Annotation of mandoc/roff.c, Revision 1.321
1.321 ! schwarze 1: /* $Id: roff.c,v 1.320 2017/07/08 15:28:09 schwarze Exp $ */
1.1 kristaps 2: /*
1.267 schwarze 3: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.287 schwarze 4: * Copyright (c) 2010-2015, 2017 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)
49:
1.266 schwarze 50: /* --- data types --------------------------------------------------------- */
51:
1.147 kristaps 52: /*
1.167 kristaps 53: * An incredibly-simple string buffer.
54: */
1.94 kristaps 55: struct roffstr {
1.167 kristaps 56: char *p; /* nil-terminated buffer */
57: size_t sz; /* saved strlen(p) */
1.166 kristaps 58: };
59:
60: /*
1.167 kristaps 61: * A key-value roffstr pair as part of a singly-linked list.
1.166 kristaps 62: */
63: struct roffkv {
64: struct roffstr key;
65: struct roffstr val;
66: struct roffkv *next; /* next in list */
1.94 kristaps 67: };
68:
1.180 schwarze 69: /*
70: * A single number register as part of a singly-linked list.
71: */
72: struct roffreg {
73: struct roffstr key;
1.181 schwarze 74: int val;
1.180 schwarze 75: struct roffreg *next;
76: };
77:
1.295 schwarze 78: /*
79: * Association of request and macro names with token IDs.
80: */
81: struct roffreq {
82: enum roff_tok tok;
83: char name[];
84: };
85:
1.67 kristaps 86: struct roff {
1.128 kristaps 87: struct mparse *parse; /* parse point */
1.296 schwarze 88: struct roff_man *man; /* mdoc or man parser */
1.67 kristaps 89: struct roffnode *last; /* leaf of stack */
1.223 schwarze 90: int *rstack; /* stack of inverted `ie' values */
1.295 schwarze 91: struct ohash *reqtab; /* request lookup table */
1.180 schwarze 92: struct roffreg *regtab; /* number registers */
1.166 kristaps 93: struct roffkv *strtab; /* user-defined strings & macros */
1.306 schwarze 94: struct roffkv *rentab; /* renamed strings & macros */
1.167 kristaps 95: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
96: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.106 kristaps 97: const char *current_string; /* value of last called user macro */
1.118 kristaps 98: struct tbl_node *first_tbl; /* first table parsed */
99: struct tbl_node *last_tbl; /* last table parsed */
100: struct tbl_node *tbl; /* current table being parsed */
1.319 schwarze 101: struct eqn_node *last_eqn; /* equation parser */
102: struct eqn_node *eqn; /* active equation parser */
1.230 schwarze 103: int eqn_inline; /* current equation is inline */
1.223 schwarze 104: int options; /* parse options */
105: int rstacksz; /* current size limit of rstack */
106: int rstackpos; /* position in rstack */
1.227 schwarze 107: int format; /* current file in mdoc or man format */
1.273 schwarze 108: int argc; /* number of args of the last macro */
1.223 schwarze 109: char control; /* control character */
1.303 schwarze 110: char escape; /* escape character */
1.79 kristaps 111: };
112:
1.67 kristaps 113: struct roffnode {
1.294 schwarze 114: enum roff_tok tok; /* type of node */
1.67 kristaps 115: struct roffnode *parent; /* up one in stack */
116: int line; /* parse line */
117: int col; /* parse col */
1.106 kristaps 118: char *name; /* node name, e.g. macro name */
1.79 kristaps 119: char *end; /* end-rules: custom token */
120: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 121: int rule; /* current evaluation rule */
1.67 kristaps 122: };
123:
124: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.294 schwarze 125: enum roff_tok tok, /* tok of macro */ \
1.238 schwarze 126: struct buf *buf, /* input buffer */ \
1.67 kristaps 127: int ln, /* parse line */ \
1.75 kristaps 128: int ppos, /* original pos in buffer */ \
129: int pos, /* current pos in buffer */ \
1.74 kristaps 130: int *offs /* reset offset of buffer data */
1.67 kristaps 131:
132: typedef enum rofferr (*roffproc)(ROFF_ARGS);
133:
134: struct roffmac {
1.79 kristaps 135: roffproc proc; /* process new macro */
136: roffproc text; /* process as child text of macro */
137: roffproc sub; /* process as child of macro */
138: int flags;
139: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.67 kristaps 140: };
141:
1.141 kristaps 142: struct predef {
143: const char *name; /* predefined input name */
144: const char *str; /* replacement symbol */
145: };
146:
147: #define PREDEF(__name, __str) \
148: { (__name), (__str) },
149:
1.266 schwarze 150: /* --- function prototypes ------------------------------------------------ */
151:
1.155 kristaps 152: static void roffnode_cleanscope(struct roff *);
153: static void roffnode_pop(struct roff *);
1.294 schwarze 154: static void roffnode_push(struct roff *, enum roff_tok,
1.155 kristaps 155: const char *, int, int);
1.321 ! schwarze 156: static void roff_addtbl(struct roff_man *, struct tbl_node *);
1.311 schwarze 157: static enum rofferr roff_als(ROFF_ARGS);
1.80 kristaps 158: static enum rofferr roff_block(ROFF_ARGS);
159: static enum rofferr roff_block_text(ROFF_ARGS);
160: static enum rofferr roff_block_sub(ROFF_ARGS);
1.296 schwarze 161: static enum rofferr roff_br(ROFF_ARGS);
1.80 kristaps 162: static enum rofferr roff_cblock(ROFF_ARGS);
1.174 kristaps 163: static enum rofferr roff_cc(ROFF_ARGS);
1.195 schwarze 164: static void roff_ccond(struct roff *, int, int);
1.82 kristaps 165: static enum rofferr roff_cond(ROFF_ARGS);
166: static enum rofferr roff_cond_text(ROFF_ARGS);
167: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.92 schwarze 168: static enum rofferr roff_ds(ROFF_ARGS);
1.303 schwarze 169: static enum rofferr roff_ec(ROFF_ARGS);
170: static enum rofferr roff_eo(ROFF_ARGS);
1.238 schwarze 171: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.271 schwarze 172: static int roff_evalcond(struct roff *r, int, char *, int *);
1.234 kristaps 173: static int roff_evalnum(struct roff *, int,
174: const char *, int *, int *, int);
175: static int roff_evalpar(struct roff *, int,
1.261 schwarze 176: const char *, int *, int *, int);
1.198 schwarze 177: static int roff_evalstrcond(const char *, int *);
1.155 kristaps 178: static void roff_free1(struct roff *);
1.180 schwarze 179: static void roff_freereg(struct roffreg *);
1.167 kristaps 180: static void roff_freestr(struct roffkv *);
1.212 schwarze 181: static size_t roff_getname(struct roff *, char **, int, int);
1.261 schwarze 182: static int roff_getnum(const char *, int *, int *, int);
1.184 schwarze 183: static int roff_getop(const char *, int *, char *);
1.181 schwarze 184: static int roff_getregn(const struct roff *,
185: const char *, size_t);
1.273 schwarze 186: static int roff_getregro(const struct roff *,
187: const char *name);
1.207 schwarze 188: static const char *roff_getstrn(const struct roff *,
1.315 schwarze 189: const char *, size_t, int *);
1.271 schwarze 190: static int roff_hasregn(const struct roff *,
191: const char *, size_t);
1.251 schwarze 192: static enum rofferr roff_insec(ROFF_ARGS);
1.178 schwarze 193: static enum rofferr roff_it(ROFF_ARGS);
1.103 kristaps 194: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.265 schwarze 195: static void roff_man_alloc1(struct roff_man *);
196: static void roff_man_free1(struct roff_man *);
1.300 schwarze 197: static enum rofferr roff_manyarg(ROFF_ARGS);
1.89 kristaps 198: static enum rofferr roff_nr(ROFF_ARGS);
1.297 schwarze 199: static enum rofferr roff_onearg(ROFF_ARGS);
1.294 schwarze 200: static enum roff_tok roff_parse(struct roff *, char *, int *,
1.214 schwarze 201: int, int);
1.305 schwarze 202: static enum rofferr roff_parsetext(struct roff *, struct buf *,
203: int, int *);
1.306 schwarze 204: static enum rofferr roff_renamed(ROFF_ARGS);
1.238 schwarze 205: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.122 schwarze 206: static enum rofferr roff_rm(ROFF_ARGS);
1.306 schwarze 207: static enum rofferr roff_rn(ROFF_ARGS);
1.203 schwarze 208: static enum rofferr roff_rr(ROFF_ARGS);
1.94 kristaps 209: static void roff_setstr(struct roff *,
1.106 kristaps 210: const char *, const char *, int);
1.207 schwarze 211: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 212: size_t, const char *, size_t, int);
1.105 kristaps 213: static enum rofferr roff_so(ROFF_ARGS);
1.164 kristaps 214: static enum rofferr roff_tr(ROFF_ARGS);
1.175 schwarze 215: static enum rofferr roff_Dd(ROFF_ARGS);
1.109 kristaps 216: static enum rofferr roff_TE(ROFF_ARGS);
217: static enum rofferr roff_TS(ROFF_ARGS);
1.125 kristaps 218: static enum rofferr roff_EQ(ROFF_ARGS);
219: static enum rofferr roff_EN(ROFF_ARGS);
1.112 kristaps 220: static enum rofferr roff_T_(ROFF_ARGS);
1.251 schwarze 221: static enum rofferr roff_unsupp(ROFF_ARGS);
1.106 kristaps 222: static enum rofferr roff_userdef(ROFF_ARGS);
1.67 kristaps 223:
1.266 schwarze 224: /* --- constant data ------------------------------------------------------ */
225:
1.261 schwarze 226: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
227: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
228:
1.294 schwarze 229: const char *__roff_name[MAN_MAX + 1] = {
1.305 schwarze 230: "br", "ce", "ft", "ll",
1.312 schwarze 231: "mc", "po", "rj", "sp",
232: "ta", "ti", NULL,
1.294 schwarze 233: "ab", "ad", "af", "aln",
234: "als", "am", "am1", "ami",
235: "ami1", "as", "as1", "asciify",
236: "backtrace", "bd", "bleedat", "blm",
237: "box", "boxa", "bp", "BP",
238: "break", "breakchar", "brnl", "brp",
1.305 schwarze 239: "brpnl", "c2", "cc",
1.294 schwarze 240: "cf", "cflags", "ch", "char",
241: "chop", "class", "close", "CL",
242: "color", "composite", "continue", "cp",
243: "cropat", "cs", "cu", "da",
244: "dch", "Dd", "de", "de1",
245: "defcolor", "dei", "dei1", "device",
246: "devicem", "di", "do", "ds",
247: "ds1", "dwh", "dt", "ec",
248: "ecr", "ecs", "el", "em",
249: "EN", "eo", "EP", "EQ",
250: "errprint", "ev", "evc", "ex",
251: "fallback", "fam", "fc", "fchar",
252: "fcolor", "fdeferlig", "feature", "fkern",
253: "fl", "flig", "fp", "fps",
254: "fschar", "fspacewidth", "fspecial", "ftr",
255: "fzoom", "gcolor", "hc", "hcode",
256: "hidechar", "hla", "hlm", "hpf",
257: "hpfa", "hpfcode", "hw", "hy",
258: "hylang", "hylen", "hym", "hypp",
259: "hys", "ie", "if", "ig",
260: "index", "it", "itc", "IX",
261: "kern", "kernafter", "kernbefore", "kernpair",
262: "lc", "lc_ctype", "lds", "length",
263: "letadj", "lf", "lg", "lhang",
264: "linetabs", "lnr", "lnrf", "lpfx",
1.304 schwarze 265: "ls", "lsm", "lt",
1.294 schwarze 266: "mediasize", "minss", "mk", "mso",
267: "na", "ne", "nh", "nhychar",
268: "nm", "nn", "nop", "nr",
269: "nrf", "nroff", "ns", "nx",
270: "open", "opena", "os", "output",
271: "padj", "papersize", "pc", "pev",
272: "pi", "PI", "pl", "pm",
1.312 schwarze 273: "pn", "pnr", "ps",
1.294 schwarze 274: "psbb", "pshape", "pso", "ptr",
275: "pvs", "rchar", "rd", "recursionlimit",
1.309 schwarze 276: "return", "rfschar", "rhang",
1.294 schwarze 277: "rm", "rn", "rnn", "rr",
278: "rs", "rt", "schar", "sentchar",
279: "shc", "shift", "sizes", "so",
280: "spacewidth", "special", "spreadwarn", "ss",
281: "sty", "substring", "sv", "sy",
1.300 schwarze 282: "T&", "tc", "TE",
1.301 schwarze 283: "TH", "tkf", "tl",
1.294 schwarze 284: "tm", "tm1", "tmc", "tr",
285: "track", "transchar", "trf", "trimat",
286: "trin", "trnt", "troff", "TS",
287: "uf", "ul", "unformat", "unwatch",
288: "unwatchn", "vpt", "vs", "warn",
289: "warnscale", "watch", "watchlength", "watchn",
290: "wh", "while", "write", "writec",
291: "writem", "xflag", ".", NULL,
1.306 schwarze 292: NULL, "text",
1.294 schwarze 293: "Dd", "Dt", "Os", "Sh",
294: "Ss", "Pp", "D1", "Dl",
295: "Bd", "Ed", "Bl", "El",
296: "It", "Ad", "An", "Ap",
297: "Ar", "Cd", "Cm", "Dv",
298: "Er", "Ev", "Ex", "Fa",
299: "Fd", "Fl", "Fn", "Ft",
300: "Ic", "In", "Li", "Nd",
301: "Nm", "Op", "Ot", "Pa",
302: "Rv", "St", "Va", "Vt",
303: "Xr", "%A", "%B", "%D",
304: "%I", "%J", "%N", "%O",
305: "%P", "%R", "%T", "%V",
306: "Ac", "Ao", "Aq", "At",
307: "Bc", "Bf", "Bo", "Bq",
308: "Bsx", "Bx", "Db", "Dc",
309: "Do", "Dq", "Ec", "Ef",
310: "Em", "Eo", "Fx", "Ms",
311: "No", "Ns", "Nx", "Ox",
312: "Pc", "Pf", "Po", "Pq",
313: "Qc", "Ql", "Qo", "Qq",
314: "Re", "Rs", "Sc", "So",
315: "Sq", "Sm", "Sx", "Sy",
316: "Tn", "Ux", "Xc", "Xo",
317: "Fo", "Fc", "Oo", "Oc",
318: "Bk", "Ek", "Bt", "Hf",
319: "Fr", "Ud", "Lb", "Lp",
320: "Lk", "Mt", "Brq", "Bro",
321: "Brc", "%C", "Es", "En",
1.299 schwarze 322: "Dx", "%Q", "%U", "Ta",
323: NULL,
1.294 schwarze 324: "TH", "SH", "SS", "TP",
325: "LP", "PP", "P", "IP",
326: "HP", "SM", "SB", "BI",
327: "IB", "BR", "RB", "R",
328: "B", "I", "IR", "RI",
1.299 schwarze 329: "nf", "fi",
1.294 schwarze 330: "RE", "RS", "DT", "UC",
1.297 schwarze 331: "PD", "AT", "in",
1.294 schwarze 332: "OP", "EX", "EE", "UR",
1.317 schwarze 333: "UE", "MT", "ME", NULL
1.294 schwarze 334: };
335: const char *const *roff_name = __roff_name;
336:
337: static struct roffmac roffs[TOKEN_NONE] = {
1.296 schwarze 338: { roff_br, NULL, NULL, 0 }, /* br */
1.305 schwarze 339: { roff_onearg, NULL, NULL, 0 }, /* ce */
1.297 schwarze 340: { roff_onearg, NULL, NULL, 0 }, /* ft */
1.298 schwarze 341: { roff_onearg, NULL, NULL, 0 }, /* ll */
1.304 schwarze 342: { roff_onearg, NULL, NULL, 0 }, /* mc */
1.312 schwarze 343: { roff_onearg, NULL, NULL, 0 }, /* po */
1.309 schwarze 344: { roff_onearg, NULL, NULL, 0 }, /* rj */
1.299 schwarze 345: { roff_onearg, NULL, NULL, 0 }, /* sp */
1.300 schwarze 346: { roff_manyarg, NULL, NULL, 0 }, /* ta */
1.301 schwarze 347: { roff_onearg, NULL, NULL, 0 }, /* ti */
1.296 schwarze 348: { NULL, NULL, NULL, 0 }, /* ROFF_MAX */
1.295 schwarze 349: { roff_unsupp, NULL, NULL, 0 }, /* ab */
350: { roff_line_ignore, NULL, NULL, 0 }, /* ad */
351: { roff_line_ignore, NULL, NULL, 0 }, /* af */
352: { roff_unsupp, NULL, NULL, 0 }, /* aln */
1.311 schwarze 353: { roff_als, NULL, NULL, 0 }, /* als */
1.295 schwarze 354: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am */
355: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am1 */
356: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami */
