL, NULL, 0 }, /* trin */
562: { roff_unsupp, NULL, NULL, 0 }, /* trnt */
563: { roff_line_ignore, NULL, NULL, 0 }, /* troff */
564: { roff_TS, NULL, NULL, 0 }, /* TS */
565: { roff_line_ignore, NULL, NULL, 0 }, /* uf */
566: { roff_line_ignore, NULL, NULL, 0 }, /* ul */
567: { roff_unsupp, NULL, NULL, 0 }, /* unformat */
568: { roff_line_ignore, NULL, NULL, 0 }, /* unwatch */
569: { roff_line_ignore, NULL, NULL, 0 }, /* unwatchn */
570: { roff_line_ignore, NULL, NULL, 0 }, /* vpt */
571: { roff_line_ignore, NULL, NULL, 0 }, /* vs */
572: { roff_line_ignore, NULL, NULL, 0 }, /* warn */
573: { roff_line_ignore, NULL, NULL, 0 }, /* warnscale */
574: { roff_line_ignore, NULL, NULL, 0 }, /* watch */
575: { roff_line_ignore, NULL, NULL, 0 }, /* watchlength */
576: { roff_line_ignore, NULL, NULL, 0 }, /* watchn */
577: { roff_unsupp, NULL, NULL, 0 }, /* wh */
578: { roff_unsupp, NULL, NULL, 0 }, /* while */
579: { roff_insec, NULL, NULL, 0 }, /* write */
580: { roff_insec, NULL, NULL, 0 }, /* writec */
581: { roff_insec, NULL, NULL, 0 }, /* writem */
582: { roff_line_ignore, NULL, NULL, 0 }, /* xflag */
583: { roff_cblock, NULL, NULL, 0 }, /* . */
1.306 schwarze 584: { roff_renamed, NULL, NULL, 0 },
1.295 schwarze 585: { roff_userdef, NULL, NULL, 0 }
1.67 kristaps 586: };
587:
1.141 kristaps 588: /* Array of injected predefined strings. */
589: #define PREDEFS_MAX 38
590: static const struct predef predefs[PREDEFS_MAX] = {
591: #include "predefs.in"
592: };
593:
1.305 schwarze 594: static int roffce_lines; /* number of input lines to center */
595: static struct roff_node *roffce_node; /* active request */
1.178 schwarze 596: static int roffit_lines; /* number of lines to delay */
597: static char *roffit_macro; /* nil-terminated macro line */
598:
1.207 schwarze 599:
1.266 schwarze 600: /* --- request table ------------------------------------------------------ */
601:
1.295 schwarze 602: struct ohash *
603: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85 kristaps 604: {
1.295 schwarze 605: struct ohash *htab;
606: struct roffreq *req;
607: enum roff_tok tok;
608: size_t sz;
609: unsigned int slot;
1.85 kristaps 610:
1.295 schwarze 611: htab = mandoc_malloc(sizeof(*htab));
612: mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85 kristaps 613:
1.295 schwarze 614: for (tok = mintok; tok < maxtok; tok++) {
1.296 schwarze 615: if (roff_name[tok] == NULL)
616: continue;
1.295 schwarze 617: sz = strlen(roff_name[tok]);
618: req = mandoc_malloc(sizeof(*req) + sz + 1);
619: req->tok = tok;
620: memcpy(req->name, roff_name[tok], sz + 1);
621: slot = ohash_qlookup(htab, req->name);
622: ohash_insert(htab, slot, req);
1.85 kristaps 623: }
1.295 schwarze 624: return htab;
1.85 kristaps 625: }
626:
1.295 schwarze 627: void
628: roffhash_free(struct ohash *htab)
1.67 kristaps 629: {
1.295 schwarze 630: struct roffreq *req;
631: unsigned int slot;
1.67 kristaps 632:
1.295 schwarze 633: if (htab == NULL)
634: return;
635: for (req = ohash_first(htab, &slot); req != NULL;
636: req = ohash_next(htab, &slot))
637: free(req);
638: ohash_delete(htab);
639: free(htab);
640: }
641:
642: enum roff_tok
643: roffhash_find(struct ohash *htab, const char *name, size_t sz)
644: {
645: struct roffreq *req;
646: const char *end;
647:
648: if (sz) {
649: end = name + sz;
650: req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
651: } else
652: req = ohash_find(htab, ohash_qlookup(htab, name));
653: return req == NULL ? TOKEN_NONE : req->tok;
1.67 kristaps 654: }
655:
1.266 schwarze 656: /* --- stack of request blocks -------------------------------------------- */
657:
1.67 kristaps 658: /*
659: * Pop the current node off of the stack of roff instructions currently
660: * pending.
661: */
662: static void
663: roffnode_pop(struct roff *r)
664: {
665: struct roffnode *p;
666:
1.75 kristaps 667: assert(r->last);
1.207 schwarze 668: p = r->last;
1.82 kristaps 669:
1.75 kristaps 670: r->last = r->last->parent;
1.106 kristaps 671: free(p->name);
672: free(p->end);
1.67 kristaps 673: free(p);
674: }
675:
676: /*
677: * Push a roff node onto the instruction stack. This must later be
678: * removed with roffnode_pop().
679: */
1.98 schwarze 680: static void
1.294 schwarze 681: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106 kristaps 682: int line, int col)
1.67 kristaps 683: {
684: struct roffnode *p;
685:
1.98 schwarze 686: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 687: p->tok = tok;
1.106 kristaps 688: if (name)
689: p->name = mandoc_strdup(name);
1.67 kristaps 690: p->parent = r->last;
691: p->line = line;
692: p->col = col;
1.198 schwarze 693: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 694:
695: r->last = p;
696: }
697:
1.266 schwarze 698: /* --- roff parser state data management ---------------------------------- */
699:
1.67 kristaps 700: static void
701: roff_free1(struct roff *r)
702: {
1.176 schwarze 703: struct tbl_node *tbl;
1.167 kristaps 704: int i;
1.67 kristaps 705:
1.176 schwarze 706: while (NULL != (tbl = r->first_tbl)) {
707: r->first_tbl = tbl->next;
708: tbl_free(tbl);
1.109 kristaps 709: }
1.113 kristaps 710: r->first_tbl = r->last_tbl = r->tbl = NULL;
711:
1.319 schwarze 712: if (r->last_eqn != NULL)
713: eqn_free(r->last_eqn);
714: r->last_eqn = r->eqn = NULL;
1.125 kristaps 715:
1.67 kristaps 716: while (r->last)
717: roffnode_pop(r);
1.109 kristaps 718:
1.223 schwarze 719: free (r->rstack);
720: r->rstack = NULL;
721: r->rstacksz = 0;
722: r->rstackpos = -1;
723:
724: roff_freereg(r->regtab);
725: r->regtab = NULL;
726:
1.167 kristaps 727: roff_freestr(r->strtab);
1.306 schwarze 728: roff_freestr(r->rentab);
1.167 kristaps 729: roff_freestr(r->xmbtab);
1.306 schwarze 730: r->strtab = r->rentab = r->xmbtab = NULL;
1.167 kristaps 731:
732: if (r->xtab)
733: for (i = 0; i < 128; i++)
734: free(r->xtab[i].p);
735: free(r->xtab);
736: r->xtab = NULL;
1.67 kristaps 737: }
738:
739: void
740: roff_reset(struct roff *r)
741: {
742: roff_free1(r);
1.227 schwarze 743: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.303 schwarze 744: r->control = '\0';
745: r->escape = '\\';
1.307 schwarze 746: roffce_lines = 0;
747: roffce_node = NULL;
748: roffit_lines = 0;
749: roffit_macro = NULL;
1.67 kristaps 750: }
751:
752: void
753: roff_free(struct roff *r)
754: {
755: roff_free1(r);
1.295 schwarze 756: roffhash_free(r->reqtab);
1.67 kristaps 757: free(r);
758: }
759:
760: struct roff *
1.279 schwarze 761: roff_alloc(struct mparse *parse, int options)
1.67 kristaps 762: {
763: struct roff *r;
764:
1.98 schwarze 765: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 766: r->parse = parse;
1.328 schwarze 767: r->reqtab = roffhash_alloc(0, ROFF_RENAMED);
1.199 schwarze 768: r->options = options;
1.227 schwarze 769: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82 kristaps 770: r->rstackpos = -1;
1.303 schwarze 771: r->escape = '\\';
1.277 schwarze 772: return r;
1.265 schwarze 773: }
774:
1.266 schwarze 775: /* --- syntax tree state data management ---------------------------------- */
776:
1.265 schwarze 777: static void
778: roff_man_free1(struct roff_man *man)
779: {
780:
1.266 schwarze 781: if (man->first != NULL)
782: roff_node_delete(man, man->first);
1.265 schwarze 783: free(man->meta.msec);
784: free(man->meta.vol);
785: free(man->meta.os);
786: free(man->meta.arch);
787: free(man->meta.title);
788: free(man->meta.name);
789: free(man->meta.date);
790: }
791:
792: static void
793: roff_man_alloc1(struct roff_man *man)
794: {
795:
796: memset(&man->meta, 0, sizeof(man->meta));
797: man->first = mandoc_calloc(1, sizeof(*man->first));
798: man->first->type = ROFFT_ROOT;
799: man->last = man->first;
800: man->last_es = NULL;
801: man->flags = 0;
802: man->macroset = MACROSET_NONE;
803: man->lastsec = man->lastnamed = SEC_NONE;
804: man->next = ROFF_NEXT_CHILD;
805: }
806:
807: void
808: roff_man_reset(struct roff_man *man)
809: {
810:
811: roff_man_free1(man);
812: roff_man_alloc1(man);
813: }
814:
815: void
816: roff_man_free(struct roff_man *man)
817: {
818:
819: roff_man_free1(man);
820: free(man);
821: }
822:
823: struct roff_man *
824: roff_man_alloc(struct roff *roff, struct mparse *parse,
1.316 schwarze 825: const char *os_s, int quick)
1.265 schwarze 826: {
827: struct roff_man *man;
828:
829: man = mandoc_calloc(1, sizeof(*man));
830: man->parse = parse;
831: man->roff = roff;
1.316 schwarze 832: man->os_s = os_s;
1.265 schwarze 833: man->quick = quick;
834: roff_man_alloc1(man);
1.296 schwarze 835: roff->man = man;
1.277 schwarze 836: return man;
1.67 kristaps 837: }
838:
1.266 schwarze 839: /* --- syntax tree handling ----------------------------------------------- */
840:
841: struct roff_node *
842: roff_node_alloc(struct roff_man *man, int line, int pos,
843: enum roff_type type, int tok)
844: {
845: struct roff_node *n;
846:
847: n = mandoc_calloc(1, sizeof(*n));
848: n->line = line;
849: n->pos = pos;
850: n->tok = tok;
851: n->type = type;
852: n->sec = man->lastsec;
853:
854: if (man->flags & MDOC_SYNOPSIS)
1.285 schwarze 855: n->flags |= NODE_SYNPRETTY;
1.266 schwarze 856: else
1.285 schwarze 857: n->flags &= ~NODE_SYNPRETTY;
1.266 schwarze 858: if (man->flags & MDOC_NEWLINE)
1.285 schwarze 859: n->flags |= NODE_LINE;
1.266 schwarze 860: man->flags &= ~MDOC_NEWLINE;
861:
1.277 schwarze 862: return n;
1.266 schwarze 863: }
864:
865: void
866: roff_node_append(struct roff_man *man, struct roff_node *n)
867: {
868:
869: switch (man->next) {
870: case ROFF_NEXT_SIBLING:
1.281 schwarze 871: if (man->last->next != NULL) {
872: n->next = man->last->next;
873: man->last->next->prev = n;
874: } else
875: man->last->parent->last = n;
1.266 schwarze 876: man->last->next = n;
877: n->prev = man->last;
878: n->parent = man->last->parent;
879: break;
880: case ROFF_NEXT_CHILD:
1.287 schwarze 881: if (man->last->child != NULL) {
882: n->next = man->last->child;
883: man->last->child->prev = n;
884: } else
885: man->last->last = n;
1.266 schwarze 886: man->last->child = n;
887: n->parent = man->last;
888: break;
889: default:
890: abort();
891: }
1.282 schwarze 892: man->last = n;
1.266 schwarze 893:
894: switch (n->type) {
895: case ROFFT_HEAD:
896: n->parent->head = n;
897: break;
898: case ROFFT_BODY:
1.282 schwarze 899: if (n->end != ENDBODY_NOT)
900: return;
1.266 schwarze 901: n->parent->body = n;
902: break;
903: case ROFFT_TAIL:
904: n->parent->tail = n;
905: break;
906: default:
1.282 schwarze 907: return;
1.266 schwarze 908: }
1.282 schwarze 909:
910: /*
911: * Copy over the normalised-data pointer of our parent. Not
912: * everybody has one, but copying a null pointer is fine.
