Annotation of mandoc/roff.c, Revision 1.327
1.327 ! schwarze 1: /* $Id: roff.c,v 1.326 2018/04/09 22:27:04 schwarze Exp $ */
1.1 kristaps 2: /*
1.267 schwarze 3: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.325 schwarze 4: * Copyright (c) 2010-2015, 2017, 2018 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.66 kristaps 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.106 kristaps 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.66 kristaps 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106 kristaps 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.66 kristaps 13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1 kristaps 17: */
1.66 kristaps 18: #include "config.h"
1.225 schwarze 19:
20: #include <sys/types.h>
1.30 kristaps 21:
1.67 kristaps 22: #include <assert.h>
1.85 kristaps 23: #include <ctype.h>
1.245 schwarze 24: #include <limits.h>
1.295 schwarze 25: #include <stddef.h>
26: #include <stdint.h>
1.178 schwarze 27: #include <stdio.h>
1.1 kristaps 28: #include <stdlib.h>
1.67 kristaps 29: #include <string.h>
1.1 kristaps 30:
1.67 kristaps 31: #include "mandoc.h"
1.201 schwarze 32: #include "mandoc_aux.h"
1.295 schwarze 33: #include "mandoc_ohash.h"
1.265 schwarze 34: #include "roff.h"
1.226 schwarze 35: #include "libmandoc.h"
1.266 schwarze 36: #include "roff_int.h"
1.109 kristaps 37: #include "libroff.h"
1.82 kristaps 38:
1.170 schwarze 39: /* Maximum number of string expansions per line, to break infinite loops. */
40: #define EXPAND_LIMIT 1000
41:
1.315 schwarze 42: /* Types of definitions of macros and strings. */
43: #define ROFFDEF_USER (1 << 1) /* User-defined. */
44: #define ROFFDEF_PRE (1 << 2) /* Predefined. */
45: #define ROFFDEF_REN (1 << 3) /* Renamed standard macro. */
46: #define ROFFDEF_STD (1 << 4) /* mdoc(7) or man(7) macro. */
47: #define ROFFDEF_ANY (ROFFDEF_USER | ROFFDEF_PRE | \
48: ROFFDEF_REN | ROFFDEF_STD)
1.325 schwarze 49: #define ROFFDEF_UNDEF (1 << 5) /* Completely undefined. */
1.315 schwarze 50:
1.266 schwarze 51: /* --- data types --------------------------------------------------------- */
52:
1.147 kristaps 53: /*
1.167 kristaps 54: * An incredibly-simple string buffer.
55: */
1.94 kristaps 56: struct roffstr {
1.167 kristaps 57: char *p; /* nil-terminated buffer */
58: size_t sz; /* saved strlen(p) */
1.166 kristaps 59: };
60:
61: /*
1.167 kristaps 62: * A key-value roffstr pair as part of a singly-linked list.
1.166 kristaps 63: */
64: struct roffkv {
65: struct roffstr key;
66: struct roffstr val;
67: struct roffkv *next; /* next in list */
1.94 kristaps 68: };
69:
1.180 schwarze 70: /*
71: * A single number register as part of a singly-linked list.
72: */
73: struct roffreg {
74: struct roffstr key;
1.181 schwarze 75: int val;
1.327 ! schwarze 76: int step;
1.180 schwarze 77: struct roffreg *next;
78: };
79:
1.295 schwarze 80: /*
81: * Association of request and macro names with token IDs.