357: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami1 */
358: { roff_ds, NULL, NULL, 0 }, /* as */
359: { roff_ds, NULL, NULL, 0 }, /* as1 */
360: { roff_unsupp, NULL, NULL, 0 }, /* asciify */
361: { roff_line_ignore, NULL, NULL, 0 }, /* backtrace */
362: { roff_line_ignore, NULL, NULL, 0 }, /* bd */
363: { roff_line_ignore, NULL, NULL, 0 }, /* bleedat */
364: { roff_unsupp, NULL, NULL, 0 }, /* blm */
365: { roff_unsupp, NULL, NULL, 0 }, /* box */
366: { roff_unsupp, NULL, NULL, 0 }, /* boxa */
367: { roff_line_ignore, NULL, NULL, 0 }, /* bp */
368: { roff_unsupp, NULL, NULL, 0 }, /* BP */
369: { roff_unsupp, NULL, NULL, 0 }, /* break */
370: { roff_line_ignore, NULL, NULL, 0 }, /* breakchar */
371: { roff_line_ignore, NULL, NULL, 0 }, /* brnl */
1.296 schwarze 372: { roff_br, NULL, NULL, 0 }, /* brp */
1.295 schwarze 373: { roff_line_ignore, NULL, NULL, 0 }, /* brpnl */
374: { roff_unsupp, NULL, NULL, 0 }, /* c2 */
375: { roff_cc, NULL, NULL, 0 }, /* cc */
376: { roff_insec, NULL, NULL, 0 }, /* cf */
377: { roff_line_ignore, NULL, NULL, 0 }, /* cflags */
378: { roff_line_ignore, NULL, NULL, 0 }, /* ch */
379: { roff_unsupp, NULL, NULL, 0 }, /* char */
380: { roff_unsupp, NULL, NULL, 0 }, /* chop */
381: { roff_line_ignore, NULL, NULL, 0 }, /* class */
382: { roff_insec, NULL, NULL, 0 }, /* close */
383: { roff_unsupp, NULL, NULL, 0 }, /* CL */
384: { roff_line_ignore, NULL, NULL, 0 }, /* color */
385: { roff_unsupp, NULL, NULL, 0 }, /* composite */
386: { roff_unsupp, NULL, NULL, 0 }, /* continue */
387: { roff_line_ignore, NULL, NULL, 0 }, /* cp */
388: { roff_line_ignore, NULL, NULL, 0 }, /* cropat */
389: { roff_line_ignore, NULL, NULL, 0 }, /* cs */
390: { roff_line_ignore, NULL, NULL, 0 }, /* cu */
391: { roff_unsupp, NULL, NULL, 0 }, /* da */
392: { roff_unsupp, NULL, NULL, 0 }, /* dch */
393: { roff_Dd, NULL, NULL, 0 }, /* Dd */
394: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de */
395: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de1 */
396: { roff_line_ignore, NULL, NULL, 0 }, /* defcolor */
397: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei */
398: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei1 */
399: { roff_unsupp, NULL, NULL, 0 }, /* device */
400: { roff_unsupp, NULL, NULL, 0 }, /* devicem */
401: { roff_unsupp, NULL, NULL, 0 }, /* di */
402: { roff_unsupp, NULL, NULL, 0 }, /* do */
403: { roff_ds, NULL, NULL, 0 }, /* ds */
404: { roff_ds, NULL, NULL, 0 }, /* ds1 */
405: { roff_unsupp, NULL, NULL, 0 }, /* dwh */
406: { roff_unsupp, NULL, NULL, 0 }, /* dt */
1.303 schwarze 407: { roff_ec, NULL, NULL, 0 }, /* ec */
1.295 schwarze 408: { roff_unsupp, NULL, NULL, 0 }, /* ecr */
409: { roff_unsupp, NULL, NULL, 0 }, /* ecs */
410: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* el */
411: { roff_unsupp, NULL, NULL, 0 }, /* em */
412: { roff_EN, NULL, NULL, 0 }, /* EN */
1.303 schwarze 413: { roff_eo, NULL, NULL, 0 }, /* eo */
1.295 schwarze 414: { roff_unsupp, NULL, NULL, 0 }, /* EP */
415: { roff_EQ, NULL, NULL, 0 }, /* EQ */
416: { roff_line_ignore, NULL, NULL, 0 }, /* errprint */
417: { roff_unsupp, NULL, NULL, 0 }, /* ev */
418: { roff_unsupp, NULL, NULL, 0 }, /* evc */
419: { roff_unsupp, NULL, NULL, 0 }, /* ex */
420: { roff_line_ignore, NULL, NULL, 0 }, /* fallback */
421: { roff_line_ignore, NULL, NULL, 0 }, /* fam */
422: { roff_unsupp, NULL, NULL, 0 }, /* fc */
423: { roff_unsupp, NULL, NULL, 0 }, /* fchar */
424: { roff_line_ignore, NULL, NULL, 0 }, /* fcolor */
425: { roff_line_ignore, NULL, NULL, 0 }, /* fdeferlig */
426: { roff_line_ignore, NULL, NULL, 0 }, /* feature */
427: { roff_line_ignore, NULL, NULL, 0 }, /* fkern */
428: { roff_line_ignore, NULL, NULL, 0 }, /* fl */
429: { roff_line_ignore, NULL, NULL, 0 }, /* flig */
430: { roff_line_ignore, NULL, NULL, 0 }, /* fp */
431: { roff_line_ignore, NULL, NULL, 0 }, /* fps */
432: { roff_unsupp, NULL, NULL, 0 }, /* fschar */
433: { roff_line_ignore, NULL, NULL, 0 }, /* fspacewidth */
434: { roff_line_ignore, NULL, NULL, 0 }, /* fspecial */
435: { roff_line_ignore, NULL, NULL, 0 }, /* ftr */
436: { roff_line_ignore, NULL, NULL, 0 }, /* fzoom */
437: { roff_line_ignore, NULL, NULL, 0 }, /* gcolor */
438: { roff_line_ignore, NULL, NULL, 0 }, /* hc */
439: { roff_line_ignore, NULL, NULL, 0 }, /* hcode */
440: { roff_line_ignore, NULL, NULL, 0 }, /* hidechar */
441: { roff_line_ignore, NULL, NULL, 0 }, /* hla */
442: { roff_line_ignore, NULL, NULL, 0 }, /* hlm */
443: { roff_line_ignore, NULL, NULL, 0 }, /* hpf */
444: { roff_line_ignore, NULL, NULL, 0 }, /* hpfa */
445: { roff_line_ignore, NULL, NULL, 0 }, /* hpfcode */
446: { roff_line_ignore, NULL, NULL, 0 }, /* hw */
447: { roff_line_ignore, NULL, NULL, 0 }, /* hy */
448: { roff_line_ignore, NULL, NULL, 0 }, /* hylang */
449: { roff_line_ignore, NULL, NULL, 0 }, /* hylen */
450: { roff_line_ignore, NULL, NULL, 0 }, /* hym */
451: { roff_line_ignore, NULL, NULL, 0 }, /* hypp */
452: { roff_line_ignore, NULL, NULL, 0 }, /* hys */
453: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* ie */
454: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* if */
455: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ig */
456: { roff_unsupp, NULL, NULL, 0 }, /* index */
457: { roff_it, NULL, NULL, 0 }, /* it */
458: { roff_unsupp, NULL, NULL, 0 }, /* itc */
459: { roff_line_ignore, NULL, NULL, 0 }, /* IX */
460: { roff_line_ignore, NULL, NULL, 0 }, /* kern */
461: { roff_line_ignore, NULL, NULL, 0 }, /* kernafter */
462: { roff_line_ignore, NULL, NULL, 0 }, /* kernbefore */
463: { roff_line_ignore, NULL, NULL, 0 }, /* kernpair */
464: { roff_unsupp, NULL, NULL, 0 }, /* lc */
465: { roff_unsupp, NULL, NULL, 0 }, /* lc_ctype */
466: { roff_unsupp, NULL, NULL, 0 }, /* lds */
467: { roff_unsupp, NULL, NULL, 0 }, /* length */
468: { roff_line_ignore, NULL, NULL, 0 }, /* letadj */
469: { roff_insec, NULL, NULL, 0 }, /* lf */
470: { roff_line_ignore, NULL, NULL, 0 }, /* lg */
471: { roff_line_ignore, NULL, NULL, 0 }, /* lhang */
472: { roff_unsupp, NULL, NULL, 0 }, /* linetabs */
473: { roff_unsupp, NULL, NULL, 0 }, /* lnr */
474: { roff_unsupp, NULL, NULL, 0 }, /* lnrf */
475: { roff_unsupp, NULL, NULL, 0 }, /* lpfx */
476: { roff_line_ignore, NULL, NULL, 0 }, /* ls */
477: { roff_unsupp, NULL, NULL, 0 }, /* lsm */
478: { roff_line_ignore, NULL, NULL, 0 }, /* lt */
479: { roff_line_ignore, NULL, NULL, 0 }, /* mediasize */
480: { roff_line_ignore, NULL, NULL, 0 }, /* minss */
481: { roff_line_ignore, NULL, NULL, 0 }, /* mk */
482: { roff_insec, NULL, NULL, 0 }, /* mso */
483: { roff_line_ignore, NULL, NULL, 0 }, /* na */
484: { roff_line_ignore, NULL, NULL, 0 }, /* ne */
485: { roff_line_ignore, NULL, NULL, 0 }, /* nh */
486: { roff_line_ignore, NULL, NULL, 0 }, /* nhychar */
487: { roff_unsupp, NULL, NULL, 0 }, /* nm */
488: { roff_unsupp, NULL, NULL, 0 }, /* nn */
489: { roff_unsupp, NULL, NULL, 0 }, /* nop */
490: { roff_nr, NULL, NULL, 0 }, /* nr */
491: { roff_unsupp, NULL, NULL, 0 }, /* nrf */
492: { roff_line_ignore, NULL, NULL, 0 }, /* nroff */
493: { roff_line_ignore, NULL, NULL, 0 }, /* ns */
494: { roff_insec, NULL, NULL, 0 }, /* nx */
495: { roff_insec, NULL, NULL, 0 }, /* open */
496: { roff_insec, NULL, NULL, 0 }, /* opena */
497: { roff_line_ignore, NULL, NULL, 0 }, /* os */
498: { roff_unsupp, NULL, NULL, 0 }, /* output */
499: { roff_line_ignore, NULL, NULL, 0 }, /* padj */
500: { roff_line_ignore, NULL, NULL, 0 }, /* papersize */
501: { roff_line_ignore, NULL, NULL, 0 }, /* pc */
502: { roff_line_ignore, NULL, NULL, 0 }, /* pev */
503: { roff_insec, NULL, NULL, 0 }, /* pi */
504: { roff_unsupp, NULL, NULL, 0 }, /* PI */
505: { roff_line_ignore, NULL, NULL, 0 }, /* pl */
506: { roff_line_ignore, NULL, NULL, 0 }, /* pm */
507: { roff_line_ignore, NULL, NULL, 0 }, /* pn */
508: { roff_line_ignore, NULL, NULL, 0 }, /* pnr */
509: { roff_line_ignore, NULL, NULL, 0 }, /* ps */
510: { roff_unsupp, NULL, NULL, 0 }, /* psbb */
511: { roff_unsupp, NULL, NULL, 0 }, /* pshape */
512: { roff_insec, NULL, NULL, 0 }, /* pso */
513: { roff_line_ignore, NULL, NULL, 0 }, /* ptr */
514: { roff_line_ignore, NULL, NULL, 0 }, /* pvs */
515: { roff_unsupp, NULL, NULL, 0 }, /* rchar */
516: { roff_line_ignore, NULL, NULL, 0 }, /* rd */
517: { roff_line_ignore, NULL, NULL, 0 }, /* recursionlimit */
518: { roff_unsupp, NULL, NULL, 0 }, /* return */
519: { roff_unsupp, NULL, NULL, 0 }, /* rfschar */
520: { roff_line_ignore, NULL, NULL, 0 }, /* rhang */
521: { roff_rm, NULL, NULL, 0 }, /* rm */
1.306 schwarze 522: { roff_rn, NULL, NULL, 0 }, /* rn */
1.295 schwarze 523: { roff_unsupp, NULL, NULL, 0 }, /* rnn */
524: { roff_rr, NULL, NULL, 0 }, /* rr */
525: { roff_line_ignore, NULL, NULL, 0 }, /* rs */
526: { roff_line_ignore, NULL, NULL, 0 }, /* rt */
527: { roff_unsupp, NULL, NULL, 0 }, /* schar */
528: { roff_line_ignore, NULL, NULL, 0 }, /* sentchar */
529: { roff_line_ignore, NULL, NULL, 0 }, /* shc */
530: { roff_unsupp, NULL, NULL, 0 }, /* shift */
531: { roff_line_ignore, NULL, NULL, 0 }, /* sizes */
532: { roff_so, NULL, NULL, 0 }, /* so */
533: { roff_line_ignore, NULL, NULL, 0 }, /* spacewidth */
534: { roff_line_ignore, NULL, NULL, 0 }, /* special */
535: { roff_line_ignore, NULL, NULL, 0 }, /* spreadwarn */
536: { roff_line_ignore, NULL, NULL, 0 }, /* ss */
537: { roff_line_ignore, NULL, NULL, 0 }, /* sty */
538: { roff_unsupp, NULL, NULL, 0 }, /* substring */
539: { roff_line_ignore, NULL, NULL, 0 }, /* sv */
540: { roff_insec, NULL, NULL, 0 }, /* sy */
541: { roff_T_, NULL, NULL, 0 }, /* T& */
542: { roff_unsupp, NULL, NULL, 0 }, /* tc */
543: { roff_TE, NULL, NULL, 0 }, /* TE */
1.315 schwarze 544: { roff_Dd, NULL, NULL, 0 }, /* TH */
1.295 schwarze 545: { roff_line_ignore, NULL, NULL, 0 }, /* tkf */
546: { roff_unsupp, NULL, NULL, 0 }, /* tl */
547: { roff_line_ignore, NULL, NULL, 0 }, /* tm */
548: { roff_line_ignore, NULL, NULL, 0 }, /* tm1 */
549: { roff_line_ignore, NULL, NULL, 0 }, /* tmc */
550: { roff_tr, NULL, NULL, 0 }, /* tr */
551: { roff_line_ignore, NULL, NULL, 0 }, /* track */
552: { roff_line_ignore, NULL, NULL, 0 }, /* transchar */
553: { roff_insec, NULL, NULL, 0 }, /* trf */
554: { roff_line_ignore, NULL, NULL, 0 }, /* trimat */
555: { roff_unsupp, NULL, NULL, 0 }, /* trin */
556: { roff_unsupp, NULL, NULL, 0 }, /* trnt */
557: { roff_line_ignore, NULL, NULL, 0 }, /* troff */
558: { roff_TS, NULL, NULL, 0 }, /* TS */
559: { roff_line_ignore, NULL, NULL, 0 }, /* uf */
560: { roff_line_ignore, NULL, NULL, 0 }, /* ul */
561: { roff_unsupp, NULL, NULL, 0 }, /* unformat */
562: { roff_line_ignore, NULL, NULL, 0 }, /* unwatch */
563: { roff_line_ignore, NULL, NULL, 0 }, /* unwatchn */
564: { roff_line_ignore, NULL, NULL, 0 }, /* vpt */
565: { roff_line_ignore, NULL, NULL, 0 }, /* vs */
566: { roff_line_ignore, NULL, NULL, 0 }, /* warn */
567: { roff_line_ignore, NULL, NULL, 0 }, /* warnscale */
568: { roff_line_ignore, NULL, NULL, 0 }, /* watch */
569: { roff_line_ignore, NULL, NULL, 0 }, /* watchlength */
570: { roff_line_ignore, NULL, NULL, 0 }, /* watchn */
571: { roff_unsupp, NULL, NULL, 0 }, /* wh */
572: { roff_unsupp, NULL, NULL, 0 }, /* while */
573: { roff_insec, NULL, NULL, 0 }, /* write */
574: { roff_insec, NULL, NULL, 0 }, /* writec */
575: { roff_insec, NULL, NULL, 0 }, /* writem */
576: { roff_line_ignore, NULL, NULL, 0 }, /* xflag */
577: { roff_cblock, NULL, NULL, 0 }, /* . */
1.306 schwarze 578: { roff_renamed, NULL, NULL, 0 },
1.295 schwarze 579: { roff_userdef, NULL, NULL, 0 }
1.67 kristaps 580: };
581:
1.141 kristaps 582: /* Array of injected predefined strings. */
583: #define PREDEFS_MAX 38
584: static const struct predef predefs[PREDEFS_MAX] = {
585: #include "predefs.in"
586: };
587:
1.305 schwarze 588: static int roffce_lines; /* number of input lines to center */
589: static struct roff_node *roffce_node; /* active request */
1.178 schwarze 590: static int roffit_lines; /* number of lines to delay */
591: static char *roffit_macro; /* nil-terminated macro line */
592:
1.207 schwarze 593:
1.266 schwarze 594: /* --- request table ------------------------------------------------------ */
595:
1.295 schwarze 596: struct ohash *
597: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85 kristaps 598: {
1.295 schwarze 599: struct ohash *htab;
600: struct roffreq *req;
601: enum roff_tok tok;
602: size_t sz;
603: unsigned int slot;
1.85 kristaps 604:
1.295 schwarze 605: htab = mandoc_malloc(sizeof(*htab));
606: mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85 kristaps 607:
1.295 schwarze 608: for (tok = mintok; tok < maxtok; tok++) {
1.296 schwarze 609: if (roff_name[tok] == NULL)
610: continue;
1.295 schwarze 611: sz = strlen(roff_name[tok]);
612: req = mandoc_malloc(sizeof(*req) + sz + 1);
613: req->tok = tok;
614: memcpy(req->name, roff_name[tok], sz + 1);
615: slot = ohash_qlookup(htab, req->name);
616: ohash_insert(htab, slot, req);
1.85 kristaps 617: }
1.295 schwarze 618: return htab;
1.85 kristaps 619: }
620:
1.295 schwarze 621: void
622: roffhash_free(struct ohash *htab)
1.67 kristaps 623: {
1.295 schwarze 624: struct roffreq *req;
625: unsigned int slot;
1.67 kristaps 626:
1.295 schwarze 627: if (htab == NULL)
628: return;
629: for (req = ohash_first(htab, &slot); req != NULL;
630: req = ohash_next(htab, &slot))
631: free(req);
632: ohash_delete(htab);
633: free(htab);
634: }
635:
636: enum roff_tok
637: roffhash_find(struct ohash *htab, const char *name, size_t sz)
638: {
639: struct roffreq *req;
640: const char *end;
641:
642: if (sz) {
643: end = name + sz;
644: req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
645: } else
646: req = ohash_find(htab, ohash_qlookup(htab, name));
647: return req == NULL ? TOKEN_NONE : req->tok;
1.67 kristaps 648: }
649:
1.266 schwarze 650: /* --- stack of request blocks -------------------------------------------- */
651:
1.67 kristaps 652: /*
653: * Pop the current node off of the stack of roff instructions currently
654: * pending.