913: */
914:
915: n->norm = n->parent->norm;
916: assert(n->parent->type == ROFFT_BLOCK);
1.266 schwarze 917: }
918:
1.267 schwarze 919: void
920: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
921: {
922: struct roff_node *n;
923:
924: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
925: n->string = roff_strdup(man->roff, word);
926: roff_node_append(man, n);
1.286 schwarze 927: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 928: man->next = ROFF_NEXT_SIBLING;
929: }
930:
931: void
932: roff_word_append(struct roff_man *man, const char *word)
933: {
934: struct roff_node *n;
935: char *addstr, *newstr;
936:
937: n = man->last;
938: addstr = roff_strdup(man->roff, word);
939: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
940: free(addstr);
941: free(n->string);
942: n->string = newstr;
943: man->next = ROFF_NEXT_SIBLING;
1.268 schwarze 944: }
945:
946: void
947: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
948: {
949: struct roff_node *n;
950:
951: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
952: roff_node_append(man, n);
953: man->next = ROFF_NEXT_CHILD;
954: }
955:
956: struct roff_node *
957: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
958: {
959: struct roff_node *n;
960:
961: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
962: roff_node_append(man, n);
963: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 964: return n;
1.267 schwarze 965: }
966:
1.266 schwarze 967: struct roff_node *
968: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
969: {
970: struct roff_node *n;
971:
972: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
973: roff_node_append(man, n);
974: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 975: return n;
1.266 schwarze 976: }
977:
978: struct roff_node *
979: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
980: {
981: struct roff_node *n;
982:
983: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
984: roff_node_append(man, n);
985: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 986: return n;
1.267 schwarze 987: }
988:
1.321 schwarze 989: static void
990: roff_addtbl(struct roff_man *man, struct tbl_node *tbl)
1.267 schwarze 991: {
992: struct roff_node *n;
1.321 schwarze 993: const struct tbl_span *span;
1.267 schwarze 994:
995: if (man->macroset == MACROSET_MAN)
1.302 schwarze 996: man_breakscope(man, ROFF_TS);
1.321 schwarze 997: while ((span = tbl_span(tbl)) != NULL) {
998: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
999: n->span = span;
1000: roff_node_append(man, n);
1001: n->flags |= NODE_VALID | NODE_ENDED;
1002: man->next = ROFF_NEXT_SIBLING;
1003: }
1.266 schwarze 1004: }
1005:
1006: void
1007: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1008: {
1009:
1010: /* Adjust siblings. */
1011:
1012: if (n->prev)
1013: n->prev->next = n->next;
1014: if (n->next)
1015: n->next->prev = n->prev;
1016:
1017: /* Adjust parent. */
1018:
1019: if (n->parent != NULL) {
1020: if (n->parent->child == n)
1021: n->parent->child = n->next;
1022: if (n->parent->last == n)
1023: n->parent->last = n->prev;
1024: }
1025:
1026: /* Adjust parse point. */
1027:
1028: if (man == NULL)
1029: return;
1030: if (man->last == n) {
1031: if (n->prev == NULL) {
1032: man->last = n->parent;
1033: man->next = ROFF_NEXT_CHILD;
1034: } else {
1035: man->last = n->prev;
1036: man->next = ROFF_NEXT_SIBLING;
1037: }
1038: }
1039: if (man->first == n)
1040: man->first = NULL;
1041: }
1042:
1043: void
1044: roff_node_free(struct roff_node *n)
1045: {
1046:
1047: if (n->args != NULL)
1048: mdoc_argv_free(n->args);
1049: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1050: free(n->norm);
1.319 schwarze 1051: if (n->eqn != NULL)
1052: eqn_box_free(n->eqn);
1.266 schwarze 1053: free(n->string);
1054: free(n);
1055: }
1056:
1057: void
1058: roff_node_delete(struct roff_man *man, struct roff_node *n)
1059: {
1060:
1061: while (n->child != NULL)
1062: roff_node_delete(man, n->child);
1063: roff_node_unlink(man, n);
1064: roff_node_free(n);
1.269 schwarze 1065: }
1066:
1067: void
1068: deroff(char **dest, const struct roff_node *n)
1069: {
1070: char *cp;
1071: size_t sz;
1072:
1073: if (n->type != ROFFT_TEXT) {
1074: for (n = n->child; n != NULL; n = n->next)
1075: deroff(dest, n);
1076: return;
1077: }
1078:
1.288 schwarze 1079: /* Skip leading whitespace. */
1.269 schwarze 1080:
1.288 schwarze 1081: for (cp = n->string; *cp != '\0'; cp++) {
1.289 schwarze 1082: if (cp[0] == '\\' && cp[1] != '\0' &&
1083: strchr(" %&0^|~", cp[1]) != NULL)
1.269 schwarze 1084: cp++;
1.288 schwarze 1085: else if ( ! isspace((unsigned char)*cp))
1.269 schwarze 1086: break;
1087: }
1088:
1.289 schwarze 1089: /* Skip trailing backslash. */
1090:
1091: sz = strlen(cp);
1.290 schwarze 1092: if (sz > 0 && cp[sz - 1] == '\\')
1.289 schwarze 1093: sz--;
1094:
1.269 schwarze 1095: /* Skip trailing whitespace. */
1096:
1.289 schwarze 1097: for (; sz; sz--)
1.269 schwarze 1098: if ( ! isspace((unsigned char)cp[sz-1]))
1099: break;
1100:
1101: /* Skip empty strings. */
1102:
1103: if (sz == 0)
1104: return;
1105:
1106: if (*dest == NULL) {
1107: *dest = mandoc_strndup(cp, sz);
1108: return;
1109: }
1110:
1111: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1112: free(*dest);
1113: *dest = cp;
1.266 schwarze 1114: }
1115:
1116: /* --- main functions of the roff parser ---------------------------------- */
1117:
1.94 kristaps 1118: /*
1.206 schwarze 1119: * In the current line, expand escape sequences that tend to get
1120: * used in numerical expressions and conditional requests.
1121: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 1122: */
1.172 schwarze 1123: static enum rofferr
1.238 schwarze 1124: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 1125: {
1.208 schwarze 1126: char ubuf[24]; /* buffer to print the number */
1.328 schwarze 1127: struct roff_node *n; /* used for header comments */
1.205 schwarze 1128: const char *start; /* start of the string to process */
1.209 schwarze 1129: char *stesc; /* start of an escape sequence ('\\') */
1.328 schwarze 1130: char *ep; /* end of comment string */
1.108 schwarze 1131: const char *stnam; /* start of the name, after "[(*" */
1132: const char *cp; /* end of the name, e.g. before ']' */
1133: const char *res; /* the string to be substituted */
1.238 schwarze 1134: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 1135: size_t maxl; /* expected length of the escape name */
1136: size_t naml; /* actual length of the escape name */
1.237 schwarze 1137: enum mandoc_esc esc; /* type of the escape sequence */
1138: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 1139: int expand_count; /* to avoid infinite loops */
1.206 schwarze 1140: int npos; /* position in numeric expression */
1.218 schwarze 1141: int arg_complete; /* argument not interrupted by eol */
1.303 schwarze 1142: int done; /* no more input available */
1.315 schwarze 1143: int deftype; /* type of definition to paste */
1.323 schwarze 1144: int rcsid; /* kind of RCS id seen */
1.327 schwarze 1145: char sign; /* increment number register */
1.206 schwarze 1146: char term; /* character terminating the escape */
1.94 kristaps 1147:
1.303 schwarze 1148: /* Search forward for comments. */
1149:
1150: done = 0;
1151: start = buf->buf + pos;
1152: for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
1153: if (stesc[0] != r->escape || stesc[1] == '\0')
1154: continue;
1155: stesc++;
1156: if (*stesc != '"' && *stesc != '#')
1157: continue;
1.313 schwarze 1158:
1159: /* Comment found, look for RCS id. */
1160:
1.323 schwarze 1161: rcsid = 0;
1.313 schwarze 1162: if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.323 schwarze 1163: rcsid = 1 << MANDOC_OS_OPENBSD;
1.313 schwarze 1164: cp += 8;
1165: } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.323 schwarze 1166: rcsid = 1 << MANDOC_OS_NETBSD;
1.313 schwarze 1167: cp += 7;
1168: }
1169: if (cp != NULL &&
1170: isalnum((unsigned char)*cp) == 0 &&
1.314 schwarze 1171: strchr(cp, '$') != NULL) {
1.323 schwarze 1172: if (r->man->meta.rcsids & rcsid)
1.314 schwarze 1173: mandoc_msg(MANDOCERR_RCS_REP, r->parse,
1174: ln, stesc + 1 - buf->buf, stesc + 1);
1.323 schwarze 1175: r->man->meta.rcsids |= rcsid;
1.314 schwarze 1176: }
1.313 schwarze 1177:
1178: /* Handle trailing whitespace. */
1179:
1.328 schwarze 1180: ep = strchr(stesc--, '\0') - 1;
1181: if (*ep == '\n') {
1.303 schwarze 1182: done = 1;
1.328 schwarze 1183: ep--;
1.303 schwarze 1184: }
1.328 schwarze 1185: if (*ep == ' ' || *ep == '\t')
1.303 schwarze 1186: mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
1.328 schwarze 1187: ln, ep - buf->buf, NULL);
1188:
1189: /*
1190: * Save comments preceding the title macro
1191: * in the syntax tree.
1192: */
1193:
1194: if (r->format == 0) {
1195: while (*ep == ' ' || *ep == '\t')
1196: ep--;
1197: ep[1] = '\0';
1198: n = roff_node_alloc(r->man,
1199: ln, stesc + 1 - buf->buf,
1200: ROFFT_COMMENT, TOKEN_NONE);
1201: n->string = mandoc_strdup(stesc + 2);
1202: roff_node_append(r->man, n);
1203: n->flags |= NODE_VALID | NODE_ENDED;
1204: r->man->next = ROFF_NEXT_SIBLING;
1205: }
1206:
1.334 schwarze 1207: /* Line continuation with comment. */
1208:
1209: if (stesc[1] == '#') {
1210: *stesc = '\0';
1211: return ROFF_APPEND;
1212: }
1213:
1214: /* Discard normal comments. */
1.328 schwarze 1215:
1.335 schwarze 1216: while (stesc > start && stesc[-1] == ' ' &&
1217: (stesc == start + 1 || stesc[-2] != '\\'))
1.303 schwarze 1218: stesc--;
1219: *stesc = '\0';
1220: break;
1221: }
1222: if (stesc == start)
1223: return ROFF_CONT;
1224: stesc--;
1225:
1226: /* Notice the end of the input. */
1227:
1228: if (*stesc == '\n') {
1229: *stesc-- = '\0';
1230: done = 1;
1231: }
1232:
1.170 schwarze 1233: expand_count = 0;
1.303 schwarze 1234: while (stesc >= start) {
1.170 schwarze 1235:
1.205 schwarze 1236: /* Search backwards for the next backslash. */
1237:
1.303 schwarze 1238: if (*stesc != r->escape) {
1239: if (*stesc == '\\') {
1240: *stesc = '\0';
1241: buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
1242: buf->buf, stesc + 1) + 1;
1243: start = nbuf + pos;
1244: stesc = nbuf + (stesc - buf->buf);
1245: free(buf->buf);
1246: buf->buf = nbuf;
1247: }
1248: stesc--;
1.205 schwarze 1249: continue;
1.303 schwarze 1250: }
1.205 schwarze 1251:
1252: /* If it is escaped, skip it. */
1253:
1254: for (cp = stesc - 1; cp >= start; cp--)
1.303 schwarze 1255: if (*cp != r->escape)
1.205 schwarze 1256: break;
1257:
1.238 schwarze 1258: if ((stesc - cp) % 2 == 0) {
1.303 schwarze 1259: while (stesc > cp)
1260: *stesc-- = '\\';
1.205 schwarze 1261: continue;
1.303 schwarze 1262: } else if (stesc[1] != '\0') {
1263: *stesc = '\\';
1264: } else {
1265: *stesc-- = '\0';
1266: if (done)
1267: continue;
1268: else
1269: return ROFF_APPEND;
1.205 schwarze 1270: }
1.108 schwarze 1271:
1.206 schwarze 1272: /* Decide whether to expand or to check only. */
1.108 schwarze 1273:
1.206 schwarze 1274: term = '\0';
1.205 schwarze 1275: cp = stesc + 1;
1.181 schwarze 1276: switch (*cp) {
1.207 schwarze 1277: case '*':
1.181 schwarze 1278: res = NULL;
1279: break;
1.207 schwarze 1280: case 'B':
1281: case 'w':
1.206 schwarze 1282: term = cp[1];
1283: /* FALLTHROUGH */
1.207 schwarze 1284: case 'n':
1.327 schwarze 1285: sign = cp[1];
1286: if (sign == '+' || sign == '-')
1287: cp++;
1.181 schwarze 1288: res = ubuf;
1289: break;
1290: default:
1.237 schwarze 1291: esc = mandoc_escape(&cp, &stnam, &inaml);
1292: if (esc == ESCAPE_ERROR ||
1293: (esc == ESCAPE_SPECIAL &&
1.279 schwarze 1294: mchars_spec2cp(stnam, inaml) < 0))
1.219 schwarze 1295: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1296: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1297: "%.*s", (int)(cp - stesc), stesc);
1.303 schwarze 1298: stesc--;
1.205 schwarze 1299: continue;
1.152 kristaps 1300: }
1301:
1.205 schwarze 1302: if (EXPAND_LIMIT < ++expand_count) {
1303: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1304: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1305: return ROFF_IGN;
1.205 schwarze 1306: }
1.108 schwarze 1307:
1308: /*
1309: * The third character decides the length
1.181 schwarze 1310: * of the name of the string or register.
1.108 schwarze 1311: * Save a pointer to the name.
1312: */
1313:
1.238 schwarze 1314: if (term == '\0') {
1.206 schwarze 1315: switch (*++cp) {
1.207 schwarze 1316: case '\0':
1.206 schwarze 1317: maxl = 0;
1318: break;
1.207 schwarze 1319: case '(':
1.206 schwarze 1320: cp++;
1321: maxl = 2;
1322: break;
1.207 schwarze 1323: case '[':
1.206 schwarze 1324: cp++;
1325: term = ']';
1326: maxl = 0;
1327: break;
1328: default:
1329: maxl = 1;
1330: break;
1331: }
1332: } else {
1333: cp += 2;
1.94 kristaps 1334: maxl = 0;
1335: }
1.108 schwarze 1336: stnam = cp;
1.94 kristaps 1337:
1.108 schwarze 1338: /* Advance to the end of the name. */
1.94 kristaps 1339:
1.253 schwarze 1340: naml = 0;
1.218 schwarze 1341: arg_complete = 1;
1.253 schwarze 1342: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1343: if (*cp == '\0') {
1.219 schwarze 1344: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1345: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1346: arg_complete = 0;
1.206 schwarze 1347: break;
1.153 kristaps 1348: }
1.238 schwarze 1349: if (maxl == 0 && *cp == term) {
1.206 schwarze 1350: cp++;
1.253 schwarze 1351: break;
1352: }
1353: if (*cp++ != '\\' || stesc[1] != 'w') {
1354: naml++;
1355: continue;
1356: }
1357: switch (mandoc_escape(&cp, NULL, NULL)) {
1358: case ESCAPE_SPECIAL:
1359: case ESCAPE_UNICODE:
1360: case ESCAPE_NUMBERED:
1361: case ESCAPE_OVERSTRIKE:
1362: naml++;
1363: break;
1364: default:
1.94 kristaps 1365: break;
1.206 schwarze 1366: }
1.94 kristaps 1367: }
1368:
1.108 schwarze 1369: /*
1370: * Retrieve the replacement string; if it is
1371: * undefined, resume searching for escapes.
1372: */
1373:
1.206 schwarze 1374: switch (stesc[1]) {
1.207 schwarze 1375: case '*':
1.315 schwarze 1376: if (arg_complete) {
1377: deftype = ROFFDEF_USER | ROFFDEF_PRE;
1378: res = roff_getstrn(r, stnam, naml, &deftype);
1.331 schwarze 1379:
1380: /*
1381: * If not overriden, let \*(.T
1382: * through to the formatters.
1383: */
1384:
1385: if (res == NULL && naml == 2 &&
1386: stnam[0] == '.' && stnam[1] == 'T') {
1387: roff_setstrn(&r->strtab,
1388: ".T", 2, NULL, 0, 0);
1389: stesc--;
1390: continue;
1391: }
1.315 schwarze 1392: }
1.206 schwarze 1393: break;
1.207 schwarze 1394: case 'B':
1.206 schwarze 1395: npos = 0;
1.218 schwarze 1396: ubuf[0] = arg_complete &&
1.261 schwarze 1397: roff_evalnum(r, ln, stnam, &npos,
1398: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1399: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1400: ubuf[1] = '\0';
1401: break;
1.207 schwarze 1402: case 'n':
1.218 schwarze 1403: if (arg_complete)
1404: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.327 schwarze 1405: roff_getregn(r, stnam, naml, sign));
1.218 schwarze 1406: else
1407: ubuf[0] = '\0';
1.206 schwarze 1408: break;
1.207 schwarze 1409: case 'w':
1.218 schwarze 1410: /* use even incomplete args */
1.208 schwarze 1411: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1412: 24 * (int)naml);
1413: break;
1414: }
1.94 kristaps 1415:
1.238 schwarze 1416: if (res == NULL) {
1.219 schwarze 1417: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1418: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1419: "%.*s", (int)naml, stnam);
1.142 kristaps 1420: res = "";
1.246 schwarze 1421: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1422: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1423: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1424: return ROFF_IGN;
1.94 kristaps 1425: }
1426:
1.108 schwarze 1427: /* Replace the escape sequence by the string. */
1428:
1.209 schwarze 1429: *stesc = '\0';
1.238 schwarze 1430: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1431: buf->buf, res, cp) + 1;
1.94 kristaps 1432:
1.205 schwarze 1433: /* Prepare for the next replacement. */
1.94 kristaps 1434:
1.205 schwarze 1435: start = nbuf + pos;
1.238 schwarze 1436: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1437: free(buf->buf);
1438: buf->buf = nbuf;
1.154 kristaps 1439: }
1.277 schwarze 1440: return ROFF_CONT;
1.154 kristaps 1441: }
1442:
1443: /*
1.275 schwarze 1444: * Process text streams.