82: */
83: struct roffreq {
84: enum roff_tok tok;
85: char name[];
86: };
87:
1.67 kristaps 88: struct roff {
1.128 kristaps 89: struct mparse *parse; /* parse point */
1.296 schwarze 90: struct roff_man *man; /* mdoc or man parser */
1.67 kristaps 91: struct roffnode *last; /* leaf of stack */
1.223 schwarze 92: int *rstack; /* stack of inverted `ie' values */
1.295 schwarze 93: struct ohash *reqtab; /* request lookup table */
1.180 schwarze 94: struct roffreg *regtab; /* number registers */
1.166 kristaps 95: struct roffkv *strtab; /* user-defined strings & macros */
1.306 schwarze 96: struct roffkv *rentab; /* renamed strings & macros */
1.167 kristaps 97: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
98: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.106 kristaps 99: const char *current_string; /* value of last called user macro */
1.118 kristaps 100: struct tbl_node *first_tbl; /* first table parsed */
101: struct tbl_node *last_tbl; /* last table parsed */
102: struct tbl_node *tbl; /* current table being parsed */
1.319 schwarze 103: struct eqn_node *last_eqn; /* equation parser */
104: struct eqn_node *eqn; /* active equation parser */
1.230 schwarze 105: int eqn_inline; /* current equation is inline */
1.223 schwarze 106: int options; /* parse options */
107: int rstacksz; /* current size limit of rstack */
108: int rstackpos; /* position in rstack */
1.227 schwarze 109: int format; /* current file in mdoc or man format */
1.273 schwarze 110: int argc; /* number of args of the last macro */
1.223 schwarze 111: char control; /* control character */
1.303 schwarze 112: char escape; /* escape character */
1.79 kristaps 113: };
114:
1.67 kristaps 115: struct roffnode {
1.294 schwarze 116: enum roff_tok tok; /* type of node */
1.67 kristaps 117: struct roffnode *parent; /* up one in stack */
118: int line; /* parse line */
119: int col; /* parse col */
1.106 kristaps 120: char *name; /* node name, e.g. macro name */
1.79 kristaps 121: char *end; /* end-rules: custom token */
122: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 123: int rule; /* current evaluation rule */
1.67 kristaps 124: };
125:
126: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.294 schwarze 127: enum roff_tok tok, /* tok of macro */ \
1.238 schwarze 128: struct buf *buf, /* input buffer */ \
1.67 kristaps 129: int ln, /* parse line */ \
1.75 kristaps 130: int ppos, /* original pos in buffer */ \
131: int pos, /* current pos in buffer */ \
1.74 kristaps 132: int *offs /* reset offset of buffer data */
1.67 kristaps 133:
134: typedef enum rofferr (*roffproc)(ROFF_ARGS);
135:
136: struct roffmac {
1.79 kristaps 137: roffproc proc; /* process new macro */
138: roffproc text; /* process as child text of macro */
139: roffproc sub; /* process as child of macro */
140: int flags;
141: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.67 kristaps 142: };
143:
1.141 kristaps 144: struct predef {
145: const char *name; /* predefined input name */
146: const char *str; /* replacement symbol */
147: };
148:
149: #define PREDEF(__name, __str) \
150: { (__name), (__str) },
151:
1.266 schwarze 152: /* --- function prototypes ------------------------------------------------ */