655: */
656: static void
657: roffnode_pop(struct roff *r)
658: {
659: struct roffnode *p;
660:
1.75 kristaps 661: assert(r->last);
1.207 schwarze 662: p = r->last;
1.82 kristaps 663:
1.75 kristaps 664: r->last = r->last->parent;
1.106 kristaps 665: free(p->name);
666: free(p->end);
1.67 kristaps 667: free(p);
668: }
669:
670: /*
671: * Push a roff node onto the instruction stack. This must later be
672: * removed with roffnode_pop().
673: */
1.98 schwarze 674: static void
1.294 schwarze 675: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106 kristaps 676: int line, int col)
1.67 kristaps 677: {
678: struct roffnode *p;
679:
1.98 schwarze 680: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 681: p->tok = tok;
1.106 kristaps 682: if (name)
683: p->name = mandoc_strdup(name);
1.67 kristaps 684: p->parent = r->last;
685: p->line = line;
686: p->col = col;
1.198 schwarze 687: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 688:
689: r->last = p;
690: }
691:
1.266 schwarze 692: /* --- roff parser state data management ---------------------------------- */
693:
1.67 kristaps 694: static void
695: roff_free1(struct roff *r)
696: {
1.176 schwarze 697: struct tbl_node *tbl;
1.167 kristaps 698: int i;
1.67 kristaps 699:
1.176 schwarze 700: while (NULL != (tbl = r->first_tbl)) {
701: r->first_tbl = tbl->next;
702: tbl_free(tbl);
1.109 kristaps 703: }
1.113 kristaps 704: r->first_tbl = r->last_tbl = r->tbl = NULL;
705:
1.319 schwarze 706: if (r->last_eqn != NULL)
707: eqn_free(r->last_eqn);
708: r->last_eqn = r->eqn = NULL;
1.125 kristaps 709:
1.67 kristaps 710: while (r->last)
711: roffnode_pop(r);
1.109 kristaps 712:
1.223 schwarze 713: free (r->rstack);
714: r->rstack = NULL;
715: r->rstacksz = 0;
716: r->rstackpos = -1;
717:
718: roff_freereg(r->regtab);
719: r->regtab = NULL;
720:
1.167 kristaps 721: roff_freestr(r->strtab);
1.306 schwarze 722: roff_freestr(r->rentab);
1.167 kristaps 723: roff_freestr(r->xmbtab);
1.306 schwarze 724: r->strtab = r->rentab = r->xmbtab = NULL;
1.167 kristaps 725:
726: if (r->xtab)
727: for (i = 0; i < 128; i++)
728: free(r->xtab[i].p);
729: free(r->xtab);
730: r->xtab = NULL;
1.67 kristaps 731: }
732:
733: void
734: roff_reset(struct roff *r)
735: {
736: roff_free1(r);
1.227 schwarze 737: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.303 schwarze 738: r->control = '\0';
739: r->escape = '\\';
1.307 schwarze 740: roffce_lines = 0;
741: roffce_node = NULL;
742: roffit_lines = 0;
743: roffit_macro = NULL;
1.67 kristaps 744: }
745:
746: void
747: roff_free(struct roff *r)
748: {
749: roff_free1(r);
1.295 schwarze 750: roffhash_free(r->reqtab);
1.67 kristaps 751: free(r);
752: }
753:
754: struct roff *
1.279 schwarze 755: roff_alloc(struct mparse *parse, int options)
1.67 kristaps 756: {
757: struct roff *r;
758:
1.98 schwarze 759: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 760: r->parse = parse;
1.295 schwarze 761: r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.199 schwarze 762: r->options = options;
1.227 schwarze 763: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82 kristaps 764: r->rstackpos = -1;
1.303 schwarze 765: r->escape = '\\';
1.277 schwarze 766: return r;
1.265 schwarze 767: }
768:
1.266 schwarze 769: /* --- syntax tree state data management ---------------------------------- */
770:
1.265 schwarze 771: static void
772: roff_man_free1(struct roff_man *man)
773: {
774:
1.266 schwarze 775: if (man->first != NULL)
776: roff_node_delete(man, man->first);
1.265 schwarze 777: free(man->meta.msec);
778: free(man->meta.vol);
779: free(man->meta.os);
780: free(man->meta.arch);
781: free(man->meta.title);
782: free(man->meta.name);
783: free(man->meta.date);
784: }
785:
786: static void
787: roff_man_alloc1(struct roff_man *man)
788: {
789:
790: memset(&man->meta, 0, sizeof(man->meta));
791: man->first = mandoc_calloc(1, sizeof(*man->first));
792: man->first->type = ROFFT_ROOT;
793: man->last = man->first;
794: man->last_es = NULL;
795: man->flags = 0;
796: man->macroset = MACROSET_NONE;
797: man->lastsec = man->lastnamed = SEC_NONE;
798: man->next = ROFF_NEXT_CHILD;
799: }
800:
801: void
802: roff_man_reset(struct roff_man *man)
803: {
804:
805: roff_man_free1(man);
806: roff_man_alloc1(man);
807: }
808:
809: void
810: roff_man_free(struct roff_man *man)
811: {
812:
813: roff_man_free1(man);
814: free(man);
815: }
816:
817: struct roff_man *
818: roff_man_alloc(struct roff *roff, struct mparse *parse,
1.316 schwarze 819: const char *os_s, int quick)
1.265 schwarze 820: {
821: struct roff_man *man;
822:
823: man = mandoc_calloc(1, sizeof(*man));
824: man->parse = parse;
825: man->roff = roff;
1.316 schwarze 826: man->os_s = os_s;
1.265 schwarze 827: man->quick = quick;
828: roff_man_alloc1(man);
1.296 schwarze 829: roff->man = man;
1.277 schwarze 830: return man;
1.67 kristaps 831: }
832:
1.266 schwarze 833: /* --- syntax tree handling ----------------------------------------------- */
834:
835: struct roff_node *
836: roff_node_alloc(struct roff_man *man, int line, int pos,
837: enum roff_type type, int tok)
838: {
839: struct roff_node *n;
840:
841: n = mandoc_calloc(1, sizeof(*n));
842: n->line = line;
843: n->pos = pos;
844: n->tok = tok;
845: n->type = type;
846: n->sec = man->lastsec;
847:
848: if (man->flags & MDOC_SYNOPSIS)
1.285 schwarze 849: n->flags |= NODE_SYNPRETTY;
1.266 schwarze 850: else
1.285 schwarze 851: n->flags &= ~NODE_SYNPRETTY;
1.266 schwarze 852: if (man->flags & MDOC_NEWLINE)
1.285 schwarze 853: n->flags |= NODE_LINE;
1.266 schwarze 854: man->flags &= ~MDOC_NEWLINE;
855:
1.277 schwarze 856: return n;
1.266 schwarze 857: }
858:
859: void
860: roff_node_append(struct roff_man *man, struct roff_node *n)
861: {
862:
863: switch (man->next) {
864: case ROFF_NEXT_SIBLING:
1.281 schwarze 865: if (man->last->next != NULL) {
866: n->next = man->last->next;
867: man->last->next->prev = n;
868: } else
869: man->last->parent->last = n;
1.266 schwarze 870: man->last->next = n;
871: n->prev = man->last;
872: n->parent = man->last->parent;
873: break;
874: case ROFF_NEXT_CHILD:
1.287 schwarze 875: if (man->last->child != NULL) {
876: n->next = man->last->child;
877: man->last->child->prev = n;
878: } else
879: man->last->last = n;
1.266 schwarze 880: man->last->child = n;
881: n->parent = man->last;
882: break;
883: default:
884: abort();
885: }
1.282 schwarze 886: man->last = n;
1.266 schwarze 887:
888: switch (n->type) {
889: case ROFFT_HEAD:
890: n->parent->head = n;
891: break;
892: case ROFFT_BODY:
1.282 schwarze 893: if (n->end != ENDBODY_NOT)
894: return;
1.266 schwarze 895: n->parent->body = n;
896: break;
897: case ROFFT_TAIL:
898: n->parent->tail = n;
899: break;
900: default:
1.282 schwarze 901: return;
1.266 schwarze 902: }
1.282 schwarze 903:
904: /*
905: * Copy over the normalised-data pointer of our parent. Not
906: * everybody has one, but copying a null pointer is fine.
907: */
908:
909: n->norm = n->parent->norm;
910: assert(n->parent->type == ROFFT_BLOCK);
1.266 schwarze 911: }
912:
1.267 schwarze 913: void
914: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
915: {
916: struct roff_node *n;
917:
918: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
919: n->string = roff_strdup(man->roff, word);
920: roff_node_append(man, n);
1.286 schwarze 921: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 922: man->next = ROFF_NEXT_SIBLING;
923: }
924:
925: void
926: roff_word_append(struct roff_man *man, const char *word)
927: {
928: struct roff_node *n;
929: char *addstr, *newstr;
930:
931: n = man->last;
932: addstr = roff_strdup(man->roff, word);
933: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
934: free(addstr);
935: free(n->string);
936: n->string = newstr;
937: man->next = ROFF_NEXT_SIBLING;
1.268 schwarze 938: }
939:
940: void
941: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
942: {
943: struct roff_node *n;
944:
945: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
946: roff_node_append(man, n);
947: man->next = ROFF_NEXT_CHILD;
948: }
949:
950: struct roff_node *
951: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
952: {
953: struct roff_node *n;
954:
955: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
956: roff_node_append(man, n);
957: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 958: return n;
1.267 schwarze 959: }
960:
1.266 schwarze 961: struct roff_node *
962: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
963: {
964: struct roff_node *n;
965:
966: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
967: roff_node_append(man, n);
968: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 969: return n;
1.266 schwarze 970: }
971:
972: struct roff_node *
973: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
974: {
975: struct roff_node *n;
976:
977: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
978: roff_node_append(man, n);
979: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 980: return n;
1.267 schwarze 981: }
982:
1.321 ! schwarze 983: static void
! 984: roff_addtbl(struct roff_man *man, struct tbl_node *tbl)
1.267 schwarze 985: {
986: struct roff_node *n;
1.321 ! schwarze 987: const struct tbl_span *span;
1.267 schwarze 988:
989: if (man->macroset == MACROSET_MAN)
1.302 schwarze 990: man_breakscope(man, ROFF_TS);
1.321 ! schwarze 991: while ((span = tbl_span(tbl)) != NULL) {
! 992: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
! 993: n->span = span;
! 994: roff_node_append(man, n);
! 995: n->flags |= NODE_VALID | NODE_ENDED;
! 996: man->next = ROFF_NEXT_SIBLING;
! 997: }
1.266 schwarze 998: }
999:
1000: void
1001: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1002: {
1003:
1004: /* Adjust siblings. */
1005:
1006: if (n->prev)
1007: n->prev->next = n->next;
1008: if (n->next)
1009: n->next->prev = n->prev;
1010:
1011: /* Adjust parent. */
1012:
1013: if (n->parent != NULL) {
1014: if (n->parent->child == n)
1015: n->parent->child = n->next;
1016: if (n->parent->last == n)
1017: n->parent->last = n->prev;
1018: }
1019:
1020: /* Adjust parse point. */
1021:
1022: if (man == NULL)
1023: return;
1024: if (man->last == n) {
1025: if (n->prev == NULL) {
1026: man->last = n->parent;
1027: man->next = ROFF_NEXT_CHILD;
1028: } else {
1029: man->last = n->prev;
1030: man->next = ROFF_NEXT_SIBLING;
1031: }
1032: }
1033: if (man->first == n)
1034: man->first = NULL;
1035: }
1036:
1037: void
1038: roff_node_free(struct roff_node *n)
1039: {
1040:
1041: if (n->args != NULL)
1042: mdoc_argv_free(n->args);
1043: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1044: free(n->norm);
1.319 schwarze 1045: if (n->eqn != NULL)
1046: eqn_box_free(n->eqn);
1.266 schwarze 1047: free(n->string);
1048: free(n);
1049: }
1050:
1051: void
1052: roff_node_delete(struct roff_man *man, struct roff_node *n)
1053: {
1054:
1055: while (n->child != NULL)
1056: roff_node_delete(man, n->child);
1057: roff_node_unlink(man, n);
1058: roff_node_free(n);
1.269 schwarze 1059: }
1060:
1061: void
1062: deroff(char **dest, const struct roff_node *n)
1063: {
1064: char *cp;
1065: size_t sz;
1066:
1067: if (n->type != ROFFT_TEXT) {
1068: for (n = n->child; n != NULL; n = n->next)
1069: deroff(dest, n);
1070: return;
1071: }
1072:
1.288 schwarze 1073: /* Skip leading whitespace. */
1.269 schwarze 1074:
1.288 schwarze 1075: for (cp = n->string; *cp != '\0'; cp++) {
1.289 schwarze 1076: if (cp[0] == '\\' && cp[1] != '\0' &&
1077: strchr(" %&0^|~", cp[1]) != NULL)
1.269 schwarze 1078: cp++;
1.288 schwarze 1079: else if ( ! isspace((unsigned char)*cp))
1.269 schwarze 1080: break;
1081: }
1082:
1.289 schwarze 1083: /* Skip trailing backslash. */
1084:
1085: sz = strlen(cp);
1.290 schwarze 1086: if (sz > 0 && cp[sz - 1] == '\\')
1.289 schwarze 1087: sz--;
1088:
1.269 schwarze 1089: /* Skip trailing whitespace. */
1090:
1.289 schwarze 1091: for (; sz; sz--)
1.269 schwarze 1092: if ( ! isspace((unsigned char)cp[sz-1]))
1093: break;
1094:
1095: /* Skip empty strings. */
1096:
1097: if (sz == 0)
1098: return;
1099:
1100: if (*dest == NULL) {
1101: *dest = mandoc_strndup(cp, sz);
1102: return;
1103: }
1104:
1105: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1106: free(*dest);
1107: *dest = cp;
1.266 schwarze 1108: }
1109:
1110: /* --- main functions of the roff parser ---------------------------------- */
1111:
1.94 kristaps 1112: /*
1.206 schwarze 1113: * In the current line, expand escape sequences that tend to get
1114: * used in numerical expressions and conditional requests.
1115: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 1116: */
1.172 schwarze 1117: static enum rofferr
1.238 schwarze 1118: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 1119: {
1.208 schwarze 1120: char ubuf[24]; /* buffer to print the number */
1.205 schwarze 1121: const char *start; /* start of the string to process */
1.209 schwarze 1122: char *stesc; /* start of an escape sequence ('\\') */
1.108 schwarze 1123: const char *stnam; /* start of the name, after "[(*" */
1124: const char *cp; /* end of the name, e.g. before ']' */
1125: const char *res; /* the string to be substituted */
1.238 schwarze 1126: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 1127: size_t maxl; /* expected length of the escape name */
1128: size_t naml; /* actual length of the escape name */
1.237 schwarze 1129: enum mandoc_esc esc; /* type of the escape sequence */
1.316 schwarze 1130: enum mandoc_os os_e; /* kind of RCS id seen */
1.237 schwarze 1131: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 1132: int expand_count; /* to avoid infinite loops */
1.206 schwarze 1133: int npos; /* position in numeric expression */
1.218 schwarze 1134: int arg_complete; /* argument not interrupted by eol */
1.303 schwarze 1135: int done; /* no more input available */
1.315 schwarze 1136: int deftype; /* type of definition to paste */
1.206 schwarze 1137: char term; /* character terminating the escape */
1.94 kristaps 1138:
1.303 schwarze 1139: /* Search forward for comments. */
1140:
1141: done = 0;
1142: start = buf->buf + pos;
1143: for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
1144: if (stesc[0] != r->escape || stesc[1] == '\0')
1145: continue;
1146: stesc++;
1147: if (*stesc != '"' && *stesc != '#')
1148: continue;
1.313 schwarze 1149:
1150: /* Comment found, look for RCS id. */
1151:
1152: if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.316 schwarze 1153: os_e = MANDOC_OS_OPENBSD;
1.313 schwarze 1154: cp += 8;
1155: } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.316 schwarze 1156: os_e = MANDOC_OS_NETBSD;
1.313 schwarze 1157: cp += 7;
1158: }
1159: if (cp != NULL &&
1160: isalnum((unsigned char)*cp) == 0 &&
1.314 schwarze 1161: strchr(cp, '$') != NULL) {
1162: if (r->man->meta.rcsids & (1 << os_e))
1163: mandoc_msg(MANDOCERR_RCS_REP, r->parse,
1164: ln, stesc + 1 - buf->buf, stesc + 1);
1.313 schwarze 1165: r->man->meta.rcsids |= 1 << os_e;
1.314 schwarze 1166: }
1.313 schwarze 1167:
1168: /* Handle trailing whitespace. */
1169:
1.303 schwarze 1170: cp = strchr(stesc--, '\0') - 1;
1171: if (*cp == '\n') {
1172: done = 1;
1173: cp--;
1174: }
1175: if (*cp == ' ' || *cp == '\t')
1176: mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
1177: ln, cp - buf->buf, NULL);
1178: while (stesc > start && stesc[-1] == ' ')
1179: stesc--;
1180: *stesc = '\0';
1181: break;
1182: }
1183: if (stesc == start)
1184: return ROFF_CONT;
1185: stesc--;
1186:
1187: /* Notice the end of the input. */
1188:
1189: if (*stesc == '\n') {
1190: *stesc-- = '\0';
1191: done = 1;
1192: }
1193:
1.170 schwarze 1194: expand_count = 0;
1.303 schwarze 1195: while (stesc >= start) {
1.170 schwarze 1196:
1.205 schwarze 1197: /* Search backwards for the next backslash. */
1198:
1.303 schwarze 1199: if (*stesc != r->escape) {
1200: if (*stesc == '\\') {
1201: *stesc = '\0';
1202: buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
1203: buf->buf, stesc + 1) + 1;
1204: start = nbuf + pos;
1205: stesc = nbuf + (stesc - buf->buf);
1206: free(buf->buf);
1207: buf->buf = nbuf;
1208: }
1209: stesc--;
1.205 schwarze 1210: continue;
1.303 schwarze 1211: }
1.205 schwarze 1212:
1213: /* If it is escaped, skip it. */
1214:
1215: for (cp = stesc - 1; cp >= start; cp--)
1.303 schwarze 1216: if (*cp != r->escape)
1.205 schwarze 1217: break;
1218:
1.238 schwarze 1219: if ((stesc - cp) % 2 == 0) {
1.303 schwarze 1220: while (stesc > cp)
1221: *stesc-- = '\\';
1.205 schwarze 1222: continue;
1.303 schwarze 1223: } else if (stesc[1] != '\0') {
1224: *stesc = '\\';
1225: } else {
1226: *stesc-- = '\0';
1227: if (done)
1228: continue;
1229: else
1230: return ROFF_APPEND;
1.205 schwarze 1231: }
1.108 schwarze 1232:
1.206 schwarze 1233: /* Decide whether to expand or to check only. */
1.108 schwarze 1234:
1.206 schwarze 1235: term = '\0';
1.205 schwarze 1236: cp = stesc + 1;
1.181 schwarze 1237: switch (*cp) {
1.207 schwarze 1238: case '*':
1.181 schwarze 1239: res = NULL;
1240: break;
1.207 schwarze 1241: case 'B':
1242: case 'w':
1.206 schwarze 1243: term = cp[1];
1244: /* FALLTHROUGH */
1.207 schwarze 1245: case 'n':
1.181 schwarze 1246: res = ubuf;
1247: break;
1248: default:
1.237 schwarze 1249: esc = mandoc_escape(&cp, &stnam, &inaml);
1250: if (esc == ESCAPE_ERROR ||
1251: (esc == ESCAPE_SPECIAL &&
1.279 schwarze 1252: mchars_spec2cp(stnam, inaml) < 0))
1.219 schwarze 1253: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1254: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1255: "%.*s", (int)(cp - stesc), stesc);
1.303 schwarze 1256: stesc--;
1.205 schwarze 1257: continue;
1.152 kristaps 1258: }
1259:
1.205 schwarze 1260: if (EXPAND_LIMIT < ++expand_count) {
1261: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1262: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1263: return ROFF_IGN;
1.205 schwarze 1264: }
1.108 schwarze 1265:
1266: /*
1267: * The third character decides the length
1.181 schwarze 1268: * of the name of the string or register.