1.154 kristaps 1445: */
1446: static enum rofferr
1.305 schwarze 1447: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.154 kristaps 1448: {
1449: size_t sz;
1450: const char *start;
1.178 schwarze 1451: char *p;
1452: int isz;
1.154 kristaps 1453: enum mandoc_esc esc;
1454:
1.275 schwarze 1455: /* Spring the input line trap. */
1456:
1457: if (roffit_lines == 1) {
1458: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1459: free(buf->buf);
1460: buf->buf = p;
1461: buf->sz = isz + 1;
1462: *offs = 0;
1463: free(roffit_macro);
1464: roffit_lines = 0;
1.277 schwarze 1465: return ROFF_REPARSE;
1.275 schwarze 1466: } else if (roffit_lines > 1)
1467: --roffit_lines;
1468:
1.305 schwarze 1469: if (roffce_node != NULL && buf->buf[pos] != '\0') {
1470: if (roffce_lines < 1) {
1471: r->man->last = roffce_node;
1472: r->man->next = ROFF_NEXT_SIBLING;
1473: roffce_lines = 0;
1474: roffce_node = NULL;
1475: } else
1476: roffce_lines--;
1477: }
1478:
1.275 schwarze 1479: /* Convert all breakable hyphens into ASCII_HYPH. */
1480:
1.238 schwarze 1481: start = p = buf->buf + pos;
1.154 kristaps 1482:
1.238 schwarze 1483: while (*p != '\0') {
1.154 kristaps 1484: sz = strcspn(p, "-\\");
1485: p += sz;
1486:
1.238 schwarze 1487: if (*p == '\0')
1.159 kristaps 1488: break;
1489:
1.238 schwarze 1490: if (*p == '\\') {
1.154 kristaps 1491: /* Skip over escapes. */
1492: p++;
1.189 schwarze 1493: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1494: if (esc == ESCAPE_ERROR)
1.154 kristaps 1495: break;
1.264 schwarze 1496: while (*p == '-')
1497: p++;
1.155 kristaps 1498: continue;
1.159 kristaps 1499: } else if (p == start) {
1.158 kristaps 1500: p++;
1.155 kristaps 1501: continue;
1.158 kristaps 1502: }
1.155 kristaps 1503:
1.171 schwarze 1504: if (isalpha((unsigned char)p[-1]) &&
1505: isalpha((unsigned char)p[1]))
1.155 kristaps 1506: *p = ASCII_HYPH;
1507: p++;
1.94 kristaps 1508: }
1.277 schwarze 1509: return ROFF_CONT;
1.94 kristaps 1510: }
1511:
1.67 kristaps 1512: enum rofferr
1.238 schwarze 1513: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1514: {
1.294 schwarze 1515: enum roff_tok t;
1.109 kristaps 1516: enum rofferr e;
1.238 schwarze 1517: int pos; /* parse point */
1.243 schwarze 1518: int spos; /* saved parse point for messages */
1.238 schwarze 1519: int ppos; /* original offset in buf->buf */
1520: int ctl; /* macro line (boolean) */
1521:
1522: ppos = pos = *offs;
1.79 kristaps 1523:
1.230 schwarze 1524: /* Handle in-line equation delimiters. */
1525:
1.236 schwarze 1526: if (r->tbl == NULL &&
1527: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1528: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1529: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1530: if (e == ROFF_REPARSE)
1.277 schwarze 1531: return e;
1.230 schwarze 1532: assert(e == ROFF_CONT);
1533: }
1534:
1535: /* Expand some escape sequences. */
1.94 kristaps 1536:
1.238 schwarze 1537: e = roff_res(r, buf, ln, pos);
1.303 schwarze 1538: if (e == ROFF_IGN || e == ROFF_APPEND)
1.277 schwarze 1539: return e;
1.238 schwarze 1540: assert(e == ROFF_CONT);
1.94 kristaps 1541:
1.238 schwarze 1542: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1543:
1.94 kristaps 1544: /*
1.79 kristaps 1545: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1546: * text through the macro's filter.
1547: * Equations process all content themselves.
1548: * Tables process almost all content themselves, but we want
1549: * to warn about macros before passing it there.
1.79 kristaps 1550: */
1.74 kristaps 1551:
1.252 schwarze 1552: if (r->last != NULL && ! ctl) {
1.78 kristaps 1553: t = r->last->tok;
1.238 schwarze 1554: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.294 schwarze 1555: if (e == ROFF_IGN)
1.277 schwarze 1556: return e;
1.294 schwarze 1557: assert(e == ROFF_CONT);
1.182 schwarze 1558: }
1.319 schwarze 1559: if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
1560: eqn_read(r->eqn, buf->buf + ppos);
1561: return ROFF_IGN;
1562: }
1.321 schwarze 1563: if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
1564: tbl_read(r->tbl, ln, buf->buf, ppos);
1565: roff_addtbl(r->man, r->tbl);
1566: return ROFF_IGN;
1567: }
1.252 schwarze 1568: if ( ! ctl)
1.305 schwarze 1569: return roff_parsetext(r, buf, pos, offs);
1.228 schwarze 1570:
1571: /* Skip empty request lines. */
1572:
1.238 schwarze 1573: if (buf->buf[pos] == '"') {
1.228 schwarze 1574: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1575: ln, pos, NULL);
1.277 schwarze 1576: return ROFF_IGN;
1.238 schwarze 1577: } else if (buf->buf[pos] == '\0')
1.277 schwarze 1578: return ROFF_IGN;
1.67 kristaps 1579:
1.79 kristaps 1580: /*
1581: * If a scope is open, go to the child handler for that macro,
1582: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1583: * Don't do so if an equation is open.
1.79 kristaps 1584: */
1.78 kristaps 1585:
1.79 kristaps 1586: if (r->last) {
1587: t = r->last->tok;
1.277 schwarze 1588: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1589: }
1.78 kristaps 1590:
1.243 schwarze 1591: /* No scope is open. This is a new request or macro. */
1592:
1593: spos = pos;
1594: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1595:
1596: /* Tables ignore most macros. */
1597:
1.308 schwarze 1598: if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.309 schwarze 1599: t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.243 schwarze 1600: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1601: ln, pos, buf->buf + spos);
1.308 schwarze 1602: if (t != TOKEN_NONE)
1.277 schwarze 1603: return ROFF_IGN;
1.257 schwarze 1604: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1605: pos++;
1.291 schwarze 1606: while (buf->buf[pos] == ' ')
1.257 schwarze 1607: pos++;
1.321 schwarze 1608: tbl_read(r->tbl, ln, buf->buf, pos);
1609: roff_addtbl(r->man, r->tbl);
1610: return ROFF_IGN;
1.243 schwarze 1611: }
1612:
1.305 schwarze 1613: /* For now, let high level macros abort .ce mode. */
1614:
1615: if (ctl && roffce_node != NULL &&
1.324 schwarze 1616: (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
1617: t == ROFF_TH || t == ROFF_TS)) {
1.305 schwarze 1618: r->man->last = roffce_node;
1619: r->man->next = ROFF_NEXT_SIBLING;
1620: roffce_lines = 0;
1621: roffce_node = NULL;
1622: }
1623:
1.79 kristaps 1624: /*
1.243 schwarze 1625: * This is neither a roff request nor a user-defined macro.
1626: * Let the standard macro set parsers handle it.
1.79 kristaps 1627: */
1.67 kristaps 1628:
1.294 schwarze 1629: if (t == TOKEN_NONE)
1.277 schwarze 1630: return ROFF_CONT;
1.243 schwarze 1631:
1632: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1633:
1.296 schwarze 1634: return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.74 kristaps 1635: }
1636:
1.117 kristaps 1637: void
1.74 kristaps 1638: roff_endparse(struct roff *r)
1639: {
1.321 schwarze 1640: if (r->last != NULL)
1.221 schwarze 1641: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1642: r->last->line, r->last->col,
1.295 schwarze 1643: roff_name[r->last->tok]);
1.117 kristaps 1644:
1.321 schwarze 1645: if (r->eqn != NULL) {
1.221 schwarze 1646: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1.319 schwarze 1647: r->eqn->node->line, r->eqn->node->pos, "EQ");
1648: eqn_parse(r->eqn);
1649: r->eqn = NULL;
1.125 kristaps 1650: }
1651:
1.321 schwarze 1652: if (r->tbl != NULL) {
1.221 schwarze 1653: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1654: r->tbl->line, r->tbl->pos, "TS");
1.321 schwarze 1655: tbl_end(r->tbl);
1656: r->tbl = NULL;
1.117 kristaps 1657: }
1.67 kristaps 1658: }
1659:
1660: /*
1661: * Parse a roff node's type from the input buffer. This must be in the
1662: * form of ".foo xxx" in the usual way.
1663: */
1.294 schwarze 1664: static enum roff_tok
1.214 schwarze 1665: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1666: {
1.214 schwarze 1667: char *cp;
1.106 kristaps 1668: const char *mac;
1669: size_t maclen;
1.315 schwarze 1670: int deftype;
1.294 schwarze 1671: enum roff_tok t;
1.67 kristaps 1672:
1.214 schwarze 1673: cp = buf + *pos;
1674:
1675: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294 schwarze 1676: return TOKEN_NONE;
1.67 kristaps 1677:
1.214 schwarze 1678: mac = cp;
1679: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1680:
1.315 schwarze 1681: deftype = ROFFDEF_USER | ROFFDEF_REN;
1682: r->current_string = roff_getstrn(r, mac, maclen, &deftype);
1683: switch (deftype) {
1684: case ROFFDEF_USER:
1685: t = ROFF_USERDEF;
1686: break;
1687: case ROFFDEF_REN:
1688: t = ROFF_RENAMED;
1689: break;
1690: default:
1691: t = roffhash_find(r->reqtab, mac, maclen);
1692: break;
1693: }
1.294 schwarze 1694: if (t != TOKEN_NONE)
1.214 schwarze 1695: *pos = cp - buf;
1.325 schwarze 1696: else if (deftype == ROFFDEF_UNDEF) {
1697: /* Using an undefined macro defines it to be empty. */
1698: roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
1699: roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
1700: }
1.277 schwarze 1701: return t;
1.67 kristaps 1702: }
1703:
1.266 schwarze 1704: /* --- handling of request blocks ----------------------------------------- */
1705:
1.67 kristaps 1706: static enum rofferr
1.76 kristaps 1707: roff_cblock(ROFF_ARGS)
1.67 kristaps 1708: {
1709:
1.79 kristaps 1710: /*
1711: * A block-close `..' should only be invoked as a child of an
1712: * ignore macro, otherwise raise a warning and just ignore it.
1713: */
1714:
1.238 schwarze 1715: if (r->last == NULL) {
1.221 schwarze 1716: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1717: ln, ppos, "..");
1.277 schwarze 1718: return ROFF_IGN;
1.76 kristaps 1719: }
1.67 kristaps 1720:
1.81 kristaps 1721: switch (r->last->tok) {
1.207 schwarze 1722: case ROFF_am:
1.220 schwarze 1723: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.207 schwarze 1724: case ROFF_ami:
1725: case ROFF_de:
1.108 schwarze 1726: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.207 schwarze 1727: case ROFF_dei:
1728: case ROFF_ig:
1.81 kristaps 1729: break;
1730: default:
1.221 schwarze 1731: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1732: ln, ppos, "..");
1.277 schwarze 1733: return ROFF_IGN;
1.76 kristaps 1734: }
1.67 kristaps 1735:
1.238 schwarze 1736: if (buf->buf[pos] != '\0')
1.217 schwarze 1737: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1738: ".. %s", buf->buf + pos);
1.71 kristaps 1739:
1740: roffnode_pop(r);
1.76 kristaps 1741: roffnode_cleanscope(r);
1.277 schwarze 1742: return ROFF_IGN;
1.71 kristaps 1743:
1.67 kristaps 1744: }
1745:
1.76 kristaps 1746: static void
1747: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1748: {
1749:
1.76 kristaps 1750: while (r->last) {
1.173 schwarze 1751: if (--r->last->endspan != 0)
1.76 kristaps 1752: break;
1753: roffnode_pop(r);
1754: }
1.67 kristaps 1755: }
1756:
1.195 schwarze 1757: static void
1758: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1759: {
1760:
1.76 kristaps 1761: if (NULL == r->last) {
1.221 schwarze 1762: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1763: ln, ppos, "\\}");
1.195 schwarze 1764: return;
1.76 kristaps 1765: }
1766:
1.82 kristaps 1767: switch (r->last->tok) {
1.207 schwarze 1768: case ROFF_el:
1769: case ROFF_ie:
1770: case ROFF_if:
1.82 kristaps 1771: break;
1772: default:
1.221 schwarze 1773: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1774: ln, ppos, "\\}");
1.195 schwarze 1775: return;
1.75 kristaps 1776: }
1777:
1.76 kristaps 1778: if (r->last->endspan > -1) {
1.221 schwarze 1779: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1780: ln, ppos, "\\}");
1.195 schwarze 1781: return;
1.76 kristaps 1782: }
1783:
1.75 kristaps 1784: roffnode_pop(r);
1.76 kristaps 1785: roffnode_cleanscope(r);
1.195 schwarze 1786: return;
1.76 kristaps 1787: }
1788:
1789: static enum rofferr
1.80 kristaps 1790: roff_block(ROFF_ARGS)
1.76 kristaps 1791: {
1.315 schwarze 1792: const char *name, *value;
1793: char *call, *cp, *iname, *rname;
1794: size_t csz, namesz, rsz;
1795: int deftype;
1.106 kristaps 1796:
1.220 schwarze 1797: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1798:
1.238 schwarze 1799: if (tok == ROFF_de1)
1.220 schwarze 1800: tok = ROFF_de;
1.251 schwarze 1801: else if (tok == ROFF_dei1)
1802: tok = ROFF_dei;
1.238 schwarze 1803: else if (tok == ROFF_am1)
1.220 schwarze 1804: tok = ROFF_am;
1.251 schwarze 1805: else if (tok == ROFF_ami1)
1806: tok = ROFF_ami;
1.220 schwarze 1807:
1808: /* Parse the macro name argument. */
1809:
1.238 schwarze 1810: cp = buf->buf + pos;
1811: if (tok == ROFF_ig) {
1.220 schwarze 1812: iname = NULL;
1813: namesz = 0;
1814: } else {
1815: iname = cp;
1816: namesz = roff_getname(r, &cp, ln, ppos);
1817: iname[namesz] = '\0';
1818: }
1.107 kristaps 1819:
1.220 schwarze 1820: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1821:
1.238 schwarze 1822: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315 schwarze 1823: deftype = ROFFDEF_USER;
1824: name = roff_getstrn(r, iname, namesz, &deftype);
1825: if (name == NULL) {
1.220 schwarze 1826: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1827: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1828: "%.*s", (int)namesz, iname);
1829: namesz = 0;
1830: } else
1831: namesz = strlen(name);
1832: } else
1833: name = iname;
1.107 kristaps 1834:
1.238 schwarze 1835: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1836: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.295 schwarze 1837: ln, ppos, roff_name[tok]);
1.277 schwarze 1838: return ROFF_IGN;
1.220 schwarze 1839: }
1.80 kristaps 1840:
1.106 kristaps 1841: roffnode_push(r, tok, name, ln, ppos);
1842:
1843: /*
1844: * At the beginning of a `de' macro, clear the existing string
1845: * with the same name, if there is one. New content will be
1.193 schwarze 1846: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1847: */
1.107 kristaps 1848:
1.311 schwarze 1849: if (tok == ROFF_de || tok == ROFF_dei) {
1.213 schwarze 1850: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.311 schwarze 1851: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.315 schwarze 1852: } else if (tok == ROFF_am || tok == ROFF_ami) {
1853: deftype = ROFFDEF_ANY;
1854: value = roff_getstrn(r, iname, namesz, &deftype);
1855: switch (deftype) { /* Before appending, ... */
1856: case ROFFDEF_PRE: /* copy predefined to user-defined. */
1857: roff_setstrn(&r->strtab, name, namesz,
1858: value, strlen(value), 0);
1859: break;
1860: case ROFFDEF_REN: /* call original standard macro. */
1861: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1862: (int)strlen(value), value);
1863: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1864: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1865: free(call);
1866: break;
1867: case ROFFDEF_STD: /* rename and call standard macro. */
1868: rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
1869: roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
1870: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1871: (int)rsz, rname);
1872: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1873: free(call);
1874: free(rname);
1875: break;
1876: default:
1877: break;
1878: }
1.311 schwarze 1879: }
1.76 kristaps 1880:
1.238 schwarze 1881: if (*cp == '\0')
1.277 schwarze 1882: return ROFF_IGN;
1.78 kristaps 1883:
1.220 schwarze 1884: /* Get the custom end marker. */
1.107 kristaps 1885:
1.220 schwarze 1886: iname = cp;
1.213 schwarze 1887: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1888:
1889: /* Resolve the end marker if it is indirect. */
1890:
1.238 schwarze 1891: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315 schwarze 1892: deftype = ROFFDEF_USER;
1893: name = roff_getstrn(r, iname, namesz, &deftype);
1894: if (name == NULL) {
1.220 schwarze 1895: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1896: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1897: "%.*s", (int)namesz, iname);
1898: namesz = 0;
1899: } else
1900: namesz = strlen(name);
1901: } else
1902: name = iname;
1903:
1.213 schwarze 1904: if (namesz)
1905: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1906:
1.238 schwarze 1907: if (*cp != '\0')
1.217 schwarze 1908: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.295 schwarze 1909: ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74 kristaps 1910:
1.277 schwarze 1911: return ROFF_IGN;
1.78 kristaps 1912: }
1913:
1914: static enum rofferr
1.80 kristaps 1915: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1916: {
1.294 schwarze 1917: enum roff_tok t;
1.79 kristaps 1918: int i, j;
1919:
1920: /*
1921: * First check whether a custom macro exists at this level. If
1922: * it does, then check against it. This is some of groff's
1923: * stranger behaviours. If we encountered a custom end-scope
1924: * tag and that tag also happens to be a "real" macro, then we
1925: * need to try interpreting it again as a real macro. If it's
1926: * not, then return ignore. Else continue.