153:
1.155 kristaps 154: static void roffnode_cleanscope(struct roff *);
155: static void roffnode_pop(struct roff *);
1.294 schwarze 156: static void roffnode_push(struct roff *, enum roff_tok,
1.155 kristaps 157: const char *, int, int);
1.321 schwarze 158: static void roff_addtbl(struct roff_man *, struct tbl_node *);
1.311 schwarze 159: static enum rofferr roff_als(ROFF_ARGS);
1.80 kristaps 160: static enum rofferr roff_block(ROFF_ARGS);
161: static enum rofferr roff_block_text(ROFF_ARGS);
162: static enum rofferr roff_block_sub(ROFF_ARGS);
1.296 schwarze 163: static enum rofferr roff_br(ROFF_ARGS);
1.80 kristaps 164: static enum rofferr roff_cblock(ROFF_ARGS);
1.174 kristaps 165: static enum rofferr roff_cc(ROFF_ARGS);
1.195 schwarze 166: static void roff_ccond(struct roff *, int, int);
1.82 kristaps 167: static enum rofferr roff_cond(ROFF_ARGS);
168: static enum rofferr roff_cond_text(ROFF_ARGS);
169: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.92 schwarze 170: static enum rofferr roff_ds(ROFF_ARGS);
1.303 schwarze 171: static enum rofferr roff_ec(ROFF_ARGS);
172: static enum rofferr roff_eo(ROFF_ARGS);
1.238 schwarze 173: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.271 schwarze 174: static int roff_evalcond(struct roff *r, int, char *, int *);
1.234 kristaps 175: static int roff_evalnum(struct roff *, int,
176: const char *, int *, int *, int);
177: static int roff_evalpar(struct roff *, int,
1.261 schwarze 178: const char *, int *, int *, int);
1.198 schwarze 179: static int roff_evalstrcond(const char *, int *);
1.155 kristaps 180: static void roff_free1(struct roff *);
1.180 schwarze 181: static void roff_freereg(struct roffreg *);
1.167 kristaps 182: static void roff_freestr(struct roffkv *);
1.212 schwarze 183: static size_t roff_getname(struct roff *, char **, int, int);
1.261 schwarze 184: static int roff_getnum(const char *, int *, int *, int);
1.184 schwarze 185: static int roff_getop(const char *, int *, char *);
1.327 ! schwarze 186: static int roff_getregn(struct roff *,
! 187: const char *, size_t, char);
1.273 schwarze 188: static int roff_getregro(const struct roff *,
189: const char *name);
1.325 schwarze 190: static const char *roff_getstrn(struct roff *,
1.315 schwarze 191: const char *, size_t, int *);
1.271 schwarze 192: static int roff_hasregn(const struct roff *,
193: const char *, size_t);
1.251 schwarze 194: static enum rofferr roff_insec(ROFF_ARGS);
1.178 schwarze 195: static enum rofferr roff_it(ROFF_ARGS);
1.103 kristaps 196: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.265 schwarze 197: static void roff_man_alloc1(struct roff_man *);
198: static void roff_man_free1(struct roff_man *);
1.300 schwarze 199: static enum rofferr roff_manyarg(ROFF_ARGS);
1.89 kristaps 200: static enum rofferr roff_nr(ROFF_ARGS);
1.297 schwarze 201: static enum rofferr roff_onearg(ROFF_ARGS);
1.294 schwarze 202: static enum roff_tok roff_parse(struct roff *, char *, int *,
1.214 schwarze 203: int, int);
1.305 schwarze 204: static enum rofferr roff_parsetext(struct roff *, struct buf *,
205: int, int *);
1.306 schwarze 206: static enum rofferr roff_renamed(ROFF_ARGS);
