Annotation of mandoc/roff.c, Revision 1.326
1.326 ! schwarze 1: /* $Id: roff.c,v 1.325 2018/04/09 02:31:42 schwarze Exp $ */
1.1 kristaps 2: /*
1.267 schwarze 3: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.325 schwarze 4: * Copyright (c) 2010-2015, 2017, 2018 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.66 kristaps 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.106 kristaps 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.66 kristaps 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106 kristaps 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.66 kristaps 13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1 kristaps 17: */
1.66 kristaps 18: #include "config.h"
1.225 schwarze 19:
20: #include <sys/types.h>
1.30 kristaps 21:
1.67 kristaps 22: #include <assert.h>
1.85 kristaps 23: #include <ctype.h>
1.245 schwarze 24: #include <limits.h>
1.295 schwarze 25: #include <stddef.h>
26: #include <stdint.h>
1.178 schwarze 27: #include <stdio.h>
1.1 kristaps 28: #include <stdlib.h>
1.67 kristaps 29: #include <string.h>
1.1 kristaps 30:
1.67 kristaps 31: #include "mandoc.h"
1.201 schwarze 32: #include "mandoc_aux.h"
1.295 schwarze 33: #include "mandoc_ohash.h"
1.265 schwarze 34: #include "roff.h"
1.226 schwarze 35: #include "libmandoc.h"
1.266 schwarze 36: #include "roff_int.h"
1.109 kristaps 37: #include "libroff.h"
1.82 kristaps 38:
1.170 schwarze 39: /* Maximum number of string expansions per line, to break infinite loops. */
40: #define EXPAND_LIMIT 1000
41:
1.315 schwarze 42: /* Types of definitions of macros and strings. */
43: #define ROFFDEF_USER (1 << 1) /* User-defined. */
44: #define ROFFDEF_PRE (1 << 2) /* Predefined. */
45: #define ROFFDEF_REN (1 << 3) /* Renamed standard macro. */
46: #define ROFFDEF_STD (1 << 4) /* mdoc(7) or man(7) macro. */
47: #define ROFFDEF_ANY (ROFFDEF_USER | ROFFDEF_PRE | \
48: ROFFDEF_REN | ROFFDEF_STD)
1.325 schwarze 49: #define ROFFDEF_UNDEF (1 << 5) /* Completely undefined. */
1.315 schwarze 50:
1.266 schwarze 51: /* --- data types --------------------------------------------------------- */
52:
1.147 kristaps 53: /*
1.167 kristaps 54: * An incredibly-simple string buffer.
55: */
1.94 kristaps 56: struct roffstr {
1.167 kristaps 57: char *p; /* nil-terminated buffer */
58: size_t sz; /* saved strlen(p) */
1.166 kristaps 59: };
60:
61: /*
1.167 kristaps 62: * A key-value roffstr pair as part of a singly-linked list.
1.166 kristaps 63: */
64: struct roffkv {
65: struct roffstr key;
66: struct roffstr val;
67: struct roffkv *next; /* next in list */
1.94 kristaps 68: };
69:
1.180 schwarze 70: /*
71: * A single number register as part of a singly-linked list.
72: */
73: struct roffreg {
74: struct roffstr key;
1.181 schwarze 75: int val;
1.180 schwarze 76: struct roffreg *next;
77: };
78:
1.295 schwarze 79: /*
80: * Association of request and macro names with token IDs.
81: */
82: struct roffreq {
83: enum roff_tok tok;
84: char name[];
85: };
86:
1.67 kristaps 87: struct roff {
1.128 kristaps 88: struct mparse *parse; /* parse point */
1.296 schwarze 89: struct roff_man *man; /* mdoc or man parser */
1.67 kristaps 90: struct roffnode *last; /* leaf of stack */
1.223 schwarze 91: int *rstack; /* stack of inverted `ie' values */
1.295 schwarze 92: struct ohash *reqtab; /* request lookup table */
1.180 schwarze 93: struct roffreg *regtab; /* number registers */
1.166 kristaps 94: struct roffkv *strtab; /* user-defined strings & macros */
1.306 schwarze 95: struct roffkv *rentab; /* renamed strings & macros */
1.167 kristaps 96: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
97: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.106 kristaps 98: const char *current_string; /* value of last called user macro */
1.118 kristaps 99: struct tbl_node *first_tbl; /* first table parsed */
100: struct tbl_node *last_tbl; /* last table parsed */
101: struct tbl_node *tbl; /* current table being parsed */
1.319 schwarze 102: struct eqn_node *last_eqn; /* equation parser */
103: struct eqn_node *eqn; /* active equation parser */
1.230 schwarze 104: int eqn_inline; /* current equation is inline */
1.223 schwarze 105: int options; /* parse options */
106: int rstacksz; /* current size limit of rstack */
107: int rstackpos; /* position in rstack */
1.227 schwarze 108: int format; /* current file in mdoc or man format */
1.273 schwarze 109: int argc; /* number of args of the last macro */
1.223 schwarze 110: char control; /* control character */
1.303 schwarze 111: char escape; /* escape character */
1.79 kristaps 112: };
113:
1.67 kristaps 114: struct roffnode {
1.294 schwarze 115: enum roff_tok tok; /* type of node */
1.67 kristaps 116: struct roffnode *parent; /* up one in stack */
117: int line; /* parse line */
118: int col; /* parse col */
1.106 kristaps 119: char *name; /* node name, e.g. macro name */
1.79 kristaps 120: char *end; /* end-rules: custom token */
121: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 122: int rule; /* current evaluation rule */
1.67 kristaps 123: };
124:
125: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.294 schwarze 126: enum roff_tok tok, /* tok of macro */ \
1.238 schwarze 127: struct buf *buf, /* input buffer */ \
1.67 kristaps 128: int ln, /* parse line */ \
1.75 kristaps 129: int ppos, /* original pos in buffer */ \
130: int pos, /* current pos in buffer */ \
1.74 kristaps 131: int *offs /* reset offset of buffer data */
1.67 kristaps 132:
133: typedef enum rofferr (*roffproc)(ROFF_ARGS);
134:
135: struct roffmac {
1.79 kristaps 136: roffproc proc; /* process new macro */
137: roffproc text; /* process as child text of macro */
138: roffproc sub; /* process as child of macro */
139: int flags;
140: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.67 kristaps 141: };
142:
1.141 kristaps 143: struct predef {
144: const char *name; /* predefined input name */
145: const char *str; /* replacement symbol */
146: };
147:
148: #define PREDEF(__name, __str) \
149: { (__name), (__str) },
150:
1.266 schwarze 151: /* --- function prototypes ------------------------------------------------ */
152:
1.155 kristaps 153: static void roffnode_cleanscope(struct roff *);
154: static void roffnode_pop(struct roff *);
1.294 schwarze 155: static void roffnode_push(struct roff *, enum roff_tok,
1.155 kristaps 156: const char *, int, int);
1.321 schwarze 157: static void roff_addtbl(struct roff_man *, struct tbl_node *);
1.311 schwarze 158: static enum rofferr roff_als(ROFF_ARGS);
1.80 kristaps 159: static enum rofferr roff_block(ROFF_ARGS);
160: static enum rofferr roff_block_text(ROFF_ARGS);
161: static enum rofferr roff_block_sub(ROFF_ARGS);
1.296 schwarze 162: static enum rofferr roff_br(ROFF_ARGS);
1.80 kristaps 163: static enum rofferr roff_cblock(ROFF_ARGS);
1.174 kristaps 164: static enum rofferr roff_cc(ROFF_ARGS);
1.195 schwarze 165: static void roff_ccond(struct roff *, int, int);
1.82 kristaps 166: static enum rofferr roff_cond(ROFF_ARGS);
167: static enum rofferr roff_cond_text(ROFF_ARGS);
168: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.92 schwarze 169: static enum rofferr roff_ds(ROFF_ARGS);
1.303 schwarze 170: static enum rofferr roff_ec(ROFF_ARGS);
171: static enum rofferr roff_eo(ROFF_ARGS);
1.238 schwarze 172: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.271 schwarze 173: static int roff_evalcond(struct roff *r, int, char *, int *);
1.234 kristaps 174: static int roff_evalnum(struct roff *, int,
175: const char *, int *, int *, int);
176: static int roff_evalpar(struct roff *, int,
1.261 schwarze 177: const char *, int *, int *, int);
1.198 schwarze 178: static int roff_evalstrcond(const char *, int *);
1.155 kristaps 179: static void roff_free1(struct roff *);
1.180 schwarze 180: static void roff_freereg(struct roffreg *);
1.167 kristaps 181: static void roff_freestr(struct roffkv *);
1.212 schwarze 182: static size_t roff_getname(struct roff *, char **, int, int);
1.261 schwarze 183: static int roff_getnum(const char *, int *, int *, int);
1.184 schwarze 184: static int roff_getop(const char *, int *, char *);
1.326 ! schwarze 185: static int roff_getregn(struct roff *, const char *, size_t);
1.273 schwarze 186: static int roff_getregro(const struct roff *,
187: const char *name);
1.325 schwarze 188: static const char *roff_getstrn(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);
1.326 ! schwarze 209: static void roff_setregn(struct roff *, const char *,
! 210: size_t, int, char);
1.94 kristaps 211: static void roff_setstr(struct roff *,
1.106 kristaps 212: const char *, const char *, int);
1.207 schwarze 213: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 214: size_t, const char *, size_t, int);
1.105 kristaps 215: static enum rofferr roff_so(ROFF_ARGS);
1.164 kristaps 216: static enum rofferr roff_tr(ROFF_ARGS);
1.175 schwarze 217: static enum rofferr roff_Dd(ROFF_ARGS);
1.109 kristaps 218: static enum rofferr roff_TE(ROFF_ARGS);
219: static enum rofferr roff_TS(ROFF_ARGS);
1.125 kristaps 220: static enum rofferr roff_EQ(ROFF_ARGS);
221: static enum rofferr roff_EN(ROFF_ARGS);
1.112 kristaps 222: static enum rofferr roff_T_(ROFF_ARGS);
1.251 schwarze 223: static enum rofferr roff_unsupp(ROFF_ARGS);
1.106 kristaps 224: static enum rofferr roff_userdef(ROFF_ARGS);
1.67 kristaps 225:
1.266 schwarze 226: /* --- constant data ------------------------------------------------------ */
227:
1.261 schwarze 228: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
229: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
230:
1.294 schwarze 231: const char *__roff_name[MAN_MAX + 1] = {
1.305 schwarze 232: "br", "ce", "ft", "ll",
1.312 schwarze 233: "mc", "po", "rj", "sp",
234: "ta", "ti", NULL,
1.294 schwarze 235: "ab", "ad", "af", "aln",
236: "als", "am", "am1", "ami",
237: "ami1", "as", "as1", "asciify",
238: "backtrace", "bd", "bleedat", "blm",
239: "box", "boxa", "bp", "BP",
240: "break", "breakchar", "brnl", "brp",
1.305 schwarze 241: "brpnl", "c2", "cc",
1.294 schwarze 242: "cf", "cflags", "ch", "char",
243: "chop", "class", "close", "CL",
244: "color", "composite", "continue", "cp",
245: "cropat", "cs", "cu", "da",
246: "dch", "Dd", "de", "de1",
247: "defcolor", "dei", "dei1", "device",
248: "devicem", "di", "do", "ds",
249: "ds1", "dwh", "dt", "ec",
250: "ecr", "ecs", "el", "em",
251: "EN", "eo", "EP", "EQ",
252: "errprint", "ev", "evc", "ex",
253: "fallback", "fam", "fc", "fchar",
254: "fcolor", "fdeferlig", "feature", "fkern",
255: "fl", "flig", "fp", "fps",
256: "fschar", "fspacewidth", "fspecial", "ftr",
257: "fzoom", "gcolor", "hc", "hcode",
258: "hidechar", "hla", "hlm", "hpf",
259: "hpfa", "hpfcode", "hw", "hy",
260: "hylang", "hylen", "hym", "hypp",
261: "hys", "ie", "if", "ig",
262: "index", "it", "itc", "IX",
263: "kern", "kernafter", "kernbefore", "kernpair",
264: "lc", "lc_ctype", "lds", "length",
265: "letadj", "lf", "lg", "lhang",
266: "linetabs", "lnr", "lnrf", "lpfx",
1.304 schwarze 267: "ls", "lsm", "lt",
1.294 schwarze 268: "mediasize", "minss", "mk", "mso",
269: "na", "ne", "nh", "nhychar",
270: "nm", "nn", "nop", "nr",
271: "nrf", "nroff", "ns", "nx",
272: "open", "opena", "os", "output",
273: "padj", "papersize", "pc", "pev",
274: "pi", "PI", "pl", "pm",
1.312 schwarze 275: "pn", "pnr", "ps",
1.294 schwarze 276: "psbb", "pshape", "pso", "ptr",
277: "pvs", "rchar", "rd", "recursionlimit",
1.309 schwarze 278: "return", "rfschar", "rhang",
1.294 schwarze 279: "rm", "rn", "rnn", "rr",
280: "rs", "rt", "schar", "sentchar",
281: "shc", "shift", "sizes", "so",
282: "spacewidth", "special", "spreadwarn", "ss",
283: "sty", "substring", "sv", "sy",
1.300 schwarze 284: "T&", "tc", "TE",
1.301 schwarze 285: "TH", "tkf", "tl",
1.294 schwarze 286: "tm", "tm1", "tmc", "tr",
287: "track", "transchar", "trf", "trimat",
288: "trin", "trnt", "troff", "TS",
289: "uf", "ul", "unformat", "unwatch",
290: "unwatchn", "vpt", "vs", "warn",
291: "warnscale", "watch", "watchlength", "watchn",
292: "wh", "while", "write", "writec",
293: "writem", "xflag", ".", NULL,
1.306 schwarze 294: NULL, "text",
1.294 schwarze 295: "Dd", "Dt", "Os", "Sh",
296: "Ss", "Pp", "D1", "Dl",
297: "Bd", "Ed", "Bl", "El",
298: "It", "Ad", "An", "Ap",
299: "Ar", "Cd", "Cm", "Dv",
300: "Er", "Ev", "Ex", "Fa",
301: "Fd", "Fl", "Fn", "Ft",
302: "Ic", "In", "Li", "Nd",
303: "Nm", "Op", "Ot", "Pa",
304: "Rv", "St", "Va", "Vt",
305: "Xr", "%A", "%B", "%D",
306: "%I", "%J", "%N", "%O",
307: "%P", "%R", "%T", "%V",
308: "Ac", "Ao", "Aq", "At",
309: "Bc", "Bf", "Bo", "Bq",
310: "Bsx", "Bx", "Db", "Dc",
311: "Do", "Dq", "Ec", "Ef",
312: "Em", "Eo", "Fx", "Ms",
313: "No", "Ns", "Nx", "Ox",
314: "Pc", "Pf", "Po", "Pq",
315: "Qc", "Ql", "Qo", "Qq",
316: "Re", "Rs", "Sc", "So",
317: "Sq", "Sm", "Sx", "Sy",
318: "Tn", "Ux", "Xc", "Xo",
319: "Fo", "Fc", "Oo", "Oc",
320: "Bk", "Ek", "Bt", "Hf",
321: "Fr", "Ud", "Lb", "Lp",
322: "Lk", "Mt", "Brq", "Bro",
323: "Brc", "%C", "Es", "En",
1.299 schwarze 324: "Dx", "%Q", "%U", "Ta",
325: NULL,
1.294 schwarze 326: "TH", "SH", "SS", "TP",
327: "LP", "PP", "P", "IP",
328: "HP", "SM", "SB", "BI",
329: "IB", "BR", "RB", "R",
330: "B", "I", "IR", "RI",
1.299 schwarze 331: "nf", "fi",
1.294 schwarze 332: "RE", "RS", "DT", "UC",
1.297 schwarze 333: "PD", "AT", "in",
1.294 schwarze 334: "OP", "EX", "EE", "UR",
1.317 schwarze 335: "UE", "MT", "ME", NULL
1.294 schwarze 336: };
337: const char *const *roff_name = __roff_name;
338:
339: static struct roffmac roffs[TOKEN_NONE] = {
1.296 schwarze 340: { roff_br, NULL, NULL, 0 }, /* br */
1.305 schwarze 341: { roff_onearg, NULL, NULL, 0 }, /* ce */
1.297 schwarze 342: { roff_onearg, NULL, NULL, 0 }, /* ft */
1.298 schwarze 343: { roff_onearg, NULL, NULL, 0 }, /* ll */
1.304 schwarze 344: { roff_onearg, NULL, NULL, 0 }, /* mc */
1.312 schwarze 345: { roff_onearg, NULL, NULL, 0 }, /* po */
1.309 schwarze 346: { roff_onearg, NULL, NULL, 0 }, /* rj */
1.299 schwarze 347: { roff_onearg, NULL, NULL, 0 }, /* sp */
1.300 schwarze 348: { roff_manyarg, NULL, NULL, 0 }, /* ta */
1.301 schwarze 349: { roff_onearg, NULL, NULL, 0 }, /* ti */
1.296 schwarze 350: { NULL, NULL, NULL, 0 }, /* ROFF_MAX */
1.295 schwarze 351: { roff_unsupp, NULL, NULL, 0 }, /* ab */
352: { roff_line_ignore, NULL, NULL, 0 }, /* ad */
353: { roff_line_ignore, NULL, NULL, 0 }, /* af */