1927: */
1928:
1929: if (r->last->end) {
1.130 kristaps 1930: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1931: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1932: break;
1933:
1.238 schwarze 1934: if (r->last->end[j] == '\0' &&
1935: (buf->buf[i] == '\0' ||
1936: buf->buf[i] == ' ' ||
1937: buf->buf[i] == '\t')) {
1.79 kristaps 1938: roffnode_pop(r);
1939: roffnode_cleanscope(r);
1940:
1.238 schwarze 1941: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1942: i++;
1943:
1944: pos = i;
1.238 schwarze 1945: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294 schwarze 1946: TOKEN_NONE)
1.277 schwarze 1947: return ROFF_RERUN;
1948: return ROFF_IGN;
1.79 kristaps 1949: }
1950: }
1951:
1952: /*
1953: * If we have no custom end-query or lookup failed, then try
1954: * pulling it out of the hashtable.
1955: */
1956:
1.238 schwarze 1957: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1958:
1.238 schwarze 1959: if (t != ROFF_cblock) {
1960: if (tok != ROFF_ig)
1961: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.277 schwarze 1962: return ROFF_IGN;
1.106 kristaps 1963: }
1.79 kristaps 1964:
1.277 schwarze 1965: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1966: }
1967:
1968: static enum rofferr
1.80 kristaps 1969: roff_block_text(ROFF_ARGS)
1.78 kristaps 1970: {
1971:
1.238 schwarze 1972: if (tok != ROFF_ig)
1973: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1974:
1.277 schwarze 1975: return ROFF_IGN;
1.78 kristaps 1976: }
1977:
1978: static enum rofferr
1.82 kristaps 1979: roff_cond_sub(ROFF_ARGS)
1980: {
1.294 schwarze 1981: enum roff_tok t;
1.139 kristaps 1982: char *ep;
1.198 schwarze 1983: int rr;
1.82 kristaps 1984:
1985: rr = r->last->rule;
1.139 kristaps 1986: roffnode_cleanscope(r);
1.144 kristaps 1987:
1.196 schwarze 1988: /*
1989: * If `\}' occurs on a macro line without a preceding macro,
1990: * drop the line completely.
1991: */
1992:
1.238 schwarze 1993: ep = buf->buf + pos;
1994: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1995: rr = 0;
1.196 schwarze 1996:
1.177 schwarze 1997: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1998:
1.238 schwarze 1999: while ((ep = strchr(ep, '\\')) != NULL) {
1.318 schwarze 2000: switch (ep[1]) {
2001: case '}':
2002: memmove(ep, ep + 2, strlen(ep + 2) + 1);
2003: roff_ccond(r, ln, ep - buf->buf);
2004: break;
2005: case '\0':
2006: ++ep;
2007: break;
2008: default:
2009: ep += 2;
2010: break;
1.197 schwarze 2011: }
1.177 schwarze 2012: }
1.318 schwarze 2013:
2014: /*
2015: * Fully handle known macros when they are structurally
2016: * required or when the conditional evaluated to true.
2017: */
2018:
2019: t = roff_parse(r, buf->buf, &pos, ln, ppos);
2020: return t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT)
2021: ? (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs) : rr
2022: ? ROFF_CONT : ROFF_IGN;
1.82 kristaps 2023: }
2024:
2025: static enum rofferr
2026: roff_cond_text(ROFF_ARGS)
1.78 kristaps 2027: {
1.140 kristaps 2028: char *ep;
1.198 schwarze 2029: int rr;
1.82 kristaps 2030:
2031: rr = r->last->rule;
1.140 kristaps 2032: roffnode_cleanscope(r);
1.82 kristaps 2033:
1.238 schwarze 2034: ep = buf->buf + pos;
2035: while ((ep = strchr(ep, '\\')) != NULL) {
2036: if (*(++ep) == '}') {
1.197 schwarze 2037: *ep = '&';
1.238 schwarze 2038: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 2039: }
1.247 schwarze 2040: if (*ep != '\0')
2041: ++ep;
1.78 kristaps 2042: }
1.277 schwarze 2043: return rr ? ROFF_CONT : ROFF_IGN;
1.74 kristaps 2044: }
2045:
1.266 schwarze 2046: /* --- handling of numeric and conditional expressions -------------------- */
2047:
1.204 schwarze 2048: /*
2049: * Parse a single signed integer number. Stop at the first non-digit.
2050: * If there is at least one digit, return success and advance the
2051: * parse point, else return failure and let the parse point unchanged.
2052: * Ignore overflows, treat them just like the C language.
2053: */
1.184 schwarze 2054: static int
1.261 schwarze 2055: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 2056: {
1.261 schwarze 2057: int myres, scaled, n, p;
1.206 schwarze 2058:
2059: if (NULL == res)
2060: res = &myres;
1.184 schwarze 2061:
2062: p = *pos;
2063: n = v[p] == '-';
1.261 schwarze 2064: if (n || v[p] == '+')
1.184 schwarze 2065: p++;
2066:
1.261 schwarze 2067: if (flags & ROFFNUM_WHITE)
2068: while (isspace((unsigned char)v[p]))
2069: p++;
2070:
1.184 schwarze 2071: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 2072: *res = 10 * *res + v[p] - '0';
1.184 schwarze 2073: if (p == *pos + n)
2074: return 0;
2075:
2076: if (n)
2077: *res = -*res;
2078:
1.254 schwarze 2079: /* Each number may be followed by one optional scaling unit. */
2080:
2081: switch (v[p]) {
2082: case 'f':
1.261 schwarze 2083: scaled = *res * 65536;
1.254 schwarze 2084: break;
2085: case 'i':
1.261 schwarze 2086: scaled = *res * 240;
1.254 schwarze 2087: break;
2088: case 'c':
1.261 schwarze 2089: scaled = *res * 240 / 2.54;
1.254 schwarze 2090: break;
2091: case 'v':
2092: case 'P':
1.261 schwarze 2093: scaled = *res * 40;
1.254 schwarze 2094: break;
2095: case 'm':
2096: case 'n':
1.261 schwarze 2097: scaled = *res * 24;
1.254 schwarze 2098: break;
2099: case 'p':
1.261 schwarze 2100: scaled = *res * 10 / 3;
1.254 schwarze 2101: break;
2102: case 'u':
1.261 schwarze 2103: scaled = *res;
1.254 schwarze 2104: break;
2105: case 'M':
1.261 schwarze 2106: scaled = *res * 6 / 25;
1.254 schwarze 2107: break;
2108: default:
1.261 schwarze 2109: scaled = *res;
1.254 schwarze 2110: p--;
2111: break;
2112: }
1.261 schwarze 2113: if (flags & ROFFNUM_SCALE)
2114: *res = scaled;
1.254 schwarze 2115:
2116: *pos = p + 1;
1.277 schwarze 2117: return 1;
1.184 schwarze 2118: }
2119:
1.198 schwarze 2120: /*
2121: * Evaluate a string comparison condition.
2122: * The first character is the delimiter.
2123: * Succeed if the string up to its second occurrence
2124: * matches the string up to its third occurence.
2125: * Advance the cursor after the third occurrence
2126: * or lacking that, to the end of the line.
2127: */
2128: static int
2129: roff_evalstrcond(const char *v, int *pos)
2130: {
2131: const char *s1, *s2, *s3;
2132: int match;
2133:
2134: match = 0;
2135: s1 = v + *pos; /* initial delimiter */
2136: s2 = s1 + 1; /* for scanning the first string */
2137: s3 = strchr(s2, *s1); /* for scanning the second string */
2138:
2139: if (NULL == s3) /* found no middle delimiter */
2140: goto out;
2141:
2142: while ('\0' != *++s3) {
2143: if (*s2 != *s3) { /* mismatch */
2144: s3 = strchr(s3, *s1);
2145: break;
2146: }
2147: if (*s3 == *s1) { /* found the final delimiter */
2148: match = 1;
2149: break;
2150: }
2151: s2++;
2152: }
2153:
2154: out:
2155: if (NULL == s3)
2156: s3 = strchr(s2, '\0');
1.242 schwarze 2157: else if (*s3 != '\0')
1.198 schwarze 2158: s3++;
2159: *pos = s3 - v;
1.277 schwarze 2160: return match;
1.198 schwarze 2161: }
2162:
1.204 schwarze 2163: /*
2164: * Evaluate an optionally negated single character, numerical,
2165: * or string condition.
2166: */
1.198 schwarze 2167: static int
1.271 schwarze 2168: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88 kristaps 2169: {
1.336 schwarze 2170: const char *start, *end;
2171: char *cp, *name;
2172: size_t sz;
2173: int deftype, len, number, savepos, istrue, wanttrue;
1.88 kristaps 2174:
1.198 schwarze 2175: if ('!' == v[*pos]) {
2176: wanttrue = 0;
2177: (*pos)++;
2178: } else
2179: wanttrue = 1;
2180:
1.88 kristaps 2181: switch (v[*pos]) {
1.240 schwarze 2182: case '\0':
1.277 schwarze 2183: return 0;
1.207 schwarze 2184: case 'n':
2185: case 'o':
1.88 kristaps 2186: (*pos)++;
1.277 schwarze 2187: return wanttrue;
1.207 schwarze 2188: case 'e':
2189: case 't':
1.239 schwarze 2190: case 'v':
1.88 kristaps 2191: (*pos)++;
1.277 schwarze 2192: return !wanttrue;
1.336 schwarze 2193: case 'c':
2194: do {
2195: (*pos)++;
2196: } while (v[*pos] == ' ');
2197:
2198: /*
2199: * Quirk for groff compatibility:
2200: * The horizontal tab is neither available nor unavailable.
2201: */
2202:
2203: if (v[*pos] == '\t') {
2204: (*pos)++;
2205: return 0;
2206: }
2207:
2208: /* Printable ASCII characters are available. */
2209:
2210: if (v[*pos] != '\\') {
2211: (*pos)++;
2212: return wanttrue;
2213: }
2214:
2215: end = v + ++*pos;
2216: switch (mandoc_escape(&end, &start, &len)) {
2217: case ESCAPE_SPECIAL:
2218: istrue = mchars_spec2cp(start, len) != -1;
2219: break;
2220: case ESCAPE_UNICODE:
2221: istrue = 1;
2222: break;
2223: case ESCAPE_NUMBERED:
2224: istrue = mchars_num2char(start, len) != -1;
2225: break;
2226: default:
2227: istrue = !wanttrue;
2228: break;
2229: }
2230: *pos = end - v;
2231: return istrue == wanttrue;
1.310 schwarze 2232: case 'd':
1.271 schwarze 2233: case 'r':
1.310 schwarze 2234: cp = v + *pos + 1;
2235: while (*cp == ' ')
2236: cp++;
2237: name = cp;
2238: sz = roff_getname(r, &cp, ln, cp - v);
1.315 schwarze 2239: if (sz == 0)
2240: istrue = 0;
2241: else if (v[*pos] == 'r')
2242: istrue = roff_hasregn(r, name, sz);
2243: else {
2244: deftype = ROFFDEF_ANY;
2245: roff_getstrn(r, name, sz, &deftype);
2246: istrue = !!deftype;
2247: }
1.271 schwarze 2248: *pos = cp - v;
1.310 schwarze 2249: return istrue == wanttrue;
1.88 kristaps 2250: default:
2251: break;
2252: }
2253:
1.241 schwarze 2254: savepos = *pos;
1.261 schwarze 2255: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277 schwarze 2256: return (number > 0) == wanttrue;
1.241 schwarze 2257: else if (*pos == savepos)
1.277 schwarze 2258: return roff_evalstrcond(v, pos) == wanttrue;
1.204 schwarze 2259: else
1.277 schwarze 2260: return 0;
1.88 kristaps 2261: }
2262:
1.74 kristaps 2263: static enum rofferr
1.103 kristaps 2264: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 2265: {
1.123 schwarze 2266:
1.277 schwarze 2267: return ROFF_IGN;
1.89 kristaps 2268: }
2269:
1.104 kristaps 2270: static enum rofferr
1.251 schwarze 2271: roff_insec(ROFF_ARGS)
2272: {
2273:
2274: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.295 schwarze 2275: ln, ppos, roff_name[tok]);
1.277 schwarze 2276: return ROFF_IGN;
1.251 schwarze 2277: }
2278:
2279: static enum rofferr
2280: roff_unsupp(ROFF_ARGS)
2281: {
2282:
2283: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.295 schwarze 2284: ln, ppos, roff_name[tok]);
1.277 schwarze 2285: return ROFF_IGN;
1.251 schwarze 2286: }
2287:
2288: static enum rofferr
1.82 kristaps 2289: roff_cond(ROFF_ARGS)
1.74 kristaps 2290: {
1.173 schwarze 2291:
2292: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 2293:
1.207 schwarze 2294: /*
1.134 kristaps 2295: * An `.el' has no conditional body: it will consume the value
2296: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 2297: * defaults to DENY.