1.238 schwarze 207: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.122 schwarze 208: static enum rofferr roff_rm(ROFF_ARGS);
1.306 schwarze 209: static enum rofferr roff_rn(ROFF_ARGS);
1.203 schwarze 210: static enum rofferr roff_rr(ROFF_ARGS);
1.326 schwarze 211: static void roff_setregn(struct roff *, const char *,
1.327 ! schwarze 212: size_t, int, char, int);
1.94 kristaps 213: static void roff_setstr(struct roff *,
1.106 kristaps 214: const char *, const char *, int);
1.207 schwarze 215: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 216: size_t, const char *, size_t, int);
1.105 kristaps 217: static enum rofferr roff_so(ROFF_ARGS);
1.164 kristaps 218: static enum rofferr roff_tr(ROFF_ARGS);
1.175 schwarze 219: static enum rofferr roff_Dd(ROFF_ARGS);
1.109 kristaps 220: static enum rofferr roff_TE(ROFF_ARGS);
221: static enum rofferr roff_TS(ROFF_ARGS);
1.125 kristaps 222: static enum rofferr roff_EQ(ROFF_ARGS);
223: static enum rofferr roff_EN(ROFF_ARGS);
1.112 kristaps 224: static enum rofferr roff_T_(ROFF_ARGS);
1.251 schwarze 225: static enum rofferr roff_unsupp(ROFF_ARGS);
1.106 kristaps 226: static enum rofferr roff_userdef(ROFF_ARGS);
1.67 kristaps 227:
1.266 schwarze 228: /* --- constant data ------------------------------------------------------ */
229:
1.261 schwarze 230: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
231: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
232:
1.294 schwarze 233: const char *__roff_name[MAN_MAX + 1] = {
1.305 schwarze 234: "br", "ce", "ft", "ll",
1.312 schwarze 235: "mc", "po", "rj", "sp",
236: "ta", "ti", NULL,
1.294 schwarze 237: "ab", "ad", "af", "aln",
238: "als", "am", "am1", "ami",
239: "ami1", "as", "as1", "asciify",
240: "backtrace", "bd", "bleedat", "blm",
241: "box", "boxa", "bp", "BP",
242: "break", "breakchar", "brnl", "brp",
1.305 schwarze 243: "brpnl", "c2", "cc",
1.294 schwarze 244: "cf", "cflags", "ch", "char",
245: "chop", "class", "close", "CL",
246: "color", "composite", "continue", "cp",
247: "cropat", "cs", "cu", "da",
248: "dch", "Dd", "de", "de1",
249: "defcolor", "dei", "dei1", "device",
250: "devicem", "di", "do", "ds",
251: "ds1", "dwh", "dt", "ec",
252: "ecr", "ecs", "el", "em",
253: "EN", "eo", "EP", "EQ",
254: "errprint", "ev", "evc", "ex",
255: "fallback", "fam", "fc", "fchar",
256: "fcolor", "fdeferlig", "feature", "fkern",
257: "fl", "flig", "fp", "fps",
258: "fschar", "fspacewidth", "fspecial", "ftr",
259: "fzoom", "gcolor", "hc", "hcode",
260: "hidechar", "hla", "hlm", "hpf",
261: "hpfa", "hpfcode", "hw", "hy",
262: "hylang", "hylen", "hym", "hypp",
263: "hys", "ie", "if", "ig",
264: "index", "it", "itc", "IX",
265: "kern", "kernafter", "kernbefore", "kernpair",
266: "lc", "lc_ctype", "lds", "length",
267: "letadj", "lf", "lg", "lhang",
268: "linetabs", "lnr", "lnrf", "lpfx",
1.304 schwarze 269: "ls", "lsm", "lt",
1.294 schwarze 270: "mediasize", "minss", "mk", "mso",
271: "na", "ne", "nh", "nhychar",
272: "nm", "nn", "nop", "nr",
273: "nrf", "nroff", "ns", "nx",
274: "open", "opena", "os", "output",
275: "padj", "papersize", "pc", "pev",
276: "pi", "PI", "pl", "pm",
1.312 schwarze 277: "pn", "pnr", "ps",