354: { roff_unsupp, NULL, NULL, 0 }, /* aln */
1.311 schwarze 355: { roff_als, NULL, NULL, 0 }, /* als */
1.295 schwarze 356: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am */
357: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am1 */
358: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami */
359: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami1 */
360: { roff_ds, NULL, NULL, 0 }, /* as */
361: { roff_ds, NULL, NULL, 0 }, /* as1 */
362: { roff_unsupp, NULL, NULL, 0 }, /* asciify */
363: { roff_line_ignore, NULL, NULL, 0 }, /* backtrace */
364: { roff_line_ignore, NULL, NULL, 0 }, /* bd */
365: { roff_line_ignore, NULL, NULL, 0 }, /* bleedat */
366: { roff_unsupp, NULL, NULL, 0 }, /* blm */
367: { roff_unsupp, NULL, NULL, 0 }, /* box */
368: { roff_unsupp, NULL, NULL, 0 }, /* boxa */
369: { roff_line_ignore, NULL, NULL, 0 }, /* bp */
370: { roff_unsupp, NULL, NULL, 0 }, /* BP */
371: { roff_unsupp, NULL, NULL, 0 }, /* break */
372: { roff_line_ignore, NULL, NULL, 0 }, /* breakchar */
373: { roff_line_ignore, NULL, NULL, 0 }, /* brnl */
1.296 schwarze 374: { roff_br, NULL, NULL, 0 }, /* brp */
1.295 schwarze 375: { roff_line_ignore, NULL, NULL, 0 }, /* brpnl */
376: { roff_unsupp, NULL, NULL, 0 }, /* c2 */
377: { roff_cc, NULL, NULL, 0 }, /* cc */
378: { roff_insec, NULL, NULL, 0 }, /* cf */
379: { roff_line_ignore, NULL, NULL, 0 }, /* cflags */
380: { roff_line_ignore, NULL, NULL, 0 }, /* ch */
381: { roff_unsupp, NULL, NULL, 0 }, /* char */
382: { roff_unsupp, NULL, NULL, 0 }, /* chop */
383: { roff_line_ignore, NULL, NULL, 0 }, /* class */
384: { roff_insec, NULL, NULL, 0 }, /* close */
385: { roff_unsupp, NULL, NULL, 0 }, /* CL */
386: { roff_line_ignore, NULL, NULL, 0 }, /* color */
387: { roff_unsupp, NULL, NULL, 0 }, /* composite */
388: { roff_unsupp, NULL, NULL, 0 }, /* continue */
389: { roff_line_ignore, NULL, NULL, 0 }, /* cp */
390: { roff_line_ignore, NULL, NULL, 0 }, /* cropat */
391: { roff_line_ignore, NULL, NULL, 0 }, /* cs */
392: { roff_line_ignore, NULL, NULL, 0 }, /* cu */
393: { roff_unsupp, NULL, NULL, 0 }, /* da */
394: { roff_unsupp, NULL, NULL, 0 }, /* dch */
395: { roff_Dd, NULL, NULL, 0 }, /* Dd */
396: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de */
397: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de1 */
398: { roff_line_ignore, NULL, NULL, 0 }, /* defcolor */
399: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei */
400: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei1 */
401: { roff_unsupp, NULL, NULL, 0 }, /* device */
402: { roff_unsupp, NULL, NULL, 0 }, /* devicem */
403: { roff_unsupp, NULL, NULL, 0 }, /* di */
404: { roff_unsupp, NULL, NULL, 0 }, /* do */
405: { roff_ds, NULL, NULL, 0 }, /* ds */
406: { roff_ds, NULL, NULL, 0 }, /* ds1 */
407: { roff_unsupp, NULL, NULL, 0 }, /* dwh */
408: { roff_unsupp, NULL, NULL, 0 }, /* dt */
1.303 schwarze 409: { roff_ec, NULL, NULL, 0 }, /* ec */
1.295 schwarze 410: { roff_unsupp, NULL, NULL, 0 }, /* ecr */
411: { roff_unsupp, NULL, NULL, 0 }, /* ecs */
412: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* el */
413: { roff_unsupp, NULL, NULL, 0 }, /* em */
414: { roff_EN, NULL, NULL, 0 }, /* EN */
1.303 schwarze 415: { roff_eo, NULL, NULL, 0 }, /* eo */
1.295 schwarze 416: { roff_unsupp, NULL, NULL, 0 }, /* EP */
417: { roff_EQ, NULL, NULL, 0 }, /* EQ */
418: { roff_line_ignore, NULL, NULL, 0 }, /* errprint */
419: { roff_unsupp, NULL, NULL, 0 }, /* ev */
420: { roff_unsupp, NULL, NULL, 0 }, /* evc */
421: { roff_unsupp, NULL, NULL, 0 }, /* ex */
422: { roff_line_ignore, NULL, NULL, 0 }, /* fallback */
423: { roff_line_ignore, NULL, NULL, 0 }, /* fam */
424: { roff_unsupp, NULL, NULL, 0 }, /* fc */
425: { roff_unsupp, NULL, NULL, 0 }, /* fchar */
426: { roff_line_ignore, NULL, NULL, 0 }, /* fcolor */
427: { roff_line_ignore, NULL, NULL, 0 }, /* fdeferlig */
428: { roff_line_ignore, NULL, NULL, 0 }, /* feature */
429: { roff_line_ignore, NULL, NULL, 0 }, /* fkern */
430: { roff_line_ignore, NULL, NULL, 0 }, /* fl */
431: { roff_line_ignore, NULL, NULL, 0 }, /* flig */
432: { roff_line_ignore, NULL, NULL, 0 }, /* fp */
433: { roff_line_ignore, NULL, NULL, 0 }, /* fps */
434: { roff_unsupp, NULL, NULL, 0 }, /* fschar */
435: { roff_line_ignore, NULL, NULL, 0 }, /* fspacewidth */
436: { roff_line_ignore, NULL, NULL, 0 }, /* fspecial */
437: { roff_line_ignore, NULL, NULL, 0 }, /* ftr */
438: { roff_line_ignore, NULL, NULL, 0 }, /* fzoom */
439: { roff_line_ignore, NULL, NULL, 0 }, /* gcolor */
440: { roff_line_ignore, NULL, NULL, 0 }, /* hc */
441: { roff_line_ignore, NULL, NULL, 0 }, /* hcode */
442: { roff_line_ignore, NULL, NULL, 0 }, /* hidechar */
443: { roff_line_ignore, NULL, NULL, 0 }, /* hla */
444: { roff_line_ignore, NULL, NULL, 0 }, /* hlm */
445: { roff_line_ignore, NULL, NULL, 0 }, /* hpf */
446: { roff_line_ignore, NULL, NULL, 0 }, /* hpfa */
447: { roff_line_ignore, NULL, NULL, 0 }, /* hpfcode */
448: { roff_line_ignore, NULL, NULL, 0 }, /* hw */
449: { roff_line_ignore, NULL, NULL, 0 }, /* hy */
450: { roff_line_ignore, NULL, NULL, 0 }, /* hylang */
451: { roff_line_ignore, NULL, NULL, 0 }, /* hylen */
452: { roff_line_ignore, NULL, NULL, 0 }, /* hym */
453: { roff_line_ignore, NULL, NULL, 0 }, /* hypp */
454: { roff_line_ignore, NULL, NULL, 0 }, /* hys */
455: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* ie */
456: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* if */
457: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ig */
458: { roff_unsupp, NULL, NULL, 0 }, /* index */
459: { roff_it, NULL, NULL, 0 }, /* it */
460: { roff_unsupp, NULL, NULL, 0 }, /* itc */
461: { roff_line_ignore, NULL, NULL, 0 }, /* IX */
462: { roff_line_ignore, NULL, NULL, 0 }, /* kern */
463: { roff_line_ignore, NULL, NULL, 0 }, /* kernafter */
464: { roff_line_ignore, NULL, NULL, 0 }, /* kernbefore */
465: { roff_line_ignore, NULL, NULL, 0 }, /* kernpair */
466: { roff_unsupp, NULL, NULL, 0 }, /* lc */
467: { roff_unsupp, NULL, NULL, 0 }, /* lc_ctype */
468: { roff_unsupp, NULL, NULL, 0 }, /* lds */
469: { roff_unsupp, NULL, NULL, 0 }, /* length */
470: { roff_line_ignore, NULL, NULL, 0 }, /* letadj */
471: { roff_insec, NULL, NULL, 0 }, /* lf */
472: { roff_line_ignore, NULL, NULL, 0 }, /* lg */
473: { roff_line_ignore, NULL, NULL, 0 }, /* lhang */
474: { roff_unsupp, NULL, NULL, 0 }, /* linetabs */
475: { roff_unsupp, NULL, NULL, 0 }, /* lnr */
476: { roff_unsupp, NULL, NULL, 0 }, /* lnrf */
477: { roff_unsupp, NULL, NULL, 0 }, /* lpfx */
478: { roff_line_ignore, NULL, NULL, 0 }, /* ls */
479: { roff_unsupp, NULL, NULL, 0 }, /* lsm */
480: { roff_line_ignore, NULL, NULL, 0 }, /* lt */
481: { roff_line_ignore, NULL, NULL, 0 }, /* mediasize */
482: { roff_line_ignore, NULL, NULL, 0 }, /* minss */
483: { roff_line_ignore, NULL, NULL, 0 }, /* mk */
484: { roff_insec, NULL, NULL, 0 }, /* mso */
485: { roff_line_ignore, NULL, NULL, 0 }, /* na */
486: { roff_line_ignore, NULL, NULL, 0 }, /* ne */
487: { roff_line_ignore, NULL, NULL, 0 }, /* nh */
488: { roff_line_ignore, NULL, NULL, 0 }, /* nhychar */
489: { roff_unsupp, NULL, NULL, 0 }, /* nm */
490: { roff_unsupp, NULL, NULL, 0 }, /* nn */
491: { roff_unsupp, NULL, NULL, 0 }, /* nop */
492: { roff_nr, NULL, NULL, 0 }, /* nr */
493: { roff_unsupp, NULL, NULL, 0 }, /* nrf */
494: { roff_line_ignore, NULL, NULL, 0 }, /* nroff */
495: { roff_line_ignore, NULL, NULL, 0 }, /* ns */
496: { roff_insec, NULL, NULL, 0 }, /* nx */
497: { roff_insec, NULL, NULL, 0 }, /* open */
498: { roff_insec, NULL, NULL, 0 }, /* opena */
499: { roff_line_ignore, NULL, NULL, 0 }, /* os */
500: { roff_unsupp, NULL, NULL, 0 }, /* output */
501: { roff_line_ignore, NULL, NULL, 0 }, /* padj */
502: { roff_line_ignore, NULL, NULL, 0 }, /* papersize */
503: { roff_line_ignore, NULL, NULL, 0 }, /* pc */
504: { roff_line_ignore, NULL, NULL, 0 }, /* pev */
505: { roff_insec, NULL, NULL, 0 }, /* pi */
506: { roff_unsupp, NULL, NULL, 0 }, /* PI */
507: { roff_line_ignore, NULL, NULL, 0 }, /* pl */
508: { roff_line_ignore, NULL, NULL, 0 }, /* pm */
509: { roff_line_ignore, NULL, NULL, 0 }, /* pn */
510: { roff_line_ignore, NULL, NULL, 0 }, /* pnr */
511: { roff_line_ignore, NULL, NULL, 0 }, /* ps */
512: { roff_unsupp, NULL, NULL, 0 }, /* psbb */
513: { roff_unsupp, NULL, NULL, 0 }, /* pshape */
514: { roff_insec, NULL, NULL, 0 }, /* pso */
515: { roff_line_ignore, NULL, NULL, 0 }, /* ptr */
516: { roff_line_ignore, NULL, NULL, 0 }, /* pvs */
517: { roff_unsupp, NULL, NULL, 0 }, /* rchar */
518: { roff_line_ignore, NULL, NULL, 0 }, /* rd */
519: { roff_line_ignore, NULL, NULL, 0 }, /* recursionlimit */
520: { roff_unsupp, NULL, NULL, 0 }, /* return */
521: { roff_unsupp, NULL, NULL, 0 }, /* rfschar */
522: { roff_line_ignore, NULL, NULL, 0 }, /* rhang */
523: { roff_rm, NULL, NULL, 0 }, /* rm */
1.306 schwarze 524: { roff_rn, NULL, NULL, 0 }, /* rn */
1.295 schwarze 525: { roff_unsupp, NULL, NULL, 0 }, /* rnn */
526: { roff_rr, NULL, NULL, 0 }, /* rr */
527: { roff_line_ignore, NULL, NULL, 0 }, /* rs */
528: { roff_line_ignore, NULL, NULL, 0 }, /* rt */
529: { roff_unsupp, NULL, NULL, 0 }, /* schar */
530: { roff_line_ignore, NULL, NULL, 0 }, /* sentchar */
531: { roff_line_ignore, NULL, NULL, 0 }, /* shc */
532: { roff_unsupp, NULL, NULL, 0 }, /* shift */
533: { roff_line_ignore, NULL, NULL, 0 }, /* sizes */
534: { roff_so, NULL, NULL, 0 }, /* so */
535: { roff_line_ignore, NULL, NULL, 0 }, /* spacewidth */
536: { roff_line_ignore, NULL, NULL, 0 }, /* special */
537: { roff_line_ignore, NULL, NULL, 0 }, /* spreadwarn */
538: { roff_line_ignore, NULL, NULL, 0 }, /* ss */
539: { roff_line_ignore, NULL, NULL, 0 }, /* sty */
540: { roff_unsupp, NULL, NULL, 0 }, /* substring */
541: { roff_line_ignore, NULL, NULL, 0 }, /* sv */
542: { roff_insec, NULL, NULL, 0 }, /* sy */
543: { roff_T_, NULL, NULL, 0 }, /* T& */
544: { roff_unsupp, NULL, NULL, 0 }, /* tc */
545: { roff_TE, NULL, NULL, 0 }, /* TE */
1.315 schwarze 546: { roff_Dd, NULL, NULL, 0 }, /* TH */
1.295 schwarze 547: { roff_line_ignore, NULL, NULL, 0 }, /* tkf */
548: { roff_unsupp, NULL, NULL, 0 }, /* tl */
549: { roff_line_ignore, NULL, NULL, 0 }, /* tm */
550: { roff_line_ignore, NULL, NULL, 0 }, /* tm1 */
551: { roff_line_ignore, NULL, NULL, 0 }, /* tmc */
552: { roff_tr, NULL, NULL, 0 }, /* tr */
553: { roff_line_ignore, NULL, NULL, 0 }, /* track */
554: { roff_line_ignore, NULL, NULL, 0 }, /* transchar */
555: { roff_insec, NULL, NULL, 0 }, /* trf */
556: { roff_line_ignore, NULL, NULL, 0 }, /* trimat */
557: { roff_unsupp, NULL, NULL, 0 }, /* trin */
558: { roff_unsupp, NULL, NULL, 0 }, /* trnt */
559: { roff_line_ignore, NULL, NULL, 0 }, /* troff */
560: { roff_TS, NULL, NULL, 0 }, /* TS */
561: { roff_line_ignore, NULL, NULL, 0 }, /* uf */
562: { roff_line_ignore, NULL, NULL, 0 }, /* ul */
563: { roff_unsupp, NULL, NULL, 0 }, /* unformat */
564: { roff_line_ignore, NULL, NULL, 0 }, /* unwatch */
565: { roff_line_ignore, NULL, NULL, 0 }, /* unwatchn */
566: { roff_line_ignore, NULL, NULL, 0 }, /* vpt */
567: { roff_line_ignore, NULL, NULL, 0 }, /* vs */
568: { roff_line_ignore, NULL, NULL, 0 }, /* warn */
569: { roff_line_ignore, NULL, NULL, 0 }, /* warnscale */
570: { roff_line_ignore, NULL, NULL, 0 }, /* watch */
571: { roff_line_ignore, NULL, NULL, 0 }, /* watchlength */
572: { roff_line_ignore, NULL, NULL, 0 }, /* watchn */
573: { roff_unsupp, NULL, NULL, 0 }, /* wh */
574: { roff_unsupp, NULL, NULL, 0 }, /* while */
575: { roff_insec, NULL, NULL, 0 }, /* write */
576: { roff_insec, NULL, NULL, 0 }, /* writec */
577: { roff_insec, NULL, NULL, 0 }, /* writem */
578: { roff_line_ignore, NULL, NULL, 0 }, /* xflag */
579: { roff_cblock, NULL, NULL, 0 }, /* . */
1.306 schwarze 580: { roff_renamed, NULL, NULL, 0 },
1.295 schwarze 581: { roff_userdef, NULL, NULL, 0 }
1.67 kristaps 582: };
583:
1.141 kristaps 584: /* Array of injected predefined strings. */
585: #define PREDEFS_MAX 38
586: static const struct predef predefs[PREDEFS_MAX] = {
587: #include "predefs.in"
588: };
589:
1.305 schwarze 590: static int roffce_lines; /* number of input lines to center */
591: static struct roff_node *roffce_node; /* active request */
1.178 schwarze 592: static int roffit_lines; /* number of lines to delay */
593: static char *roffit_macro; /* nil-terminated macro line */
594:
1.207 schwarze 595:
1.266 schwarze 596: /* --- request table ------------------------------------------------------ */
597:
1.295 schwarze 598: struct ohash *
599: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85 kristaps 600: {
1.295 schwarze 601: struct ohash *htab;
602: struct roffreq *req;
603: enum roff_tok tok;
604: size_t sz;
605: unsigned int slot;
1.85 kristaps 606:
1.295 schwarze 607: htab = mandoc_malloc(sizeof(*htab));
608: mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85 kristaps 609:
1.295 schwarze 610: for (tok = mintok; tok < maxtok; tok++) {
1.296 schwarze 611: if (roff_name[tok] == NULL)
612: continue;
1.295 schwarze 613: sz = strlen(roff_name[tok]);
614: req = mandoc_malloc(sizeof(*req) + sz + 1);
615: req->tok = tok;
616: memcpy(req->name, roff_name[tok], sz + 1);
617: slot = ohash_qlookup(htab, req->name);
618: ohash_insert(htab, slot, req);
1.85 kristaps 619: }
1.295 schwarze 620: return htab;
1.85 kristaps 621: }
622:
1.295 schwarze 623: void
624: roffhash_free(struct ohash *htab)
1.67 kristaps 625: {
1.295 schwarze 626: struct roffreq *req;
627: unsigned int slot;
1.67 kristaps 628:
1.295 schwarze 629: if (htab == NULL)
630: return;
631: for (req = ohash_first(htab, &slot); req != NULL;
632: req = ohash_next(htab, &slot))
633: free(req);
634: ohash_delete(htab);
635: free(htab);
636: }
637:
638: enum roff_tok
639: roffhash_find(struct ohash *htab, const char *name, size_t sz)
640: {
641: struct roffreq *req;
642: const char *end;
643:
644: if (sz) {
645: end = name + sz;
646: req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
647: } else
648: req = ohash_find(htab, ohash_qlookup(htab, name));
649: return req == NULL ? TOKEN_NONE : req->tok;
1.67 kristaps 650: }
651:
1.266 schwarze 652: /* --- stack of request blocks -------------------------------------------- */
653:
1.67 kristaps 654: /*
655: * Pop the current node off of the stack of roff instructions currently
656: * pending.