1.134 kristaps 2298: *
2299: * If we're not an `el', however, then evaluate the conditional.
2300: */
1.133 kristaps 2301:
1.238 schwarze 2302: r->last->rule = tok == ROFF_el ?
1.207 schwarze 2303: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 2304: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 2305:
1.134 kristaps 2306: /*
2307: * An if-else will put the NEGATION of the current evaluated
2308: * conditional into the stack of rules.
2309: */
2310:
1.238 schwarze 2311: if (tok == ROFF_ie) {
1.223 schwarze 2312: if (r->rstackpos + 1 == r->rstacksz) {
2313: r->rstacksz += 16;
2314: r->rstack = mandoc_reallocarray(r->rstack,
2315: r->rstacksz, sizeof(int));
1.134 kristaps 2316: }
1.198 schwarze 2317: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 2318: }
1.88 kristaps 2319:
2320: /* If the parent has false as its rule, then so do we. */
2321:
1.198 schwarze 2322: if (r->last->parent && !r->last->parent->rule)
2323: r->last->rule = 0;
1.88 kristaps 2324:
2325: /*
1.173 schwarze 2326: * Determine scope.
2327: * If there is nothing on the line after the conditional,
2328: * not even whitespace, use next-line scope.
1.88 kristaps 2329: */
1.74 kristaps 2330:
1.238 schwarze 2331: if (buf->buf[pos] == '\0') {
1.173 schwarze 2332: r->last->endspan = 2;
2333: goto out;
2334: }
2335:
1.238 schwarze 2336: while (buf->buf[pos] == ' ')
1.173 schwarze 2337: pos++;
2338:
2339: /* An opening brace requests multiline scope. */
1.75 kristaps 2340:
1.238 schwarze 2341: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 2342: r->last->endspan = -1;
2343: pos += 2;
1.272 schwarze 2344: while (buf->buf[pos] == ' ')
2345: pos++;
1.173 schwarze 2346: goto out;
1.207 schwarze 2347: }
1.74 kristaps 2348:
1.77 kristaps 2349: /*
1.173 schwarze 2350: * Anything else following the conditional causes
2351: * single-line scope. Warn if the scope contains
2352: * nothing but trailing whitespace.
1.77 kristaps 2353: */
2354:
1.238 schwarze 2355: if (buf->buf[pos] == '\0')
1.216 schwarze 2356: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.295 schwarze 2357: ln, ppos, roff_name[tok]);
1.77 kristaps 2358:
1.173 schwarze 2359: r->last->endspan = 1;
1.74 kristaps 2360:
1.173 schwarze 2361: out:
1.75 kristaps 2362: *offs = pos;
1.277 schwarze 2363: return ROFF_RERUN;
1.83 schwarze 2364: }
2365:
2366: static enum rofferr
1.92 schwarze 2367: roff_ds(ROFF_ARGS)
2368: {
1.212 schwarze 2369: char *string;
2370: const char *name;
2371: size_t namesz;
1.96 kristaps 2372:
1.251 schwarze 2373: /* Ignore groff compatibility mode for now. */
2374:
2375: if (tok == ROFF_ds1)
2376: tok = ROFF_ds;
2377: else if (tok == ROFF_as1)
2378: tok = ROFF_as;
2379:
1.96 kristaps 2380: /*
1.212 schwarze 2381: * The first word is the name of the string.
2382: * If it is empty or terminated by an escape sequence,
2383: * abort the `ds' request without defining anything.
1.96 kristaps 2384: */
1.92 schwarze 2385:
1.238 schwarze 2386: name = string = buf->buf + pos;
2387: if (*name == '\0')
1.277 schwarze 2388: return ROFF_IGN;
1.92 schwarze 2389:
1.212 schwarze 2390: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 2391: if (name[namesz] == '\\')
1.277 schwarze 2392: return ROFF_IGN;
1.212 schwarze 2393:
2394: /* Read past the initial double-quote, if any. */
1.238 schwarze 2395: if (*string == '"')
1.92 schwarze 2396: string++;
2397:
1.96 kristaps 2398: /* The rest is the value. */
1.212 schwarze 2399: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2400: ROFF_as == tok);
1.311 schwarze 2401: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.277 schwarze 2402: return ROFF_IGN;
1.92 schwarze 2403: }
2404:
1.204 schwarze 2405: /*
2406: * Parse a single operator, one or two characters long.
2407: * If the operator is recognized, return success and advance the
2408: * parse point, else return failure and let the parse point unchanged.
2409: */
2410: static int
2411: roff_getop(const char *v, int *pos, char *res)
2412: {
2413:
2414: *res = v[*pos];
2415:
2416: switch (*res) {
1.207 schwarze 2417: case '+':
2418: case '-':
2419: case '*':
2420: case '/':
2421: case '%':
2422: case '&':
2423: case ':':
1.204 schwarze 2424: break;
2425: case '<':
2426: switch (v[*pos + 1]) {
1.207 schwarze 2427: case '=':
1.204 schwarze 2428: *res = 'l';
2429: (*pos)++;
2430: break;
1.207 schwarze 2431: case '>':
1.204 schwarze 2432: *res = '!';
2433: (*pos)++;
2434: break;
1.207 schwarze 2435: case '?':
1.204 schwarze 2436: *res = 'i';
2437: (*pos)++;
2438: break;
2439: default:
2440: break;
2441: }
2442: break;
2443: case '>':
2444: switch (v[*pos + 1]) {
1.207 schwarze 2445: case '=':
1.204 schwarze 2446: *res = 'g';
2447: (*pos)++;
2448: break;
1.207 schwarze 2449: case '?':
1.204 schwarze 2450: *res = 'a';
2451: (*pos)++;
2452: break;
2453: default:
2454: break;
2455: }
2456: break;
2457: case '=':
2458: if ('=' == v[*pos + 1])
2459: (*pos)++;
2460: break;
2461: default:
1.277 schwarze 2462: return 0;
1.204 schwarze 2463: }
2464: (*pos)++;
2465:
1.277 schwarze 2466: return *res;
1.204 schwarze 2467: }
2468:
2469: /*
2470: * Evaluate either a parenthesized numeric expression
2471: * or a single signed integer number.
2472: */
2473: static int
1.235 schwarze 2474: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2475: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2476: {
2477:
2478: if ('(' != v[*pos])
1.277 schwarze 2479: return roff_getnum(v, pos, res, flags);
1.204 schwarze 2480:
2481: (*pos)++;
1.261 schwarze 2482: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277 schwarze 2483: return 0;
1.204 schwarze 2484:
1.206 schwarze 2485: /*
2486: * Omission of the closing parenthesis
2487: * is an error in validation mode,
2488: * but ignored in evaluation mode.
2489: */
2490:
1.204 schwarze 2491: if (')' == v[*pos])
2492: (*pos)++;
1.206 schwarze 2493: else if (NULL == res)
1.277 schwarze 2494: return 0;
1.204 schwarze 2495:
1.277 schwarze 2496: return 1;
1.204 schwarze 2497: }
2498:
2499: /*
2500: * Evaluate a complete numeric expression.
2501: * Proceed left to right, there is no concept of precedence.
2502: */
2503: static int
1.235 schwarze 2504: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2505: int *pos, int *res, int flags)
1.204 schwarze 2506: {
2507: int mypos, operand2;
2508: char operator;
2509:
2510: if (NULL == pos) {
2511: mypos = 0;
2512: pos = &mypos;
2513: }
2514:
1.261 schwarze 2515: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2516: while (isspace((unsigned char)v[*pos]))
2517: (*pos)++;
2518:
1.261 schwarze 2519: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277 schwarze 2520: return 0;
1.204 schwarze 2521:
2522: while (1) {
1.261 schwarze 2523: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2524: while (isspace((unsigned char)v[*pos]))
2525: (*pos)++;
2526:
2527: if ( ! roff_getop(v, pos, &operator))
2528: break;
2529:
1.261 schwarze 2530: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2531: while (isspace((unsigned char)v[*pos]))
2532: (*pos)++;
2533:
1.261 schwarze 2534: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277 schwarze 2535: return 0;
1.204 schwarze 2536:
1.261 schwarze 2537: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2538: while (isspace((unsigned char)v[*pos]))
2539: (*pos)++;
1.206 schwarze 2540:
2541: if (NULL == res)
2542: continue;
1.204 schwarze 2543:
2544: switch (operator) {
1.207 schwarze 2545: case '+':
1.204 schwarze 2546: *res += operand2;
2547: break;
1.207 schwarze 2548: case '-':
1.204 schwarze 2549: *res -= operand2;
2550: break;
1.207 schwarze 2551: case '*':
1.204 schwarze 2552: *res *= operand2;
2553: break;
1.207 schwarze 2554: case '/':
1.244 schwarze 2555: if (operand2 == 0) {
1.235 schwarze 2556: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2557: r->parse, ln, *pos, v);
2558: *res = 0;
2559: break;
2560: }
1.204 schwarze 2561: *res /= operand2;
2562: break;
1.207 schwarze 2563: case '%':
1.244 schwarze 2564: if (operand2 == 0) {
2565: mandoc_msg(MANDOCERR_DIVZERO,
2566: r->parse, ln, *pos, v);
2567: *res = 0;
2568: break;
2569: }
1.204 schwarze 2570: *res %= operand2;
2571: break;
1.207 schwarze 2572: case '<':
1.204 schwarze 2573: *res = *res < operand2;
2574: break;
1.207 schwarze 2575: case '>':
1.204 schwarze 2576: *res = *res > operand2;
2577: break;
1.207 schwarze 2578: case 'l':
1.204 schwarze 2579: *res = *res <= operand2;
2580: break;
1.207 schwarze 2581: case 'g':
1.204 schwarze 2582: *res = *res >= operand2;
2583: break;
1.207 schwarze 2584: case '=':
1.204 schwarze 2585: *res = *res == operand2;
2586: break;
1.207 schwarze 2587: case '!':
1.204 schwarze 2588: *res = *res != operand2;
2589: break;
1.207 schwarze 2590: case '&':
1.204 schwarze 2591: *res = *res && operand2;
2592: break;
1.207 schwarze 2593: case ':':
1.204 schwarze 2594: *res = *res || operand2;
2595: break;
1.207 schwarze 2596: case 'i':
1.204 schwarze 2597: if (operand2 < *res)
2598: *res = operand2;
2599: break;
1.207 schwarze 2600: case 'a':
1.204 schwarze 2601: if (operand2 > *res)
2602: *res = operand2;
2603: break;
2604: default:
2605: abort();
2606: }
2607: }
1.277 schwarze 2608: return 1;
1.204 schwarze 2609: }
2610:
1.266 schwarze 2611: /* --- register management ------------------------------------------------ */
2612:
1.180 schwarze 2613: void
1.187 schwarze 2614: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2615: {
1.327 schwarze 2616: roff_setregn(r, name, strlen(name), val, sign, INT_MIN);
1.326 schwarze 2617: }
2618:
2619: static void
2620: roff_setregn(struct roff *r, const char *name, size_t len,
1.327 schwarze 2621: int val, char sign, int step)
1.326 schwarze 2622: {
1.180 schwarze 2623: struct roffreg *reg;
2624:
2625: /* Search for an existing register with the same name. */
2626: reg = r->regtab;
2627:
1.326 schwarze 2628: while (reg != NULL && (reg->key.sz != len ||
2629: strncmp(reg->key.p, name, len) != 0))
1.180 schwarze 2630: reg = reg->next;
1.147 kristaps 2631:
1.180 schwarze 2632: if (NULL == reg) {
2633: /* Create a new register. */
2634: reg = mandoc_malloc(sizeof(struct roffreg));
1.326 schwarze 2635: reg->key.p = mandoc_strndup(name, len);
2636: reg->key.sz = len;
1.187 schwarze 2637: reg->val = 0;
1.327 schwarze 2638: reg->step = 0;
1.180 schwarze 2639: reg->next = r->regtab;
2640: r->regtab = reg;
2641: }
2642:
1.187 schwarze 2643: if ('+' == sign)
2644: reg->val += val;
2645: else if ('-' == sign)
2646: reg->val -= val;
2647: else
2648: reg->val = val;
1.327 schwarze 2649: if (step != INT_MIN)
2650: reg->step = step;
1.147 kristaps 2651: }
2652:
1.192 schwarze 2653: /*
2654: * Handle some predefined read-only number registers.
2655: * For now, return -1 if the requested register is not predefined;
2656: * in case a predefined read-only register having the value -1
2657: * were to turn up, another special value would have to be chosen.