1.294 schwarze 278: "psbb", "pshape", "pso", "ptr",
279: "pvs", "rchar", "rd", "recursionlimit",
1.309 schwarze 280: "return", "rfschar", "rhang",
1.294 schwarze 281: "rm", "rn", "rnn", "rr",
282: "rs", "rt", "schar", "sentchar",
283: "shc", "shift", "sizes", "so",
284: "spacewidth", "special", "spreadwarn", "ss",
285: "sty", "substring", "sv", "sy",
1.300 schwarze 286: "T&", "tc", "TE",
1.301 schwarze 287: "TH", "tkf", "tl",
1.294 schwarze 288: "tm", "tm1", "tmc", "tr",
289: "track", "transchar", "trf", "trimat",
290: "trin", "trnt", "troff", "TS",
291: "uf", "ul", "unformat", "unwatch",
292: "unwatchn", "vpt", "vs", "warn",
293: "warnscale", "watch", "watchlength", "watchn",
294: "wh", "while", "write", "writec",
295: "writem", "xflag", ".", NULL,
1.306 schwarze 296: NULL, "text",
1.294 schwarze 297: "Dd", "Dt", "Os", "Sh",
298: "Ss", "Pp", "D1", "Dl",
299: "Bd", "Ed", "Bl", "El",
300: "It", "Ad", "An", "Ap",
301: "Ar", "Cd", "Cm", "Dv",
302: "Er", "Ev", "Ex", "Fa",
303: "Fd", "Fl", "Fn", "Ft",
304: "Ic", "In", "Li", "Nd",
305: "Nm", "Op", "Ot", "Pa",
306: "Rv", "St", "Va", "Vt",
307: "Xr", "%A", "%B", "%D",
308: "%I", "%J", "%N", "%O",
309: "%P", "%R", "%T", "%V",
310: "Ac", "Ao", "Aq", "At",
311: "Bc", "Bf", "Bo", "Bq",
312: "Bsx", "Bx", "Db", "Dc",
313: "Do", "Dq", "Ec", "Ef",
314: "Em", "Eo", "Fx", "Ms",
315: "No", "Ns", "Nx", "Ox",
316: "Pc", "Pf", "Po", "Pq",
317: "Qc", "Ql", "Qo", "Qq",
318: "Re", "Rs", "Sc", "So",
319: "Sq", "Sm", "Sx", "Sy",
320: "Tn", "Ux", "Xc", "Xo",
321: "Fo", "Fc", "Oo", "Oc",
322: "Bk", "Ek", "Bt", "Hf",
323: "Fr", "Ud", "Lb", "Lp",
324: "Lk", "Mt", "Brq", "Bro",
325: "Brc", "%C", "Es", "En",
1.299 schwarze 326: "Dx", "%Q", "%U", "Ta",
327: NULL,
1.294 schwarze 328: "TH", "SH", "SS", "TP",
329: "LP", "PP", "P", "IP",
330: "HP", "SM", "SB", "BI",
331: "IB", "BR", "RB", "R",
332: "B", "I", "IR", "RI",
1.299 schwarze 333: "nf", "fi",
1.294 schwarze 334: "RE", "RS", "DT", "UC",
1.297 schwarze 335: "PD", "AT", "in",
1.294 schwarze 336: "OP", "EX", "EE", "UR",
1.317 schwarze 337: "UE", "MT", "ME", NULL
1.294 schwarze 338: };
339: const char *const *roff_name = __roff_name;
340:
341: static struct roffmac roffs[TOKEN_NONE] = {
1.296 schwarze 342: { roff_br, NULL, NULL, 0 }, /* br */
1.305 schwarze 343: { roff_onearg, NULL, NULL, 0 }, /* ce */
1.297 schwarze 344: { roff_onearg, NULL, NULL, 0 }, /* ft */
1.298 schwarze 345: { roff_onearg, NULL, NULL, 0 }, /* ll */
1.304 schwarze 346: { roff_onearg, NULL, NULL, 0 }, /* mc */
1.312 schwarze 347: { roff_onearg, NULL, NULL, 0 }, /* po */
1.309 schwarze 348: { roff_onearg, NULL, NULL, 0 }, /* rj */
1.299 schwarze 349: { roff_onearg, NULL, NULL, 0 }, /* sp */
1.300 schwarze 350: { roff_manyarg, NULL, NULL, 0 }, /* ta */
1.301 schwarze 351: { roff_onearg, NULL, NULL, 0 }, /* ti */
1.296 schwarze 352: { NULL, NULL, NULL, 0 }, /* ROFF_MAX */
1.295 schwarze 353: { roff_unsupp, NULL, NULL, 0 }, /* ab */
354: { roff_line_ignore, NULL, NULL, 0 }, /* ad */
355: { roff_line_ignore, NULL, NULL, 0 }, /* af */
356: { roff_unsupp, NULL, NULL, 0 }, /* aln */
1.311 schwarze 357: { roff_als, NULL, NULL, 0 }, /* als */
1.295 schwarze 358: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am */