657: */
658: static void
659: roffnode_pop(struct roff *r)
660: {
661: struct roffnode *p;
662:
1.75 kristaps 663: assert(r->last);
1.207 schwarze 664: p = r->last;
1.82 kristaps 665:
1.75 kristaps 666: r->last = r->last->parent;
1.106 kristaps 667: free(p->name);
668: free(p->end);
1.67 kristaps 669: free(p);
670: }
671:
672: /*
673: * Push a roff node onto the instruction stack. This must later be
674: * removed with roffnode_pop().
675: */
1.98 schwarze 676: static void
1.294 schwarze 677: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106 kristaps 678: int line, int col)
1.67 kristaps 679: {
680: struct roffnode *p;
681:
1.98 schwarze 682: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 683: p->tok = tok;
1.106 kristaps 684: if (name)
685: p->name = mandoc_strdup(name);
1.67 kristaps 686: p->parent = r->last;
687: p->line = line;
688: p->col = col;
1.198 schwarze 689: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 690:
691: r->last = p;
692: }
693:
1.266 schwarze 694: /* --- roff parser state data management ---------------------------------- */
695:
1.67 kristaps 696: static void
697: roff_free1(struct roff *r)
698: {
1.176 schwarze 699: struct tbl_node *tbl;
1.167 kristaps 700: int i;
1.67 kristaps 701:
1.176 schwarze 702: while (NULL != (tbl = r->first_tbl)) {
703: r->first_tbl = tbl->next;
704: tbl_free(tbl);
1.109 kristaps 705: }
1.113 kristaps 706: r->first_tbl = r->last_tbl = r->tbl = NULL;
707:
1.319 schwarze 708: if (r->last_eqn != NULL)
709: eqn_free(r->last_eqn);
710: r->last_eqn = r->eqn = NULL;
1.125 kristaps 711:
1.67 kristaps 712: while (r->last)
713: roffnode_pop(r);
1.109 kristaps 714:
1.223 schwarze 715: free (r->rstack);
716: r->rstack = NULL;
717: r->rstacksz = 0;
718: r->rstackpos = -1;
719:
720: roff_freereg(r->regtab);
721: r->regtab = NULL;
722:
1.167 kristaps 723: roff_freestr(r->strtab);
1.306 schwarze 724: roff_freestr(r->rentab);
1.167 kristaps 725: roff_freestr(r->xmbtab);
1.306 schwarze 726: r->strtab = r->rentab = r->xmbtab = NULL;
1.167 kristaps 727:
728: if (r->xtab)
729: for (i = 0; i < 128; i++)
730: free(r->xtab[i].p);
731: free(r->xtab);
732: r->xtab = NULL;
1.67 kristaps 733: }
734:
735: void
736: roff_reset(struct roff *r)
737: {
738: roff_free1(r);
1.227 schwarze 739: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.303 schwarze 740: r->control = '\0';
741: r->escape = '\\';
1.307 schwarze 742: roffce_lines = 0;
743: roffce_node = NULL;
744: roffit_lines = 0;
745: roffit_macro = NULL;
1.67 kristaps 746: }
747:
748: void
749: roff_free(struct roff *r)
750: {
751: roff_free1(r);
1.295 schwarze 752: roffhash_free(r->reqtab);
1.67 kristaps 753: free(r);
754: }
755:
756: struct roff *
1.279 schwarze 757: roff_alloc(struct mparse *parse, int options)
1.67 kristaps 758: {
759: struct roff *r;
760:
1.98 schwarze 761: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 762: r->parse = parse;
1.295 schwarze 763: r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.199 schwarze 764: r->options = options;
1.227 schwarze 765: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82 kristaps 766: r->rstackpos = -1;
1.303 schwarze 767: r->escape = '\\';
1.277 schwarze 768: return r;
1.265 schwarze 769: }
770:
1.266 schwarze 771: /* --- syntax tree state data management ---------------------------------- */
772:
1.265 schwarze 773: static void
774: roff_man_free1(struct roff_man *man)
775: {
776:
1.266 schwarze 777: if (man->first != NULL)
778: roff_node_delete(man, man->first);
1.265 schwarze 779: free(man->meta.msec);
780: free(man->meta.vol);
781: free(man->meta.os);
782: free(man->meta.arch);
783: free(man->meta.title);
784: free(man->meta.name);
785: free(man->meta.date);
786: }
787:
788: static void
789: roff_man_alloc1(struct roff_man *man)
790: {
791:
792: memset(&man->meta, 0, sizeof(man->meta));
793: man->first = mandoc_calloc(1, sizeof(*man->first));
794: man->first->type = ROFFT_ROOT;
795: man->last = man->first;
796: man->last_es = NULL;
797: man->flags = 0;
798: man->macroset = MACROSET_NONE;
799: man->lastsec = man->lastnamed = SEC_NONE;
800: man->next = ROFF_NEXT_CHILD;
801: }
802:
803: void
804: roff_man_reset(struct roff_man *man)
805: {
806:
807: roff_man_free1(man);
808: roff_man_alloc1(man);
809: }
810:
811: void
812: roff_man_free(struct roff_man *man)
813: {
814:
815: roff_man_free1(man);
816: free(man);
817: }
818:
819: struct roff_man *
820: roff_man_alloc(struct roff *roff, struct mparse *parse,
1.316 schwarze 821: const char *os_s, int quick)
1.265 schwarze 822: {
823: struct roff_man *man;
824:
825: man = mandoc_calloc(1, sizeof(*man));
826: man->parse = parse;
827: man->roff = roff;
1.316 schwarze 828: man->os_s = os_s;
1.265 schwarze 829: man->quick = quick;
830: roff_man_alloc1(man);
1.296 schwarze 831: roff->man = man;
1.277 schwarze 832: return man;
1.67 kristaps 833: }
834:
1.266 schwarze 835: /* --- syntax tree handling ----------------------------------------------- */
836:
837: struct roff_node *
838: roff_node_alloc(struct roff_man *man, int line, int pos,
839: enum roff_type type, int tok)
840: {
841: struct roff_node *n;
842:
843: n = mandoc_calloc(1, sizeof(*n));
844: n->line = line;
845: n->pos = pos;
846: n->tok = tok;
847: n->type = type;
848: n->sec = man->lastsec;
849:
850: if (man->flags & MDOC_SYNOPSIS)
1.285 schwarze 851: n->flags |= NODE_SYNPRETTY;
1.266 schwarze 852: else
1.285 schwarze 853: n->flags &= ~NODE_SYNPRETTY;
1.266 schwarze 854: if (man->flags & MDOC_NEWLINE)
1.285 schwarze 855: n->flags |= NODE_LINE;
1.266 schwarze 856: man->flags &= ~MDOC_NEWLINE;
857:
1.277 schwarze 858: return n;
1.266 schwarze 859: }
860:
861: void
862: roff_node_append(struct roff_man *man, struct roff_node *n)
863: {
864:
865: switch (man->next) {
866: case ROFF_NEXT_SIBLING:
1.281 schwarze 867: if (man->last->next != NULL) {
868: n->next = man->last->next;
869: man->last->next->prev = n;
870: } else
871: man->last->parent->last = n;
1.266 schwarze 872: man->last->next = n;
873: n->prev = man->last;
874: n->parent = man->last->parent;
875: break;
876: case ROFF_NEXT_CHILD:
1.287 schwarze 877: if (man->last->child != NULL) {
878: n->next = man->last->child;
879: man->last->child->prev = n;
880: } else
881: man->last->last = n;
1.266 schwarze 882: man->last->child = n;
883: n->parent = man->last;
884: break;
885: default:
886: abort();
887: }
1.282 schwarze 888: man->last = n;
1.266 schwarze 889:
890: switch (n->type) {
891: case ROFFT_HEAD:
892: n->parent->head = n;
893: break;
894: case ROFFT_BODY:
1.282 schwarze 895: if (n->end != ENDBODY_NOT)
896: return;
1.266 schwarze 897: n->parent->body = n;
898: break;
899: case ROFFT_TAIL:
900: n->parent->tail = n;
901: break;
902: default:
1.282 schwarze 903: return;
1.266 schwarze 904: }
1.282 schwarze 905:
906: /*
907: * Copy over the normalised-data pointer of our parent. Not
908: * everybody has one, but copying a null pointer is fine.
909: */
910:
911: n->norm = n->parent->norm;
912: assert(n->parent->type == ROFFT_BLOCK);
1.266 schwarze 913: }
914:
1.267 schwarze 915: void
916: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
917: {
918: struct roff_node *n;
919:
920: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
921: n->string = roff_strdup(man->roff, word);
922: roff_node_append(man, n);
1.286 schwarze 923: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 924: man->next = ROFF_NEXT_SIBLING;
925: }
926:
927: void
928: roff_word_append(struct roff_man *man, const char *word)
929: {
930: struct roff_node *n;
931: char *addstr, *newstr;
932:
933: n = man->last;
934: addstr = roff_strdup(man->roff, word);
935: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
936: free(addstr);
937: free(n->string);
938: n->string = newstr;
939: man->next = ROFF_NEXT_SIBLING;
1.268 schwarze 940: }
941:
942: void
943: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
944: {
945: struct roff_node *n;
946:
947: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
948: roff_node_append(man, n);
949: man->next = ROFF_NEXT_CHILD;
950: }
951:
952: struct roff_node *
953: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
954: {
955: struct roff_node *n;
956:
957: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
958: roff_node_append(man, n);
959: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 960: return n;
1.267 schwarze 961: }
962:
1.266 schwarze 963: struct roff_node *
964: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
965: {
966: struct roff_node *n;
967:
968: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
969: roff_node_append(man, n);
970: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 971: return n;
1.266 schwarze 972: }
973:
974: struct roff_node *
975: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
976: {
977: struct roff_node *n;
978:
979: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
980: roff_node_append(man, n);
981: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 982: return n;
1.267 schwarze 983: }
984:
1.321 schwarze 985: static void
986: roff_addtbl(struct roff_man *man, struct tbl_node *tbl)
1.267 schwarze 987: {
988: struct roff_node *n;
1.321 schwarze 989: const struct tbl_span *span;
1.267 schwarze 990:
991: if (man->macroset == MACROSET_MAN)
1.302 schwarze 992: man_breakscope(man, ROFF_TS);
1.321 schwarze 993: while ((span = tbl_span(tbl)) != NULL) {
994: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
995: n->span = span;
996: roff_node_append(man, n);
997: n->flags |= NODE_VALID | NODE_ENDED;
998: man->next = ROFF_NEXT_SIBLING;
999: }
1.266 schwarze 1000: }
1001:
1002: void
1003: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1004: {
1005:
1006: /* Adjust siblings. */
1007:
1008: if (n->prev)
1009: n->prev->next = n->next;
1010: if (n->next)
1011: n->next->prev = n->prev;
1012:
1013: /* Adjust parent. */
1014:
1015: if (n->parent != NULL) {
1016: if (n->parent->child == n)
1017: n->parent->child = n->next;
1018: if (n->parent->last == n)
1019: n->parent->last = n->prev;
1020: }
1021:
1022: /* Adjust parse point. */
1023:
1024: if (man == NULL)
1025: return;
1026: if (man->last == n) {
1027: if (n->prev == NULL) {
1028: man->last = n->parent;
1029: man->next = ROFF_NEXT_CHILD;
1030: } else {
1031: man->last = n->prev;
1032: man->next = ROFF_NEXT_SIBLING;
1033: }
1034: }
1035: if (man->first == n)
1036: man->first = NULL;
1037: }
1038:
1039: void
1040: roff_node_free(struct roff_node *n)
1041: {
1042:
1043: if (n->args != NULL)
1044: mdoc_argv_free(n->args);
1045: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1046: free(n->norm);
1.319 schwarze 1047: if (n->eqn != NULL)
1048: eqn_box_free(n->eqn);
1.266 schwarze 1049: free(n->string);
1050: free(n);
1051: }
1052:
1053: void
1054: roff_node_delete(struct roff_man *man, struct roff_node *n)
1055: {
1056:
1057: while (n->child != NULL)
1058: roff_node_delete(man, n->child);
1059: roff_node_unlink(man, n);
1060: roff_node_free(n);
1.269 schwarze 1061: }
1062:
1063: void
1064: deroff(char **dest, const struct roff_node *n)
1065: {
1066: char *cp;
1067: size_t sz;
1068:
1069: if (n->type != ROFFT_TEXT) {
1070: for (n = n->child; n != NULL; n = n->next)
1071: deroff(dest, n);
1072: return;
1073: }
1074:
1.288 schwarze 1075: /* Skip leading whitespace. */
1.269 schwarze 1076:
1.288 schwarze 1077: for (cp = n->string; *cp != '\0'; cp++) {
1.289 schwarze 1078: if (cp[0] == '\\' && cp[1] != '\0' &&
1079: strchr(" %&0^|~", cp[1]) != NULL)
1.269 schwarze 1080: cp++;
1.288 schwarze 1081: else if ( ! isspace((unsigned char)*cp))
1.269 schwarze 1082: break;
1083: }
1084:
1.289 schwarze 1085: /* Skip trailing backslash. */
1086:
1087: sz = strlen(cp);
1.290 schwarze 1088: if (sz > 0 && cp[sz - 1] == '\\')
1.289 schwarze 1089: sz--;
1090:
1.269 schwarze 1091: /* Skip trailing whitespace. */
1092:
1.289 schwarze 1093: for (; sz; sz--)
1.269 schwarze 1094: if ( ! isspace((unsigned char)cp[sz-1]))
1095: break;
1096:
1097: /* Skip empty strings. */
1098:
1099: if (sz == 0)
1100: return;
1101:
1102: if (*dest == NULL) {
1103: *dest = mandoc_strndup(cp, sz);
1104: return;
1105: }
1106:
1107: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1108: free(*dest);
1109: *dest = cp;
1.266 schwarze 1110: }
1111:
1112: /* --- main functions of the roff parser ---------------------------------- */
1113:
1.94 kristaps 1114: /*
1.206 schwarze 1115: * In the current line, expand escape sequences that tend to get
1116: * used in numerical expressions and conditional requests.
1117: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 1118: */
1.172 schwarze 1119: static enum rofferr
1.238 schwarze 1120: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 1121: {
1.208 schwarze 1122: char ubuf[24]; /* buffer to print the number */
1.205 schwarze 1123: const char *start; /* start of the string to process */
1.209 schwarze 1124: char *stesc; /* start of an escape sequence ('\\') */
1.108 schwarze 1125: const char *stnam; /* start of the name, after "[(*" */
1126: const char *cp; /* end of the name, e.g. before ']' */
1127: const char *res; /* the string to be substituted */
1.238 schwarze 1128: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 1129: size_t maxl; /* expected length of the escape name */
1130: size_t naml; /* actual length of the escape name */
1.237 schwarze 1131: enum mandoc_esc esc; /* type of the escape sequence */
1132: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 1133: int expand_count; /* to avoid infinite loops */
1.206 schwarze 1134: int npos; /* position in numeric expression */
1.218 schwarze 1135: int arg_complete; /* argument not interrupted by eol */
1.303 schwarze 1136: int done; /* no more input available */
1.315 schwarze 1137: int deftype; /* type of definition to paste */
1.323 schwarze 1138: int rcsid; /* kind of RCS id seen */
1.206 schwarze 1139: char term; /* character terminating the escape */
1.94 kristaps 1140:
1.303 schwarze 1141: /* Search forward for comments. */
1142:
1143: done = 0;
1144: start = buf->buf + pos;
1145: for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
1146: if (stesc[0] != r->escape || stesc[1] == '\0')
1147: continue;
1148: stesc++;
1149: if (*stesc != '"' && *stesc != '#')
1150: continue;
1.313 schwarze 1151:
1152: /* Comment found, look for RCS id. */
1153:
1.323 schwarze 1154: rcsid = 0;
1.313 schwarze 1155: if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.323 schwarze 1156: rcsid = 1 << MANDOC_OS_OPENBSD;
1.313 schwarze 1157: cp += 8;
1158: } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.323 schwarze 1159: rcsid = 1 << MANDOC_OS_NETBSD;
1.313 schwarze 1160: cp += 7;
1161: }
1162: if (cp != NULL &&
1163: isalnum((unsigned char)*cp) == 0 &&
1.314 schwarze 1164: strchr(cp, '$') != NULL) {
1.323 schwarze 1165: if (r->man->meta.rcsids & rcsid)
1.314 schwarze 1166: mandoc_msg(MANDOCERR_RCS_REP, r->parse,
1167: ln, stesc + 1 - buf->buf, stesc + 1);
1.323 schwarze 1168: r->man->meta.rcsids |= rcsid;
1.314 schwarze 1169: }
1.313 schwarze 1170:
1171: /* Handle trailing whitespace. */
1172:
1.303 schwarze 1173: cp = strchr(stesc--, '\0') - 1;
1174: if (*cp == '\n') {
1175: done = 1;
1176: cp--;
1177: }
1178: if (*cp == ' ' || *cp == '\t')
1179: mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
1180: ln, cp - buf->buf, NULL);
1181: while (stesc > start && stesc[-1] == ' ')
1182: stesc--;
1183: *stesc = '\0';
1184: break;
1185: }
1186: if (stesc == start)
1187: return ROFF_CONT;
1188: stesc--;
1189:
1190: /* Notice the end of the input. */
1191:
1192: if (*stesc == '\n') {
1193: *stesc-- = '\0';
1194: done = 1;
1195: }
1196:
1.170 schwarze 1197: expand_count = 0;
1.303 schwarze 1198: while (stesc >= start) {
1.170 schwarze 1199:
1.205 schwarze 1200: /* Search backwards for the next backslash. */
1201:
1.303 schwarze 1202: if (*stesc != r->escape) {
1203: if (*stesc == '\\') {
1204: *stesc = '\0';
1205: buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
1206: buf->buf, stesc + 1) + 1;
1207: start = nbuf + pos;
1208: stesc = nbuf + (stesc - buf->buf);
1209: free(buf->buf);
1210: buf->buf = nbuf;
1211: }
1212: stesc--;
1.205 schwarze 1213: continue;
1.303 schwarze 1214: }
1.205 schwarze 1215:
1216: /* If it is escaped, skip it. */
1217:
1218: for (cp = stesc - 1; cp >= start; cp--)
1.303 schwarze 1219: if (*cp != r->escape)
1.205 schwarze 1220: break;
1221:
1.238 schwarze 1222: if ((stesc - cp) % 2 == 0) {
1.303 schwarze 1223: while (stesc > cp)
1224: *stesc-- = '\\';
1.205 schwarze 1225: continue;
1.303 schwarze 1226: } else if (stesc[1] != '\0') {
1227: *stesc = '\\';
1228: } else {
1229: *stesc-- = '\0';
1230: if (done)
1231: continue;
1232: else
1233: return ROFF_APPEND;
1.205 schwarze 1234: }
1.108 schwarze 1235:
1.206 schwarze 1236: /* Decide whether to expand or to check only. */
1.108 schwarze 1237:
1.206 schwarze 1238: term = '\0';
1.205 schwarze 1239: cp = stesc + 1;
1.181 schwarze 1240: switch (*cp) {
1.207 schwarze 1241: case '*':
1.181 schwarze 1242: res = NULL;
1243: break;
1.207 schwarze 1244: case 'B':
1245: case 'w':
1.206 schwarze 1246: term = cp[1];
1247: /* FALLTHROUGH */
1.207 schwarze 1248: case 'n':
1.181 schwarze 1249: res = ubuf;
1250: break;
1251: default:
1.237 schwarze 1252: esc = mandoc_escape(&cp, &stnam, &inaml);
1253: if (esc == ESCAPE_ERROR ||
1254: (esc == ESCAPE_SPECIAL &&
1.279 schwarze 1255: mchars_spec2cp(stnam, inaml) < 0))
1.219 schwarze 1256: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1257: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1258: "%.*s", (int)(cp - stesc), stesc);
1.303 schwarze 1259: stesc--;
1.205 schwarze 1260: continue;
1.152 kristaps 1261: }
1262:
1.205 schwarze 1263: if (EXPAND_LIMIT < ++expand_count) {
1264: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1265: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1266: return ROFF_IGN;
1.205 schwarze 1267: }
1.108 schwarze 1268:
1269: /*
1270: * The third character decides the length
1.181 schwarze 1271: * of the name of the string or register.