1.108 schwarze 1269: * Save a pointer to the name.
1270: */
1271:
1.238 schwarze 1272: if (term == '\0') {
1.206 schwarze 1273: switch (*++cp) {
1.207 schwarze 1274: case '\0':
1.206 schwarze 1275: maxl = 0;
1276: break;
1.207 schwarze 1277: case '(':
1.206 schwarze 1278: cp++;
1279: maxl = 2;
1280: break;
1.207 schwarze 1281: case '[':
1.206 schwarze 1282: cp++;
1283: term = ']';
1284: maxl = 0;
1285: break;
1286: default:
1287: maxl = 1;
1288: break;
1289: }
1290: } else {
1291: cp += 2;
1.94 kristaps 1292: maxl = 0;
1293: }
1.108 schwarze 1294: stnam = cp;
1.94 kristaps 1295:
1.108 schwarze 1296: /* Advance to the end of the name. */
1.94 kristaps 1297:
1.253 schwarze 1298: naml = 0;
1.218 schwarze 1299: arg_complete = 1;
1.253 schwarze 1300: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1301: if (*cp == '\0') {
1.219 schwarze 1302: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1303: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1304: arg_complete = 0;
1.206 schwarze 1305: break;
1.153 kristaps 1306: }
1.238 schwarze 1307: if (maxl == 0 && *cp == term) {
1.206 schwarze 1308: cp++;
1.253 schwarze 1309: break;
1310: }
1311: if (*cp++ != '\\' || stesc[1] != 'w') {
1312: naml++;
1313: continue;
1314: }
1315: switch (mandoc_escape(&cp, NULL, NULL)) {
1316: case ESCAPE_SPECIAL:
1317: case ESCAPE_UNICODE:
1318: case ESCAPE_NUMBERED:
1319: case ESCAPE_OVERSTRIKE:
1320: naml++;
1321: break;
1322: default:
1.94 kristaps 1323: break;
1.206 schwarze 1324: }
1.94 kristaps 1325: }
1326:
1.108 schwarze 1327: /*
1328: * Retrieve the replacement string; if it is
1329: * undefined, resume searching for escapes.
1330: */
1331:
1.206 schwarze 1332: switch (stesc[1]) {
1.207 schwarze 1333: case '*':
1.315 schwarze 1334: if (arg_complete) {
1335: deftype = ROFFDEF_USER | ROFFDEF_PRE;
1336: res = roff_getstrn(r, stnam, naml, &deftype);
1337: }
1.206 schwarze 1338: break;
1.207 schwarze 1339: case 'B':
1.206 schwarze 1340: npos = 0;
1.218 schwarze 1341: ubuf[0] = arg_complete &&
1.261 schwarze 1342: roff_evalnum(r, ln, stnam, &npos,
1343: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1344: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1345: ubuf[1] = '\0';
1346: break;
1.207 schwarze 1347: case 'n':
1.218 schwarze 1348: if (arg_complete)
1349: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1350: roff_getregn(r, stnam, naml));
1351: else
1352: ubuf[0] = '\0';
1.206 schwarze 1353: break;
1.207 schwarze 1354: case 'w':
1.218 schwarze 1355: /* use even incomplete args */
1.208 schwarze 1356: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1357: 24 * (int)naml);
1358: break;
1359: }
1.94 kristaps 1360:
1.238 schwarze 1361: if (res == NULL) {
1.219 schwarze 1362: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1363: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1364: "%.*s", (int)naml, stnam);
1.142 kristaps 1365: res = "";
1.246 schwarze 1366: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1367: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1368: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1369: return ROFF_IGN;
1.94 kristaps 1370: }
1371:
1.108 schwarze 1372: /* Replace the escape sequence by the string. */
1373:
1.209 schwarze 1374: *stesc = '\0';
1.238 schwarze 1375: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1376: buf->buf, res, cp) + 1;
1.94 kristaps 1377:
1.205 schwarze 1378: /* Prepare for the next replacement. */
1.94 kristaps 1379:
1.205 schwarze 1380: start = nbuf + pos;
1.238 schwarze 1381: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1382: free(buf->buf);
1383: buf->buf = nbuf;
1.154 kristaps 1384: }
1.277 schwarze 1385: return ROFF_CONT;
1.154 kristaps 1386: }
1387:
1388: /*
1.275 schwarze 1389: * Process text streams.
1.154 kristaps 1390: */
1391: static enum rofferr
1.305 schwarze 1392: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.154 kristaps 1393: {
1394: size_t sz;
1395: const char *start;
1.178 schwarze 1396: char *p;
1397: int isz;
1.154 kristaps 1398: enum mandoc_esc esc;
1399:
1.275 schwarze 1400: /* Spring the input line trap. */
1401:
1402: if (roffit_lines == 1) {
1403: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1404: free(buf->buf);
1405: buf->buf = p;
1406: buf->sz = isz + 1;
1407: *offs = 0;
1408: free(roffit_macro);
1409: roffit_lines = 0;
1.277 schwarze 1410: return ROFF_REPARSE;
1.275 schwarze 1411: } else if (roffit_lines > 1)
1412: --roffit_lines;
1413:
1.305 schwarze 1414: if (roffce_node != NULL && buf->buf[pos] != '\0') {
1415: if (roffce_lines < 1) {
1416: r->man->last = roffce_node;
1417: r->man->next = ROFF_NEXT_SIBLING;
1418: roffce_lines = 0;
1419: roffce_node = NULL;
1420: } else
1421: roffce_lines--;
1422: }
1423:
1.275 schwarze 1424: /* Convert all breakable hyphens into ASCII_HYPH. */
1425:
1.238 schwarze 1426: start = p = buf->buf + pos;
1.154 kristaps 1427:
1.238 schwarze 1428: while (*p != '\0') {
1.154 kristaps 1429: sz = strcspn(p, "-\\");
1430: p += sz;
1431:
1.238 schwarze 1432: if (*p == '\0')
1.159 kristaps 1433: break;
1434:
1.238 schwarze 1435: if (*p == '\\') {
1.154 kristaps 1436: /* Skip over escapes. */
1437: p++;
1.189 schwarze 1438: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1439: if (esc == ESCAPE_ERROR)
1.154 kristaps 1440: break;
1.264 schwarze 1441: while (*p == '-')
1442: p++;
1.155 kristaps 1443: continue;
1.159 kristaps 1444: } else if (p == start) {
1.158 kristaps 1445: p++;
1.155 kristaps 1446: continue;
1.158 kristaps 1447: }
1.155 kristaps 1448:
1.171 schwarze 1449: if (isalpha((unsigned char)p[-1]) &&
1450: isalpha((unsigned char)p[1]))
1.155 kristaps 1451: *p = ASCII_HYPH;
1452: p++;
1.94 kristaps 1453: }
1.277 schwarze 1454: return ROFF_CONT;
1.94 kristaps 1455: }
1456:
1.67 kristaps 1457: enum rofferr
1.238 schwarze 1458: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1459: {
1.294 schwarze 1460: enum roff_tok t;
1.109 kristaps 1461: enum rofferr e;
1.238 schwarze 1462: int pos; /* parse point */
1.243 schwarze 1463: int spos; /* saved parse point for messages */
1.238 schwarze 1464: int ppos; /* original offset in buf->buf */
1465: int ctl; /* macro line (boolean) */
1466:
1467: ppos = pos = *offs;
1.79 kristaps 1468:
1.230 schwarze 1469: /* Handle in-line equation delimiters. */
1470:
1.236 schwarze 1471: if (r->tbl == NULL &&
1472: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1473: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1474: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1475: if (e == ROFF_REPARSE)
1.277 schwarze 1476: return e;
1.230 schwarze 1477: assert(e == ROFF_CONT);
1478: }
1479:
1480: /* Expand some escape sequences. */
1.94 kristaps 1481:
1.238 schwarze 1482: e = roff_res(r, buf, ln, pos);
1.303 schwarze 1483: if (e == ROFF_IGN || e == ROFF_APPEND)
1.277 schwarze 1484: return e;
1.238 schwarze 1485: assert(e == ROFF_CONT);
1.94 kristaps 1486:
1.238 schwarze 1487: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1488:
1.94 kristaps 1489: /*
1.79 kristaps 1490: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1491: * text through the macro's filter.
1492: * Equations process all content themselves.
1493: * Tables process almost all content themselves, but we want
1494: * to warn about macros before passing it there.
1.79 kristaps 1495: */
1.74 kristaps 1496:
1.252 schwarze 1497: if (r->last != NULL && ! ctl) {
1.78 kristaps 1498: t = r->last->tok;
1.238 schwarze 1499: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.294 schwarze 1500: if (e == ROFF_IGN)
1.277 schwarze 1501: return e;
1.294 schwarze 1502: assert(e == ROFF_CONT);
1.182 schwarze 1503: }
1.319 schwarze 1504: if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
1505: eqn_read(r->eqn, buf->buf + ppos);
1506: return ROFF_IGN;
1507: }
1.321 ! schwarze 1508: if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
! 1509: tbl_read(r->tbl, ln, buf->buf, ppos);
! 1510: roff_addtbl(r->man, r->tbl);
! 1511: return ROFF_IGN;
! 1512: }
1.252 schwarze 1513: if ( ! ctl)
1.305 schwarze 1514: return roff_parsetext(r, buf, pos, offs);
1.228 schwarze 1515:
1516: /* Skip empty request lines. */
1517:
1.238 schwarze 1518: if (buf->buf[pos] == '"') {
1.228 schwarze 1519: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1520: ln, pos, NULL);
1.277 schwarze 1521: return ROFF_IGN;
1.238 schwarze 1522: } else if (buf->buf[pos] == '\0')
1.277 schwarze 1523: return ROFF_IGN;
1.67 kristaps 1524:
1.79 kristaps 1525: /*
1526: * If a scope is open, go to the child handler for that macro,
1527: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1528: * Don't do so if an equation is open.
1.79 kristaps 1529: */
1.78 kristaps 1530:
1.79 kristaps 1531: if (r->last) {
1532: t = r->last->tok;
1.277 schwarze 1533: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1534: }
1.78 kristaps 1535:
1.243 schwarze 1536: /* No scope is open. This is a new request or macro. */
1537:
1538: spos = pos;
1539: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1540:
1541: /* Tables ignore most macros. */
1542:
1.308 schwarze 1543: if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.309 schwarze 1544: t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.243 schwarze 1545: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1546: ln, pos, buf->buf + spos);
1.308 schwarze 1547: if (t != TOKEN_NONE)
1.277 schwarze 1548: return ROFF_IGN;
1.257 schwarze 1549: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1550: pos++;
1.291 schwarze 1551: while (buf->buf[pos] == ' ')
1.257 schwarze 1552: pos++;
1.321 ! schwarze 1553: tbl_read(r->tbl, ln, buf->buf, pos);
! 1554: roff_addtbl(r->man, r->tbl);
! 1555: return ROFF_IGN;
1.243 schwarze 1556: }
1557:
1.305 schwarze 1558: /* For now, let high level macros abort .ce mode. */
1559:
1560: if (ctl && roffce_node != NULL &&
1561: (t == TOKEN_NONE || t == ROFF_EQ || t == ROFF_TS)) {
1562: r->man->last = roffce_node;
1563: r->man->next = ROFF_NEXT_SIBLING;
1564: roffce_lines = 0;
1565: roffce_node = NULL;
1566: }
1567:
1.79 kristaps 1568: /*
1.243 schwarze 1569: * This is neither a roff request nor a user-defined macro.
1570: * Let the standard macro set parsers handle it.
1.79 kristaps 1571: */
1.67 kristaps 1572:
1.294 schwarze 1573: if (t == TOKEN_NONE)
1.277 schwarze 1574: return ROFF_CONT;
1.243 schwarze 1575:
1576: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1577:
1.296 schwarze 1578: return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.74 kristaps 1579: }
1580:
1.117 kristaps 1581: void
1.74 kristaps 1582: roff_endparse(struct roff *r)
1583: {
1.321 ! schwarze 1584: if (r->last != NULL)
1.221 schwarze 1585: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1586: r->last->line, r->last->col,
1.295 schwarze 1587: roff_name[r->last->tok]);
1.117 kristaps 1588:
1.321 ! schwarze 1589: if (r->eqn != NULL) {
1.221 schwarze 1590: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1.319 schwarze 1591: r->eqn->node->line, r->eqn->node->pos, "EQ");
1592: eqn_parse(r->eqn);
1593: r->eqn = NULL;
1.125 kristaps 1594: }
1595:
1.321 ! schwarze 1596: if (r->tbl != NULL) {
1.221 schwarze 1597: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1598: r->tbl->line, r->tbl->pos, "TS");
1.321 ! schwarze 1599: tbl_end(r->tbl);
! 1600: r->tbl = NULL;
1.117 kristaps 1601: }
1.67 kristaps 1602: }
1603:
1604: /*
1605: * Parse a roff node's type from the input buffer. This must be in the
1606: * form of ".foo xxx" in the usual way.
1607: */
1.294 schwarze 1608: static enum roff_tok
1.214 schwarze 1609: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1610: {
1.214 schwarze 1611: char *cp;
1.106 kristaps 1612: const char *mac;
1613: size_t maclen;
1.315 schwarze 1614: int deftype;
1.294 schwarze 1615: enum roff_tok t;
1.67 kristaps 1616:
1.214 schwarze 1617: cp = buf + *pos;
1618:
1619: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294 schwarze 1620: return TOKEN_NONE;
1.67 kristaps 1621:
1.214 schwarze 1622: mac = cp;
1623: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1624:
1.315 schwarze 1625: deftype = ROFFDEF_USER | ROFFDEF_REN;
1626: r->current_string = roff_getstrn(r, mac, maclen, &deftype);
1627: switch (deftype) {
1628: case ROFFDEF_USER:
1629: t = ROFF_USERDEF;
1630: break;
1631: case ROFFDEF_REN:
1632: t = ROFF_RENAMED;
1633: break;
1634: default:
1635: t = roffhash_find(r->reqtab, mac, maclen);
1636: break;
1637: }
1.294 schwarze 1638: if (t != TOKEN_NONE)
1.214 schwarze 1639: *pos = cp - buf;
1.277 schwarze 1640: return t;
1.67 kristaps 1641: }
1642:
1.266 schwarze 1643: /* --- handling of request blocks ----------------------------------------- */
1644:
1.67 kristaps 1645: static enum rofferr
1.76 kristaps 1646: roff_cblock(ROFF_ARGS)
1.67 kristaps 1647: {
1648:
1.79 kristaps 1649: /*
1650: * A block-close `..' should only be invoked as a child of an
1651: * ignore macro, otherwise raise a warning and just ignore it.