2658: */
2659: static int
1.273 schwarze 2660: roff_getregro(const struct roff *r, const char *name)
1.192 schwarze 2661: {
2662:
2663: switch (*name) {
1.273 schwarze 2664: case '$': /* Number of arguments of the last macro evaluated. */
1.337 ! schwarze 2665: return 0;
1.207 schwarze 2666: case 'A': /* ASCII approximation mode is always off. */
1.277 schwarze 2667: return 0;
1.207 schwarze 2668: case 'g': /* Groff compatibility mode is always on. */
1.277 schwarze 2669: return 1;
1.207 schwarze 2670: case 'H': /* Fixed horizontal resolution. */
1.277 schwarze 2671: return 24;
1.207 schwarze 2672: case 'j': /* Always adjust left margin only. */
1.277 schwarze 2673: return 0;
1.207 schwarze 2674: case 'T': /* Some output device is always defined. */
1.277 schwarze 2675: return 1;
1.207 schwarze 2676: case 'V': /* Fixed vertical resolution. */
1.277 schwarze 2677: return 40;
1.192 schwarze 2678: default:
1.277 schwarze 2679: return -1;
1.192 schwarze 2680: }
2681: }
2682:
1.181 schwarze 2683: int
1.326 schwarze 2684: roff_getreg(struct roff *r, const char *name)
1.147 kristaps 2685: {
1.327 schwarze 2686: return roff_getregn(r, name, strlen(name), '\0');
1.181 schwarze 2687: }
2688:
2689: static int
1.327 schwarze 2690: roff_getregn(struct roff *r, const char *name, size_t len, char sign)
1.181 schwarze 2691: {
2692: struct roffreg *reg;
1.192 schwarze 2693: int val;
2694:
2695: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2696: val = roff_getregro(r, name + 1);
1.192 schwarze 2697: if (-1 != val)
1.277 schwarze 2698: return val;
1.192 schwarze 2699: }
1.181 schwarze 2700:
1.327 schwarze 2701: for (reg = r->regtab; reg; reg = reg->next) {
1.181 schwarze 2702: if (len == reg->key.sz &&
1.327 schwarze 2703: 0 == strncmp(name, reg->key.p, len)) {
2704: switch (sign) {
2705: case '+':
2706: reg->val += reg->step;
2707: break;
2708: case '-':
2709: reg->val -= reg->step;
2710: break;
2711: default:
2712: break;
2713: }
1.277 schwarze 2714: return reg->val;
1.327 schwarze 2715: }
2716: }
1.271 schwarze 2717:
1.327 schwarze 2718: roff_setregn(r, name, len, 0, '\0', INT_MIN);
1.277 schwarze 2719: return 0;
1.271 schwarze 2720: }
2721:
2722: static int
2723: roff_hasregn(const struct roff *r, const char *name, size_t len)
2724: {
2725: struct roffreg *reg;
2726: int val;
2727:
2728: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2729: val = roff_getregro(r, name + 1);
1.271 schwarze 2730: if (-1 != val)
1.277 schwarze 2731: return 1;
1.271 schwarze 2732: }
2733:
2734: for (reg = r->regtab; reg; reg = reg->next)
2735: if (len == reg->key.sz &&
2736: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2737: return 1;
1.147 kristaps 2738:
1.277 schwarze 2739: return 0;
1.147 kristaps 2740: }
2741:
1.180 schwarze 2742: static void
2743: roff_freereg(struct roffreg *reg)
1.147 kristaps 2744: {
1.180 schwarze 2745: struct roffreg *old_reg;
1.147 kristaps 2746:
1.180 schwarze 2747: while (NULL != reg) {
2748: free(reg->key.p);
2749: old_reg = reg;
2750: reg = reg->next;
2751: free(old_reg);
2752: }
1.147 kristaps 2753: }
1.92 schwarze 2754:
2755: static enum rofferr
1.89 kristaps 2756: roff_nr(ROFF_ARGS)
1.83 schwarze 2757: {
1.327 schwarze 2758: char *key, *val, *step;
1.212 schwarze 2759: size_t keysz;
1.327 schwarze 2760: int iv, is, len;
1.187 schwarze 2761: char sign;
1.89 kristaps 2762:
1.238 schwarze 2763: key = val = buf->buf + pos;
2764: if (*key == '\0')
1.277 schwarze 2765: return ROFF_IGN;
1.212 schwarze 2766:
2767: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2768: if (key[keysz] == '\\')
1.277 schwarze 2769: return ROFF_IGN;
1.89 kristaps 2770:
1.187 schwarze 2771: sign = *val;
1.238 schwarze 2772: if (sign == '+' || sign == '-')
1.187 schwarze 2773: val++;
2774:
1.327 schwarze 2775: len = 0;
2776: if (roff_evalnum(r, ln, val, &len, &iv, ROFFNUM_SCALE) == 0)
2777: return ROFF_IGN;
2778:
2779: step = val + len;
2780: while (isspace((unsigned char)*step))
2781: step++;
2782: if (roff_evalnum(r, ln, step, NULL, &is, 0) == 0)
2783: is = INT_MIN;
1.109 kristaps 2784:
1.327 schwarze 2785: roff_setregn(r, key, keysz, iv, sign, is);
1.277 schwarze 2786: return ROFF_IGN;
1.203 schwarze 2787: }
2788:
2789: static enum rofferr
2790: roff_rr(ROFF_ARGS)
2791: {
2792: struct roffreg *reg, **prev;
1.212 schwarze 2793: char *name, *cp;
2794: size_t namesz;
1.203 schwarze 2795:
1.238 schwarze 2796: name = cp = buf->buf + pos;
2797: if (*name == '\0')
1.277 schwarze 2798: return ROFF_IGN;
1.212 schwarze 2799: namesz = roff_getname(r, &cp, ln, pos);
2800: name[namesz] = '\0';
1.203 schwarze 2801:
2802: prev = &r->regtab;
2803: while (1) {
2804: reg = *prev;
1.238 schwarze 2805: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2806: break;
2807: prev = ®->next;
2808: }
1.238 schwarze 2809: if (reg != NULL) {
1.203 schwarze 2810: *prev = reg->next;
2811: free(reg->key.p);
2812: free(reg);
2813: }
1.277 schwarze 2814: return ROFF_IGN;
1.122 schwarze 2815: }
2816:
1.266 schwarze 2817: /* --- handler functions for roff requests -------------------------------- */
2818:
1.122 schwarze 2819: static enum rofferr
2820: roff_rm(ROFF_ARGS)
2821: {
2822: const char *name;
2823: char *cp;
1.212 schwarze 2824: size_t namesz;
1.122 schwarze 2825:
1.238 schwarze 2826: cp = buf->buf + pos;
2827: while (*cp != '\0') {
1.212 schwarze 2828: name = cp;
1.238 schwarze 2829: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2830: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.311 schwarze 2831: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.238 schwarze 2832: if (name[namesz] == '\\')
1.212 schwarze 2833: break;
1.122 schwarze 2834: }
1.277 schwarze 2835: return ROFF_IGN;
1.178 schwarze 2836: }
2837:
2838: static enum rofferr
2839: roff_it(ROFF_ARGS)
2840: {
2841: int iv;
2842:
2843: /* Parse the number of lines. */
1.261 schwarze 2844:
2845: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 2846: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2847: ln, ppos, buf->buf + 1);
1.277 schwarze 2848: return ROFF_IGN;
1.178 schwarze 2849: }
2850:
1.262 schwarze 2851: while (isspace((unsigned char)buf->buf[pos]))
2852: pos++;
2853:
2854: /*
2855: * Arm the input line trap.
2856: * Special-casing "an-trap" is an ugly workaround to cope
2857: * with DocBook stupidly fiddling with man(7) internals.
2858: */
1.261 schwarze 2859:
1.178 schwarze 2860: roffit_lines = iv;
1.262 schwarze 2861: roffit_macro = mandoc_strdup(iv != 1 ||
2862: strcmp(buf->buf + pos, "an-trap") ?
2863: buf->buf + pos : "br");
1.277 schwarze 2864: return ROFF_IGN;
1.175 schwarze 2865: }
2866:
2867: static enum rofferr
2868: roff_Dd(ROFF_ARGS)
2869: {
1.315 schwarze 2870: int mask;
2871: enum roff_tok t, te;
1.227 schwarze 2872:
1.315 schwarze 2873: switch (tok) {
2874: case ROFF_Dd:
2875: tok = MDOC_Dd;
2876: te = MDOC_MAX;
2877: if (r->format == 0)
2878: r->format = MPARSE_MDOC;
2879: mask = MPARSE_MDOC | MPARSE_QUICK;
2880: break;
2881: case ROFF_TH:
2882: tok = MAN_TH;
2883: te = MAN_MAX;
2884: if (r->format == 0)
2885: r->format = MPARSE_MAN;
2886: mask = MPARSE_QUICK;
2887: break;
2888: default:
2889: abort();
2890: }
2891: if ((r->options & mask) == 0)
2892: for (t = tok; t < te; t++)
2893: roff_setstr(r, roff_name[t], NULL, 0);
1.277 schwarze 2894: return ROFF_CONT;
1.109 kristaps 2895: }
2896:
2897: static enum rofferr
2898: roff_TE(ROFF_ARGS)
2899: {
1.321 schwarze 2900: if (r->tbl == NULL) {
1.221 schwarze 2901: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2902: ln, ppos, "TE");
1.321 schwarze 2903: return ROFF_IGN;
2904: }
2905: if (tbl_end(r->tbl) == 0) {
2906: r->tbl = NULL;
1.258 schwarze 2907: free(buf->buf);
2908: buf->buf = mandoc_strdup(".sp");
2909: buf->sz = 4;
1.329 schwarze 2910: *offs = 0;
1.277 schwarze 2911: return ROFF_REPARSE;
1.258 schwarze 2912: }
1.321 schwarze 2913: r->tbl = NULL;
1.277 schwarze 2914: return ROFF_IGN;
1.112 kristaps 2915: }
2916:
2917: static enum rofferr
2918: roff_T_(ROFF_ARGS)
2919: {
2920:
2921: if (NULL == r->tbl)
1.221 schwarze 2922: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2923: ln, ppos, "T&");
1.112 kristaps 2924: else
1.296 schwarze 2925: tbl_restart(ln, ppos, r->tbl);
1.112 kristaps 2926:
1.277 schwarze 2927: return ROFF_IGN;
1.109 kristaps 2928: }
2929:
1.230 schwarze 2930: /*
2931: * Handle in-line equation delimiters.
2932: */
2933: static enum rofferr
1.238 schwarze 2934: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2935: {
1.233 schwarze 2936: char *cp1, *cp2;
2937: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2938:
1.230 schwarze 2939: /*
2940: * Outside equations, look for an opening delimiter.
2941: * If we are inside an equation, we already know it is
2942: * in-line, or this function wouldn't have been called;
2943: * so look for a closing delimiter.
2944: */
2945:
1.238 schwarze 2946: cp1 = buf->buf + pos;
1.230 schwarze 2947: cp2 = strchr(cp1, r->eqn == NULL ?
2948: r->last_eqn->odelim : r->last_eqn->cdelim);
2949: if (cp2 == NULL)
1.277 schwarze 2950: return ROFF_CONT;
1.230 schwarze 2951:
1.233 schwarze 2952: *cp2++ = '\0';
2953: bef_pr = bef_nl = aft_nl = aft_pr = "";
2954:
2955: /* Handle preceding text, protecting whitespace. */
2956:
1.238 schwarze 2957: if (*buf->buf != '\0') {
1.233 schwarze 2958: if (r->eqn == NULL)
2959: bef_pr = "\\&";
2960: bef_nl = "\n";
2961: }
2962:
2963: /*
2964: * Prepare replacing the delimiter with an equation macro
2965: * and drop leading white space from the equation.
2966: */
1.230 schwarze 2967:
1.233 schwarze 2968: if (r->eqn == NULL) {
2969: while (*cp2 == ' ')
2970: cp2++;
2971: mac = ".EQ";
2972: } else
2973: mac = ".EN";
2974:
2975: /* Handle following text, protecting whitespace. */
2976:
2977: if (*cp2 != '\0') {
2978: aft_nl = "\n";
2979: if (r->eqn != NULL)
2980: aft_pr = "\\&";
2981: }
2982:
2983: /* Do the actual replacement. */
2984:
1.238 schwarze 2985: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2986: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2987: free(buf->buf);
2988: buf->buf = cp1;
1.230 schwarze 2989:
2990: /* Toggle the in-line state of the eqn subsystem. */
2991:
2992: r->eqn_inline = r->eqn == NULL;
1.277 schwarze 2993: return ROFF_REPARSE;
1.151 kristaps 2994: }
2995:
1.235 schwarze 2996: static enum rofferr
2997: roff_EQ(ROFF_ARGS)
1.125 kristaps 2998: {
1.319 schwarze 2999: struct roff_node *n;
3000:
1.322 schwarze 3001: if (r->man->macroset == MACROSET_MAN)
3002: man_breakscope(r->man, ROFF_EQ);
1.319 schwarze 3003: n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
3004: if (ln > r->man->last->line)
3005: n->flags |= NODE_LINE;
3006: n->eqn = mandoc_calloc(1, sizeof(*n->eqn));
3007: n->eqn->expectargs = UINT_MAX;
3008: roff_node_append(r->man, n);
3009: r->man->next = ROFF_NEXT_SIBLING;
1.125 kristaps 3010:
1.238 schwarze 3011: assert(r->eqn == NULL);
1.319 schwarze 3012: if (r->last_eqn == NULL)
3013: r->last_eqn = eqn_alloc(r->parse);
3014: else
3015: eqn_reset(r->last_eqn);
3016: r->eqn = r->last_eqn;
3017: r->eqn->node = n;
1.151 kristaps 3018:
1.238 schwarze 3019: if (buf->buf[pos] != '\0')
1.235 schwarze 3020: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 3021: ".EQ %s", buf->buf + pos);
1.151 kristaps 3022:
1.277 schwarze 3023: return ROFF_IGN;
1.125 kristaps 3024: }
3025:
3026: static enum rofferr
3027: roff_EN(ROFF_ARGS)
3028: {
1.319 schwarze 3029: if (r->eqn != NULL) {
3030: eqn_parse(r->eqn);
3031: r->eqn = NULL;
3032: } else
3033: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
3034: if (buf->buf[pos] != '\0')
3035: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
3036: "EN %s", buf->buf + pos);
1.277 schwarze 3037: return ROFF_IGN;
1.125 kristaps 3038: }
3039:
3040: static enum rofferr
1.109 kristaps 3041: roff_TS(ROFF_ARGS)
3042: {
1.321 schwarze 3043: if (r->tbl != NULL) {
1.221 schwarze 3044: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
3045: ln, ppos, "TS breaks TS");
1.321 schwarze 3046: tbl_end(r->tbl);
1.115 kristaps 3047: }
1.321 schwarze 3048: r->tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 3049: if (r->last_tbl)
1.321 schwarze 3050: r->last_tbl->next = r->tbl;
1.113 kristaps 3051: else
1.321 schwarze 3052: r->first_tbl = r->tbl;
3053: r->last_tbl = r->tbl;
1.297 schwarze 3054: return ROFF_IGN;
3055: }
3056:
3057: static enum rofferr
3058: roff_onearg(ROFF_ARGS)
3059: {
3060: struct roff_node *n;
3061: char *cp;
1.305 schwarze 3062: int npos;
1.297 schwarze 3063:
1.302 schwarze 3064: if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.320 schwarze 3065: (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
3066: tok == ROFF_ti))
1.302 schwarze 3067: man_breakscope(r->man, tok);
3068:
1.309 schwarze 3069: if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.305 schwarze 3070: r->man->last = roffce_node;
3071: r->man->next = ROFF_NEXT_SIBLING;
3072: }
3073:
1.297 schwarze 3074: roff_elem_alloc(r->man, ln, ppos, tok);
3075: n = r->man->last;
3076:
3077: cp = buf->buf + pos;
3078: if (*cp != '\0') {
3079: while (*cp != '\0' && *cp != ' ')
3080: cp++;
3081: while (*cp == ' ')
3082: *cp++ = '\0';
3083: if (*cp != '\0')
3084: mandoc_vmsg(MANDOCERR_ARG_EXCESS,
3085: r->parse, ln, cp - buf->buf,
3086: "%s ... %s", roff_name[tok], cp);
3087: roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.300 schwarze 3088: }
3089:
1.309 schwarze 3090: if (tok == ROFF_ce || tok == ROFF_rj) {
3091: if (r->man->last->type == ROFFT_ELEM) {
1.305 schwarze 3092: roff_word_alloc(r->man, ln, pos, "1");
3093: r->man->last->flags |= NODE_NOSRC;
3094: }
3095: npos = 0;
3096: if (roff_evalnum(r, ln, r->man->last->string, &npos,
3097: &roffce_lines, 0) == 0) {
3098: mandoc_vmsg(MANDOCERR_CE_NONUM,
3099: r->parse, ln, pos, "ce %s", buf->buf + pos);
3100: roffce_lines = 1;
3101: }
3102: if (roffce_lines < 1) {
3103: r->man->last = r->man->last->parent;
3104: roffce_node = NULL;
3105: roffce_lines = 0;
3106: } else
3107: roffce_node = r->man->last->parent;
3108: } else {
3109: n->flags |= NODE_VALID | NODE_ENDED;
3110: r->man->last = n;
3111: }
3112: n->flags |= NODE_LINE;
1.300 schwarze 3113: r->man->next = ROFF_NEXT_SIBLING;
3114: return ROFF_IGN;
3115: }
3116:
3117: static enum rofferr
3118: roff_manyarg(ROFF_ARGS)
3119: {
3120: struct roff_node *n;
3121: char *sp, *ep;
3122:
3123: roff_elem_alloc(r->man, ln, ppos, tok);
3124: n = r->man->last;
3125:
3126: for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
3127: while (*ep != '\0' && *ep != ' ')
3128: ep++;
3129: while (*ep == ' ')
3130: *ep++ = '\0';
3131: roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.297 schwarze 3132: }
3133:
3134: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3135: r->man->last = n;
3136: r->man->next = ROFF_NEXT_SIBLING;
1.277 schwarze 3137: return ROFF_IGN;
1.251 schwarze 3138: }
3139:
3140: static enum rofferr
1.311 schwarze 3141: roff_als(ROFF_ARGS)
3142: {
3143: char *oldn, *newn, *end, *value;
3144: size_t oldsz, newsz, valsz;
3145:
3146: newn = oldn = buf->buf + pos;
3147: if (*newn == '\0')
3148: return ROFF_IGN;
3149:
3150: newsz = roff_getname(r, &oldn, ln, pos);
3151: if (newn[newsz] == '\\' || *oldn == '\0')
3152: return ROFF_IGN;
3153:
3154: end = oldn;
3155: oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
3156: if (oldsz == 0)
3157: return ROFF_IGN;
3158:
1.315 schwarze 3159: valsz = mandoc_asprintf(&value, ".%.*s \\$*\\\"\n",
3160: (int)oldsz, oldn);
1.311 schwarze 3161: roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
3162: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3163: free(value);
3164: return ROFF_IGN;
3165: }
3166:
3167: static enum rofferr
1.296 schwarze 3168: roff_br(ROFF_ARGS)
1.251 schwarze 3169: {
1.302 schwarze 3170: if (r->man->flags & (MAN_BLINE | MAN_ELINE))
3171: man_breakscope(r->man, ROFF_br);
1.296 schwarze 3172: roff_elem_alloc(r->man, ln, ppos, ROFF_br);
3173: if (buf->buf[pos] != '\0')
3174: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
3175: "%s %s", roff_name[tok], buf->buf + pos);
3176: r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3177: r->man->next = ROFF_NEXT_SIBLING;
3178: return ROFF_IGN;
1.92 schwarze 3179: }
3180:
1.105 kristaps 3181: static enum rofferr
1.174 kristaps 3182: roff_cc(ROFF_ARGS)
3183: {
3184: const char *p;
3185:
1.238 schwarze 3186: p = buf->buf + pos;
1.174 kristaps 3187:
1.238 schwarze 3188: if (*p == '\0' || (r->control = *p++) == '.')