1.108 schwarze 1272: * Save a pointer to the name.
1273: */
1274:
1.238 schwarze 1275: if (term == '\0') {
1.206 schwarze 1276: switch (*++cp) {
1.207 schwarze 1277: case '\0':
1.206 schwarze 1278: maxl = 0;
1279: break;
1.207 schwarze 1280: case '(':
1.206 schwarze 1281: cp++;
1282: maxl = 2;
1283: break;
1.207 schwarze 1284: case '[':
1.206 schwarze 1285: cp++;
1286: term = ']';
1287: maxl = 0;
1288: break;
1289: default:
1290: maxl = 1;
1291: break;
1292: }
1293: } else {
1294: cp += 2;
1.94 kristaps 1295: maxl = 0;
1296: }
1.108 schwarze 1297: stnam = cp;
1.94 kristaps 1298:
1.108 schwarze 1299: /* Advance to the end of the name. */
1.94 kristaps 1300:
1.253 schwarze 1301: naml = 0;
1.218 schwarze 1302: arg_complete = 1;
1.253 schwarze 1303: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1304: if (*cp == '\0') {
1.219 schwarze 1305: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1306: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1307: arg_complete = 0;
1.206 schwarze 1308: break;
1.153 kristaps 1309: }
1.238 schwarze 1310: if (maxl == 0 && *cp == term) {
1.206 schwarze 1311: cp++;
1.253 schwarze 1312: break;
1313: }
1314: if (*cp++ != '\\' || stesc[1] != 'w') {
1315: naml++;
1316: continue;
1317: }
1318: switch (mandoc_escape(&cp, NULL, NULL)) {
1319: case ESCAPE_SPECIAL:
1320: case ESCAPE_UNICODE:
1321: case ESCAPE_NUMBERED:
1322: case ESCAPE_OVERSTRIKE:
1323: naml++;
1324: break;
1325: default:
1.94 kristaps 1326: break;
1.206 schwarze 1327: }
1.94 kristaps 1328: }
1329:
1.108 schwarze 1330: /*
1331: * Retrieve the replacement string; if it is
1332: * undefined, resume searching for escapes.
1333: */
1334:
1.206 schwarze 1335: switch (stesc[1]) {
1.207 schwarze 1336: case '*':
1.315 schwarze 1337: if (arg_complete) {
1338: deftype = ROFFDEF_USER | ROFFDEF_PRE;
1339: res = roff_getstrn(r, stnam, naml, &deftype);
1340: }
1.206 schwarze 1341: break;
1.207 schwarze 1342: case 'B':
1.206 schwarze 1343: npos = 0;
1.218 schwarze 1344: ubuf[0] = arg_complete &&
1.261 schwarze 1345: roff_evalnum(r, ln, stnam, &npos,
1346: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1347: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1348: ubuf[1] = '\0';
1349: break;
1.207 schwarze 1350: case 'n':
1.218 schwarze 1351: if (arg_complete)
1352: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1353: roff_getregn(r, stnam, naml));
1354: else
1355: ubuf[0] = '\0';
1.206 schwarze 1356: break;
1.207 schwarze 1357: case 'w':
1.218 schwarze 1358: /* use even incomplete args */
1.208 schwarze 1359: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1360: 24 * (int)naml);
1361: break;
1362: }
1.94 kristaps 1363:
1.238 schwarze 1364: if (res == NULL) {
1.219 schwarze 1365: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1366: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1367: "%.*s", (int)naml, stnam);
1.142 kristaps 1368: res = "";
1.246 schwarze 1369: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1370: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1371: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1372: return ROFF_IGN;
1.94 kristaps 1373: }
1374:
1.108 schwarze 1375: /* Replace the escape sequence by the string. */
1376:
1.209 schwarze 1377: *stesc = '\0';
1.238 schwarze 1378: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1379: buf->buf, res, cp) + 1;
1.94 kristaps 1380:
1.205 schwarze 1381: /* Prepare for the next replacement. */
1.94 kristaps 1382:
1.205 schwarze 1383: start = nbuf + pos;
1.238 schwarze 1384: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1385: free(buf->buf);
1386: buf->buf = nbuf;
1.154 kristaps 1387: }
1.277 schwarze 1388: return ROFF_CONT;
1.154 kristaps 1389: }
1390:
1391: /*
1.275 schwarze 1392: * Process text streams.
1.154 kristaps 1393: */
1394: static enum rofferr
1.305 schwarze 1395: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.154 kristaps 1396: {
1397: size_t sz;
1398: const char *start;
1.178 schwarze 1399: char *p;
1400: int isz;
1.154 kristaps 1401: enum mandoc_esc esc;
1402:
1.275 schwarze 1403: /* Spring the input line trap. */
1404:
1405: if (roffit_lines == 1) {
1406: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1407: free(buf->buf);
1408: buf->buf = p;
1409: buf->sz = isz + 1;
1410: *offs = 0;
1411: free(roffit_macro);
1412: roffit_lines = 0;
1.277 schwarze 1413: return ROFF_REPARSE;
1.275 schwarze 1414: } else if (roffit_lines > 1)
1415: --roffit_lines;
1416:
1.305 schwarze 1417: if (roffce_node != NULL && buf->buf[pos] != '\0') {
1418: if (roffce_lines < 1) {
1419: r->man->last = roffce_node;
1420: r->man->next = ROFF_NEXT_SIBLING;
1421: roffce_lines = 0;
1422: roffce_node = NULL;
1423: } else
1424: roffce_lines--;
1425: }
1426:
1.275 schwarze 1427: /* Convert all breakable hyphens into ASCII_HYPH. */
1428:
1.238 schwarze 1429: start = p = buf->buf + pos;
1.154 kristaps 1430:
1.238 schwarze 1431: while (*p != '\0') {
1.154 kristaps 1432: sz = strcspn(p, "-\\");
1433: p += sz;
1434:
1.238 schwarze 1435: if (*p == '\0')
1.159 kristaps 1436: break;
1437:
1.238 schwarze 1438: if (*p == '\\') {
1.154 kristaps 1439: /* Skip over escapes. */
1440: p++;
1.189 schwarze 1441: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1442: if (esc == ESCAPE_ERROR)
1.154 kristaps 1443: break;
1.264 schwarze 1444: while (*p == '-')
1445: p++;
1.155 kristaps 1446: continue;
1.159 kristaps 1447: } else if (p == start) {
1.158 kristaps 1448: p++;
1.155 kristaps 1449: continue;
1.158 kristaps 1450: }
1.155 kristaps 1451:
1.171 schwarze 1452: if (isalpha((unsigned char)p[-1]) &&
1453: isalpha((unsigned char)p[1]))
1.155 kristaps 1454: *p = ASCII_HYPH;
1455: p++;
1.94 kristaps 1456: }
1.277 schwarze 1457: return ROFF_CONT;
1.94 kristaps 1458: }
1459:
1.67 kristaps 1460: enum rofferr
1.238 schwarze 1461: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1462: {
1.294 schwarze 1463: enum roff_tok t;
1.109 kristaps 1464: enum rofferr e;
1.238 schwarze 1465: int pos; /* parse point */
1.243 schwarze 1466: int spos; /* saved parse point for messages */
1.238 schwarze 1467: int ppos; /* original offset in buf->buf */
1468: int ctl; /* macro line (boolean) */
1469:
1470: ppos = pos = *offs;
1.79 kristaps 1471:
1.230 schwarze 1472: /* Handle in-line equation delimiters. */
1473:
1.236 schwarze 1474: if (r->tbl == NULL &&
1475: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1476: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1477: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1478: if (e == ROFF_REPARSE)
1.277 schwarze 1479: return e;
1.230 schwarze 1480: assert(e == ROFF_CONT);
1481: }
1482:
1483: /* Expand some escape sequences. */
1.94 kristaps 1484:
1.238 schwarze 1485: e = roff_res(r, buf, ln, pos);
1.303 schwarze 1486: if (e == ROFF_IGN || e == ROFF_APPEND)
1.277 schwarze 1487: return e;
1.238 schwarze 1488: assert(e == ROFF_CONT);
1.94 kristaps 1489:
1.238 schwarze 1490: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1491:
1.94 kristaps 1492: /*
1.79 kristaps 1493: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1494: * text through the macro's filter.
1495: * Equations process all content themselves.
1496: * Tables process almost all content themselves, but we want
1497: * to warn about macros before passing it there.
1.79 kristaps 1498: */
1.74 kristaps 1499:
1.252 schwarze 1500: if (r->last != NULL && ! ctl) {
1.78 kristaps 1501: t = r->last->tok;
1.238 schwarze 1502: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.294 schwarze 1503: if (e == ROFF_IGN)
1.277 schwarze 1504: return e;
1.294 schwarze 1505: assert(e == ROFF_CONT);
1.182 schwarze 1506: }
1.319 schwarze 1507: if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
1508: eqn_read(r->eqn, buf->buf + ppos);
1509: return ROFF_IGN;
1510: }
1.321 schwarze 1511: if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
1512: tbl_read(r->tbl, ln, buf->buf, ppos);
1513: roff_addtbl(r->man, r->tbl);
1514: return ROFF_IGN;
1515: }
1.252 schwarze 1516: if ( ! ctl)
1.305 schwarze 1517: return roff_parsetext(r, buf, pos, offs);
1.228 schwarze 1518:
1519: /* Skip empty request lines. */
1520:
1.238 schwarze 1521: if (buf->buf[pos] == '"') {
1.228 schwarze 1522: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1523: ln, pos, NULL);
1.277 schwarze 1524: return ROFF_IGN;
1.238 schwarze 1525: } else if (buf->buf[pos] == '\0')
1.277 schwarze 1526: return ROFF_IGN;
1.67 kristaps 1527:
1.79 kristaps 1528: /*
1529: * If a scope is open, go to the child handler for that macro,
1530: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1531: * Don't do so if an equation is open.
1.79 kristaps 1532: */
1.78 kristaps 1533:
1.79 kristaps 1534: if (r->last) {
1535: t = r->last->tok;
1.277 schwarze 1536: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1537: }
1.78 kristaps 1538:
1.243 schwarze 1539: /* No scope is open. This is a new request or macro. */
1540:
1541: spos = pos;
1542: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1543:
1544: /* Tables ignore most macros. */
1545:
1.308 schwarze 1546: if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.309 schwarze 1547: t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.243 schwarze 1548: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1549: ln, pos, buf->buf + spos);
1.308 schwarze 1550: if (t != TOKEN_NONE)
1.277 schwarze 1551: return ROFF_IGN;
1.257 schwarze 1552: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1553: pos++;
1.291 schwarze 1554: while (buf->buf[pos] == ' ')
1.257 schwarze 1555: pos++;
1.321 schwarze 1556: tbl_read(r->tbl, ln, buf->buf, pos);
1557: roff_addtbl(r->man, r->tbl);
1558: return ROFF_IGN;
1.243 schwarze 1559: }
1560:
1.305 schwarze 1561: /* For now, let high level macros abort .ce mode. */
1562:
1563: if (ctl && roffce_node != NULL &&
1.324 schwarze 1564: (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
1565: t == ROFF_TH || t == ROFF_TS)) {
1.305 schwarze 1566: r->man->last = roffce_node;
1567: r->man->next = ROFF_NEXT_SIBLING;
1568: roffce_lines = 0;
1569: roffce_node = NULL;
1570: }
1571:
1.79 kristaps 1572: /*
1.243 schwarze 1573: * This is neither a roff request nor a user-defined macro.
1574: * Let the standard macro set parsers handle it.
1.79 kristaps 1575: */
1.67 kristaps 1576:
1.294 schwarze 1577: if (t == TOKEN_NONE)
1.277 schwarze 1578: return ROFF_CONT;
1.243 schwarze 1579:
1580: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1581:
1.296 schwarze 1582: return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.74 kristaps 1583: }
1584:
1.117 kristaps 1585: void
1.74 kristaps 1586: roff_endparse(struct roff *r)
1587: {
1.321 schwarze 1588: if (r->last != NULL)
1.221 schwarze 1589: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1590: r->last->line, r->last->col,
1.295 schwarze 1591: roff_name[r->last->tok]);
1.117 kristaps 1592:
1.321 schwarze 1593: if (r->eqn != NULL) {
1.221 schwarze 1594: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1.319 schwarze 1595: r->eqn->node->line, r->eqn->node->pos, "EQ");
1596: eqn_parse(r->eqn);
1597: r->eqn = NULL;
1.125 kristaps 1598: }
1599:
1.321 schwarze 1600: if (r->tbl != NULL) {
1.221 schwarze 1601: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1602: r->tbl->line, r->tbl->pos, "TS");
1.321 schwarze 1603: tbl_end(r->tbl);
1604: r->tbl = NULL;
1.117 kristaps 1605: }
1.67 kristaps 1606: }
1607:
1608: /*
1609: * Parse a roff node's type from the input buffer. This must be in the
1610: * form of ".foo xxx" in the usual way.
1611: */
1.294 schwarze 1612: static enum roff_tok
1.214 schwarze 1613: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1614: {
1.214 schwarze 1615: char *cp;
1.106 kristaps 1616: const char *mac;
1617: size_t maclen;
1.315 schwarze 1618: int deftype;
1.294 schwarze 1619: enum roff_tok t;
1.67 kristaps 1620:
1.214 schwarze 1621: cp = buf + *pos;
1622:
1623: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294 schwarze 1624: return TOKEN_NONE;
1.67 kristaps 1625:
1.214 schwarze 1626: mac = cp;
1627: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1628:
1.315 schwarze 1629: deftype = ROFFDEF_USER | ROFFDEF_REN;
1630: r->current_string = roff_getstrn(r, mac, maclen, &deftype);
1631: switch (deftype) {
1632: case ROFFDEF_USER:
1633: t = ROFF_USERDEF;
1634: break;
1635: case ROFFDEF_REN:
1636: t = ROFF_RENAMED;
1637: break;
1638: default:
1639: t = roffhash_find(r->reqtab, mac, maclen);
1640: break;
1641: }
1.294 schwarze 1642: if (t != TOKEN_NONE)
1.214 schwarze 1643: *pos = cp - buf;
1.325 schwarze 1644: else if (deftype == ROFFDEF_UNDEF) {
1645: /* Using an undefined macro defines it to be empty. */
1646: roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
1647: roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
1648: }
1.277 schwarze 1649: return t;
1.67 kristaps 1650: }
1651:
1.266 schwarze 1652: /* --- handling of request blocks ----------------------------------------- */
1653:
1.67 kristaps 1654: static enum rofferr
1.76 kristaps 1655: roff_cblock(ROFF_ARGS)
1.67 kristaps 1656: {
1657:
1.79 kristaps 1658: /*
1659: * A block-close `..' should only be invoked as a child of an
1660: * ignore macro, otherwise raise a warning and just ignore it.