1652: */
1653:
1.238 schwarze 1654: if (r->last == NULL) {
1.221 schwarze 1655: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1656: ln, ppos, "..");
1.277 schwarze 1657: return ROFF_IGN;
1.76 kristaps 1658: }
1.67 kristaps 1659:
1.81 kristaps 1660: switch (r->last->tok) {
1.207 schwarze 1661: case ROFF_am:
1.220 schwarze 1662: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.207 schwarze 1663: case ROFF_ami:
1664: case ROFF_de:
1.108 schwarze 1665: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.207 schwarze 1666: case ROFF_dei:
1667: case ROFF_ig:
1.81 kristaps 1668: break;
1669: default:
1.221 schwarze 1670: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1671: ln, ppos, "..");
1.277 schwarze 1672: return ROFF_IGN;
1.76 kristaps 1673: }
1.67 kristaps 1674:
1.238 schwarze 1675: if (buf->buf[pos] != '\0')
1.217 schwarze 1676: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1677: ".. %s", buf->buf + pos);
1.71 kristaps 1678:
1679: roffnode_pop(r);
1.76 kristaps 1680: roffnode_cleanscope(r);
1.277 schwarze 1681: return ROFF_IGN;
1.71 kristaps 1682:
1.67 kristaps 1683: }
1684:
1.76 kristaps 1685: static void
1686: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1687: {
1688:
1.76 kristaps 1689: while (r->last) {
1.173 schwarze 1690: if (--r->last->endspan != 0)
1.76 kristaps 1691: break;
1692: roffnode_pop(r);
1693: }
1.67 kristaps 1694: }
1695:
1.195 schwarze 1696: static void
1697: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1698: {
1699:
1.76 kristaps 1700: if (NULL == r->last) {
1.221 schwarze 1701: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1702: ln, ppos, "\\}");
1.195 schwarze 1703: return;
1.76 kristaps 1704: }
1705:
1.82 kristaps 1706: switch (r->last->tok) {
1.207 schwarze 1707: case ROFF_el:
1708: case ROFF_ie:
1709: case ROFF_if:
1.82 kristaps 1710: break;
1711: default:
1.221 schwarze 1712: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1713: ln, ppos, "\\}");
1.195 schwarze 1714: return;
1.75 kristaps 1715: }
1716:
1.76 kristaps 1717: if (r->last->endspan > -1) {
1.221 schwarze 1718: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1719: ln, ppos, "\\}");
1.195 schwarze 1720: return;
1.76 kristaps 1721: }
1722:
1.75 kristaps 1723: roffnode_pop(r);
1.76 kristaps 1724: roffnode_cleanscope(r);
1.195 schwarze 1725: return;
1.76 kristaps 1726: }
1727:
1728: static enum rofferr
1.80 kristaps 1729: roff_block(ROFF_ARGS)
1.76 kristaps 1730: {
1.315 schwarze 1731: const char *name, *value;
1732: char *call, *cp, *iname, *rname;
1733: size_t csz, namesz, rsz;
1734: int deftype;
1.106 kristaps 1735:
1.220 schwarze 1736: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1737:
1.238 schwarze 1738: if (tok == ROFF_de1)
1.220 schwarze 1739: tok = ROFF_de;
1.251 schwarze 1740: else if (tok == ROFF_dei1)
1741: tok = ROFF_dei;
1.238 schwarze 1742: else if (tok == ROFF_am1)
1.220 schwarze 1743: tok = ROFF_am;
1.251 schwarze 1744: else if (tok == ROFF_ami1)
1745: tok = ROFF_ami;
1.220 schwarze 1746:
1747: /* Parse the macro name argument. */
1748:
1.238 schwarze 1749: cp = buf->buf + pos;
1750: if (tok == ROFF_ig) {
1.220 schwarze 1751: iname = NULL;
1752: namesz = 0;
1753: } else {
1754: iname = cp;
1755: namesz = roff_getname(r, &cp, ln, ppos);
1756: iname[namesz] = '\0';
1757: }
1.107 kristaps 1758:
1.220 schwarze 1759: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1760:
1.238 schwarze 1761: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315 schwarze 1762: deftype = ROFFDEF_USER;
1763: name = roff_getstrn(r, iname, namesz, &deftype);
1764: if (name == NULL) {
1.220 schwarze 1765: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1766: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1767: "%.*s", (int)namesz, iname);
1768: namesz = 0;
1769: } else
1770: namesz = strlen(name);
1771: } else
1772: name = iname;
1.107 kristaps 1773:
1.238 schwarze 1774: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1775: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.295 schwarze 1776: ln, ppos, roff_name[tok]);
1.277 schwarze 1777: return ROFF_IGN;
1.220 schwarze 1778: }
1.80 kristaps 1779:
1.106 kristaps 1780: roffnode_push(r, tok, name, ln, ppos);
1781:
1782: /*
1783: * At the beginning of a `de' macro, clear the existing string
1784: * with the same name, if there is one. New content will be
1.193 schwarze 1785: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1786: */
1.107 kristaps 1787:
1.311 schwarze 1788: if (tok == ROFF_de || tok == ROFF_dei) {
1.213 schwarze 1789: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.311 schwarze 1790: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.315 schwarze 1791: } else if (tok == ROFF_am || tok == ROFF_ami) {
1792: deftype = ROFFDEF_ANY;
1793: value = roff_getstrn(r, iname, namesz, &deftype);
1794: switch (deftype) { /* Before appending, ... */
1795: case ROFFDEF_PRE: /* copy predefined to user-defined. */
1796: roff_setstrn(&r->strtab, name, namesz,
1797: value, strlen(value), 0);
1798: break;
1799: case ROFFDEF_REN: /* call original standard macro. */
1800: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1801: (int)strlen(value), value);
1802: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1803: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1804: free(call);
1805: break;
1806: case ROFFDEF_STD: /* rename and call standard macro. */
1807: rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
1808: roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
1809: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1810: (int)rsz, rname);
1811: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1812: free(call);
1813: free(rname);
1814: break;
1815: default:
1816: break;
1817: }
1.311 schwarze 1818: }
1.76 kristaps 1819:
1.238 schwarze 1820: if (*cp == '\0')
1.277 schwarze 1821: return ROFF_IGN;
1.78 kristaps 1822:
1.220 schwarze 1823: /* Get the custom end marker. */
1.107 kristaps 1824:
1.220 schwarze 1825: iname = cp;
1.213 schwarze 1826: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1827:
1828: /* Resolve the end marker if it is indirect. */
1829:
1.238 schwarze 1830: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315 schwarze 1831: deftype = ROFFDEF_USER;
1832: name = roff_getstrn(r, iname, namesz, &deftype);
1833: if (name == NULL) {
1.220 schwarze 1834: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1835: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1836: "%.*s", (int)namesz, iname);
1837: namesz = 0;
1838: } else
1839: namesz = strlen(name);
1840: } else
1841: name = iname;
1842:
1.213 schwarze 1843: if (namesz)
1844: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1845:
1.238 schwarze 1846: if (*cp != '\0')
1.217 schwarze 1847: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.295 schwarze 1848: ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74 kristaps 1849:
1.277 schwarze 1850: return ROFF_IGN;
1.78 kristaps 1851: }
1852:
1853: static enum rofferr
1.80 kristaps 1854: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1855: {
1.294 schwarze 1856: enum roff_tok t;
1.79 kristaps 1857: int i, j;
1858:
1859: /*
1860: * First check whether a custom macro exists at this level. If
1861: * it does, then check against it. This is some of groff's
1862: * stranger behaviours. If we encountered a custom end-scope
1863: * tag and that tag also happens to be a "real" macro, then we
1864: * need to try interpreting it again as a real macro. If it's
1865: * not, then return ignore. Else continue.
1866: */
1867:
1868: if (r->last->end) {
1.130 kristaps 1869: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1870: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1871: break;
1872:
1.238 schwarze 1873: if (r->last->end[j] == '\0' &&
1874: (buf->buf[i] == '\0' ||
1875: buf->buf[i] == ' ' ||
1876: buf->buf[i] == '\t')) {
1.79 kristaps 1877: roffnode_pop(r);
1878: roffnode_cleanscope(r);
1879:
1.238 schwarze 1880: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1881: i++;
1882:
1883: pos = i;
1.238 schwarze 1884: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294 schwarze 1885: TOKEN_NONE)
1.277 schwarze 1886: return ROFF_RERUN;
1887: return ROFF_IGN;
1.79 kristaps 1888: }
1889: }
1890:
1891: /*
1892: * If we have no custom end-query or lookup failed, then try
1893: * pulling it out of the hashtable.
1894: */
1895:
1.238 schwarze 1896: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1897:
1.238 schwarze 1898: if (t != ROFF_cblock) {
1899: if (tok != ROFF_ig)
1900: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.277 schwarze 1901: return ROFF_IGN;
1.106 kristaps 1902: }
1.79 kristaps 1903:
1.277 schwarze 1904: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1905: }
1906:
1907: static enum rofferr
1.80 kristaps 1908: roff_block_text(ROFF_ARGS)
1.78 kristaps 1909: {
1910:
1.238 schwarze 1911: if (tok != ROFF_ig)
1912: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1913:
1.277 schwarze 1914: return ROFF_IGN;
1.78 kristaps 1915: }
1916:
1917: static enum rofferr
1.82 kristaps 1918: roff_cond_sub(ROFF_ARGS)
1919: {
1.294 schwarze 1920: enum roff_tok t;
1.139 kristaps 1921: char *ep;
1.198 schwarze 1922: int rr;
1.82 kristaps 1923:
1924: rr = r->last->rule;
1.139 kristaps 1925: roffnode_cleanscope(r);
1.144 kristaps 1926:
1.196 schwarze 1927: /*
1928: * If `\}' occurs on a macro line without a preceding macro,
1929: * drop the line completely.
1930: */
1931:
1.238 schwarze 1932: ep = buf->buf + pos;
1933: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1934: rr = 0;
1.196 schwarze 1935:
1.177 schwarze 1936: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1937:
1.238 schwarze 1938: while ((ep = strchr(ep, '\\')) != NULL) {
1.318 schwarze 1939: switch (ep[1]) {
1940: case '}':
1941: memmove(ep, ep + 2, strlen(ep + 2) + 1);
1942: roff_ccond(r, ln, ep - buf->buf);
1943: break;
1944: case '\0':
1945: ++ep;
1946: break;
1947: default:
1948: ep += 2;
1949: break;
1.197 schwarze 1950: }
1.177 schwarze 1951: }
1.318 schwarze 1952:
1953: /*
1954: * Fully handle known macros when they are structurally
1955: * required or when the conditional evaluated to true.
1956: */
1957:
1958: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1959: return t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT)
1960: ? (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs) : rr
1961: ? ROFF_CONT : ROFF_IGN;
1.82 kristaps 1962: }
1963:
1964: static enum rofferr
1965: roff_cond_text(ROFF_ARGS)
1.78 kristaps 1966: {
1.140 kristaps 1967: char *ep;
1.198 schwarze 1968: int rr;
1.82 kristaps 1969:
1970: rr = r->last->rule;
1.140 kristaps 1971: roffnode_cleanscope(r);
1.82 kristaps 1972:
1.238 schwarze 1973: ep = buf->buf + pos;
1974: while ((ep = strchr(ep, '\\')) != NULL) {
1975: if (*(++ep) == '}') {
1.197 schwarze 1976: *ep = '&';
1.238 schwarze 1977: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1978: }
1.247 schwarze 1979: if (*ep != '\0')
1980: ++ep;
1.78 kristaps 1981: }
1.277 schwarze 1982: return rr ? ROFF_CONT : ROFF_IGN;
1.74 kristaps 1983: }
1984:
1.266 schwarze 1985: /* --- handling of numeric and conditional expressions -------------------- */
1986:
1.204 schwarze 1987: /*
1988: * Parse a single signed integer number. Stop at the first non-digit.
1989: * If there is at least one digit, return success and advance the
1990: * parse point, else return failure and let the parse point unchanged.
1991: * Ignore overflows, treat them just like the C language.
1992: */
1.184 schwarze 1993: static int
1.261 schwarze 1994: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 1995: {
1.261 schwarze 1996: int myres, scaled, n, p;
1.206 schwarze 1997:
1998: if (NULL == res)
1999: res = &myres;
1.184 schwarze 2000:
2001: p = *pos;
2002: n = v[p] == '-';
1.261 schwarze 2003: if (n || v[p] == '+')
1.184 schwarze 2004: p++;
2005:
1.261 schwarze 2006: if (flags & ROFFNUM_WHITE)
2007: while (isspace((unsigned char)v[p]))
2008: p++;
2009:
1.184 schwarze 2010: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 2011: *res = 10 * *res + v[p] - '0';
1.184 schwarze 2012: if (p == *pos + n)
2013: return 0;
2014:
2015: if (n)
2016: *res = -*res;
2017:
1.254 schwarze 2018: /* Each number may be followed by one optional scaling unit. */
2019:
2020: switch (v[p]) {
2021: case 'f':
1.261 schwarze 2022: scaled = *res * 65536;
1.254 schwarze 2023: break;
2024: case 'i':
1.261 schwarze 2025: scaled = *res * 240;
1.254 schwarze 2026: break;
2027: case 'c':
1.261 schwarze 2028: scaled = *res * 240 / 2.54;
1.254 schwarze 2029: break;
2030: case 'v':
2031: case 'P':
1.261 schwarze 2032: scaled = *res * 40;
1.254 schwarze 2033: break;
2034: case 'm':
2035: case 'n':
1.261 schwarze 2036: scaled = *res * 24;
1.254 schwarze 2037: break;
2038: case 'p':
1.261 schwarze 2039: scaled = *res * 10 / 3;
1.254 schwarze 2040: break;
2041: case 'u':
1.261 schwarze 2042: scaled = *res;
1.254 schwarze 2043: break;
2044: case 'M':
1.261 schwarze 2045: scaled = *res * 6 / 25;
1.254 schwarze 2046: break;
2047: default:
1.261 schwarze 2048: scaled = *res;
1.254 schwarze 2049: p--;
2050: break;
2051: }
1.261 schwarze 2052: if (flags & ROFFNUM_SCALE)
2053: *res = scaled;
1.254 schwarze 2054:
2055: *pos = p + 1;
1.277 schwarze 2056: return 1;
1.184 schwarze 2057: }
2058:
1.198 schwarze 2059: /*
2060: * Evaluate a string comparison condition.
2061: * The first character is the delimiter.
2062: * Succeed if the string up to its second occurrence
2063: * matches the string up to its third occurence.
2064: * Advance the cursor after the third occurrence
2065: * or lacking that, to the end of the line.
2066: */
2067: static int
2068: roff_evalstrcond(const char *v, int *pos)
2069: {
2070: const char *s1, *s2, *s3;
2071: int match;
2072:
2073: match = 0;
2074: s1 = v + *pos; /* initial delimiter */
2075: s2 = s1 + 1; /* for scanning the first string */
2076: s3 = strchr(s2, *s1); /* for scanning the second string */
2077:
2078: if (NULL == s3) /* found no middle delimiter */
2079: goto out;
2080:
2081: while ('\0' != *++s3) {
2082: if (*s2 != *s3) { /* mismatch */
2083: s3 = strchr(s3, *s1);
2084: break;
2085: }
2086: if (*s3 == *s1) { /* found the final delimiter */
2087: match = 1;
2088: break;
2089: }
2090: s2++;
2091: }
2092:
2093: out:
2094: if (NULL == s3)
2095: s3 = strchr(s2, '\0');
1.242 schwarze 2096: else if (*s3 != '\0')
1.198 schwarze 2097: s3++;
2098: *pos = s3 - v;
1.277 schwarze 2099: return match;
1.198 schwarze 2100: }
2101:
1.204 schwarze 2102: /*
2103: * Evaluate an optionally negated single character, numerical,
2104: * or string condition.
2105: */
1.198 schwarze 2106: static int
1.271 schwarze 2107: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88 kristaps 2108: {
1.271 schwarze 2109: char *cp, *name;
2110: size_t sz;
1.315 schwarze 2111: int deftype, number, savepos, istrue, wanttrue;
1.88 kristaps 2112:
1.198 schwarze 2113: if ('!' == v[*pos]) {
2114: wanttrue = 0;
2115: (*pos)++;
2116: } else
2117: wanttrue = 1;
2118:
1.88 kristaps 2119: switch (v[*pos]) {
1.240 schwarze 2120: case '\0':
1.277 schwarze 2121: return 0;
1.207 schwarze 2122: case 'n':
2123: case 'o':
1.88 kristaps 2124: (*pos)++;
1.277 schwarze 2125: return wanttrue;
1.207 schwarze 2126: case 'c':
2127: case 'e':
2128: case 't':
1.239 schwarze 2129: case 'v':
1.88 kristaps 2130: (*pos)++;
1.277 schwarze 2131: return !wanttrue;
1.310 schwarze 2132: case 'd':
1.271 schwarze 2133: case 'r':
1.310 schwarze 2134: cp = v + *pos + 1;
2135: while (*cp == ' ')
2136: cp++;
2137: name = cp;
2138: sz = roff_getname(r, &cp, ln, cp - v);
1.315 schwarze 2139: if (sz == 0)
2140: istrue = 0;
2141: else if (v[*pos] == 'r')
2142: istrue = roff_hasregn(r, name, sz);
2143: else {
2144: deftype = ROFFDEF_ANY;
2145: roff_getstrn(r, name, sz, &deftype);
2146: istrue = !!deftype;
2147: }
1.271 schwarze 2148: *pos = cp - v;
1.310 schwarze 2149: return istrue == wanttrue;
1.88 kristaps 2150: default:
2151: break;
2152: }
2153:
1.241 schwarze 2154: savepos = *pos;
1.261 schwarze 2155: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277 schwarze 2156: return (number > 0) == wanttrue;
1.241 schwarze 2157: else if (*pos == savepos)
1.277 schwarze 2158: return roff_evalstrcond(v, pos) == wanttrue;
1.204 schwarze 2159: else
1.277 schwarze 2160: return 0;
1.88 kristaps 2161: }
2162:
1.74 kristaps 2163: static enum rofferr
1.103 kristaps 2164: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 2165: {
1.123 schwarze 2166:
1.277 schwarze 2167: return ROFF_IGN;
1.89 kristaps 2168: }
2169:
1.104 kristaps 2170: static enum rofferr
1.251 schwarze 2171: roff_insec(ROFF_ARGS)
2172: {
2173:
2174: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.295 schwarze 2175: ln, ppos, roff_name[tok]);
1.277 schwarze 2176: return ROFF_IGN;
1.251 schwarze 2177: }
2178:
2179: static enum rofferr
2180: roff_unsupp(ROFF_ARGS)
2181: {
2182:
2183: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.295 schwarze 2184: ln, ppos, roff_name[tok]);
1.277 schwarze 2185: return ROFF_IGN;
1.251 schwarze 2186: }
2187:
2188: static enum rofferr
1.82 kristaps 2189: roff_cond(ROFF_ARGS)
1.74 kristaps 2190: {
1.173 schwarze 2191:
2192: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 2193:
1.207 schwarze 2194: /*
1.134 kristaps 2195: * An `.el' has no conditional body: it will consume the value
2196: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 2197: * defaults to DENY.
1.134 kristaps 2198: *
2199: * If we're not an `el', however, then evaluate the conditional.
2200: */
1.133 kristaps 2201:
1.238 schwarze 2202: r->last->rule = tok == ROFF_el ?
1.207 schwarze 2203: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 2204: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 2205:
1.134 kristaps 2206: /*
2207: * An if-else will put the NEGATION of the current evaluated
2208: * conditional into the stack of rules.