1.303 schwarze 3189: r->control = '\0';
1.174 kristaps 3190:
1.238 schwarze 3191: if (*p != '\0')
1.260 schwarze 3192: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3193: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 3194:
1.277 schwarze 3195: return ROFF_IGN;
1.174 kristaps 3196: }
3197:
3198: static enum rofferr
1.303 schwarze 3199: roff_ec(ROFF_ARGS)
3200: {
3201: const char *p;
3202:
3203: p = buf->buf + pos;
3204: if (*p == '\0')
3205: r->escape = '\\';
3206: else {
3207: r->escape = *p;
3208: if (*++p != '\0')
3209: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3210: ln, p - buf->buf, "ec ... %s", p);
3211: }
3212: return ROFF_IGN;
3213: }
3214:
3215: static enum rofferr
3216: roff_eo(ROFF_ARGS)
3217: {
3218: r->escape = '\0';
3219: if (buf->buf[pos] != '\0')
3220: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
3221: ln, pos, "eo %s", buf->buf + pos);
3222: return ROFF_IGN;
1.330 schwarze 3223: }
3224:
3225: static enum rofferr
3226: roff_nop(ROFF_ARGS)
3227: {
3228: while (buf->buf[pos] == ' ')
3229: pos++;
3230: *offs = pos;
3231: return ROFF_RERUN;
1.303 schwarze 3232: }
3233:
3234: static enum rofferr
1.164 kristaps 3235: roff_tr(ROFF_ARGS)
3236: {
3237: const char *p, *first, *second;
3238: size_t fsz, ssz;
3239: enum mandoc_esc esc;
3240:
1.238 schwarze 3241: p = buf->buf + pos;
1.164 kristaps 3242:
1.238 schwarze 3243: if (*p == '\0') {
1.260 schwarze 3244: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.277 schwarze 3245: return ROFF_IGN;
1.164 kristaps 3246: }
3247:
1.238 schwarze 3248: while (*p != '\0') {
1.164 kristaps 3249: fsz = ssz = 1;
3250:
3251: first = p++;
1.238 schwarze 3252: if (*first == '\\') {
1.164 kristaps 3253: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3254: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3255: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3256: ln, (int)(p - buf->buf), first);
1.277 schwarze 3257: return ROFF_IGN;
1.164 kristaps 3258: }
3259: fsz = (size_t)(p - first);
3260: }
3261:
3262: second = p++;
1.238 schwarze 3263: if (*second == '\\') {
1.164 kristaps 3264: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3265: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3266: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3267: ln, (int)(p - buf->buf), second);
1.277 schwarze 3268: return ROFF_IGN;
1.164 kristaps 3269: }
3270: ssz = (size_t)(p - second);
1.238 schwarze 3271: } else if (*second == '\0') {
1.260 schwarze 3272: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
3273: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 3274: second = " ";
1.165 kristaps 3275: p--;
1.164 kristaps 3276: }
3277:
1.167 kristaps 3278: if (fsz > 1) {
1.207 schwarze 3279: roff_setstrn(&r->xmbtab, first, fsz,
3280: second, ssz, 0);
1.167 kristaps 3281: continue;
3282: }
3283:
1.238 schwarze 3284: if (r->xtab == NULL)
1.207 schwarze 3285: r->xtab = mandoc_calloc(128,
3286: sizeof(struct roffstr));
1.167 kristaps 3287:
3288: free(r->xtab[(int)*first].p);
3289: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
3290: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 3291: }
3292:
1.277 schwarze 3293: return ROFF_IGN;
1.164 kristaps 3294: }
3295:
3296: static enum rofferr
1.306 schwarze 3297: roff_rn(ROFF_ARGS)
3298: {
3299: const char *value;
3300: char *oldn, *newn, *end;
3301: size_t oldsz, newsz;
1.315 schwarze 3302: int deftype;
1.306 schwarze 3303:
3304: oldn = newn = buf->buf + pos;
3305: if (*oldn == '\0')
3306: return ROFF_IGN;
3307:
3308: oldsz = roff_getname(r, &newn, ln, pos);
3309: if (oldn[oldsz] == '\\' || *newn == '\0')
3310: return ROFF_IGN;
3311:
3312: end = newn;
3313: newsz = roff_getname(r, &end, ln, newn - buf->buf);
3314: if (newsz == 0)
3315: return ROFF_IGN;
3316:
1.315 schwarze 3317: deftype = ROFFDEF_ANY;
3318: value = roff_getstrn(r, oldn, oldsz, &deftype);
3319: switch (deftype) {
3320: case ROFFDEF_USER:
1.306 schwarze 3321: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3322: roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
3323: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.315 schwarze 3324: break;
3325: case ROFFDEF_PRE:
3326: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3327: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3328: break;
3329: case ROFFDEF_REN:
1.306 schwarze 3330: roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
3331: roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.315 schwarze 3332: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3333: break;
3334: case ROFFDEF_STD:
1.306 schwarze 3335: roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.315 schwarze 3336: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3337: break;
3338: default:
3339: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3340: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3341: break;
3342: }
1.306 schwarze 3343: return ROFF_IGN;
3344: }
3345:
3346: static enum rofferr
1.105 kristaps 3347: roff_so(ROFF_ARGS)
3348: {
1.249 schwarze 3349: char *name, *cp;
1.105 kristaps 3350:
1.238 schwarze 3351: name = buf->buf + pos;
1.224 schwarze 3352: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 3353:
3354: /*
3355: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3356: * opening anything that's not in our cwd or anything beneath
3357: * it. Thus, explicitly disallow traversing up the file-system
3358: * or using absolute paths.
3359: */
3360:
1.238 schwarze 3361: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 3362: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3363: ".so %s", name);
1.249 schwarze 3364: buf->sz = mandoc_asprintf(&cp,
3365: ".sp\nSee the file %s.\n.sp", name) + 1;
3366: free(buf->buf);
3367: buf->buf = cp;
3368: *offs = 0;
1.277 schwarze 3369: return ROFF_REPARSE;
1.105 kristaps 3370: }
3371:
3372: *offs = pos;
1.277 schwarze 3373: return ROFF_SO;
1.105 kristaps 3374: }
1.92 schwarze 3375:
1.266 schwarze 3376: /* --- user defined strings and macros ------------------------------------ */
3377:
1.106 kristaps 3378: static enum rofferr
3379: roff_userdef(ROFF_ARGS)
1.99 kristaps 3380: {
1.311 schwarze 3381: const char *arg[16], *ap;
1.106 kristaps 3382: char *cp, *n1, *n2;
1.337 ! schwarze 3383: int argc, expand_count, i, ib, ie;
! 3384: size_t asz, esz, rsz;
1.106 kristaps 3385:
3386: /*
3387: * Collect pointers to macro argument strings
1.188 schwarze 3388: * and NUL-terminate them.
1.106 kristaps 3389: */
1.263 schwarze 3390:
1.337 ! schwarze 3391: argc = 0;
1.238 schwarze 3392: cp = buf->buf + pos;
1.311 schwarze 3393: for (i = 0; i < 16; i++) {
1.273 schwarze 3394: if (*cp == '\0')
3395: arg[i] = "";
3396: else {
3397: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
1.337 ! schwarze 3398: argc = i + 1;
1.273 schwarze 3399: }
3400: }
1.99 kristaps 3401:
1.106 kristaps 3402: /*
3403: * Expand macro arguments.
1.99 kristaps 3404: */
1.263 schwarze 3405:
3406: buf->sz = strlen(r->current_string) + 1;
1.292 schwarze 3407: n1 = n2 = cp = mandoc_malloc(buf->sz);
1.263 schwarze 3408: memcpy(n1, r->current_string, buf->sz);
1.292 schwarze 3409: expand_count = 0;
1.263 schwarze 3410: while (*cp != '\0') {
3411:
3412: /* Scan ahead for the next argument invocation. */
3413:
3414: if (*cp++ != '\\')
3415: continue;
3416: if (*cp++ != '$')
3417: continue;
1.274 schwarze 3418: if (*cp == '*') { /* \\$* inserts all arguments */
3419: ib = 0;
1.337 ! schwarze 3420: ie = argc - 1;
1.274 schwarze 3421: } else { /* \\$1 .. \\$9 insert one argument */
3422: ib = ie = *cp - '1';
3423: if (ib < 0 || ib > 8)
3424: continue;
3425: }
1.263 schwarze 3426: cp -= 2;
1.292 schwarze 3427:
3428: /*
3429: * Prevent infinite recursion.
3430: */
3431:
3432: if (cp >= n2)
3433: expand_count = 1;
3434: else if (++expand_count > EXPAND_LIMIT) {
3435: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
3436: ln, (int)(cp - n1), NULL);
1.293 schwarze 3437: free(buf->buf);
3438: buf->buf = n1;
1.329 schwarze 3439: *offs = 0;
1.292 schwarze 3440: return ROFF_IGN;
3441: }
1.263 schwarze 3442:
3443: /*
3444: * Determine the size of the expanded argument,
3445: * taking escaping of quotes into account.
3446: */
3447:
1.274 schwarze 3448: asz = ie > ib ? ie - ib : 0; /* for blanks */
3449: for (i = ib; i <= ie; i++) {
3450: for (ap = arg[i]; *ap != '\0'; ap++) {
3451: asz++;
3452: if (*ap == '"')
3453: asz += 3;
3454: }
1.263 schwarze 3455: }
3456: if (asz != 3) {
3457:
3458: /*
3459: * Determine the size of the rest of the
3460: * unexpanded macro, including the NUL.
3461: */
3462:
3463: rsz = buf->sz - (cp - n1) - 3;
3464:
3465: /*
3466: * When shrinking, move before
3467: * releasing the storage.
3468: */
3469:
3470: if (asz < 3)
3471: memmove(cp + asz, cp + 3, rsz);
3472:
3473: /*
3474: * Resize the storage for the macro
3475: * and readjust the parse pointer.
3476: */
3477:
3478: buf->sz += asz - 3;
3479: n2 = mandoc_realloc(n1, buf->sz);
3480: cp = n2 + (cp - n1);
3481: n1 = n2;
3482:
3483: /*
3484: * When growing, make room
3485: * for the expanded argument.
3486: */
3487:
3488: if (asz > 3)
3489: memmove(cp + asz, cp + 3, rsz);
3490: }
3491:
3492: /* Copy the expanded argument, escaping quotes. */
3493:
3494: n2 = cp;
1.274 schwarze 3495: for (i = ib; i <= ie; i++) {
3496: for (ap = arg[i]; *ap != '\0'; ap++) {
3497: if (*ap == '"') {
3498: memcpy(n2, "\\(dq", 4);
3499: n2 += 4;
3500: } else
3501: *n2++ = *ap;
3502: }
3503: if (i < ie)
3504: *n2++ = ' ';
1.106 kristaps 3505: }
1.337 ! schwarze 3506: }
! 3507:
! 3508: /*
! 3509: * Expand the number of arguments, if it is used.
! 3510: * This never makes the expanded macro longer.
! 3511: */
! 3512:
! 3513: for (cp = n1; *cp != '\0'; cp++) {
! 3514: if (cp[0] != '\\')
! 3515: continue;
! 3516: if (cp[1] == '\\') {
! 3517: cp++;
! 3518: continue;
! 3519: }
! 3520: if (strncmp(cp + 1, "n(.$", 4) == 0)
! 3521: esz = 5;
! 3522: else if (strncmp(cp + 1, "n[.$]", 5) == 0)
! 3523: esz = 6;
! 3524: else
! 3525: continue;
! 3526: asz = snprintf(cp, esz, "%d", argc);
! 3527: assert(asz < esz);
! 3528: rsz = buf->sz - (cp - n1) - esz;
! 3529: memmove(cp + asz, cp + esz, rsz);
! 3530: buf->sz -= esz - asz;
! 3531: n2 = mandoc_realloc(n1, buf->sz);
! 3532: cp = n2 + (cp - n1) + asz;
! 3533: n1 = n2;
1.99 kristaps 3534: }
3535:
1.106 kristaps 3536: /*
3537: * Replace the macro invocation
3538: * by the expanded macro.