1661: */
1662:
1.238 schwarze 1663: if (r->last == NULL) {
1.221 schwarze 1664: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1665: ln, ppos, "..");
1.277 schwarze 1666: return ROFF_IGN;
1.76 kristaps 1667: }
1.67 kristaps 1668:
1.81 kristaps 1669: switch (r->last->tok) {
1.207 schwarze 1670: case ROFF_am:
1.220 schwarze 1671: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.207 schwarze 1672: case ROFF_ami:
1673: case ROFF_de:
1.108 schwarze 1674: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.207 schwarze 1675: case ROFF_dei:
1676: case ROFF_ig:
1.81 kristaps 1677: break;
1678: default:
1.221 schwarze 1679: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1680: ln, ppos, "..");
1.277 schwarze 1681: return ROFF_IGN;
1.76 kristaps 1682: }
1.67 kristaps 1683:
1.238 schwarze 1684: if (buf->buf[pos] != '\0')
1.217 schwarze 1685: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1686: ".. %s", buf->buf + pos);
1.71 kristaps 1687:
1688: roffnode_pop(r);
1.76 kristaps 1689: roffnode_cleanscope(r);
1.277 schwarze 1690: return ROFF_IGN;
1.71 kristaps 1691:
1.67 kristaps 1692: }
1693:
1.76 kristaps 1694: static void
1695: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1696: {
1697:
1.76 kristaps 1698: while (r->last) {
1.173 schwarze 1699: if (--r->last->endspan != 0)
1.76 kristaps 1700: break;
1701: roffnode_pop(r);
1702: }
1.67 kristaps 1703: }
1704:
1.195 schwarze 1705: static void
1706: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1707: {
1708:
1.76 kristaps 1709: if (NULL == r->last) {
1.221 schwarze 1710: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1711: ln, ppos, "\\}");
1.195 schwarze 1712: return;
1.76 kristaps 1713: }
1714:
1.82 kristaps 1715: switch (r->last->tok) {
1.207 schwarze 1716: case ROFF_el:
1717: case ROFF_ie:
1718: case ROFF_if:
1.82 kristaps 1719: break;
1720: default:
1.221 schwarze 1721: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1722: ln, ppos, "\\}");
1.195 schwarze 1723: return;
1.75 kristaps 1724: }
1725:
1.76 kristaps 1726: if (r->last->endspan > -1) {
1.221 schwarze 1727: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1728: ln, ppos, "\\}");
1.195 schwarze 1729: return;
1.76 kristaps 1730: }
1731:
1.75 kristaps 1732: roffnode_pop(r);
1.76 kristaps 1733: roffnode_cleanscope(r);
1.195 schwarze 1734: return;
1.76 kristaps 1735: }
1736:
1737: static enum rofferr
1.80 kristaps 1738: roff_block(ROFF_ARGS)
1.76 kristaps 1739: {
1.315 schwarze 1740: const char *name, *value;
1741: char *call, *cp, *iname, *rname;
1742: size_t csz, namesz, rsz;
1743: int deftype;
1.106 kristaps 1744:
1.220 schwarze 1745: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1746:
1.238 schwarze 1747: if (tok == ROFF_de1)
1.220 schwarze 1748: tok = ROFF_de;
1.251 schwarze 1749: else if (tok == ROFF_dei1)
1750: tok = ROFF_dei;
1.238 schwarze 1751: else if (tok == ROFF_am1)
1.220 schwarze 1752: tok = ROFF_am;
1.251 schwarze 1753: else if (tok == ROFF_ami1)
1754: tok = ROFF_ami;
1.220 schwarze 1755:
1756: /* Parse the macro name argument. */
1757:
1.238 schwarze 1758: cp = buf->buf + pos;
1759: if (tok == ROFF_ig) {
1.220 schwarze 1760: iname = NULL;
1761: namesz = 0;
1762: } else {
1763: iname = cp;
1764: namesz = roff_getname(r, &cp, ln, ppos);
1765: iname[namesz] = '\0';
1766: }
1.107 kristaps 1767:
1.220 schwarze 1768: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1769:
1.238 schwarze 1770: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315 schwarze 1771: deftype = ROFFDEF_USER;
1772: name = roff_getstrn(r, iname, namesz, &deftype);
1773: if (name == NULL) {
1.220 schwarze 1774: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1775: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1776: "%.*s", (int)namesz, iname);
1777: namesz = 0;
1778: } else
1779: namesz = strlen(name);
1780: } else
1781: name = iname;
1.107 kristaps 1782:
1.238 schwarze 1783: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1784: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.295 schwarze 1785: ln, ppos, roff_name[tok]);
1.277 schwarze 1786: return ROFF_IGN;
1.220 schwarze 1787: }
1.80 kristaps 1788:
1.106 kristaps 1789: roffnode_push(r, tok, name, ln, ppos);
1790:
1791: /*
1792: * At the beginning of a `de' macro, clear the existing string
1793: * with the same name, if there is one. New content will be
1.193 schwarze 1794: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1795: */
1.107 kristaps 1796:
1.311 schwarze 1797: if (tok == ROFF_de || tok == ROFF_dei) {
1.213 schwarze 1798: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.311 schwarze 1799: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.315 schwarze 1800: } else if (tok == ROFF_am || tok == ROFF_ami) {
1801: deftype = ROFFDEF_ANY;
1802: value = roff_getstrn(r, iname, namesz, &deftype);
1803: switch (deftype) { /* Before appending, ... */
1804: case ROFFDEF_PRE: /* copy predefined to user-defined. */
1805: roff_setstrn(&r->strtab, name, namesz,
1806: value, strlen(value), 0);
1807: break;
1808: case ROFFDEF_REN: /* call original standard macro. */
1809: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1810: (int)strlen(value), value);
1811: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1812: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1813: free(call);
1814: break;
1815: case ROFFDEF_STD: /* rename and call standard macro. */
1816: rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
1817: roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
1818: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1819: (int)rsz, rname);
1820: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1821: free(call);
1822: free(rname);
1823: break;
1824: default:
1825: break;
1826: }
1.311 schwarze 1827: }
1.76 kristaps 1828:
1.238 schwarze 1829: if (*cp == '\0')
1.277 schwarze 1830: return ROFF_IGN;
1.78 kristaps 1831:
1.220 schwarze 1832: /* Get the custom end marker. */
1.107 kristaps 1833:
1.220 schwarze 1834: iname = cp;
1.213 schwarze 1835: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1836:
1837: /* Resolve the end marker if it is indirect. */
1838:
1.238 schwarze 1839: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315 schwarze 1840: deftype = ROFFDEF_USER;
1841: name = roff_getstrn(r, iname, namesz, &deftype);
1842: if (name == NULL) {
1.220 schwarze 1843: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1844: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1845: "%.*s", (int)namesz, iname);
1846: namesz = 0;
1847: } else
1848: namesz = strlen(name);
1849: } else
1850: name = iname;
1851:
1.213 schwarze 1852: if (namesz)
1853: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1854:
1.238 schwarze 1855: if (*cp != '\0')
1.217 schwarze 1856: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.295 schwarze 1857: ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74 kristaps 1858:
1.277 schwarze 1859: return ROFF_IGN;
1.78 kristaps 1860: }
1861:
1862: static enum rofferr
1.80 kristaps 1863: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1864: {
1.294 schwarze 1865: enum roff_tok t;
1.79 kristaps 1866: int i, j;
1867:
1868: /*
1869: * First check whether a custom macro exists at this level. If
1870: * it does, then check against it. This is some of groff's
1871: * stranger behaviours. If we encountered a custom end-scope
1872: * tag and that tag also happens to be a "real" macro, then we
1873: * need to try interpreting it again as a real macro. If it's
1874: * not, then return ignore. Else continue.
1875: */
1876:
1877: if (r->last->end) {
1.130 kristaps 1878: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1879: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1880: break;
1881:
1.238 schwarze 1882: if (r->last->end[j] == '\0' &&
1883: (buf->buf[i] == '\0' ||
1884: buf->buf[i] == ' ' ||
1885: buf->buf[i] == '\t')) {
1.79 kristaps 1886: roffnode_pop(r);
1887: roffnode_cleanscope(r);
1888:
1.238 schwarze 1889: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1890: i++;
1891:
1892: pos = i;
1.238 schwarze 1893: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294 schwarze 1894: TOKEN_NONE)
1.277 schwarze 1895: return ROFF_RERUN;
1896: return ROFF_IGN;
1.79 kristaps 1897: }
1898: }
1899:
1900: /*
1901: * If we have no custom end-query or lookup failed, then try
1902: * pulling it out of the hashtable.
1903: */
1904:
1.238 schwarze 1905: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1906:
1.238 schwarze 1907: if (t != ROFF_cblock) {
1908: if (tok != ROFF_ig)
1909: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.277 schwarze 1910: return ROFF_IGN;
1.106 kristaps 1911: }
1.79 kristaps 1912:
1.277 schwarze 1913: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1914: }
1915:
1916: static enum rofferr
1.80 kristaps 1917: roff_block_text(ROFF_ARGS)
1.78 kristaps 1918: {
1919:
1.238 schwarze 1920: if (tok != ROFF_ig)
1921: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1922:
1.277 schwarze 1923: return ROFF_IGN;
1.78 kristaps 1924: }
1925:
1926: static enum rofferr
1.82 kristaps 1927: roff_cond_sub(ROFF_ARGS)
1928: {
1.294 schwarze 1929: enum roff_tok t;
1.139 kristaps 1930: char *ep;
1.198 schwarze 1931: int rr;
1.82 kristaps 1932:
1933: rr = r->last->rule;
1.139 kristaps 1934: roffnode_cleanscope(r);
1.144 kristaps 1935:
1.196 schwarze 1936: /*
1937: * If `\}' occurs on a macro line without a preceding macro,
1938: * drop the line completely.
1939: */
1940:
1.238 schwarze 1941: ep = buf->buf + pos;
1942: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1943: rr = 0;
1.196 schwarze 1944:
1.177 schwarze 1945: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1946:
1.238 schwarze 1947: while ((ep = strchr(ep, '\\')) != NULL) {
1.318 schwarze 1948: switch (ep[1]) {
1949: case '}':
1950: memmove(ep, ep + 2, strlen(ep + 2) + 1);
1951: roff_ccond(r, ln, ep - buf->buf);
1952: break;
1953: case '\0':
1954: ++ep;
1955: break;
1956: default:
1957: ep += 2;
1958: break;
1.197 schwarze 1959: }
1.177 schwarze 1960: }
1.318 schwarze 1961:
1962: /*
1963: * Fully handle known macros when they are structurally
1964: * required or when the conditional evaluated to true.
1965: */
1966:
1967: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1968: return t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT)
1969: ? (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs) : rr
1970: ? ROFF_CONT : ROFF_IGN;
1.82 kristaps 1971: }
1972:
1973: static enum rofferr
1974: roff_cond_text(ROFF_ARGS)
1.78 kristaps 1975: {
1.140 kristaps 1976: char *ep;
1.198 schwarze 1977: int rr;
1.82 kristaps 1978:
1979: rr = r->last->rule;
1.140 kristaps 1980: roffnode_cleanscope(r);
1.82 kristaps 1981:
1.238 schwarze 1982: ep = buf->buf + pos;
1983: while ((ep = strchr(ep, '\\')) != NULL) {
1984: if (*(++ep) == '}') {
1.197 schwarze 1985: *ep = '&';
1.238 schwarze 1986: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1987: }
1.247 schwarze 1988: if (*ep != '\0')
1989: ++ep;
1.78 kristaps 1990: }
1.277 schwarze 1991: return rr ? ROFF_CONT : ROFF_IGN;
1.74 kristaps 1992: }
1993:
1.266 schwarze 1994: /* --- handling of numeric and conditional expressions -------------------- */
1995:
1.204 schwarze 1996: /*
1997: * Parse a single signed integer number. Stop at the first non-digit.
1998: * If there is at least one digit, return success and advance the
1999: * parse point, else return failure and let the parse point unchanged.
2000: * Ignore overflows, treat them just like the C language.
2001: */
1.184 schwarze 2002: static int
1.261 schwarze 2003: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 2004: {
1.261 schwarze 2005: int myres, scaled, n, p;
1.206 schwarze 2006:
2007: if (NULL == res)
2008: res = &myres;
1.184 schwarze 2009:
2010: p = *pos;
2011: n = v[p] == '-';
1.261 schwarze 2012: if (n || v[p] == '+')
1.184 schwarze 2013: p++;
2014:
1.261 schwarze 2015: if (flags & ROFFNUM_WHITE)
2016: while (isspace((unsigned char)v[p]))
2017: p++;
2018:
1.184 schwarze 2019: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 2020: *res = 10 * *res + v[p] - '0';
1.184 schwarze 2021: if (p == *pos + n)
2022: return 0;
2023:
2024: if (n)
2025: *res = -*res;
2026:
1.254 schwarze 2027: /* Each number may be followed by one optional scaling unit. */
2028:
2029: switch (v[p]) {
2030: case 'f':
1.261 schwarze 2031: scaled = *res * 65536;
1.254 schwarze 2032: break;
2033: case 'i':
1.261 schwarze 2034: scaled = *res * 240;
1.254 schwarze 2035: break;
2036: case 'c':
1.261 schwarze 2037: scaled = *res * 240 / 2.54;
1.254 schwarze 2038: break;
2039: case 'v':
2040: case 'P':
1.261 schwarze 2041: scaled = *res * 40;
1.254 schwarze 2042: break;
2043: case 'm':
2044: case 'n':
1.261 schwarze 2045: scaled = *res * 24;
1.254 schwarze 2046: break;
2047: case 'p':
1.261 schwarze 2048: scaled = *res * 10 / 3;
1.254 schwarze 2049: break;
2050: case 'u':
1.261 schwarze 2051: scaled = *res;
1.254 schwarze 2052: break;
2053: case 'M':
1.261 schwarze 2054: scaled = *res * 6 / 25;
1.254 schwarze 2055: break;
2056: default:
1.261 schwarze 2057: scaled = *res;
1.254 schwarze 2058: p--;
2059: break;
2060: }
1.261 schwarze 2061: if (flags & ROFFNUM_SCALE)
2062: *res = scaled;
1.254 schwarze 2063:
2064: *pos = p + 1;
1.277 schwarze 2065: return 1;
1.184 schwarze 2066: }
2067:
1.198 schwarze 2068: /*
2069: * Evaluate a string comparison condition.
2070: * The first character is the delimiter.
2071: * Succeed if the string up to its second occurrence
2072: * matches the string up to its third occurence.
2073: * Advance the cursor after the third occurrence
2074: * or lacking that, to the end of the line.
2075: */
2076: static int
2077: roff_evalstrcond(const char *v, int *pos)
2078: {
2079: const char *s1, *s2, *s3;
2080: int match;
2081:
2082: match = 0;
2083: s1 = v + *pos; /* initial delimiter */
2084: s2 = s1 + 1; /* for scanning the first string */
2085: s3 = strchr(s2, *s1); /* for scanning the second string */
2086:
2087: if (NULL == s3) /* found no middle delimiter */
2088: goto out;
2089:
2090: while ('\0' != *++s3) {
2091: if (*s2 != *s3) { /* mismatch */
2092: s3 = strchr(s3, *s1);
2093: break;
2094: }
2095: if (*s3 == *s1) { /* found the final delimiter */
2096: match = 1;
2097: break;
2098: }
2099: s2++;
2100: }
2101:
2102: out:
2103: if (NULL == s3)
2104: s3 = strchr(s2, '\0');
1.242 schwarze 2105: else if (*s3 != '\0')
1.198 schwarze 2106: s3++;
2107: *pos = s3 - v;
1.277 schwarze 2108: return match;
1.198 schwarze 2109: }
2110:
1.204 schwarze 2111: /*
2112: * Evaluate an optionally negated single character, numerical,
2113: * or string condition.
2114: */
1.198 schwarze 2115: static int
1.271 schwarze 2116: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88 kristaps 2117: {
1.271 schwarze 2118: char *cp, *name;
2119: size_t sz;
1.315 schwarze 2120: int deftype, number, savepos, istrue, wanttrue;
1.88 kristaps 2121:
1.198 schwarze 2122: if ('!' == v[*pos]) {
2123: wanttrue = 0;
2124: (*pos)++;
2125: } else
2126: wanttrue = 1;
2127:
1.88 kristaps 2128: switch (v[*pos]) {
1.240 schwarze 2129: case '\0':
1.277 schwarze 2130: return 0;
1.207 schwarze 2131: case 'n':
2132: case 'o':
1.88 kristaps 2133: (*pos)++;
1.277 schwarze 2134: return wanttrue;
1.207 schwarze 2135: case 'c':
2136: case 'e':
2137: case 't':
1.239 schwarze 2138: case 'v':
1.88 kristaps 2139: (*pos)++;
1.277 schwarze 2140: return !wanttrue;
1.310 schwarze 2141: case 'd':
1.271 schwarze 2142: case 'r':
1.310 schwarze 2143: cp = v + *pos + 1;
2144: while (*cp == ' ')
2145: cp++;
2146: name = cp;
2147: sz = roff_getname(r, &cp, ln, cp - v);
1.315 schwarze 2148: if (sz == 0)
2149: istrue = 0;
2150: else if (v[*pos] == 'r')
2151: istrue = roff_hasregn(r, name, sz);
2152: else {
2153: deftype = ROFFDEF_ANY;
2154: roff_getstrn(r, name, sz, &deftype);
2155: istrue = !!deftype;
2156: }
1.271 schwarze 2157: *pos = cp - v;
1.310 schwarze 2158: return istrue == wanttrue;
1.88 kristaps 2159: default:
2160: break;
2161: }
2162:
1.241 schwarze 2163: savepos = *pos;
1.261 schwarze 2164: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277 schwarze 2165: return (number > 0) == wanttrue;
1.241 schwarze 2166: else if (*pos == savepos)
1.277 schwarze 2167: return roff_evalstrcond(v, pos) == wanttrue;
1.204 schwarze 2168: else
1.277 schwarze 2169: return 0;
1.88 kristaps 2170: }
2171:
1.74 kristaps 2172: static enum rofferr
1.103 kristaps 2173: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 2174: {
1.123 schwarze 2175:
1.277 schwarze 2176: return ROFF_IGN;
1.89 kristaps 2177: }
2178:
1.104 kristaps 2179: static enum rofferr
1.251 schwarze 2180: roff_insec(ROFF_ARGS)
2181: {
2182:
2183: mandoc_msg(MANDOCERR_REQ_INSEC, 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
2189: roff_unsupp(ROFF_ARGS)
2190: {
2191:
2192: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.295 schwarze 2193: ln, ppos, roff_name[tok]);
1.277 schwarze 2194: return ROFF_IGN;
1.251 schwarze 2195: }
2196:
2197: static enum rofferr
1.82 kristaps 2198: roff_cond(ROFF_ARGS)
1.74 kristaps 2199: {
1.173 schwarze 2200:
2201: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 2202:
1.207 schwarze 2203: /*
1.134 kristaps 2204: * An `.el' has no conditional body: it will consume the value
2205: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 2206: * defaults to DENY.
1.134 kristaps 2207: *
2208: * If we're not an `el', however, then evaluate the conditional.
2209: */
1.133 kristaps 2210:
1.238 schwarze 2211: r->last->rule = tok == ROFF_el ?
1.207 schwarze 2212: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 2213: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 2214:
1.134 kristaps 2215: /*
2216: * An if-else will put the NEGATION of the current evaluated
2217: * conditional into the stack of rules.
2218: */
2219:
1.238 schwarze 2220: if (tok == ROFF_ie) {
1.223 schwarze 2221: if (r->rstackpos + 1 == r->rstacksz) {
2222: r->rstacksz += 16;
2223: r->rstack = mandoc_reallocarray(r->rstack,
2224: r->rstacksz, sizeof(int));
1.134 kristaps 2225: }
1.198 schwarze 2226: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 2227: }
1.88 kristaps 2228:
2229: /* If the parent has false as its rule, then so do we. */
2230:
1.198 schwarze 2231: if (r->last->parent && !r->last->parent->rule)
2232: r->last->rule = 0;
1.88 kristaps 2233:
2234: /*
1.173 schwarze 2235: * Determine scope.