359: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am1 */
360: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami */
361: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami1 */
362: { roff_ds, NULL, NULL, 0 }, /* as */
363: { roff_ds, NULL, NULL, 0 }, /* as1 */
364: { roff_unsupp, NULL, NULL, 0 }, /* asciify */
365: { roff_line_ignore, NULL, NULL, 0 }, /* backtrace */
366: { roff_line_ignore, NULL, NULL, 0 }, /* bd */
367: { roff_line_ignore, NULL, NULL, 0 }, /* bleedat */
368: { roff_unsupp, NULL, NULL, 0 }, /* blm */
369: { roff_unsupp, NULL, NULL, 0 }, /* box */
370: { roff_unsupp, NULL, NULL, 0 }, /* boxa */
371: { roff_line_ignore, NULL, NULL, 0 }, /* bp */
372: { roff_unsupp, NULL, NULL, 0 }, /* BP */
373: { roff_unsupp, NULL, NULL, 0 }, /* break */
374: { roff_line_ignore, NULL, NULL, 0 }, /* breakchar */
375: { roff_line_ignore, NULL, NULL, 0 }, /* brnl */
1.296 schwarze 376: { roff_br, NULL, NULL, 0 }, /* brp */
1.295 schwarze 377: { roff_line_ignore, NULL, NULL, 0 }, /* brpnl */
378: { roff_unsupp, NULL, NULL, 0 }, /* c2 */
379: { roff_cc, NULL, NULL, 0 }, /* cc */
380: { roff_insec, NULL, NULL, 0 }, /* cf */
381: { roff_line_ignore, NULL, NULL, 0 }, /* cflags */
382: { roff_line_ignore, NULL, NULL, 0 }, /* ch */
383: { roff_unsupp, NULL, NULL, 0 }, /* char */
384: { roff_unsupp, NULL, NULL, 0 }, /* chop */
385: { roff_line_ignore, NULL, NULL, 0 }, /* class */
386: { roff_insec, NULL, NULL, 0 }, /* close */
387: { roff_unsupp, NULL, NULL, 0 }, /* CL */
388: { roff_line_ignore, NULL, NULL, 0 }, /* color */
389: { roff_unsupp, NULL, NULL, 0 }, /* composite */
390: { roff_unsupp, NULL, NULL, 0 }, /* continue */
391: { roff_line_ignore, NULL, NULL, 0 }, /* cp */
392: { roff_line_ignore, NULL, NULL, 0 }, /* cropat */
393: { roff_line_ignore, NULL, NULL, 0 }, /* cs */
394: { roff_line_ignore, NULL, NULL, 0 }, /* cu */
395: { roff_unsupp, NULL, NULL, 0 }, /* da */
396: { roff_unsupp, NULL, NULL, 0 }, /* dch */
397: { roff_Dd, NULL, NULL, 0 }, /* Dd */
398: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de */
399: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de1 */
400: { roff_line_ignore, NULL, NULL, 0 }, /* defcolor */
401: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei */
402: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei1 */
403: { roff_unsupp, NULL, NULL, 0 }, /* device */
404: { roff_unsupp, NULL, NULL, 0 }, /* devicem */
405: { roff_unsupp, NULL, NULL, 0 }, /* di */
406: { roff_unsupp, NULL, NULL, 0 }, /* do */
407: { roff_ds, NULL, NULL, 0 }, /* ds */
408: { roff_ds, NULL, NULL, 0 }, /* ds1 */
409: { roff_unsupp, NULL, NULL, 0 }, /* dwh */
410: { roff_unsupp, NULL, NULL, 0 }, /* dt */
1.303 schwarze 411: { roff_ec, NULL, NULL, 0 }, /* ec */
1.295 schwarze 412: { roff_unsupp, NULL, NULL, 0 }, /* ecr */
413: { roff_unsupp, NULL, NULL, 0 }, /* ecs */
414: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* el */
415: { roff_unsupp, NULL, NULL, 0 }, /* em */
416: { roff_EN, NULL, NULL, 0 }, /* EN */
1.303 schwarze 417: { roff_eo, NULL, NULL, 0 }, /* eo */
1.295 schwarze 418: { roff_unsupp, NULL, NULL, 0 }, /* EP */