2209: */
2210:
1.238 schwarze 2211: if (tok == ROFF_ie) {
1.223 schwarze 2212: if (r->rstackpos + 1 == r->rstacksz) {
2213: r->rstacksz += 16;
2214: r->rstack = mandoc_reallocarray(r->rstack,
2215: r->rstacksz, sizeof(int));
1.134 kristaps 2216: }
1.198 schwarze 2217: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 2218: }
1.88 kristaps 2219:
2220: /* If the parent has false as its rule, then so do we. */
2221:
1.198 schwarze 2222: if (r->last->parent && !r->last->parent->rule)
2223: r->last->rule = 0;
1.88 kristaps 2224:
2225: /*
1.173 schwarze 2226: * Determine scope.
2227: * If there is nothing on the line after the conditional,
2228: * not even whitespace, use next-line scope.
1.88 kristaps 2229: */
1.74 kristaps 2230:
1.238 schwarze 2231: if (buf->buf[pos] == '\0') {
1.173 schwarze 2232: r->last->endspan = 2;
2233: goto out;
2234: }
2235:
1.238 schwarze 2236: while (buf->buf[pos] == ' ')
1.173 schwarze 2237: pos++;
2238:
2239: /* An opening brace requests multiline scope. */
1.75 kristaps 2240:
1.238 schwarze 2241: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 2242: r->last->endspan = -1;
2243: pos += 2;
1.272 schwarze 2244: while (buf->buf[pos] == ' ')
2245: pos++;
1.173 schwarze 2246: goto out;
1.207 schwarze 2247: }
1.74 kristaps 2248:
1.77 kristaps 2249: /*
1.173 schwarze 2250: * Anything else following the conditional causes
2251: * single-line scope. Warn if the scope contains
2252: * nothing but trailing whitespace.
1.77 kristaps 2253: */
2254:
1.238 schwarze 2255: if (buf->buf[pos] == '\0')
1.216 schwarze 2256: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.295 schwarze 2257: ln, ppos, roff_name[tok]);
1.77 kristaps 2258:
1.173 schwarze 2259: r->last->endspan = 1;
1.74 kristaps 2260:
1.173 schwarze 2261: out:
1.75 kristaps 2262: *offs = pos;
1.277 schwarze 2263: return ROFF_RERUN;
1.83 schwarze 2264: }
2265:
2266: static enum rofferr
1.92 schwarze 2267: roff_ds(ROFF_ARGS)
2268: {
1.212 schwarze 2269: char *string;
2270: const char *name;
2271: size_t namesz;
1.96 kristaps 2272:
1.251 schwarze 2273: /* Ignore groff compatibility mode for now. */
2274:
2275: if (tok == ROFF_ds1)
2276: tok = ROFF_ds;
2277: else if (tok == ROFF_as1)
2278: tok = ROFF_as;
2279:
1.96 kristaps 2280: /*
1.212 schwarze 2281: * The first word is the name of the string.
2282: * If it is empty or terminated by an escape sequence,
2283: * abort the `ds' request without defining anything.
1.96 kristaps 2284: */
1.92 schwarze 2285:
1.238 schwarze 2286: name = string = buf->buf + pos;
2287: if (*name == '\0')
1.277 schwarze 2288: return ROFF_IGN;
1.92 schwarze 2289:
1.212 schwarze 2290: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 2291: if (name[namesz] == '\\')
1.277 schwarze 2292: return ROFF_IGN;
1.212 schwarze 2293:
2294: /* Read past the initial double-quote, if any. */
1.238 schwarze 2295: if (*string == '"')
1.92 schwarze 2296: string++;
2297:
1.96 kristaps 2298: /* The rest is the value. */
1.212 schwarze 2299: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2300: ROFF_as == tok);
1.311 schwarze 2301: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.277 schwarze 2302: return ROFF_IGN;
1.92 schwarze 2303: }
2304:
1.204 schwarze 2305: /*
2306: * Parse a single operator, one or two characters long.
2307: * If the operator is recognized, return success and advance the
2308: * parse point, else return failure and let the parse point unchanged.
2309: */
2310: static int
2311: roff_getop(const char *v, int *pos, char *res)
2312: {
2313:
2314: *res = v[*pos];
2315:
2316: switch (*res) {
1.207 schwarze 2317: case '+':
2318: case '-':
2319: case '*':
2320: case '/':
2321: case '%':
2322: case '&':
2323: case ':':
1.204 schwarze 2324: break;
2325: case '<':
2326: switch (v[*pos + 1]) {
1.207 schwarze 2327: case '=':
1.204 schwarze 2328: *res = 'l';
2329: (*pos)++;
2330: break;
1.207 schwarze 2331: case '>':
1.204 schwarze 2332: *res = '!';
2333: (*pos)++;
2334: break;
1.207 schwarze 2335: case '?':
1.204 schwarze 2336: *res = 'i';
2337: (*pos)++;
2338: break;
2339: default:
2340: break;
2341: }
2342: break;
2343: case '>':
2344: switch (v[*pos + 1]) {
1.207 schwarze 2345: case '=':
1.204 schwarze 2346: *res = 'g';
2347: (*pos)++;
2348: break;
1.207 schwarze 2349: case '?':
1.204 schwarze 2350: *res = 'a';
2351: (*pos)++;
2352: break;
2353: default:
2354: break;
2355: }
2356: break;
2357: case '=':
2358: if ('=' == v[*pos + 1])
2359: (*pos)++;
2360: break;
2361: default:
1.277 schwarze 2362: return 0;
1.204 schwarze 2363: }
2364: (*pos)++;
2365:
1.277 schwarze 2366: return *res;
1.204 schwarze 2367: }
2368:
2369: /*
2370: * Evaluate either a parenthesized numeric expression
2371: * or a single signed integer number.
2372: */
2373: static int
1.235 schwarze 2374: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2375: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2376: {
2377:
2378: if ('(' != v[*pos])
1.277 schwarze 2379: return roff_getnum(v, pos, res, flags);
1.204 schwarze 2380:
2381: (*pos)++;
1.261 schwarze 2382: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277 schwarze 2383: return 0;
1.204 schwarze 2384:
1.206 schwarze 2385: /*
2386: * Omission of the closing parenthesis
2387: * is an error in validation mode,
2388: * but ignored in evaluation mode.
2389: */
2390:
1.204 schwarze 2391: if (')' == v[*pos])
2392: (*pos)++;
1.206 schwarze 2393: else if (NULL == res)
1.277 schwarze 2394: return 0;
1.204 schwarze 2395:
1.277 schwarze 2396: return 1;
1.204 schwarze 2397: }
2398:
2399: /*
2400: * Evaluate a complete numeric expression.
2401: * Proceed left to right, there is no concept of precedence.
2402: */
2403: static int
1.235 schwarze 2404: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2405: int *pos, int *res, int flags)
1.204 schwarze 2406: {
2407: int mypos, operand2;
2408: char operator;
2409:
2410: if (NULL == pos) {
2411: mypos = 0;
2412: pos = &mypos;
2413: }
2414:
1.261 schwarze 2415: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2416: while (isspace((unsigned char)v[*pos]))
2417: (*pos)++;
2418:
1.261 schwarze 2419: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277 schwarze 2420: return 0;
1.204 schwarze 2421:
2422: while (1) {
1.261 schwarze 2423: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2424: while (isspace((unsigned char)v[*pos]))
2425: (*pos)++;
2426:
2427: if ( ! roff_getop(v, pos, &operator))
2428: break;
2429:
1.261 schwarze 2430: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2431: while (isspace((unsigned char)v[*pos]))
2432: (*pos)++;
2433:
1.261 schwarze 2434: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277 schwarze 2435: return 0;
1.204 schwarze 2436:
1.261 schwarze 2437: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2438: while (isspace((unsigned char)v[*pos]))
2439: (*pos)++;
1.206 schwarze 2440:
2441: if (NULL == res)
2442: continue;
1.204 schwarze 2443:
2444: switch (operator) {
1.207 schwarze 2445: case '+':
1.204 schwarze 2446: *res += operand2;
2447: break;
1.207 schwarze 2448: case '-':
1.204 schwarze 2449: *res -= operand2;
2450: break;
1.207 schwarze 2451: case '*':
1.204 schwarze 2452: *res *= operand2;
2453: break;
1.207 schwarze 2454: case '/':
1.244 schwarze 2455: if (operand2 == 0) {
1.235 schwarze 2456: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2457: r->parse, ln, *pos, v);
2458: *res = 0;
2459: break;
2460: }
1.204 schwarze 2461: *res /= operand2;
2462: break;
1.207 schwarze 2463: case '%':
1.244 schwarze 2464: if (operand2 == 0) {
2465: mandoc_msg(MANDOCERR_DIVZERO,
2466: r->parse, ln, *pos, v);
2467: *res = 0;
2468: break;
2469: }
1.204 schwarze 2470: *res %= operand2;
2471: break;
1.207 schwarze 2472: case '<':
1.204 schwarze 2473: *res = *res < operand2;
2474: break;
1.207 schwarze 2475: case '>':
1.204 schwarze 2476: *res = *res > operand2;
2477: break;
1.207 schwarze 2478: case 'l':
1.204 schwarze 2479: *res = *res <= operand2;
2480: break;
1.207 schwarze 2481: case 'g':
1.204 schwarze 2482: *res = *res >= operand2;
2483: break;
1.207 schwarze 2484: case '=':
1.204 schwarze 2485: *res = *res == operand2;
2486: break;
1.207 schwarze 2487: case '!':
1.204 schwarze 2488: *res = *res != operand2;
2489: break;
1.207 schwarze 2490: case '&':
1.204 schwarze 2491: *res = *res && operand2;
2492: break;
1.207 schwarze 2493: case ':':
1.204 schwarze 2494: *res = *res || operand2;
2495: break;
1.207 schwarze 2496: case 'i':
1.204 schwarze 2497: if (operand2 < *res)
2498: *res = operand2;
2499: break;
1.207 schwarze 2500: case 'a':
1.204 schwarze 2501: if (operand2 > *res)
2502: *res = operand2;
2503: break;
2504: default:
2505: abort();
2506: }
2507: }
1.277 schwarze 2508: return 1;
1.204 schwarze 2509: }
2510:
1.266 schwarze 2511: /* --- register management ------------------------------------------------ */
2512:
1.180 schwarze 2513: void
1.187 schwarze 2514: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2515: {
1.180 schwarze 2516: struct roffreg *reg;
2517:
2518: /* Search for an existing register with the same name. */
2519: reg = r->regtab;
2520:
2521: while (reg && strcmp(name, reg->key.p))
2522: reg = reg->next;
1.147 kristaps 2523:
1.180 schwarze 2524: if (NULL == reg) {
2525: /* Create a new register. */
2526: reg = mandoc_malloc(sizeof(struct roffreg));
2527: reg->key.p = mandoc_strdup(name);
2528: reg->key.sz = strlen(name);
1.187 schwarze 2529: reg->val = 0;
1.180 schwarze 2530: reg->next = r->regtab;
2531: r->regtab = reg;
2532: }
2533:
1.187 schwarze 2534: if ('+' == sign)
2535: reg->val += val;
2536: else if ('-' == sign)
2537: reg->val -= val;
2538: else
2539: reg->val = val;
1.147 kristaps 2540: }
2541:
1.192 schwarze 2542: /*
2543: * Handle some predefined read-only number registers.
2544: * For now, return -1 if the requested register is not predefined;
2545: * in case a predefined read-only register having the value -1
2546: * were to turn up, another special value would have to be chosen.
2547: */
2548: static int
1.273 schwarze 2549: roff_getregro(const struct roff *r, const char *name)
1.192 schwarze 2550: {
2551:
2552: switch (*name) {
1.273 schwarze 2553: case '$': /* Number of arguments of the last macro evaluated. */
1.277 schwarze 2554: return r->argc;
1.207 schwarze 2555: case 'A': /* ASCII approximation mode is always off. */
1.277 schwarze 2556: return 0;
1.207 schwarze 2557: case 'g': /* Groff compatibility mode is always on. */
1.277 schwarze 2558: return 1;
1.207 schwarze 2559: case 'H': /* Fixed horizontal resolution. */
1.277 schwarze 2560: return 24;
1.207 schwarze 2561: case 'j': /* Always adjust left margin only. */
1.277 schwarze 2562: return 0;
1.207 schwarze 2563: case 'T': /* Some output device is always defined. */
1.277 schwarze 2564: return 1;
1.207 schwarze 2565: case 'V': /* Fixed vertical resolution. */
1.277 schwarze 2566: return 40;
1.192 schwarze 2567: default:
1.277 schwarze 2568: return -1;
1.192 schwarze 2569: }
2570: }
2571:
1.181 schwarze 2572: int
1.180 schwarze 2573: roff_getreg(const struct roff *r, const char *name)
1.147 kristaps 2574: {
1.180 schwarze 2575: struct roffreg *reg;
1.192 schwarze 2576: int val;
2577:
2578: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.273 schwarze 2579: val = roff_getregro(r, name + 1);
1.192 schwarze 2580: if (-1 != val)
1.277 schwarze 2581: return val;
1.192 schwarze 2582: }
1.180 schwarze 2583:
2584: for (reg = r->regtab; reg; reg = reg->next)
2585: if (0 == strcmp(name, reg->key.p))
1.277 schwarze 2586: return reg->val;
1.181 schwarze 2587:
1.277 schwarze 2588: return 0;
1.181 schwarze 2589: }
2590:
2591: static int
2592: roff_getregn(const struct roff *r, const char *name, size_t len)
2593: {
2594: struct roffreg *reg;
1.192 schwarze 2595: int val;
2596:
2597: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2598: val = roff_getregro(r, name + 1);
1.192 schwarze 2599: if (-1 != val)
1.277 schwarze 2600: return val;
1.192 schwarze 2601: }
1.181 schwarze 2602:
2603: for (reg = r->regtab; reg; reg = reg->next)
2604: if (len == reg->key.sz &&
2605: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2606: return reg->val;
1.271 schwarze 2607:
1.277 schwarze 2608: return 0;
1.271 schwarze 2609: }
2610:
2611: static int
2612: roff_hasregn(const struct roff *r, const char *name, size_t len)
2613: {
2614: struct roffreg *reg;
2615: int val;
2616:
2617: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2618: val = roff_getregro(r, name + 1);
1.271 schwarze 2619: if (-1 != val)
1.277 schwarze 2620: return 1;
1.271 schwarze 2621: }
2622:
2623: for (reg = r->regtab; reg; reg = reg->next)
2624: if (len == reg->key.sz &&
2625: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2626: return 1;
1.147 kristaps 2627:
1.277 schwarze 2628: return 0;
1.147 kristaps 2629: }
2630:
1.180 schwarze 2631: static void
2632: roff_freereg(struct roffreg *reg)
1.147 kristaps 2633: {
1.180 schwarze 2634: struct roffreg *old_reg;
1.147 kristaps 2635:
1.180 schwarze 2636: while (NULL != reg) {
2637: free(reg->key.p);
2638: old_reg = reg;
2639: reg = reg->next;
2640: free(old_reg);
2641: }
1.147 kristaps 2642: }
1.92 schwarze 2643:
2644: static enum rofferr
1.89 kristaps 2645: roff_nr(ROFF_ARGS)
1.83 schwarze 2646: {
1.212 schwarze 2647: char *key, *val;
2648: size_t keysz;
1.138 kristaps 2649: int iv;
1.187 schwarze 2650: char sign;
1.89 kristaps 2651:
1.238 schwarze 2652: key = val = buf->buf + pos;
2653: if (*key == '\0')
1.277 schwarze 2654: return ROFF_IGN;
1.212 schwarze 2655:
2656: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2657: if (key[keysz] == '\\')
1.277 schwarze 2658: return ROFF_IGN;
1.212 schwarze 2659: key[keysz] = '\0';
1.89 kristaps 2660:
1.187 schwarze 2661: sign = *val;
1.238 schwarze 2662: if (sign == '+' || sign == '-')
1.187 schwarze 2663: val++;
2664:
1.261 schwarze 2665: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.204 schwarze 2666: roff_setreg(r, key, iv, sign);
1.109 kristaps 2667:
1.277 schwarze 2668: return ROFF_IGN;
1.203 schwarze 2669: }
2670:
2671: static enum rofferr
2672: roff_rr(ROFF_ARGS)
2673: {
2674: struct roffreg *reg, **prev;
1.212 schwarze 2675: char *name, *cp;
2676: size_t namesz;
1.203 schwarze 2677:
1.238 schwarze 2678: name = cp = buf->buf + pos;
2679: if (*name == '\0')
1.277 schwarze 2680: return ROFF_IGN;
1.212 schwarze 2681: namesz = roff_getname(r, &cp, ln, pos);
2682: name[namesz] = '\0';
1.203 schwarze 2683:
2684: prev = &r->regtab;
2685: while (1) {
2686: reg = *prev;
1.238 schwarze 2687: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2688: break;
2689: prev = ®->next;
2690: }
1.238 schwarze 2691: if (reg != NULL) {
1.203 schwarze 2692: *prev = reg->next;
2693: free(reg->key.p);
2694: free(reg);
2695: }
1.277 schwarze 2696: return ROFF_IGN;
1.122 schwarze 2697: }
2698:
1.266 schwarze 2699: /* --- handler functions for roff requests -------------------------------- */
2700:
1.122 schwarze 2701: static enum rofferr
2702: roff_rm(ROFF_ARGS)
2703: {
2704: const char *name;
2705: char *cp;
1.212 schwarze 2706: size_t namesz;
1.122 schwarze 2707:
1.238 schwarze 2708: cp = buf->buf + pos;
2709: while (*cp != '\0') {
1.212 schwarze 2710: name = cp;
1.238 schwarze 2711: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2712: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.311 schwarze 2713: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.238 schwarze 2714: if (name[namesz] == '\\')
1.212 schwarze 2715: break;
1.122 schwarze 2716: }
1.277 schwarze 2717: return ROFF_IGN;
1.178 schwarze 2718: }
2719:
2720: static enum rofferr
2721: roff_it(ROFF_ARGS)
2722: {
2723: int iv;
2724:
2725: /* Parse the number of lines. */
1.261 schwarze 2726:
2727: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 2728: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2729: ln, ppos, buf->buf + 1);
1.277 schwarze 2730: return ROFF_IGN;
1.178 schwarze 2731: }
2732:
1.262 schwarze 2733: while (isspace((unsigned char)buf->buf[pos]))
2734: pos++;
2735:
2736: /*
2737: * Arm the input line trap.