2236: * If there is nothing on the line after the conditional,
2237: * not even whitespace, use next-line scope.
1.88 kristaps 2238: */
1.74 kristaps 2239:
1.238 schwarze 2240: if (buf->buf[pos] == '\0') {
1.173 schwarze 2241: r->last->endspan = 2;
2242: goto out;
2243: }
2244:
1.238 schwarze 2245: while (buf->buf[pos] == ' ')
1.173 schwarze 2246: pos++;
2247:
2248: /* An opening brace requests multiline scope. */
1.75 kristaps 2249:
1.238 schwarze 2250: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 2251: r->last->endspan = -1;
2252: pos += 2;
1.272 schwarze 2253: while (buf->buf[pos] == ' ')
2254: pos++;
1.173 schwarze 2255: goto out;
1.207 schwarze 2256: }
1.74 kristaps 2257:
1.77 kristaps 2258: /*
1.173 schwarze 2259: * Anything else following the conditional causes
2260: * single-line scope. Warn if the scope contains
2261: * nothing but trailing whitespace.
1.77 kristaps 2262: */
2263:
1.238 schwarze 2264: if (buf->buf[pos] == '\0')
1.216 schwarze 2265: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.295 schwarze 2266: ln, ppos, roff_name[tok]);
1.77 kristaps 2267:
1.173 schwarze 2268: r->last->endspan = 1;
1.74 kristaps 2269:
1.173 schwarze 2270: out:
1.75 kristaps 2271: *offs = pos;
1.277 schwarze 2272: return ROFF_RERUN;
1.83 schwarze 2273: }
2274:
2275: static enum rofferr
1.92 schwarze 2276: roff_ds(ROFF_ARGS)
2277: {
1.212 schwarze 2278: char *string;
2279: const char *name;
2280: size_t namesz;
1.96 kristaps 2281:
1.251 schwarze 2282: /* Ignore groff compatibility mode for now. */
2283:
2284: if (tok == ROFF_ds1)
2285: tok = ROFF_ds;
2286: else if (tok == ROFF_as1)
2287: tok = ROFF_as;
2288:
1.96 kristaps 2289: /*
1.212 schwarze 2290: * The first word is the name of the string.
2291: * If it is empty or terminated by an escape sequence,
2292: * abort the `ds' request without defining anything.
1.96 kristaps 2293: */
1.92 schwarze 2294:
1.238 schwarze 2295: name = string = buf->buf + pos;
2296: if (*name == '\0')
1.277 schwarze 2297: return ROFF_IGN;
1.92 schwarze 2298:
1.212 schwarze 2299: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 2300: if (name[namesz] == '\\')
1.277 schwarze 2301: return ROFF_IGN;
1.212 schwarze 2302:
2303: /* Read past the initial double-quote, if any. */
1.238 schwarze 2304: if (*string == '"')
1.92 schwarze 2305: string++;
2306:
1.96 kristaps 2307: /* The rest is the value. */
1.212 schwarze 2308: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2309: ROFF_as == tok);
1.311 schwarze 2310: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.277 schwarze 2311: return ROFF_IGN;
1.92 schwarze 2312: }
2313:
1.204 schwarze 2314: /*
2315: * Parse a single operator, one or two characters long.
2316: * If the operator is recognized, return success and advance the
2317: * parse point, else return failure and let the parse point unchanged.
2318: */
2319: static int
2320: roff_getop(const char *v, int *pos, char *res)
2321: {
2322:
2323: *res = v[*pos];
2324:
2325: switch (*res) {
1.207 schwarze 2326: case '+':
2327: case '-':
2328: case '*':
2329: case '/':
2330: case '%':
2331: case '&':
2332: case ':':
1.204 schwarze 2333: break;
2334: case '<':
2335: switch (v[*pos + 1]) {
1.207 schwarze 2336: case '=':
1.204 schwarze 2337: *res = 'l';
2338: (*pos)++;
2339: break;
1.207 schwarze 2340: case '>':
1.204 schwarze 2341: *res = '!';
2342: (*pos)++;
2343: break;
1.207 schwarze 2344: case '?':
1.204 schwarze 2345: *res = 'i';
2346: (*pos)++;
2347: break;
2348: default:
2349: break;
2350: }
2351: break;
2352: case '>':
2353: switch (v[*pos + 1]) {
1.207 schwarze 2354: case '=':
1.204 schwarze 2355: *res = 'g';
2356: (*pos)++;
2357: break;
1.207 schwarze 2358: case '?':
1.204 schwarze 2359: *res = 'a';
2360: (*pos)++;
2361: break;
2362: default:
2363: break;
2364: }
2365: break;
2366: case '=':
2367: if ('=' == v[*pos + 1])
2368: (*pos)++;
2369: break;
2370: default:
1.277 schwarze 2371: return 0;
1.204 schwarze 2372: }
2373: (*pos)++;
2374:
1.277 schwarze 2375: return *res;
1.204 schwarze 2376: }
2377:
2378: /*
2379: * Evaluate either a parenthesized numeric expression
2380: * or a single signed integer number.
2381: */
2382: static int
1.235 schwarze 2383: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2384: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2385: {
2386:
2387: if ('(' != v[*pos])
1.277 schwarze 2388: return roff_getnum(v, pos, res, flags);
1.204 schwarze 2389:
2390: (*pos)++;
1.261 schwarze 2391: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277 schwarze 2392: return 0;
1.204 schwarze 2393:
1.206 schwarze 2394: /*
2395: * Omission of the closing parenthesis
2396: * is an error in validation mode,
2397: * but ignored in evaluation mode.
2398: */
2399:
1.204 schwarze 2400: if (')' == v[*pos])
2401: (*pos)++;
1.206 schwarze 2402: else if (NULL == res)
1.277 schwarze 2403: return 0;
1.204 schwarze 2404:
1.277 schwarze 2405: return 1;
1.204 schwarze 2406: }
2407:
2408: /*
2409: * Evaluate a complete numeric expression.
2410: * Proceed left to right, there is no concept of precedence.
2411: */
2412: static int
1.235 schwarze 2413: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2414: int *pos, int *res, int flags)
1.204 schwarze 2415: {
2416: int mypos, operand2;
2417: char operator;
2418:
2419: if (NULL == pos) {
2420: mypos = 0;
2421: pos = &mypos;
2422: }
2423:
1.261 schwarze 2424: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2425: while (isspace((unsigned char)v[*pos]))
2426: (*pos)++;
2427:
1.261 schwarze 2428: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277 schwarze 2429: return 0;
1.204 schwarze 2430:
2431: while (1) {
1.261 schwarze 2432: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2433: while (isspace((unsigned char)v[*pos]))
2434: (*pos)++;
2435:
2436: if ( ! roff_getop(v, pos, &operator))
2437: break;
2438:
1.261 schwarze 2439: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2440: while (isspace((unsigned char)v[*pos]))
2441: (*pos)++;
2442:
1.261 schwarze 2443: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277 schwarze 2444: return 0;
1.204 schwarze 2445:
1.261 schwarze 2446: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2447: while (isspace((unsigned char)v[*pos]))
2448: (*pos)++;
1.206 schwarze 2449:
2450: if (NULL == res)
2451: continue;
1.204 schwarze 2452:
2453: switch (operator) {
1.207 schwarze 2454: case '+':
1.204 schwarze 2455: *res += operand2;
2456: break;
1.207 schwarze 2457: case '-':
1.204 schwarze 2458: *res -= operand2;
2459: break;
1.207 schwarze 2460: case '*':
1.204 schwarze 2461: *res *= operand2;
2462: break;
1.207 schwarze 2463: case '/':
1.244 schwarze 2464: if (operand2 == 0) {
1.235 schwarze 2465: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 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 '%':
1.244 schwarze 2473: if (operand2 == 0) {
2474: mandoc_msg(MANDOCERR_DIVZERO,
2475: r->parse, ln, *pos, v);
2476: *res = 0;
2477: break;
2478: }
1.204 schwarze 2479: *res %= operand2;
2480: break;
1.207 schwarze 2481: case '<':
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 'l':
1.204 schwarze 2488: *res = *res <= operand2;
2489: break;
1.207 schwarze 2490: case 'g':
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 '!':
1.204 schwarze 2497: *res = *res != operand2;
2498: break;
1.207 schwarze 2499: case '&':
1.204 schwarze 2500: *res = *res && operand2;
2501: break;
1.207 schwarze 2502: case ':':
1.204 schwarze 2503: *res = *res || operand2;
2504: break;
1.207 schwarze 2505: case 'i':
1.204 schwarze 2506: if (operand2 < *res)
2507: *res = operand2;
2508: break;
1.207 schwarze 2509: case 'a':
1.204 schwarze 2510: if (operand2 > *res)
2511: *res = operand2;
2512: break;
2513: default:
2514: abort();
2515: }
2516: }
1.277 schwarze 2517: return 1;
1.204 schwarze 2518: }
2519:
1.266 schwarze 2520: /* --- register management ------------------------------------------------ */
2521:
1.180 schwarze 2522: void
1.187 schwarze 2523: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2524: {
1.326 ! schwarze 2525: roff_setregn(r, name, strlen(name), val, sign);
! 2526: }
! 2527:
! 2528: static void
! 2529: roff_setregn(struct roff *r, const char *name, size_t len,
! 2530: int val, char sign)
! 2531: {
1.180 schwarze 2532: struct roffreg *reg;
2533:
2534: /* Search for an existing register with the same name. */
2535: reg = r->regtab;
2536:
1.326 ! schwarze 2537: while (reg != NULL && (reg->key.sz != len ||
! 2538: strncmp(reg->key.p, name, len) != 0))
1.180 schwarze 2539: reg = reg->next;
1.147 kristaps 2540:
1.180 schwarze 2541: if (NULL == reg) {
2542: /* Create a new register. */
2543: reg = mandoc_malloc(sizeof(struct roffreg));
1.326 ! schwarze 2544: reg->key.p = mandoc_strndup(name, len);
! 2545: reg->key.sz = len;
1.187 schwarze 2546: reg->val = 0;
1.180 schwarze 2547: reg->next = r->regtab;
2548: r->regtab = reg;
2549: }
2550:
1.187 schwarze 2551: if ('+' == sign)
2552: reg->val += val;
2553: else if ('-' == sign)
2554: reg->val -= val;
2555: else
2556: reg->val = val;
1.147 kristaps 2557: }
2558:
1.192 schwarze 2559: /*
2560: * Handle some predefined read-only number registers.
2561: * For now, return -1 if the requested register is not predefined;
2562: * in case a predefined read-only register having the value -1
2563: * were to turn up, another special value would have to be chosen.
2564: */
2565: static int
1.273 schwarze 2566: roff_getregro(const struct roff *r, const char *name)
1.192 schwarze 2567: {
2568:
2569: switch (*name) {
1.273 schwarze 2570: case '$': /* Number of arguments of the last macro evaluated. */
1.277 schwarze 2571: return r->argc;
1.207 schwarze 2572: case 'A': /* ASCII approximation mode is always off. */
1.277 schwarze 2573: return 0;
1.207 schwarze 2574: case 'g': /* Groff compatibility mode is always on. */
1.277 schwarze 2575: return 1;
1.207 schwarze 2576: case 'H': /* Fixed horizontal resolution. */
1.277 schwarze 2577: return 24;
1.207 schwarze 2578: case 'j': /* Always adjust left margin only. */
1.277 schwarze 2579: return 0;
1.207 schwarze 2580: case 'T': /* Some output device is always defined. */
1.277 schwarze 2581: return 1;
1.207 schwarze 2582: case 'V': /* Fixed vertical resolution. */
1.277 schwarze 2583: return 40;
1.192 schwarze 2584: default:
1.277 schwarze 2585: return -1;
1.192 schwarze 2586: }
2587: }
2588:
1.181 schwarze 2589: int
1.326 ! schwarze 2590: roff_getreg(struct roff *r, const char *name)
1.147 kristaps 2591: {
1.326 ! schwarze 2592: return roff_getregn(r, name, strlen(name));
1.181 schwarze 2593: }
2594:
2595: static int
1.326 ! schwarze 2596: roff_getregn(struct roff *r, const char *name, size_t len)
1.181 schwarze 2597: {
2598: struct roffreg *reg;
1.192 schwarze 2599: int val;
2600:
2601: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2602: val = roff_getregro(r, name + 1);
1.192 schwarze 2603: if (-1 != val)
1.277 schwarze 2604: return val;
1.192 schwarze 2605: }
1.181 schwarze 2606:
2607: for (reg = r->regtab; reg; reg = reg->next)
2608: if (len == reg->key.sz &&
2609: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2610: return reg->val;
1.271 schwarze 2611:
1.326 ! schwarze 2612: roff_setregn(r, name, len, 0, '\0');
1.277 schwarze 2613: return 0;
1.271 schwarze 2614: }
2615:
2616: static int
2617: roff_hasregn(const struct roff *r, const char *name, size_t len)
2618: {
2619: struct roffreg *reg;
2620: int val;
2621:
2622: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2623: val = roff_getregro(r, name + 1);
1.271 schwarze 2624: if (-1 != val)
1.277 schwarze 2625: return 1;
1.271 schwarze 2626: }
2627:
2628: for (reg = r->regtab; reg; reg = reg->next)
2629: if (len == reg->key.sz &&
2630: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2631: return 1;
1.147 kristaps 2632:
1.277 schwarze 2633: return 0;
1.147 kristaps 2634: }
2635:
1.180 schwarze 2636: static void
2637: roff_freereg(struct roffreg *reg)
1.147 kristaps 2638: {
1.180 schwarze 2639: struct roffreg *old_reg;
1.147 kristaps 2640:
1.180 schwarze 2641: while (NULL != reg) {
2642: free(reg->key.p);
2643: old_reg = reg;
2644: reg = reg->next;
2645: free(old_reg);
2646: }
1.147 kristaps 2647: }
1.92 schwarze 2648:
2649: static enum rofferr
1.89 kristaps 2650: roff_nr(ROFF_ARGS)
1.83 schwarze 2651: {
1.212 schwarze 2652: char *key, *val;
2653: size_t keysz;
1.138 kristaps 2654: int iv;
1.187 schwarze 2655: char sign;
1.89 kristaps 2656:
1.238 schwarze 2657: key = val = buf->buf + pos;
2658: if (*key == '\0')
1.277 schwarze 2659: return ROFF_IGN;
1.212 schwarze 2660:
2661: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2662: if (key[keysz] == '\\')
1.277 schwarze 2663: return ROFF_IGN;
1.89 kristaps 2664:
1.187 schwarze 2665: sign = *val;
1.238 schwarze 2666: if (sign == '+' || sign == '-')
1.187 schwarze 2667: val++;
2668:
1.261 schwarze 2669: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.326 ! schwarze 2670: roff_setregn(r, key, keysz, iv, sign);
1.109 kristaps 2671:
1.277 schwarze 2672: return ROFF_IGN;
1.203 schwarze 2673: }
2674:
2675: static enum rofferr
2676: roff_rr(ROFF_ARGS)
2677: {
2678: struct roffreg *reg, **prev;
1.212 schwarze 2679: char *name, *cp;
2680: size_t namesz;
1.203 schwarze 2681:
1.238 schwarze 2682: name = cp = buf->buf + pos;
2683: if (*name == '\0')
1.277 schwarze 2684: return ROFF_IGN;
1.212 schwarze 2685: namesz = roff_getname(r, &cp, ln, pos);
2686: name[namesz] = '\0';
1.203 schwarze 2687:
2688: prev = &r->regtab;
2689: while (1) {
2690: reg = *prev;
1.238 schwarze 2691: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2692: break;
2693: prev = ®->next;
2694: }
1.238 schwarze 2695: if (reg != NULL) {
1.203 schwarze 2696: *prev = reg->next;
2697: free(reg->key.p);
2698: free(reg);
2699: }
1.277 schwarze 2700: return ROFF_IGN;
1.122 schwarze 2701: }
2702:
1.266 schwarze 2703: /* --- handler functions for roff requests -------------------------------- */
2704:
1.122 schwarze 2705: static enum rofferr
2706: roff_rm(ROFF_ARGS)
2707: {
2708: const char *name;
2709: char *cp;
1.212 schwarze 2710: size_t namesz;
1.122 schwarze 2711:
1.238 schwarze 2712: cp = buf->buf + pos;
2713: while (*cp != '\0') {
1.212 schwarze 2714: name = cp;
1.238 schwarze 2715: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2716: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.311 schwarze 2717: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.238 schwarze 2718: if (name[namesz] == '\\')
1.212 schwarze 2719: break;
1.122 schwarze 2720: }
1.277 schwarze 2721: return ROFF_IGN;
1.178 schwarze 2722: }
2723:
2724: static enum rofferr
2725: roff_it(ROFF_ARGS)
2726: {
2727: int iv;
2728:
2729: /* Parse the number of lines. */
1.261 schwarze 2730:
2731: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 2732: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2733: ln, ppos, buf->buf + 1);
1.277 schwarze 2734: return ROFF_IGN;
1.178 schwarze 2735: }
2736:
1.262 schwarze 2737: while (isspace((unsigned char)buf->buf[pos]))
2738: pos++;
2739:
2740: /*
2741: * Arm the input line trap.
2742: * Special-casing "an-trap" is an ugly workaround to cope
2743: * with DocBook stupidly fiddling with man(7) internals.
2744: */
1.261 schwarze 2745:
1.178 schwarze 2746: roffit_lines = iv;
1.262 schwarze 2747: roffit_macro = mandoc_strdup(iv != 1 ||
2748: strcmp(buf->buf + pos, "an-trap") ?