419: { roff_EQ, NULL, NULL, 0 }, /* EQ */
420: { roff_line_ignore, NULL, NULL, 0 }, /* errprint */
421: { roff_unsupp, NULL, NULL, 0 }, /* ev */
422: { roff_unsupp, NULL, NULL, 0 }, /* evc */
423: { roff_unsupp, NULL, NULL, 0 }, /* ex */
424: { roff_line_ignore, NULL, NULL, 0 }, /* fallback */
425: { roff_line_ignore, NULL, NULL, 0 }, /* fam */
426: { roff_unsupp, NULL, NULL, 0 }, /* fc */
427: { roff_unsupp, NULL, NULL, 0 }, /* fchar */
428: { roff_line_ignore, NULL, NULL, 0 }, /* fcolor */
429: { roff_line_ignore, NULL, NULL, 0 }, /* fdeferlig */
430: { roff_line_ignore, NULL, NULL, 0 }, /* feature */
431: { roff_line_ignore, NULL, NULL, 0 }, /* fkern */
432: { roff_line_ignore, NULL, NULL, 0 }, /* fl */
433: { roff_line_ignore, NULL, NULL, 0 }, /* flig */
434: { roff_line_ignore, NULL, NULL, 0 }, /* fp */
435: { roff_line_ignore, NULL, NULL, 0 }, /* fps */
436: { roff_unsupp, NULL, NULL, 0 }, /* fschar */
437: { roff_line_ignore, NULL, NULL, 0 }, /* fspacewidth */
438: { roff_line_ignore, NULL, NULL, 0 }, /* fspecial */
439: { roff_line_ignore, NULL, NULL, 0 }, /* ftr */
440: { roff_line_ignore, NULL, NULL, 0 }, /* fzoom */
441: { roff_line_ignore, NULL, NULL, 0 }, /* gcolor */
442: { roff_line_ignore, NULL, NULL, 0 }, /* hc */
443: { roff_line_ignore, NULL, NULL, 0 }, /* hcode */
444: { roff_line_ignore, NULL, NULL, 0 }, /* hidechar */
445: { roff_line_ignore, NULL, NULL, 0 }, /* hla */
446: { roff_line_ignore, NULL, NULL, 0 }, /* hlm */
447: { roff_line_ignore, NULL, NULL, 0 }, /* hpf */
448: { roff_line_ignore, NULL, NULL, 0 }, /* hpfa */
449: { roff_line_ignore, NULL, NULL, 0 }, /* hpfcode */
450: { roff_line_ignore, NULL, NULL, 0 }, /* hw */
451: { roff_line_ignore, NULL, NULL, 0 }, /* hy */
452: { roff_line_ignore, NULL, NULL, 0 }, /* hylang */
453: { roff_line_ignore, NULL, NULL, 0 }, /* hylen */
454: { roff_line_ignore, NULL, NULL, 0 }, /* hym */
455: { roff_line_ignore, NULL, NULL, 0 }, /* hypp */
456: { roff_line_ignore, NULL, NULL, 0 }, /* hys */
457: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* ie */
458: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* if */
459: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ig */
460: { roff_unsupp, NULL, NULL, 0 }, /* index */
461: { roff_it, NULL, NULL, 0 }, /* it */
462: { roff_unsupp, NULL, NULL, 0 }, /* itc */
463: { roff_line_ignore, NULL, NULL, 0 }, /* IX */
464: { roff_line_ignore, NULL, NULL, 0 }, /* kern */
465: { roff_line_ignore, NULL, NULL, 0 }, /* kernafter */
466: { roff_line_ignore, NULL, NULL, 0 }, /* kernbefore */
467: { roff_line_ignore, NULL, NULL, 0 }, /* kernpair */
468: { roff_unsupp, NULL, NULL, 0 }, /* lc */
469: { roff_unsupp, NULL, NULL, 0 }, /* lc_ctype */
470: { roff_unsupp, NULL, NULL, 0 }, /* lds */
471: { roff_unsupp, NULL, NULL, 0 }, /* length */
472: { roff_line_ignore, NULL, NULL, 0 }, /* letadj */
473: { roff_insec, NULL, NULL, 0 }, /* lf */
474: { roff_line_ignore, NULL, NULL, 0 }, /* lg */
475: { roff_line_ignore, NULL, NULL, 0 }, /* lhang */
476: { roff_unsupp, NULL, NULL, 0 }, /* linetabs */
477: { roff_unsupp, NULL, NULL, 0 }, /* lnr */
478: { roff_unsupp, NULL, NULL, 0 }, /* lnrf */
479: { roff_unsupp, NULL, NULL, 0 }, /* lpfx */