3539: */
1.263 schwarze 3540:
1.238 schwarze 3541: free(buf->buf);
3542: buf->buf = n1;
1.248 schwarze 3543: *offs = 0;
1.106 kristaps 3544:
1.277 schwarze 3545: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
3546: ROFF_REPARSE : ROFF_APPEND;
1.99 kristaps 3547: }
1.121 schwarze 3548:
1.306 schwarze 3549: /*
3550: * Calling a high-level macro that was renamed with .rn.
3551: * r->current_string has already been set up by roff_parse().
3552: */
3553: static enum rofferr
3554: roff_renamed(ROFF_ARGS)
3555: {
3556: char *nbuf;
3557:
1.315 schwarze 3558: buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
3559: buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.306 schwarze 3560: free(buf->buf);
3561: buf->buf = nbuf;
1.329 schwarze 3562: *offs = 0;
1.306 schwarze 3563: return ROFF_CONT;
3564: }
3565:
1.212 schwarze 3566: static size_t
1.121 schwarze 3567: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3568: {
3569: char *name, *cp;
1.212 schwarze 3570: size_t namesz;
1.121 schwarze 3571:
3572: name = *cpp;
3573: if ('\0' == *name)
1.277 schwarze 3574: return 0;
1.121 schwarze 3575:
1.212 schwarze 3576: /* Read until end of name and terminate it with NUL. */
3577: for (cp = name; 1; cp++) {
3578: if ('\0' == *cp || ' ' == *cp) {
3579: namesz = cp - name;
3580: break;
3581: }
1.121 schwarze 3582: if ('\\' != *cp)
3583: continue;
1.215 schwarze 3584: namesz = cp - name;
3585: if ('{' == cp[1] || '}' == cp[1])
3586: break;
1.121 schwarze 3587: cp++;
3588: if ('\\' == *cp)
3589: continue;
1.224 schwarze 3590: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3591: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 3592: mandoc_escape((const char **)&cp, NULL, NULL);
3593: break;
1.121 schwarze 3594: }
3595:
3596: /* Read past spaces. */
3597: while (' ' == *cp)
3598: cp++;
3599:
3600: *cpp = cp;
1.277 schwarze 3601: return namesz;
1.121 schwarze 3602: }
3603:
1.106 kristaps 3604: /*
3605: * Store *string into the user-defined string called *name.
3606: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 3607: * append == 0: replace mode
3608: * append == 1: single-line append mode
3609: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 3610: */
1.94 kristaps 3611: static void
1.106 kristaps 3612: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 3613: int append)
1.92 schwarze 3614: {
1.315 schwarze 3615: size_t namesz;
1.164 kristaps 3616:
1.315 schwarze 3617: namesz = strlen(name);
3618: roff_setstrn(&r->strtab, name, namesz, string,
1.207 schwarze 3619: string ? strlen(string) : 0, append);
1.315 schwarze 3620: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.164 kristaps 3621: }
3622:
3623: static void
1.166 kristaps 3624: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 3625: const char *string, size_t stringsz, int append)
1.164 kristaps 3626: {
1.166 kristaps 3627: struct roffkv *n;
1.164 kristaps 3628: char *c;
3629: int i;
3630: size_t oldch, newch;
1.92 schwarze 3631:
1.106 kristaps 3632: /* Search for an existing string with the same name. */
1.164 kristaps 3633: n = *r;
3634:
1.211 schwarze 3635: while (n && (namesz != n->key.sz ||
3636: strncmp(n->key.p, name, namesz)))
1.92 schwarze 3637: n = n->next;
1.94 kristaps 3638:
3639: if (NULL == n) {
1.106 kristaps 3640: /* Create a new string table entry. */
1.166 kristaps 3641: n = mandoc_malloc(sizeof(struct roffkv));
3642: n->key.p = mandoc_strndup(name, namesz);
3643: n->key.sz = namesz;
3644: n->val.p = NULL;
3645: n->val.sz = 0;
1.164 kristaps 3646: n->next = *r;
3647: *r = n;
1.193 schwarze 3648: } else if (0 == append) {
1.166 kristaps 3649: free(n->val.p);
3650: n->val.p = NULL;
3651: n->val.sz = 0;
1.106 kristaps 3652: }
3653:
3654: if (NULL == string)
3655: return;
3656:
3657: /*
3658: * One additional byte for the '\n' in multiline mode,
3659: * and one for the terminating '\0'.
3660: */
1.193 schwarze 3661: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 3662:
1.166 kristaps 3663: if (NULL == n->val.p) {
3664: n->val.p = mandoc_malloc(newch);
3665: *n->val.p = '\0';
1.106 kristaps 3666: oldch = 0;
3667: } else {
1.166 kristaps 3668: oldch = n->val.sz;
3669: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 3670: }
3671:
3672: /* Skip existing content in the destination buffer. */
1.166 kristaps 3673: c = n->val.p + (int)oldch;
1.106 kristaps 3674:
3675: /* Append new content to the destination buffer. */
1.164 kristaps 3676: i = 0;
3677: while (i < (int)stringsz) {
1.106 kristaps 3678: /*
3679: * Rudimentary roff copy mode:
3680: * Handle escaped backslashes.
3681: */
1.164 kristaps 3682: if ('\\' == string[i] && '\\' == string[i + 1])
3683: i++;
3684: *c++ = string[i++];
1.106 kristaps 3685: }
1.94 kristaps 3686:
1.106 kristaps 3687: /* Append terminating bytes. */
1.193 schwarze 3688: if (1 < append)
1.106 kristaps 3689: *c++ = '\n';
1.163 kristaps 3690:
1.106 kristaps 3691: *c = '\0';
1.166 kristaps 3692: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 3693: }
3694:
1.94 kristaps 3695: static const char *
1.325 schwarze 3696: roff_getstrn(struct roff *r, const char *name, size_t len,
1.315 schwarze 3697: int *deftype)
1.92 schwarze 3698: {
1.315 schwarze 3699: const struct roffkv *n;
1.325 schwarze 3700: int found, i;
1.315 schwarze 3701: enum roff_tok tok;
3702:
1.325 schwarze 3703: found = 0;
3704: for (n = r->strtab; n != NULL; n = n->next) {
3705: if (strncmp(name, n->key.p, len) != 0 ||
3706: n->key.p[len] != '\0' || n->val.p == NULL)
3707: continue;
3708: if (*deftype & ROFFDEF_USER) {
3709: *deftype = ROFFDEF_USER;
3710: return n->val.p;
3711: } else {
3712: found = 1;
3713: break;
3714: }
3715: }
3716: for (n = r->rentab; n != NULL; n = n->next) {
3717: if (strncmp(name, n->key.p, len) != 0 ||
3718: n->key.p[len] != '\0' || n->val.p == NULL)
3719: continue;
3720: if (*deftype & ROFFDEF_REN) {
3721: *deftype = ROFFDEF_REN;
3722: return n->val.p;
3723: } else {
3724: found = 1;
3725: break;
1.315 schwarze 3726: }
3727: }
1.325 schwarze 3728: for (i = 0; i < PREDEFS_MAX; i++) {
3729: if (strncmp(name, predefs[i].name, len) != 0 ||
3730: predefs[i].name[len] != '\0')
3731: continue;
3732: if (*deftype & ROFFDEF_PRE) {
3733: *deftype = ROFFDEF_PRE;
3734: return predefs[i].str;
3735: } else {
3736: found = 1;
3737: break;
1.315 schwarze 3738: }
3739: }
1.325 schwarze 3740: if (r->man->macroset != MACROSET_MAN) {
3741: for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
3742: if (strncmp(name, roff_name[tok], len) != 0 ||
3743: roff_name[tok][len] != '\0')
3744: continue;
3745: if (*deftype & ROFFDEF_STD) {
3746: *deftype = ROFFDEF_STD;
3747: return NULL;
3748: } else {
3749: found = 1;
3750: break;
1.315 schwarze 3751: }
3752: }
3753: }
1.325 schwarze 3754: if (r->man->macroset != MACROSET_MDOC) {
3755: for (tok = MAN_TH; tok < MAN_MAX; tok++) {
3756: if (strncmp(name, roff_name[tok], len) != 0 ||
3757: roff_name[tok][len] != '\0')
3758: continue;
3759: if (*deftype & ROFFDEF_STD) {
3760: *deftype = ROFFDEF_STD;
3761: return NULL;
3762: } else {
3763: found = 1;
3764: break;
1.315 schwarze 3765: }
3766: }
1.325 schwarze 3767: }
3768:
3769: if (found == 0 && *deftype != ROFFDEF_ANY) {
3770: if (*deftype & ROFFDEF_REN) {
3771: /*
3772: * This might still be a request,
3773: * so do not treat it as undefined yet.
3774: */
3775: *deftype = ROFFDEF_UNDEF;
3776: return NULL;
1.315 schwarze 3777: }
1.325 schwarze 3778:
3779: /* Using an undefined string defines it to be empty. */
3780:
3781: roff_setstrn(&r->strtab, name, len, "", 0, 0);
3782: roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.315 schwarze 3783: }
1.325 schwarze 3784:
1.315 schwarze 3785: *deftype = 0;
1.277 schwarze 3786: return NULL;
1.92 schwarze 3787: }
3788:
1.94 kristaps 3789: static void
1.167 kristaps 3790: roff_freestr(struct roffkv *r)
1.92 schwarze 3791: {
1.166 kristaps 3792: struct roffkv *n, *nn;
1.92 schwarze 3793:
1.167 kristaps 3794: for (n = r; n; n = nn) {
1.166 kristaps 3795: free(n->key.p);
3796: free(n->val.p);
1.92 schwarze 3797: nn = n->next;
3798: free(n);
3799: }
1.114 kristaps 3800: }
1.266 schwarze 3801:
3802: /* --- accessors and utility functions ------------------------------------ */
1.164 kristaps 3803:
3804: /*
3805: * Duplicate an input string, making the appropriate character
3806: * conversations (as stipulated by `tr') along the way.
3807: * Returns a heap-allocated string with all the replacements made.
3808: */
3809: char *
3810: roff_strdup(const struct roff *r, const char *p)
3811: {
1.166 kristaps 3812: const struct roffkv *cp;
1.164 kristaps 3813: char *res;
3814: const char *pp;
3815: size_t ssz, sz;
3816: enum mandoc_esc esc;
3817:
1.167 kristaps 3818: if (NULL == r->xmbtab && NULL == r->xtab)
1.277 schwarze 3819: return mandoc_strdup(p);
1.164 kristaps 3820: else if ('\0' == *p)
1.277 schwarze 3821: return mandoc_strdup("");
1.164 kristaps 3822:
3823: /*
3824: * Step through each character looking for term matches
3825: * (remember that a `tr' can be invoked with an escape, which is
3826: * a glyph but the escape is multi-character).
3827: * We only do this if the character hash has been initialised
3828: * and the string is >0 length.
3829: */
3830:
3831: res = NULL;
3832: ssz = 0;
3833:
3834: while ('\0' != *p) {
1.289 schwarze 3835: assert((unsigned int)*p < 128);
3836: if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167 kristaps 3837: sz = r->xtab[(int)*p].sz;
3838: res = mandoc_realloc(res, ssz + sz + 1);
3839: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3840: ssz += sz;
3841: p++;
3842: continue;
3843: } else if ('\\' != *p) {
3844: res = mandoc_realloc(res, ssz + 2);
3845: res[ssz++] = *p++;
3846: continue;
3847: }
3848:
1.164 kristaps 3849: /* Search for term matches. */
1.167 kristaps 3850: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 3851: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 3852: break;
3853:
3854: if (NULL != cp) {
3855: /*
3856: * A match has been found.
3857: * Append the match to the array and move
3858: * forward by its keysize.
3859: */
1.207 schwarze 3860: res = mandoc_realloc(res,
3861: ssz + cp->val.sz + 1);
1.166 kristaps 3862: memcpy(res + ssz, cp->val.p, cp->val.sz);
3863: ssz += cp->val.sz;
3864: p += (int)cp->key.sz;
1.164 kristaps 3865: continue;
3866: }
3867:
1.167 kristaps 3868: /*
3869: * Handle escapes carefully: we need to copy
3870: * over just the escape itself, or else we might
3871: * do replacements within the escape itself.
3872: * Make sure to pass along the bogus string.
3873: */
3874: pp = p++;
3875: esc = mandoc_escape(&p, NULL, NULL);
3876: if (ESCAPE_ERROR == esc) {
3877: sz = strlen(pp);
1.164 kristaps 3878: res = mandoc_realloc(res, ssz + sz + 1);
3879: memcpy(res + ssz, pp, sz);
1.167 kristaps 3880: break;
1.164 kristaps 3881: }
1.207 schwarze 3882: /*
3883: * We bail out on bad escapes.
1.167 kristaps 3884: * No need to warn: we already did so when
3885: * roff_res() was called.
3886: */
3887: sz = (int)(p - pp);
3888: res = mandoc_realloc(res, ssz + sz + 1);
3889: memcpy(res + ssz, pp, sz);
3890: ssz += sz;
1.164 kristaps 3891: }
3892:
3893: res[(int)ssz] = '\0';
1.277 schwarze 3894: return res;
1.227 schwarze 3895: }
3896:
3897: int
3898: roff_getformat(const struct roff *r)
3899: {
3900:
1.277 schwarze 3901: return r->format;
1.174 kristaps 3902: }
3903:
3904: /*
1.207 schwarze 3905: * Find out whether a line is a macro line or not.
1.174 kristaps 3906: * If it is, adjust the current position and return one; if it isn't,
3907: * return zero and don't change the current position.
3908: * If the control character has been set with `.cc', then let that grain
3909: * precedence.
3910: * This is slighly contrary to groff, where using the non-breaking
3911: * control character when `cc' has been invoked will cause the
3912: * non-breaking macro contents to be printed verbatim.
3913: */
3914: int
3915: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3916: {
3917: int pos;
3918:
3919: pos = *ppos;
3920:
1.303 schwarze 3921: if (r->control != '\0' && cp[pos] == r->control)
1.174 kristaps 3922: pos++;
1.303 schwarze 3923: else if (r->control != '\0')
1.277 schwarze 3924: return 0;
1.174 kristaps 3925: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3926: pos += 2;
3927: else if ('.' == cp[pos] || '\'' == cp[pos])
3928: pos++;
3929: else
1.277 schwarze 3930: return 0;
1.174 kristaps 3931:
3932: while (' ' == cp[pos] || '\t' == cp[pos])
3933: pos++;
3934:
3935: *ppos = pos;
1.277 schwarze 3936: return 1;
1.74 kristaps 3937: }
CVSweb
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to roff.c CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [cvsweb.bsd.lv] / mandoc