2749: buf->buf + pos : "br");
1.277 schwarze 2750: return ROFF_IGN;
1.175 schwarze 2751: }
2752:
2753: static enum rofferr
2754: roff_Dd(ROFF_ARGS)
2755: {
1.315 schwarze 2756: int mask;
2757: enum roff_tok t, te;
1.227 schwarze 2758:
1.315 schwarze 2759: switch (tok) {
2760: case ROFF_Dd:
2761: tok = MDOC_Dd;
2762: te = MDOC_MAX;
2763: if (r->format == 0)
2764: r->format = MPARSE_MDOC;
2765: mask = MPARSE_MDOC | MPARSE_QUICK;
2766: break;
2767: case ROFF_TH:
2768: tok = MAN_TH;
2769: te = MAN_MAX;
2770: if (r->format == 0)
2771: r->format = MPARSE_MAN;
2772: mask = MPARSE_QUICK;
2773: break;
2774: default:
2775: abort();
2776: }
2777: if ((r->options & mask) == 0)
2778: for (t = tok; t < te; t++)
2779: roff_setstr(r, roff_name[t], NULL, 0);
1.277 schwarze 2780: return ROFF_CONT;
1.109 kristaps 2781: }
2782:
2783: static enum rofferr
2784: roff_TE(ROFF_ARGS)
2785: {
1.321 schwarze 2786: if (r->tbl == NULL) {
1.221 schwarze 2787: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2788: ln, ppos, "TE");
1.321 schwarze 2789: return ROFF_IGN;
2790: }
2791: if (tbl_end(r->tbl) == 0) {
2792: r->tbl = NULL;
1.258 schwarze 2793: free(buf->buf);
2794: buf->buf = mandoc_strdup(".sp");
2795: buf->sz = 4;
1.277 schwarze 2796: return ROFF_REPARSE;
1.258 schwarze 2797: }
1.321 schwarze 2798: r->tbl = NULL;
1.277 schwarze 2799: return ROFF_IGN;
1.112 kristaps 2800: }
2801:
2802: static enum rofferr
2803: roff_T_(ROFF_ARGS)
2804: {
2805:
2806: if (NULL == r->tbl)
1.221 schwarze 2807: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2808: ln, ppos, "T&");
1.112 kristaps 2809: else
1.296 schwarze 2810: tbl_restart(ln, ppos, r->tbl);
1.112 kristaps 2811:
1.277 schwarze 2812: return ROFF_IGN;
1.109 kristaps 2813: }
2814:
1.230 schwarze 2815: /*
2816: * Handle in-line equation delimiters.
2817: */
2818: static enum rofferr
1.238 schwarze 2819: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2820: {
1.233 schwarze 2821: char *cp1, *cp2;
2822: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2823:
1.230 schwarze 2824: /*
2825: * Outside equations, look for an opening delimiter.
2826: * If we are inside an equation, we already know it is
2827: * in-line, or this function wouldn't have been called;
2828: * so look for a closing delimiter.
2829: */
2830:
1.238 schwarze 2831: cp1 = buf->buf + pos;
1.230 schwarze 2832: cp2 = strchr(cp1, r->eqn == NULL ?
2833: r->last_eqn->odelim : r->last_eqn->cdelim);
2834: if (cp2 == NULL)
1.277 schwarze 2835: return ROFF_CONT;
1.230 schwarze 2836:
1.233 schwarze 2837: *cp2++ = '\0';
2838: bef_pr = bef_nl = aft_nl = aft_pr = "";
2839:
2840: /* Handle preceding text, protecting whitespace. */
2841:
1.238 schwarze 2842: if (*buf->buf != '\0') {
1.233 schwarze 2843: if (r->eqn == NULL)
2844: bef_pr = "\\&";
2845: bef_nl = "\n";
2846: }
2847:
2848: /*
2849: * Prepare replacing the delimiter with an equation macro
2850: * and drop leading white space from the equation.
2851: */
1.230 schwarze 2852:
1.233 schwarze 2853: if (r->eqn == NULL) {
2854: while (*cp2 == ' ')
2855: cp2++;
2856: mac = ".EQ";
2857: } else
2858: mac = ".EN";
2859:
2860: /* Handle following text, protecting whitespace. */
2861:
2862: if (*cp2 != '\0') {
2863: aft_nl = "\n";
2864: if (r->eqn != NULL)
2865: aft_pr = "\\&";
2866: }
2867:
2868: /* Do the actual replacement. */
2869:
1.238 schwarze 2870: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2871: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2872: free(buf->buf);
2873: buf->buf = cp1;
1.230 schwarze 2874:
2875: /* Toggle the in-line state of the eqn subsystem. */
2876:
2877: r->eqn_inline = r->eqn == NULL;
1.277 schwarze 2878: return ROFF_REPARSE;
1.151 kristaps 2879: }
2880:
1.235 schwarze 2881: static enum rofferr
2882: roff_EQ(ROFF_ARGS)
1.125 kristaps 2883: {
1.319 schwarze 2884: struct roff_node *n;
2885:
1.322 schwarze 2886: if (r->man->macroset == MACROSET_MAN)
2887: man_breakscope(r->man, ROFF_EQ);
1.319 schwarze 2888: n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
2889: if (ln > r->man->last->line)
2890: n->flags |= NODE_LINE;
2891: n->eqn = mandoc_calloc(1, sizeof(*n->eqn));
2892: n->eqn->expectargs = UINT_MAX;
2893: roff_node_append(r->man, n);
2894: r->man->next = ROFF_NEXT_SIBLING;
1.125 kristaps 2895:
1.238 schwarze 2896: assert(r->eqn == NULL);
1.319 schwarze 2897: if (r->last_eqn == NULL)
2898: r->last_eqn = eqn_alloc(r->parse);
2899: else
2900: eqn_reset(r->last_eqn);
2901: r->eqn = r->last_eqn;
2902: r->eqn->node = n;
1.151 kristaps 2903:
1.238 schwarze 2904: if (buf->buf[pos] != '\0')
1.235 schwarze 2905: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 2906: ".EQ %s", buf->buf + pos);
1.151 kristaps 2907:
1.277 schwarze 2908: return ROFF_IGN;
1.125 kristaps 2909: }
2910:
2911: static enum rofferr
2912: roff_EN(ROFF_ARGS)
2913: {
1.319 schwarze 2914: if (r->eqn != NULL) {
2915: eqn_parse(r->eqn);
2916: r->eqn = NULL;
2917: } else
2918: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
2919: if (buf->buf[pos] != '\0')
2920: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
2921: "EN %s", buf->buf + pos);
1.277 schwarze 2922: return ROFF_IGN;
1.125 kristaps 2923: }
2924:
2925: static enum rofferr
1.109 kristaps 2926: roff_TS(ROFF_ARGS)
2927: {
1.321 schwarze 2928: if (r->tbl != NULL) {
1.221 schwarze 2929: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2930: ln, ppos, "TS breaks TS");
1.321 schwarze 2931: tbl_end(r->tbl);
1.115 kristaps 2932: }
1.321 schwarze 2933: r->tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 2934: if (r->last_tbl)
1.321 schwarze 2935: r->last_tbl->next = r->tbl;
1.113 kristaps 2936: else
1.321 schwarze 2937: r->first_tbl = r->tbl;
2938: r->last_tbl = r->tbl;
1.297 schwarze 2939: return ROFF_IGN;
2940: }
2941:
2942: static enum rofferr
2943: roff_onearg(ROFF_ARGS)
2944: {
2945: struct roff_node *n;
2946: char *cp;
1.305 schwarze 2947: int npos;
1.297 schwarze 2948:
1.302 schwarze 2949: if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.320 schwarze 2950: (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
2951: tok == ROFF_ti))
1.302 schwarze 2952: man_breakscope(r->man, tok);
2953:
1.309 schwarze 2954: if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.305 schwarze 2955: r->man->last = roffce_node;
2956: r->man->next = ROFF_NEXT_SIBLING;
2957: }
2958:
1.297 schwarze 2959: roff_elem_alloc(r->man, ln, ppos, tok);
2960: n = r->man->last;
2961:
2962: cp = buf->buf + pos;
2963: if (*cp != '\0') {
2964: while (*cp != '\0' && *cp != ' ')
2965: cp++;
2966: while (*cp == ' ')
2967: *cp++ = '\0';
2968: if (*cp != '\0')
2969: mandoc_vmsg(MANDOCERR_ARG_EXCESS,
2970: r->parse, ln, cp - buf->buf,
2971: "%s ... %s", roff_name[tok], cp);
2972: roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.300 schwarze 2973: }
2974:
1.309 schwarze 2975: if (tok == ROFF_ce || tok == ROFF_rj) {
2976: if (r->man->last->type == ROFFT_ELEM) {
1.305 schwarze 2977: roff_word_alloc(r->man, ln, pos, "1");
2978: r->man->last->flags |= NODE_NOSRC;
2979: }
2980: npos = 0;
2981: if (roff_evalnum(r, ln, r->man->last->string, &npos,
2982: &roffce_lines, 0) == 0) {
2983: mandoc_vmsg(MANDOCERR_CE_NONUM,
2984: r->parse, ln, pos, "ce %s", buf->buf + pos);
2985: roffce_lines = 1;
2986: }
2987: if (roffce_lines < 1) {
2988: r->man->last = r->man->last->parent;
2989: roffce_node = NULL;
2990: roffce_lines = 0;
2991: } else
2992: roffce_node = r->man->last->parent;
2993: } else {
2994: n->flags |= NODE_VALID | NODE_ENDED;
2995: r->man->last = n;
2996: }
2997: n->flags |= NODE_LINE;
1.300 schwarze 2998: r->man->next = ROFF_NEXT_SIBLING;
2999: return ROFF_IGN;
3000: }
3001:
3002: static enum rofferr
3003: roff_manyarg(ROFF_ARGS)
3004: {
3005: struct roff_node *n;
3006: char *sp, *ep;
3007:
3008: roff_elem_alloc(r->man, ln, ppos, tok);
3009: n = r->man->last;
3010:
3011: for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
3012: while (*ep != '\0' && *ep != ' ')
3013: ep++;
3014: while (*ep == ' ')
3015: *ep++ = '\0';
3016: roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.297 schwarze 3017: }
3018:
3019: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3020: r->man->last = n;
3021: r->man->next = ROFF_NEXT_SIBLING;
1.277 schwarze 3022: return ROFF_IGN;
1.251 schwarze 3023: }
3024:
3025: static enum rofferr
1.311 schwarze 3026: roff_als(ROFF_ARGS)
3027: {
3028: char *oldn, *newn, *end, *value;
3029: size_t oldsz, newsz, valsz;
3030:
3031: newn = oldn = buf->buf + pos;
3032: if (*newn == '\0')
3033: return ROFF_IGN;
3034:
3035: newsz = roff_getname(r, &oldn, ln, pos);
3036: if (newn[newsz] == '\\' || *oldn == '\0')
3037: return ROFF_IGN;
3038:
3039: end = oldn;
3040: oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
3041: if (oldsz == 0)
3042: return ROFF_IGN;
3043:
1.315 schwarze 3044: valsz = mandoc_asprintf(&value, ".%.*s \\$*\\\"\n",
3045: (int)oldsz, oldn);
1.311 schwarze 3046: roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
3047: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3048: free(value);
3049: return ROFF_IGN;
3050: }
3051:
3052: static enum rofferr
1.296 schwarze 3053: roff_br(ROFF_ARGS)
1.251 schwarze 3054: {
1.302 schwarze 3055: if (r->man->flags & (MAN_BLINE | MAN_ELINE))
3056: man_breakscope(r->man, ROFF_br);
1.296 schwarze 3057: roff_elem_alloc(r->man, ln, ppos, ROFF_br);
3058: if (buf->buf[pos] != '\0')
3059: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
3060: "%s %s", roff_name[tok], buf->buf + pos);
3061: r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3062: r->man->next = ROFF_NEXT_SIBLING;
3063: return ROFF_IGN;
1.92 schwarze 3064: }
3065:
1.105 kristaps 3066: static enum rofferr
1.174 kristaps 3067: roff_cc(ROFF_ARGS)
3068: {
3069: const char *p;
3070:
1.238 schwarze 3071: p = buf->buf + pos;
1.174 kristaps 3072:
1.238 schwarze 3073: if (*p == '\0' || (r->control = *p++) == '.')
1.303 schwarze 3074: r->control = '\0';
1.174 kristaps 3075:
1.238 schwarze 3076: if (*p != '\0')
1.260 schwarze 3077: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3078: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 3079:
1.277 schwarze 3080: return ROFF_IGN;
1.174 kristaps 3081: }
3082:
3083: static enum rofferr
1.303 schwarze 3084: roff_ec(ROFF_ARGS)
3085: {
3086: const char *p;
3087:
3088: p = buf->buf + pos;
3089: if (*p == '\0')
3090: r->escape = '\\';
3091: else {
3092: r->escape = *p;
3093: if (*++p != '\0')
3094: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3095: ln, p - buf->buf, "ec ... %s", p);
3096: }
3097: return ROFF_IGN;
3098: }
3099:
3100: static enum rofferr
3101: roff_eo(ROFF_ARGS)
3102: {
3103: r->escape = '\0';
3104: if (buf->buf[pos] != '\0')
3105: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
3106: ln, pos, "eo %s", buf->buf + pos);
3107: return ROFF_IGN;
3108: }
3109:
3110: static enum rofferr
1.164 kristaps 3111: roff_tr(ROFF_ARGS)
3112: {
3113: const char *p, *first, *second;
3114: size_t fsz, ssz;
3115: enum mandoc_esc esc;
3116:
1.238 schwarze 3117: p = buf->buf + pos;
1.164 kristaps 3118:
1.238 schwarze 3119: if (*p == '\0') {
1.260 schwarze 3120: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.277 schwarze 3121: return ROFF_IGN;
1.164 kristaps 3122: }
3123:
1.238 schwarze 3124: while (*p != '\0') {
1.164 kristaps 3125: fsz = ssz = 1;
3126:
3127: first = p++;
1.238 schwarze 3128: if (*first == '\\') {
1.164 kristaps 3129: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3130: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3131: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3132: ln, (int)(p - buf->buf), first);
1.277 schwarze 3133: return ROFF_IGN;
1.164 kristaps 3134: }
3135: fsz = (size_t)(p - first);
3136: }
3137:
3138: second = p++;
1.238 schwarze 3139: if (*second == '\\') {
1.164 kristaps 3140: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3141: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3142: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3143: ln, (int)(p - buf->buf), second);
1.277 schwarze 3144: return ROFF_IGN;
1.164 kristaps 3145: }
3146: ssz = (size_t)(p - second);
1.238 schwarze 3147: } else if (*second == '\0') {
1.260 schwarze 3148: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
3149: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 3150: second = " ";
1.165 kristaps 3151: p--;
1.164 kristaps 3152: }
3153:
1.167 kristaps 3154: if (fsz > 1) {
1.207 schwarze 3155: roff_setstrn(&r->xmbtab, first, fsz,
3156: second, ssz, 0);
1.167 kristaps 3157: continue;
3158: }
3159:
1.238 schwarze 3160: if (r->xtab == NULL)
1.207 schwarze 3161: r->xtab = mandoc_calloc(128,
3162: sizeof(struct roffstr));
1.167 kristaps 3163:
3164: free(r->xtab[(int)*first].p);
3165: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
3166: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 3167: }
3168:
1.277 schwarze 3169: return ROFF_IGN;
1.164 kristaps 3170: }
3171:
3172: static enum rofferr
1.306 schwarze 3173: roff_rn(ROFF_ARGS)
3174: {
3175: const char *value;
3176: char *oldn, *newn, *end;
3177: size_t oldsz, newsz;
1.315 schwarze 3178: int deftype;
1.306 schwarze 3179:
3180: oldn = newn = buf->buf + pos;
3181: if (*oldn == '\0')
3182: return ROFF_IGN;
3183:
3184: oldsz = roff_getname(r, &newn, ln, pos);
3185: if (oldn[oldsz] == '\\' || *newn == '\0')
3186: return ROFF_IGN;
3187:
3188: end = newn;
3189: newsz = roff_getname(r, &end, ln, newn - buf->buf);
3190: if (newsz == 0)
3191: return ROFF_IGN;
3192:
1.315 schwarze 3193: deftype = ROFFDEF_ANY;
3194: value = roff_getstrn(r, oldn, oldsz, &deftype);
3195: switch (deftype) {
3196: case ROFFDEF_USER:
1.306 schwarze 3197: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3198: roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
3199: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.315 schwarze 3200: break;
3201: case ROFFDEF_PRE:
3202: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3203: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3204: break;
3205: case ROFFDEF_REN:
1.306 schwarze 3206: roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
3207: roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.315 schwarze 3208: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3209: break;
3210: case ROFFDEF_STD:
1.306 schwarze 3211: roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.315 schwarze 3212: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3213: break;
3214: default:
3215: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3216: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3217: break;
3218: }
1.306 schwarze 3219: return ROFF_IGN;
3220: }
3221:
3222: static enum rofferr
1.105 kristaps 3223: roff_so(ROFF_ARGS)
3224: {
1.249 schwarze 3225: char *name, *cp;
1.105 kristaps 3226:
1.238 schwarze 3227: name = buf->buf + pos;
1.224 schwarze 3228: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 3229:
3230: /*
3231: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3232: * opening anything that's not in our cwd or anything beneath
3233: * it. Thus, explicitly disallow traversing up the file-system
3234: * or using absolute paths.
3235: */
3236:
1.238 schwarze 3237: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 3238: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3239: ".so %s", name);
1.249 schwarze 3240: buf->sz = mandoc_asprintf(&cp,
3241: ".sp\nSee the file %s.\n.sp", name) + 1;
3242: free(buf->buf);
3243: buf->buf = cp;
3244: *offs = 0;
1.277 schwarze 3245: return ROFF_REPARSE;
1.105 kristaps 3246: }
3247:
3248: *offs = pos;
1.277 schwarze 3249: return ROFF_SO;
1.105 kristaps 3250: }
1.92 schwarze 3251:
1.266 schwarze 3252: /* --- user defined strings and macros ------------------------------------ */
3253:
1.106 kristaps 3254: static enum rofferr
3255: roff_userdef(ROFF_ARGS)
1.99 kristaps 3256: {
1.311 schwarze 3257: const char *arg[16], *ap;
1.106 kristaps 3258: char *cp, *n1, *n2;
1.292 schwarze 3259: int expand_count, i, ib, ie;
1.263 schwarze 3260: size_t asz, rsz;
1.106 kristaps 3261:
3262: /*
3263: * Collect pointers to macro argument strings
1.188 schwarze 3264: * and NUL-terminate them.