2738: * Special-casing "an-trap" is an ugly workaround to cope
2739: * with DocBook stupidly fiddling with man(7) internals.
2740: */
1.261 schwarze 2741:
1.178 schwarze 2742: roffit_lines = iv;
1.262 schwarze 2743: roffit_macro = mandoc_strdup(iv != 1 ||
2744: strcmp(buf->buf + pos, "an-trap") ?
2745: buf->buf + pos : "br");
1.277 schwarze 2746: return ROFF_IGN;
1.175 schwarze 2747: }
2748:
2749: static enum rofferr
2750: roff_Dd(ROFF_ARGS)
2751: {
1.315 schwarze 2752: int mask;
2753: enum roff_tok t, te;
1.227 schwarze 2754:
1.315 schwarze 2755: switch (tok) {
2756: case ROFF_Dd:
2757: tok = MDOC_Dd;
2758: te = MDOC_MAX;
2759: if (r->format == 0)
2760: r->format = MPARSE_MDOC;
2761: mask = MPARSE_MDOC | MPARSE_QUICK;
2762: break;
2763: case ROFF_TH:
2764: tok = MAN_TH;
2765: te = MAN_MAX;
2766: if (r->format == 0)
2767: r->format = MPARSE_MAN;
2768: mask = MPARSE_QUICK;
2769: break;
2770: default:
2771: abort();
2772: }
2773: if ((r->options & mask) == 0)
2774: for (t = tok; t < te; t++)
2775: roff_setstr(r, roff_name[t], NULL, 0);
1.277 schwarze 2776: return ROFF_CONT;
1.109 kristaps 2777: }
2778:
2779: static enum rofferr
2780: roff_TE(ROFF_ARGS)
2781: {
1.321 ! schwarze 2782: if (r->tbl == NULL) {
1.221 schwarze 2783: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2784: ln, ppos, "TE");
1.321 ! schwarze 2785: return ROFF_IGN;
! 2786: }
! 2787: if (tbl_end(r->tbl) == 0) {
! 2788: r->tbl = NULL;
1.258 schwarze 2789: free(buf->buf);
2790: buf->buf = mandoc_strdup(".sp");
2791: buf->sz = 4;
1.277 schwarze 2792: return ROFF_REPARSE;
1.258 schwarze 2793: }
1.321 ! schwarze 2794: r->tbl = NULL;
1.277 schwarze 2795: return ROFF_IGN;
1.112 kristaps 2796: }
2797:
2798: static enum rofferr
2799: roff_T_(ROFF_ARGS)
2800: {
2801:
2802: if (NULL == r->tbl)
1.221 schwarze 2803: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2804: ln, ppos, "T&");
1.112 kristaps 2805: else
1.296 schwarze 2806: tbl_restart(ln, ppos, r->tbl);
1.112 kristaps 2807:
1.277 schwarze 2808: return ROFF_IGN;
1.109 kristaps 2809: }
2810:
1.230 schwarze 2811: /*
2812: * Handle in-line equation delimiters.
2813: */
2814: static enum rofferr
1.238 schwarze 2815: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2816: {
1.233 schwarze 2817: char *cp1, *cp2;
2818: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2819:
1.230 schwarze 2820: /*
2821: * Outside equations, look for an opening delimiter.
2822: * If we are inside an equation, we already know it is
2823: * in-line, or this function wouldn't have been called;
2824: * so look for a closing delimiter.
2825: */
2826:
1.238 schwarze 2827: cp1 = buf->buf + pos;
1.230 schwarze 2828: cp2 = strchr(cp1, r->eqn == NULL ?
2829: r->last_eqn->odelim : r->last_eqn->cdelim);
2830: if (cp2 == NULL)
1.277 schwarze 2831: return ROFF_CONT;
1.230 schwarze 2832:
1.233 schwarze 2833: *cp2++ = '\0';
2834: bef_pr = bef_nl = aft_nl = aft_pr = "";
2835:
2836: /* Handle preceding text, protecting whitespace. */
2837:
1.238 schwarze 2838: if (*buf->buf != '\0') {
1.233 schwarze 2839: if (r->eqn == NULL)
2840: bef_pr = "\\&";
2841: bef_nl = "\n";
2842: }
2843:
2844: /*
2845: * Prepare replacing the delimiter with an equation macro
2846: * and drop leading white space from the equation.
2847: */
1.230 schwarze 2848:
1.233 schwarze 2849: if (r->eqn == NULL) {
2850: while (*cp2 == ' ')
2851: cp2++;
2852: mac = ".EQ";
2853: } else
2854: mac = ".EN";
2855:
2856: /* Handle following text, protecting whitespace. */
2857:
2858: if (*cp2 != '\0') {
2859: aft_nl = "\n";
2860: if (r->eqn != NULL)
2861: aft_pr = "\\&";
2862: }
2863:
2864: /* Do the actual replacement. */
2865:
1.238 schwarze 2866: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2867: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2868: free(buf->buf);
2869: buf->buf = cp1;
1.230 schwarze 2870:
2871: /* Toggle the in-line state of the eqn subsystem. */
2872:
2873: r->eqn_inline = r->eqn == NULL;
1.277 schwarze 2874: return ROFF_REPARSE;
1.151 kristaps 2875: }
2876:
1.235 schwarze 2877: static enum rofferr
2878: roff_EQ(ROFF_ARGS)
1.125 kristaps 2879: {
1.319 schwarze 2880: struct roff_node *n;
2881:
2882: n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
2883: if (ln > r->man->last->line)
2884: n->flags |= NODE_LINE;
2885: n->eqn = mandoc_calloc(1, sizeof(*n->eqn));
2886: n->eqn->expectargs = UINT_MAX;
2887: roff_node_append(r->man, n);
2888: r->man->next = ROFF_NEXT_SIBLING;
1.125 kristaps 2889:
1.238 schwarze 2890: assert(r->eqn == NULL);
1.319 schwarze 2891: if (r->last_eqn == NULL)
2892: r->last_eqn = eqn_alloc(r->parse);
2893: else
2894: eqn_reset(r->last_eqn);
2895: r->eqn = r->last_eqn;
2896: r->eqn->node = n;
1.151 kristaps 2897:
1.238 schwarze 2898: if (buf->buf[pos] != '\0')
1.235 schwarze 2899: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 2900: ".EQ %s", buf->buf + pos);
1.151 kristaps 2901:
1.277 schwarze 2902: return ROFF_IGN;
1.125 kristaps 2903: }
2904:
2905: static enum rofferr
2906: roff_EN(ROFF_ARGS)
2907: {
1.319 schwarze 2908: if (r->eqn != NULL) {
2909: eqn_parse(r->eqn);
2910: r->eqn = NULL;
2911: } else
2912: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
2913: if (buf->buf[pos] != '\0')
2914: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
2915: "EN %s", buf->buf + pos);
1.277 schwarze 2916: return ROFF_IGN;
1.125 kristaps 2917: }
2918:
2919: static enum rofferr
1.109 kristaps 2920: roff_TS(ROFF_ARGS)
2921: {
1.321 ! schwarze 2922: if (r->tbl != NULL) {
1.221 schwarze 2923: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2924: ln, ppos, "TS breaks TS");
1.321 ! schwarze 2925: tbl_end(r->tbl);
1.115 kristaps 2926: }
1.321 ! schwarze 2927: r->tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 2928: if (r->last_tbl)
1.321 ! schwarze 2929: r->last_tbl->next = r->tbl;
1.113 kristaps 2930: else
1.321 ! schwarze 2931: r->first_tbl = r->tbl;
! 2932: r->last_tbl = r->tbl;
1.297 schwarze 2933: return ROFF_IGN;
2934: }
2935:
2936: static enum rofferr
2937: roff_onearg(ROFF_ARGS)
2938: {
2939: struct roff_node *n;
2940: char *cp;
1.305 schwarze 2941: int npos;
1.297 schwarze 2942:
1.302 schwarze 2943: if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.320 schwarze 2944: (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
2945: tok == ROFF_ti))
1.302 schwarze 2946: man_breakscope(r->man, tok);
2947:
1.309 schwarze 2948: if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.305 schwarze 2949: r->man->last = roffce_node;
2950: r->man->next = ROFF_NEXT_SIBLING;
2951: }
2952:
1.297 schwarze 2953: roff_elem_alloc(r->man, ln, ppos, tok);
2954: n = r->man->last;
2955:
2956: cp = buf->buf + pos;
2957: if (*cp != '\0') {
2958: while (*cp != '\0' && *cp != ' ')
2959: cp++;
2960: while (*cp == ' ')
2961: *cp++ = '\0';
2962: if (*cp != '\0')
2963: mandoc_vmsg(MANDOCERR_ARG_EXCESS,
2964: r->parse, ln, cp - buf->buf,
2965: "%s ... %s", roff_name[tok], cp);
2966: roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.300 schwarze 2967: }
2968:
1.309 schwarze 2969: if (tok == ROFF_ce || tok == ROFF_rj) {
2970: if (r->man->last->type == ROFFT_ELEM) {
1.305 schwarze 2971: roff_word_alloc(r->man, ln, pos, "1");
2972: r->man->last->flags |= NODE_NOSRC;
2973: }
2974: npos = 0;
2975: if (roff_evalnum(r, ln, r->man->last->string, &npos,
2976: &roffce_lines, 0) == 0) {
2977: mandoc_vmsg(MANDOCERR_CE_NONUM,
2978: r->parse, ln, pos, "ce %s", buf->buf + pos);
2979: roffce_lines = 1;
2980: }
2981: if (roffce_lines < 1) {
2982: r->man->last = r->man->last->parent;
2983: roffce_node = NULL;
2984: roffce_lines = 0;
2985: } else
2986: roffce_node = r->man->last->parent;
2987: } else {
2988: n->flags |= NODE_VALID | NODE_ENDED;
2989: r->man->last = n;
2990: }
2991: n->flags |= NODE_LINE;
1.300 schwarze 2992: r->man->next = ROFF_NEXT_SIBLING;
2993: return ROFF_IGN;
2994: }
2995:
2996: static enum rofferr
2997: roff_manyarg(ROFF_ARGS)
2998: {
2999: struct roff_node *n;
3000: char *sp, *ep;
3001:
3002: roff_elem_alloc(r->man, ln, ppos, tok);
3003: n = r->man->last;
3004:
3005: for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
3006: while (*ep != '\0' && *ep != ' ')
3007: ep++;
3008: while (*ep == ' ')
3009: *ep++ = '\0';
3010: roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.297 schwarze 3011: }
3012:
3013: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3014: r->man->last = n;
3015: r->man->next = ROFF_NEXT_SIBLING;
1.277 schwarze 3016: return ROFF_IGN;
1.251 schwarze 3017: }
3018:
3019: static enum rofferr
1.311 schwarze 3020: roff_als(ROFF_ARGS)
3021: {
3022: char *oldn, *newn, *end, *value;
3023: size_t oldsz, newsz, valsz;
3024:
3025: newn = oldn = buf->buf + pos;
3026: if (*newn == '\0')
3027: return ROFF_IGN;
3028:
3029: newsz = roff_getname(r, &oldn, ln, pos);
3030: if (newn[newsz] == '\\' || *oldn == '\0')
3031: return ROFF_IGN;
3032:
3033: end = oldn;
3034: oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
3035: if (oldsz == 0)
3036: return ROFF_IGN;
3037:
1.315 schwarze 3038: valsz = mandoc_asprintf(&value, ".%.*s \\$*\\\"\n",
3039: (int)oldsz, oldn);
1.311 schwarze 3040: roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
3041: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3042: free(value);
3043: return ROFF_IGN;
3044: }
3045:
3046: static enum rofferr
1.296 schwarze 3047: roff_br(ROFF_ARGS)
1.251 schwarze 3048: {
1.302 schwarze 3049: if (r->man->flags & (MAN_BLINE | MAN_ELINE))
3050: man_breakscope(r->man, ROFF_br);
1.296 schwarze 3051: roff_elem_alloc(r->man, ln, ppos, ROFF_br);
3052: if (buf->buf[pos] != '\0')
3053: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
3054: "%s %s", roff_name[tok], buf->buf + pos);
3055: r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3056: r->man->next = ROFF_NEXT_SIBLING;
3057: return ROFF_IGN;
1.92 schwarze 3058: }
3059:
1.105 kristaps 3060: static enum rofferr
1.174 kristaps 3061: roff_cc(ROFF_ARGS)
3062: {
3063: const char *p;
3064:
1.238 schwarze 3065: p = buf->buf + pos;
1.174 kristaps 3066:
1.238 schwarze 3067: if (*p == '\0' || (r->control = *p++) == '.')
1.303 schwarze 3068: r->control = '\0';
1.174 kristaps 3069:
1.238 schwarze 3070: if (*p != '\0')
1.260 schwarze 3071: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3072: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 3073:
1.277 schwarze 3074: return ROFF_IGN;
1.174 kristaps 3075: }
3076:
3077: static enum rofferr
1.303 schwarze 3078: roff_ec(ROFF_ARGS)
3079: {
3080: const char *p;
3081:
3082: p = buf->buf + pos;
3083: if (*p == '\0')
3084: r->escape = '\\';
3085: else {
3086: r->escape = *p;
3087: if (*++p != '\0')
3088: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3089: ln, p - buf->buf, "ec ... %s", p);
3090: }
3091: return ROFF_IGN;
3092: }
3093:
3094: static enum rofferr
3095: roff_eo(ROFF_ARGS)
3096: {
3097: r->escape = '\0';
3098: if (buf->buf[pos] != '\0')
3099: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
3100: ln, pos, "eo %s", buf->buf + pos);
3101: return ROFF_IGN;
3102: }
3103:
3104: static enum rofferr
1.164 kristaps 3105: roff_tr(ROFF_ARGS)
3106: {
3107: const char *p, *first, *second;
3108: size_t fsz, ssz;
3109: enum mandoc_esc esc;
3110:
1.238 schwarze 3111: p = buf->buf + pos;
1.164 kristaps 3112:
1.238 schwarze 3113: if (*p == '\0') {
1.260 schwarze 3114: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.277 schwarze 3115: return ROFF_IGN;
1.164 kristaps 3116: }
3117:
1.238 schwarze 3118: while (*p != '\0') {
1.164 kristaps 3119: fsz = ssz = 1;
3120:
3121: first = p++;
1.238 schwarze 3122: if (*first == '\\') {
1.164 kristaps 3123: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3124: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3125: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3126: ln, (int)(p - buf->buf), first);
1.277 schwarze 3127: return ROFF_IGN;
1.164 kristaps 3128: }
3129: fsz = (size_t)(p - first);
3130: }
3131:
3132: second = p++;
1.238 schwarze 3133: if (*second == '\\') {
1.164 kristaps 3134: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3135: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3136: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3137: ln, (int)(p - buf->buf), second);
1.277 schwarze 3138: return ROFF_IGN;
1.164 kristaps 3139: }
3140: ssz = (size_t)(p - second);
1.238 schwarze 3141: } else if (*second == '\0') {
1.260 schwarze 3142: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
3143: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 3144: second = " ";
1.165 kristaps 3145: p--;
1.164 kristaps 3146: }
3147:
1.167 kristaps 3148: if (fsz > 1) {
1.207 schwarze 3149: roff_setstrn(&r->xmbtab, first, fsz,
3150: second, ssz, 0);
1.167 kristaps 3151: continue;
3152: }
3153:
1.238 schwarze 3154: if (r->xtab == NULL)
1.207 schwarze 3155: r->xtab = mandoc_calloc(128,
3156: sizeof(struct roffstr));
1.167 kristaps 3157:
3158: free(r->xtab[(int)*first].p);
3159: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
3160: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 3161: }
3162:
1.277 schwarze 3163: return ROFF_IGN;
1.164 kristaps 3164: }
3165:
3166: static enum rofferr
1.306 schwarze 3167: roff_rn(ROFF_ARGS)
3168: {
3169: const char *value;
3170: char *oldn, *newn, *end;
3171: size_t oldsz, newsz;
1.315 schwarze 3172: int deftype;
1.306 schwarze 3173:
3174: oldn = newn = buf->buf + pos;
3175: if (*oldn == '\0')
3176: return ROFF_IGN;
3177:
3178: oldsz = roff_getname(r, &newn, ln, pos);
3179: if (oldn[oldsz] == '\\' || *newn == '\0')
3180: return ROFF_IGN;
3181:
3182: end = newn;
3183: newsz = roff_getname(r, &end, ln, newn - buf->buf);
3184: if (newsz == 0)
3185: return ROFF_IGN;
3186:
1.315 schwarze 3187: deftype = ROFFDEF_ANY;
3188: value = roff_getstrn(r, oldn, oldsz, &deftype);
3189: switch (deftype) {
3190: case ROFFDEF_USER:
1.306 schwarze 3191: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3192: roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
3193: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.315 schwarze 3194: break;
3195: case ROFFDEF_PRE:
3196: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3197: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3198: break;
3199: case ROFFDEF_REN:
1.306 schwarze 3200: roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
3201: roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.315 schwarze 3202: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3203: break;
3204: case ROFFDEF_STD:
1.306 schwarze 3205: roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.315 schwarze 3206: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3207: break;
3208: default:
3209: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3210: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3211: break;
3212: }
1.306 schwarze 3213: return ROFF_IGN;
3214: }
3215:
3216: static enum rofferr
1.105 kristaps 3217: roff_so(ROFF_ARGS)
3218: {
1.249 schwarze 3219: char *name, *cp;
1.105 kristaps 3220:
1.238 schwarze 3221: name = buf->buf + pos;
1.224 schwarze 3222: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 3223:
3224: /*
3225: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3226: * opening anything that's not in our cwd or anything beneath
3227: * it. Thus, explicitly disallow traversing up the file-system
3228: * or using absolute paths.