1.106 kristaps 3265: */
1.263 schwarze 3266:
1.273 schwarze 3267: r->argc = 0;
1.238 schwarze 3268: cp = buf->buf + pos;
1.311 schwarze 3269: for (i = 0; i < 16; i++) {
1.273 schwarze 3270: if (*cp == '\0')
3271: arg[i] = "";
3272: else {
3273: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
3274: r->argc = i + 1;
3275: }
3276: }
1.99 kristaps 3277:
1.106 kristaps 3278: /*
3279: * Expand macro arguments.
1.99 kristaps 3280: */
1.263 schwarze 3281:
3282: buf->sz = strlen(r->current_string) + 1;
1.292 schwarze 3283: n1 = n2 = cp = mandoc_malloc(buf->sz);
1.263 schwarze 3284: memcpy(n1, r->current_string, buf->sz);
1.292 schwarze 3285: expand_count = 0;
1.263 schwarze 3286: while (*cp != '\0') {
3287:
3288: /* Scan ahead for the next argument invocation. */
3289:
3290: if (*cp++ != '\\')
3291: continue;
3292: if (*cp++ != '$')
3293: continue;
1.274 schwarze 3294: if (*cp == '*') { /* \\$* inserts all arguments */
3295: ib = 0;
3296: ie = r->argc - 1;
3297: } else { /* \\$1 .. \\$9 insert one argument */
3298: ib = ie = *cp - '1';
3299: if (ib < 0 || ib > 8)
3300: continue;
3301: }
1.263 schwarze 3302: cp -= 2;
1.292 schwarze 3303:
3304: /*
3305: * Prevent infinite recursion.
3306: */
3307:
3308: if (cp >= n2)
3309: expand_count = 1;
3310: else if (++expand_count > EXPAND_LIMIT) {
3311: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
3312: ln, (int)(cp - n1), NULL);
1.293 schwarze 3313: free(buf->buf);
3314: buf->buf = n1;
1.292 schwarze 3315: return ROFF_IGN;
3316: }
1.263 schwarze 3317:
3318: /*
3319: * Determine the size of the expanded argument,
3320: * taking escaping of quotes into account.
3321: */
3322:
1.274 schwarze 3323: asz = ie > ib ? ie - ib : 0; /* for blanks */
3324: for (i = ib; i <= ie; i++) {
3325: for (ap = arg[i]; *ap != '\0'; ap++) {
3326: asz++;
3327: if (*ap == '"')
3328: asz += 3;
3329: }
1.263 schwarze 3330: }
3331: if (asz != 3) {
3332:
3333: /*
3334: * Determine the size of the rest of the
3335: * unexpanded macro, including the NUL.
3336: */
3337:
3338: rsz = buf->sz - (cp - n1) - 3;
3339:
3340: /*
3341: * When shrinking, move before
3342: * releasing the storage.
3343: */
3344:
3345: if (asz < 3)
3346: memmove(cp + asz, cp + 3, rsz);
3347:
3348: /*
3349: * Resize the storage for the macro
3350: * and readjust the parse pointer.
3351: */
3352:
3353: buf->sz += asz - 3;
3354: n2 = mandoc_realloc(n1, buf->sz);
3355: cp = n2 + (cp - n1);
3356: n1 = n2;
3357:
3358: /*
3359: * When growing, make room
3360: * for the expanded argument.
3361: */
3362:
3363: if (asz > 3)
3364: memmove(cp + asz, cp + 3, rsz);
3365: }
3366:
3367: /* Copy the expanded argument, escaping quotes. */
3368:
3369: n2 = cp;
1.274 schwarze 3370: for (i = ib; i <= ie; i++) {
3371: for (ap = arg[i]; *ap != '\0'; ap++) {
3372: if (*ap == '"') {
3373: memcpy(n2, "\\(dq", 4);
3374: n2 += 4;
3375: } else
3376: *n2++ = *ap;
3377: }
3378: if (i < ie)
3379: *n2++ = ' ';
1.106 kristaps 3380: }
1.99 kristaps 3381: }
3382:
1.106 kristaps 3383: /*
3384: * Replace the macro invocation
3385: * by the expanded macro.
3386: */
1.263 schwarze 3387:
1.238 schwarze 3388: free(buf->buf);
3389: buf->buf = n1;
1.248 schwarze 3390: *offs = 0;
1.106 kristaps 3391:
1.277 schwarze 3392: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
3393: ROFF_REPARSE : ROFF_APPEND;
1.99 kristaps 3394: }
1.121 schwarze 3395:
1.306 schwarze 3396: /*
3397: * Calling a high-level macro that was renamed with .rn.
3398: * r->current_string has already been set up by roff_parse().
3399: */
3400: static enum rofferr
3401: roff_renamed(ROFF_ARGS)
3402: {
3403: char *nbuf;
3404:
1.315 schwarze 3405: buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
3406: buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.306 schwarze 3407: free(buf->buf);
3408: buf->buf = nbuf;
3409: return ROFF_CONT;
3410: }
3411:
1.212 schwarze 3412: static size_t
1.121 schwarze 3413: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3414: {
3415: char *name, *cp;
1.212 schwarze 3416: size_t namesz;
1.121 schwarze 3417:
3418: name = *cpp;
3419: if ('\0' == *name)
1.277 schwarze 3420: return 0;
1.121 schwarze 3421:
1.212 schwarze 3422: /* Read until end of name and terminate it with NUL. */
3423: for (cp = name; 1; cp++) {
3424: if ('\0' == *cp || ' ' == *cp) {
3425: namesz = cp - name;
3426: break;
3427: }
1.121 schwarze 3428: if ('\\' != *cp)
3429: continue;
1.215 schwarze 3430: namesz = cp - name;
3431: if ('{' == cp[1] || '}' == cp[1])
3432: break;
1.121 schwarze 3433: cp++;
3434: if ('\\' == *cp)
3435: continue;
1.224 schwarze 3436: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3437: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 3438: mandoc_escape((const char **)&cp, NULL, NULL);
3439: break;
1.121 schwarze 3440: }
3441:
3442: /* Read past spaces. */
3443: while (' ' == *cp)
3444: cp++;
3445:
3446: *cpp = cp;
1.277 schwarze 3447: return namesz;
1.121 schwarze 3448: }
3449:
1.106 kristaps 3450: /*
3451: * Store *string into the user-defined string called *name.
3452: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 3453: * append == 0: replace mode
3454: * append == 1: single-line append mode
3455: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 3456: */
1.94 kristaps 3457: static void
1.106 kristaps 3458: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 3459: int append)
1.92 schwarze 3460: {
1.315 schwarze 3461: size_t namesz;
1.164 kristaps 3462:
1.315 schwarze 3463: namesz = strlen(name);
3464: roff_setstrn(&r->strtab, name, namesz, string,
1.207 schwarze 3465: string ? strlen(string) : 0, append);
1.315 schwarze 3466: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.164 kristaps 3467: }
3468:
3469: static void
1.166 kristaps 3470: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 3471: const char *string, size_t stringsz, int append)
1.164 kristaps 3472: {
1.166 kristaps 3473: struct roffkv *n;
1.164 kristaps 3474: char *c;
3475: int i;
3476: size_t oldch, newch;
1.92 schwarze 3477:
1.106 kristaps 3478: /* Search for an existing string with the same name. */
1.164 kristaps 3479: n = *r;
3480:
1.211 schwarze 3481: while (n && (namesz != n->key.sz ||
3482: strncmp(n->key.p, name, namesz)))
1.92 schwarze 3483: n = n->next;
1.94 kristaps 3484:
3485: if (NULL == n) {
1.106 kristaps 3486: /* Create a new string table entry. */
1.166 kristaps 3487: n = mandoc_malloc(sizeof(struct roffkv));
3488: n->key.p = mandoc_strndup(name, namesz);
3489: n->key.sz = namesz;
3490: n->val.p = NULL;
3491: n->val.sz = 0;
1.164 kristaps 3492: n->next = *r;
3493: *r = n;
1.193 schwarze 3494: } else if (0 == append) {
1.166 kristaps 3495: free(n->val.p);
3496: n->val.p = NULL;
3497: n->val.sz = 0;
1.106 kristaps 3498: }
3499:
3500: if (NULL == string)
3501: return;
3502:
3503: /*
3504: * One additional byte for the '\n' in multiline mode,
3505: * and one for the terminating '\0'.
3506: */
1.193 schwarze 3507: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 3508:
1.166 kristaps 3509: if (NULL == n->val.p) {
3510: n->val.p = mandoc_malloc(newch);
3511: *n->val.p = '\0';
1.106 kristaps 3512: oldch = 0;
3513: } else {
1.166 kristaps 3514: oldch = n->val.sz;
3515: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 3516: }
3517:
3518: /* Skip existing content in the destination buffer. */
1.166 kristaps 3519: c = n->val.p + (int)oldch;
1.106 kristaps 3520:
3521: /* Append new content to the destination buffer. */
1.164 kristaps 3522: i = 0;
3523: while (i < (int)stringsz) {
1.106 kristaps 3524: /*
3525: * Rudimentary roff copy mode:
3526: * Handle escaped backslashes.
3527: */
1.164 kristaps 3528: if ('\\' == string[i] && '\\' == string[i + 1])
3529: i++;
3530: *c++ = string[i++];
1.106 kristaps 3531: }
1.94 kristaps 3532:
1.106 kristaps 3533: /* Append terminating bytes. */
1.193 schwarze 3534: if (1 < append)
1.106 kristaps 3535: *c++ = '\n';
1.163 kristaps 3536:
1.106 kristaps 3537: *c = '\0';
1.166 kristaps 3538: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 3539: }
3540:
1.94 kristaps 3541: static const char *
1.325 schwarze 3542: roff_getstrn(struct roff *r, const char *name, size_t len,
1.315 schwarze 3543: int *deftype)
1.92 schwarze 3544: {
1.315 schwarze 3545: const struct roffkv *n;
1.325 schwarze 3546: int found, i;
1.315 schwarze 3547: enum roff_tok tok;
3548:
1.325 schwarze 3549: found = 0;
3550: for (n = r->strtab; n != NULL; n = n->next) {
3551: if (strncmp(name, n->key.p, len) != 0 ||
3552: n->key.p[len] != '\0' || n->val.p == NULL)
3553: continue;
3554: if (*deftype & ROFFDEF_USER) {
3555: *deftype = ROFFDEF_USER;
3556: return n->val.p;
3557: } else {
3558: found = 1;
3559: break;
3560: }
3561: }
3562: for (n = r->rentab; n != NULL; n = n->next) {
3563: if (strncmp(name, n->key.p, len) != 0 ||
3564: n->key.p[len] != '\0' || n->val.p == NULL)
3565: continue;
3566: if (*deftype & ROFFDEF_REN) {
3567: *deftype = ROFFDEF_REN;
3568: return n->val.p;
3569: } else {
3570: found = 1;
3571: break;
1.315 schwarze 3572: }
3573: }
1.325 schwarze 3574: for (i = 0; i < PREDEFS_MAX; i++) {
3575: if (strncmp(name, predefs[i].name, len) != 0 ||
3576: predefs[i].name[len] != '\0')
3577: continue;
3578: if (*deftype & ROFFDEF_PRE) {
3579: *deftype = ROFFDEF_PRE;
3580: return predefs[i].str;
3581: } else {
3582: found = 1;
3583: break;
1.315 schwarze 3584: }
3585: }
1.325 schwarze 3586: if (r->man->macroset != MACROSET_MAN) {
3587: for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
3588: if (strncmp(name, roff_name[tok], len) != 0 ||
3589: roff_name[tok][len] != '\0')
3590: continue;
3591: if (*deftype & ROFFDEF_STD) {
3592: *deftype = ROFFDEF_STD;
3593: return NULL;
3594: } else {
3595: found = 1;
3596: break;
1.315 schwarze 3597: }
3598: }
3599: }
1.325 schwarze 3600: if (r->man->macroset != MACROSET_MDOC) {
3601: for (tok = MAN_TH; tok < MAN_MAX; tok++) {
3602: if (strncmp(name, roff_name[tok], len) != 0 ||
3603: roff_name[tok][len] != '\0')
3604: continue;
3605: if (*deftype & ROFFDEF_STD) {
3606: *deftype = ROFFDEF_STD;
3607: return NULL;
3608: } else {
3609: found = 1;
3610: break;
1.315 schwarze 3611: }
3612: }
1.325 schwarze 3613: }
3614:
3615: if (found == 0 && *deftype != ROFFDEF_ANY) {
3616: if (*deftype & ROFFDEF_REN) {
3617: /*
3618: * This might still be a request,
3619: * so do not treat it as undefined yet.
3620: */
3621: *deftype = ROFFDEF_UNDEF;
3622: return NULL;
1.315 schwarze 3623: }
1.325 schwarze 3624:
3625: /* Using an undefined string defines it to be empty. */
3626:
3627: roff_setstrn(&r->strtab, name, len, "", 0, 0);
3628: roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.315 schwarze 3629: }
1.325 schwarze 3630:
1.315 schwarze 3631: *deftype = 0;
1.277 schwarze 3632: return NULL;
1.92 schwarze 3633: }
3634:
1.94 kristaps 3635: static void
1.167 kristaps 3636: roff_freestr(struct roffkv *r)
1.92 schwarze 3637: {
1.166 kristaps 3638: struct roffkv *n, *nn;
1.92 schwarze 3639:
1.167 kristaps 3640: for (n = r; n; n = nn) {
1.166 kristaps 3641: free(n->key.p);
3642: free(n->val.p);
1.92 schwarze 3643: nn = n->next;
3644: free(n);
3645: }
1.114 kristaps 3646: }
1.266 schwarze 3647:
3648: /* --- accessors and utility functions ------------------------------------ */
1.164 kristaps 3649:
3650: /*
3651: * Duplicate an input string, making the appropriate character
3652: * conversations (as stipulated by `tr') along the way.
3653: * Returns a heap-allocated string with all the replacements made.
3654: */
3655: char *
3656: roff_strdup(const struct roff *r, const char *p)
3657: {
1.166 kristaps 3658: const struct roffkv *cp;
1.164 kristaps 3659: char *res;
3660: const char *pp;
3661: size_t ssz, sz;
3662: enum mandoc_esc esc;
3663:
1.167 kristaps 3664: if (NULL == r->xmbtab && NULL == r->xtab)
1.277 schwarze 3665: return mandoc_strdup(p);
1.164 kristaps 3666: else if ('\0' == *p)
1.277 schwarze 3667: return mandoc_strdup("");
1.164 kristaps 3668:
3669: /*
3670: * Step through each character looking for term matches
3671: * (remember that a `tr' can be invoked with an escape, which is
3672: * a glyph but the escape is multi-character).
3673: * We only do this if the character hash has been initialised
3674: * and the string is >0 length.
3675: */
3676:
3677: res = NULL;
3678: ssz = 0;
3679:
3680: while ('\0' != *p) {
1.289 schwarze 3681: assert((unsigned int)*p < 128);
3682: if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167 kristaps 3683: sz = r->xtab[(int)*p].sz;
3684: res = mandoc_realloc(res, ssz + sz + 1);
3685: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3686: ssz += sz;
3687: p++;
3688: continue;
3689: } else if ('\\' != *p) {
3690: res = mandoc_realloc(res, ssz + 2);
3691: res[ssz++] = *p++;
3692: continue;
3693: }
3694:
1.164 kristaps 3695: /* Search for term matches. */
1.167 kristaps 3696: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 3697: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 3698: break;
3699:
3700: if (NULL != cp) {
3701: /*
3702: * A match has been found.
3703: * Append the match to the array and move
3704: * forward by its keysize.
3705: */
1.207 schwarze 3706: res = mandoc_realloc(res,
3707: ssz + cp->val.sz + 1);
1.166 kristaps 3708: memcpy(res + ssz, cp->val.p, cp->val.sz);
3709: ssz += cp->val.sz;
3710: p += (int)cp->key.sz;
1.164 kristaps 3711: continue;
3712: }
3713:
1.167 kristaps 3714: /*
3715: * Handle escapes carefully: we need to copy
3716: * over just the escape itself, or else we might
3717: * do replacements within the escape itself.
3718: * Make sure to pass along the bogus string.
3719: */
3720: pp = p++;
3721: esc = mandoc_escape(&p, NULL, NULL);
3722: if (ESCAPE_ERROR == esc) {
3723: sz = strlen(pp);
1.164 kristaps 3724: res = mandoc_realloc(res, ssz + sz + 1);
3725: memcpy(res + ssz, pp, sz);
1.167 kristaps 3726: break;
1.164 kristaps 3727: }
1.207 schwarze 3728: /*
3729: * We bail out on bad escapes.
1.167 kristaps 3730: * No need to warn: we already did so when
3731: * roff_res() was called.
3732: */
3733: sz = (int)(p - pp);
3734: res = mandoc_realloc(res, ssz + sz + 1);
3735: memcpy(res + ssz, pp, sz);
3736: ssz += sz;
1.164 kristaps 3737: }
3738:
3739: res[(int)ssz] = '\0';
1.277 schwarze 3740: return res;
1.227 schwarze 3741: }
3742:
3743: int
3744: roff_getformat(const struct roff *r)
3745: {
3746:
1.277 schwarze 3747: return r->format;
1.174 kristaps 3748: }
3749:
3750: /*
1.207 schwarze 3751: * Find out whether a line is a macro line or not.
1.174 kristaps 3752: * If it is, adjust the current position and return one; if it isn't,
3753: * return zero and don't change the current position.
3754: * If the control character has been set with `.cc', then let that grain
3755: * precedence.
3756: * This is slighly contrary to groff, where using the non-breaking
3757: * control character when `cc' has been invoked will cause the
3758: * non-breaking macro contents to be printed verbatim.
3759: */
3760: int
3761: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3762: {
3763: int pos;
3764:
3765: pos = *ppos;
3766:
1.303 schwarze 3767: if (r->control != '\0' && cp[pos] == r->control)
1.174 kristaps 3768: pos++;
1.303 schwarze 3769: else if (r->control != '\0')
1.277 schwarze 3770: return 0;
1.174 kristaps 3771: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3772: pos += 2;
3773: else if ('.' == cp[pos] || '\'' == cp[pos])
3774: pos++;
3775: else
1.277 schwarze 3776: return 0;
1.174 kristaps 3777:
3778: while (' ' == cp[pos] || '\t' == cp[pos])
3779: pos++;
3780:
3781: *ppos = pos;
1.277 schwarze 3782: return 1;
1.74 kristaps 3783: }
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