3229: */
3230:
1.238 schwarze 3231: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 3232: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3233: ".so %s", name);
1.249 schwarze 3234: buf->sz = mandoc_asprintf(&cp,
3235: ".sp\nSee the file %s.\n.sp", name) + 1;
3236: free(buf->buf);
3237: buf->buf = cp;
3238: *offs = 0;
1.277 schwarze 3239: return ROFF_REPARSE;
1.105 kristaps 3240: }
3241:
3242: *offs = pos;
1.277 schwarze 3243: return ROFF_SO;
1.105 kristaps 3244: }
1.92 schwarze 3245:
1.266 schwarze 3246: /* --- user defined strings and macros ------------------------------------ */
3247:
1.106 kristaps 3248: static enum rofferr
3249: roff_userdef(ROFF_ARGS)
1.99 kristaps 3250: {
1.311 schwarze 3251: const char *arg[16], *ap;
1.106 kristaps 3252: char *cp, *n1, *n2;
1.292 schwarze 3253: int expand_count, i, ib, ie;
1.263 schwarze 3254: size_t asz, rsz;
1.106 kristaps 3255:
3256: /*
3257: * Collect pointers to macro argument strings
1.188 schwarze 3258: * and NUL-terminate them.
1.106 kristaps 3259: */
1.263 schwarze 3260:
1.273 schwarze 3261: r->argc = 0;
1.238 schwarze 3262: cp = buf->buf + pos;
1.311 schwarze 3263: for (i = 0; i < 16; i++) {
1.273 schwarze 3264: if (*cp == '\0')
3265: arg[i] = "";
3266: else {
3267: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
3268: r->argc = i + 1;
3269: }
3270: }
1.99 kristaps 3271:
1.106 kristaps 3272: /*
3273: * Expand macro arguments.
1.99 kristaps 3274: */
1.263 schwarze 3275:
3276: buf->sz = strlen(r->current_string) + 1;
1.292 schwarze 3277: n1 = n2 = cp = mandoc_malloc(buf->sz);
1.263 schwarze 3278: memcpy(n1, r->current_string, buf->sz);
1.292 schwarze 3279: expand_count = 0;
1.263 schwarze 3280: while (*cp != '\0') {
3281:
3282: /* Scan ahead for the next argument invocation. */
3283:
3284: if (*cp++ != '\\')
3285: continue;
3286: if (*cp++ != '$')
3287: continue;
1.274 schwarze 3288: if (*cp == '*') { /* \\$* inserts all arguments */
3289: ib = 0;
3290: ie = r->argc - 1;
3291: } else { /* \\$1 .. \\$9 insert one argument */
3292: ib = ie = *cp - '1';
3293: if (ib < 0 || ib > 8)
3294: continue;
3295: }
1.263 schwarze 3296: cp -= 2;
1.292 schwarze 3297:
3298: /*
3299: * Prevent infinite recursion.
3300: */
3301:
3302: if (cp >= n2)
3303: expand_count = 1;
3304: else if (++expand_count > EXPAND_LIMIT) {
3305: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
3306: ln, (int)(cp - n1), NULL);
1.293 schwarze 3307: free(buf->buf);
3308: buf->buf = n1;
1.292 schwarze 3309: return ROFF_IGN;
3310: }
1.263 schwarze 3311:
3312: /*
3313: * Determine the size of the expanded argument,
3314: * taking escaping of quotes into account.
3315: */
3316:
1.274 schwarze 3317: asz = ie > ib ? ie - ib : 0; /* for blanks */
3318: for (i = ib; i <= ie; i++) {
3319: for (ap = arg[i]; *ap != '\0'; ap++) {
3320: asz++;
3321: if (*ap == '"')
3322: asz += 3;
3323: }
1.263 schwarze 3324: }
3325: if (asz != 3) {
3326:
3327: /*
3328: * Determine the size of the rest of the
3329: * unexpanded macro, including the NUL.
3330: */
3331:
3332: rsz = buf->sz - (cp - n1) - 3;
3333:
3334: /*
3335: * When shrinking, move before
3336: * releasing the storage.
3337: */
3338:
3339: if (asz < 3)
3340: memmove(cp + asz, cp + 3, rsz);
3341:
3342: /*
3343: * Resize the storage for the macro
3344: * and readjust the parse pointer.
3345: */
3346:
3347: buf->sz += asz - 3;
3348: n2 = mandoc_realloc(n1, buf->sz);
3349: cp = n2 + (cp - n1);
3350: n1 = n2;
3351:
3352: /*
3353: * When growing, make room
3354: * for the expanded argument.
3355: */
3356:
3357: if (asz > 3)
3358: memmove(cp + asz, cp + 3, rsz);
3359: }
3360:
3361: /* Copy the expanded argument, escaping quotes. */
3362:
3363: n2 = cp;
1.274 schwarze 3364: for (i = ib; i <= ie; i++) {
3365: for (ap = arg[i]; *ap != '\0'; ap++) {
3366: if (*ap == '"') {
3367: memcpy(n2, "\\(dq", 4);
3368: n2 += 4;
3369: } else
3370: *n2++ = *ap;
3371: }
3372: if (i < ie)
3373: *n2++ = ' ';
1.106 kristaps 3374: }
1.99 kristaps 3375: }
3376:
1.106 kristaps 3377: /*
3378: * Replace the macro invocation
3379: * by the expanded macro.
3380: */
1.263 schwarze 3381:
1.238 schwarze 3382: free(buf->buf);
3383: buf->buf = n1;
1.248 schwarze 3384: *offs = 0;
1.106 kristaps 3385:
1.277 schwarze 3386: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
3387: ROFF_REPARSE : ROFF_APPEND;
1.99 kristaps 3388: }
1.121 schwarze 3389:
1.306 schwarze 3390: /*
3391: * Calling a high-level macro that was renamed with .rn.
3392: * r->current_string has already been set up by roff_parse().
3393: */
3394: static enum rofferr
3395: roff_renamed(ROFF_ARGS)
3396: {
3397: char *nbuf;
3398:
1.315 schwarze 3399: buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
3400: buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.306 schwarze 3401: free(buf->buf);
3402: buf->buf = nbuf;
3403: return ROFF_CONT;
3404: }
3405:
1.212 schwarze 3406: static size_t
1.121 schwarze 3407: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3408: {
3409: char *name, *cp;
1.212 schwarze 3410: size_t namesz;
1.121 schwarze 3411:
3412: name = *cpp;
3413: if ('\0' == *name)
1.277 schwarze 3414: return 0;
1.121 schwarze 3415:
1.212 schwarze 3416: /* Read until end of name and terminate it with NUL. */
3417: for (cp = name; 1; cp++) {
3418: if ('\0' == *cp || ' ' == *cp) {
3419: namesz = cp - name;
3420: break;
3421: }
1.121 schwarze 3422: if ('\\' != *cp)
3423: continue;
1.215 schwarze 3424: namesz = cp - name;
3425: if ('{' == cp[1] || '}' == cp[1])
3426: break;
1.121 schwarze 3427: cp++;
3428: if ('\\' == *cp)
3429: continue;
1.224 schwarze 3430: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3431: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 3432: mandoc_escape((const char **)&cp, NULL, NULL);
3433: break;
1.121 schwarze 3434: }
3435:
3436: /* Read past spaces. */
3437: while (' ' == *cp)
3438: cp++;
3439:
3440: *cpp = cp;
1.277 schwarze 3441: return namesz;
1.121 schwarze 3442: }
3443:
1.106 kristaps 3444: /*
3445: * Store *string into the user-defined string called *name.
3446: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 3447: * append == 0: replace mode
3448: * append == 1: single-line append mode
3449: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 3450: */
1.94 kristaps 3451: static void
1.106 kristaps 3452: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 3453: int append)
1.92 schwarze 3454: {
1.315 schwarze 3455: size_t namesz;
1.164 kristaps 3456:
1.315 schwarze 3457: namesz = strlen(name);
3458: roff_setstrn(&r->strtab, name, namesz, string,
1.207 schwarze 3459: string ? strlen(string) : 0, append);
1.315 schwarze 3460: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.164 kristaps 3461: }
3462:
3463: static void
1.166 kristaps 3464: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 3465: const char *string, size_t stringsz, int append)
1.164 kristaps 3466: {
1.166 kristaps 3467: struct roffkv *n;
1.164 kristaps 3468: char *c;
3469: int i;
3470: size_t oldch, newch;
1.92 schwarze 3471:
1.106 kristaps 3472: /* Search for an existing string with the same name. */
1.164 kristaps 3473: n = *r;
3474:
1.211 schwarze 3475: while (n && (namesz != n->key.sz ||
3476: strncmp(n->key.p, name, namesz)))
1.92 schwarze 3477: n = n->next;
1.94 kristaps 3478:
3479: if (NULL == n) {
1.106 kristaps 3480: /* Create a new string table entry. */
1.166 kristaps 3481: n = mandoc_malloc(sizeof(struct roffkv));
3482: n->key.p = mandoc_strndup(name, namesz);
3483: n->key.sz = namesz;
3484: n->val.p = NULL;
3485: n->val.sz = 0;
1.164 kristaps 3486: n->next = *r;
3487: *r = n;
1.193 schwarze 3488: } else if (0 == append) {
1.166 kristaps 3489: free(n->val.p);
3490: n->val.p = NULL;
3491: n->val.sz = 0;
1.106 kristaps 3492: }
3493:
3494: if (NULL == string)
3495: return;
3496:
3497: /*
3498: * One additional byte for the '\n' in multiline mode,
3499: * and one for the terminating '\0'.
3500: */
1.193 schwarze 3501: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 3502:
1.166 kristaps 3503: if (NULL == n->val.p) {
3504: n->val.p = mandoc_malloc(newch);
3505: *n->val.p = '\0';
1.106 kristaps 3506: oldch = 0;
3507: } else {
1.166 kristaps 3508: oldch = n->val.sz;
3509: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 3510: }
3511:
3512: /* Skip existing content in the destination buffer. */
1.166 kristaps 3513: c = n->val.p + (int)oldch;
1.106 kristaps 3514:
3515: /* Append new content to the destination buffer. */
1.164 kristaps 3516: i = 0;
3517: while (i < (int)stringsz) {
1.106 kristaps 3518: /*
3519: * Rudimentary roff copy mode:
3520: * Handle escaped backslashes.
3521: */
1.164 kristaps 3522: if ('\\' == string[i] && '\\' == string[i + 1])
3523: i++;
3524: *c++ = string[i++];
1.106 kristaps 3525: }
1.94 kristaps 3526:
1.106 kristaps 3527: /* Append terminating bytes. */
1.193 schwarze 3528: if (1 < append)
1.106 kristaps 3529: *c++ = '\n';
1.163 kristaps 3530:
1.106 kristaps 3531: *c = '\0';
1.166 kristaps 3532: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 3533: }
3534:
1.94 kristaps 3535: static const char *
1.315 schwarze 3536: roff_getstrn(const struct roff *r, const char *name, size_t len,
3537: int *deftype)
1.92 schwarze 3538: {
1.315 schwarze 3539: const struct roffkv *n;
3540: int i;
3541: enum roff_tok tok;
3542:
3543: if (*deftype & ROFFDEF_USER) {
3544: for (n = r->strtab; n != NULL; n = n->next) {
3545: if (strncmp(name, n->key.p, len) == 0 &&
3546: n->key.p[len] == '\0' &&
3547: n->val.p != NULL) {
3548: *deftype = ROFFDEF_USER;
3549: return n->val.p;
3550: }
3551: }
3552: }
3553: if (*deftype & ROFFDEF_PRE) {
3554: for (i = 0; i < PREDEFS_MAX; i++) {
3555: if (strncmp(name, predefs[i].name, len) == 0 &&
3556: predefs[i].name[len] == '\0') {
3557: *deftype = ROFFDEF_PRE;
3558: return predefs[i].str;
3559: }
3560: }
3561: }
3562: if (*deftype & ROFFDEF_REN) {
3563: for (n = r->rentab; n != NULL; n = n->next) {
3564: if (strncmp(name, n->key.p, len) == 0 &&
3565: n->key.p[len] == '\0' &&
3566: n->val.p != NULL) {
3567: *deftype = ROFFDEF_REN;
3568: return n->val.p;
3569: }
3570: }
3571: }
3572: if (*deftype & ROFFDEF_STD) {
3573: if (r->man->macroset != MACROSET_MAN) {
3574: for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
3575: if (strncmp(name, roff_name[tok], len) == 0 &&
3576: roff_name[tok][len] == '\0') {
3577: *deftype = ROFFDEF_STD;
3578: return NULL;
3579: }
3580: }
3581: }
3582: if (r->man->macroset != MACROSET_MDOC) {
3583: for (tok = MAN_TH; tok < MAN_MAX; tok++) {
3584: if (strncmp(name, roff_name[tok], len) == 0 &&
3585: roff_name[tok][len] == '\0') {
3586: *deftype = ROFFDEF_STD;
3587: return NULL;
3588: }
3589: }
3590: }
3591: }
3592: *deftype = 0;
1.277 schwarze 3593: return NULL;
1.92 schwarze 3594: }
3595:
1.94 kristaps 3596: static void
1.167 kristaps 3597: roff_freestr(struct roffkv *r)
1.92 schwarze 3598: {
1.166 kristaps 3599: struct roffkv *n, *nn;
1.92 schwarze 3600:
1.167 kristaps 3601: for (n = r; n; n = nn) {
1.166 kristaps 3602: free(n->key.p);
3603: free(n->val.p);
1.92 schwarze 3604: nn = n->next;
3605: free(n);
3606: }
1.114 kristaps 3607: }
1.266 schwarze 3608:
3609: /* --- accessors and utility functions ------------------------------------ */
1.164 kristaps 3610:
3611: /*
3612: * Duplicate an input string, making the appropriate character
3613: * conversations (as stipulated by `tr') along the way.
3614: * Returns a heap-allocated string with all the replacements made.
3615: */
3616: char *
3617: roff_strdup(const struct roff *r, const char *p)
3618: {
1.166 kristaps 3619: const struct roffkv *cp;
1.164 kristaps 3620: char *res;
3621: const char *pp;
3622: size_t ssz, sz;
3623: enum mandoc_esc esc;
3624:
1.167 kristaps 3625: if (NULL == r->xmbtab && NULL == r->xtab)
1.277 schwarze 3626: return mandoc_strdup(p);
1.164 kristaps 3627: else if ('\0' == *p)
1.277 schwarze 3628: return mandoc_strdup("");
1.164 kristaps 3629:
3630: /*
3631: * Step through each character looking for term matches
3632: * (remember that a `tr' can be invoked with an escape, which is
3633: * a glyph but the escape is multi-character).
3634: * We only do this if the character hash has been initialised
3635: * and the string is >0 length.
3636: */
3637:
3638: res = NULL;
3639: ssz = 0;
3640:
3641: while ('\0' != *p) {
1.289 schwarze 3642: assert((unsigned int)*p < 128);
3643: if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167 kristaps 3644: sz = r->xtab[(int)*p].sz;
3645: res = mandoc_realloc(res, ssz + sz + 1);
3646: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3647: ssz += sz;
3648: p++;
3649: continue;
3650: } else if ('\\' != *p) {
3651: res = mandoc_realloc(res, ssz + 2);
3652: res[ssz++] = *p++;
3653: continue;
3654: }
3655:
1.164 kristaps 3656: /* Search for term matches. */
1.167 kristaps 3657: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 3658: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 3659: break;
3660:
3661: if (NULL != cp) {
3662: /*
3663: * A match has been found.
3664: * Append the match to the array and move
3665: * forward by its keysize.
3666: */
1.207 schwarze 3667: res = mandoc_realloc(res,
3668: ssz + cp->val.sz + 1);
1.166 kristaps 3669: memcpy(res + ssz, cp->val.p, cp->val.sz);
3670: ssz += cp->val.sz;
3671: p += (int)cp->key.sz;
1.164 kristaps 3672: continue;
3673: }
3674:
1.167 kristaps 3675: /*
3676: * Handle escapes carefully: we need to copy
3677: * over just the escape itself, or else we might
3678: * do replacements within the escape itself.
3679: * Make sure to pass along the bogus string.
3680: */
3681: pp = p++;
3682: esc = mandoc_escape(&p, NULL, NULL);
3683: if (ESCAPE_ERROR == esc) {
3684: sz = strlen(pp);
1.164 kristaps 3685: res = mandoc_realloc(res, ssz + sz + 1);
3686: memcpy(res + ssz, pp, sz);
1.167 kristaps 3687: break;
1.164 kristaps 3688: }
1.207 schwarze 3689: /*
3690: * We bail out on bad escapes.
1.167 kristaps 3691: * No need to warn: we already did so when
3692: * roff_res() was called.
3693: */
3694: sz = (int)(p - pp);
3695: res = mandoc_realloc(res, ssz + sz + 1);
3696: memcpy(res + ssz, pp, sz);
3697: ssz += sz;
1.164 kristaps 3698: }
3699:
3700: res[(int)ssz] = '\0';
1.277 schwarze 3701: return res;
1.227 schwarze 3702: }
3703:
3704: int
3705: roff_getformat(const struct roff *r)
3706: {
3707:
1.277 schwarze 3708: return r->format;
1.174 kristaps 3709: }
3710:
3711: /*
1.207 schwarze 3712: * Find out whether a line is a macro line or not.
1.174 kristaps 3713: * If it is, adjust the current position and return one; if it isn't,
3714: * return zero and don't change the current position.
3715: * If the control character has been set with `.cc', then let that grain
3716: * precedence.
3717: * This is slighly contrary to groff, where using the non-breaking
3718: * control character when `cc' has been invoked will cause the
3719: * non-breaking macro contents to be printed verbatim.
3720: */
3721: int
3722: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3723: {
3724: int pos;
3725:
3726: pos = *ppos;
3727:
1.303 schwarze 3728: if (r->control != '\0' && cp[pos] == r->control)
1.174 kristaps 3729: pos++;
1.303 schwarze 3730: else if (r->control != '\0')
1.277 schwarze 3731: return 0;
1.174 kristaps 3732: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3733: pos += 2;
3734: else if ('.' == cp[pos] || '\'' == cp[pos])
3735: pos++;
3736: else
1.277 schwarze 3737: return 0;
1.174 kristaps 3738:
3739: while (' ' == cp[pos] || '\t' == cp[pos])
3740: pos++;
3741:
3742: *ppos = pos;
1.277 schwarze 3743: return 1;
1.74 kristaps 3744: }
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