480: { roff_line_ignore, NULL, NULL, 0 }, /* ls */
481: { roff_unsupp, NULL, NULL, 0 }, /* lsm */
482: { roff_line_ignore, NULL, NULL, 0 }, /* lt */
483: { roff_line_ignore, NULL, NULL, 0 }, /* mediasize */
484: { roff_line_ignore, NULL, NULL, 0 }, /* minss */
485: { roff_line_ignore, NULL, NULL, 0 }, /* mk */
486: { roff_insec, NULL, NULL, 0 }, /* mso */
487: { roff_line_ignore, NULL, NULL, 0 }, /* na */
488: { roff_line_ignore, NULL, NULL, 0 }, /* ne */
489: { roff_line_ignore, NULL, NULL, 0 }, /* nh */
490: { roff_line_ignore, NULL, NULL, 0 }, /* nhychar */
491: { roff_unsupp, NULL, NULL, 0 }, /* nm */
492: { roff_unsupp, NULL, NULL, 0 }, /* nn */
493: { roff_unsupp, NULL, NULL, 0 }, /* nop */
494: { roff_nr, NULL, NULL, 0 }, /* nr */
495: { roff_unsupp, NULL, NULL, 0 }, /* nrf */
496: { roff_line_ignore, NULL, NULL, 0 }, /* nroff */
497: { roff_line_ignore, NULL, NULL, 0 }, /* ns */
498: { roff_insec, NULL, NULL, 0 }, /* nx */
499: { roff_insec, NULL, NULL, 0 }, /* open */
500: { roff_insec, NULL, NULL, 0 }, /* opena */
501: { roff_line_ignore, NULL, NULL, 0 }, /* os */
502: { roff_unsupp, NULL, NULL, 0 }, /* output */
503: { roff_line_ignore, NULL, NULL, 0 }, /* padj */
504: { roff_line_ignore, NULL, NULL, 0 }, /* papersize */
505: { roff_line_ignore, NULL, NULL, 0 }, /* pc */
506: { roff_line_ignore, NULL, NULL, 0 }, /* pev */
507: { roff_insec, NULL, NULL, 0 }, /* pi */
508: { roff_unsupp, NULL, NULL, 0 }, /* PI */
509: { roff_line_ignore, NULL, NULL, 0 }, /* pl */
510: { roff_line_ignore, NULL, NULL, 0 }, /* pm */
511: { roff_line_ignore, NULL, NULL, 0 }, /* pn */
512: { roff_line_ignore, NULL, NULL, 0 }, /* pnr */
513: { roff_line_ignore, NULL, NULL, 0 }, /* ps */
514: { roff_unsupp, NULL, NULL, 0 }, /* psbb */
515: { roff_unsupp, NULL, NULL, 0 }, /* pshape */
516: { roff_insec, NULL, NULL, 0 }, /* pso */
517: { roff_line_ignore, NULL, NULL, 0 }, /* ptr */
518: { roff_line_ignore, NULL, NULL, 0 }, /* pvs */
519: { roff_unsupp, NULL, NULL, 0 }, /* rchar */
520: { roff_line_ignore, NULL, NULL, 0 }, /* rd */
521: { roff_line_ignore, NULL, NULL, 0 }, /* recursionlimit */
522: { roff_unsupp, NULL, NULL, 0 }, /* return */
523: { roff_unsupp, NULL, NULL, 0 }, /* rfschar */
524: { roff_line_ignore, NULL, NULL, 0 }, /* rhang */
525: { roff_rm, NULL, NULL, 0 }, /* rm */
1.306 schwarze 526: { roff_rn, NULL, NULL, 0 }, /* rn */
1.295 schwarze 527: { roff_unsupp, NULL, NULL, 0 }, /* rnn */
528: { roff_rr, NULL, NULL, 0 }, /* rr */
529: { roff_line_ignore, NULL, NULL, 0 }, /* rs */
530: { roff_line_ignore, NULL, NULL, 0 }, /* rt */
531: { roff_unsupp, NULL, NULL, 0 }, /* schar */
532: { roff_line_ignore, NULL, NULL, 0 }, /* sentchar */
533: { roff_line_ignore, NULL, NULL, 0 }, /* shc */
534: { roff_unsupp, NULL, NULL, 0 }, /* shift */
535: { roff_line_ignore, NULL, NULL, 0 }, /* sizes */
536: { roff_so, NULL, NULL, 0 }, /* so */
537: { roff_line_ignore, NULL, NULL, 0 }, /* spacewidth */
538: { roff_line_ignore, NULL, NULL, 0 }, /* special */
539: { roff_line_ignore, NULL, NULL, 0 }, /* spreadwarn */
540: { roff_line_ignore, NULL, NULL, 0 }, /* ss */
541: { roff_line_ignore, NULL, NULL, 0 }, /* sty */
542: { roff_unsupp, NULL, NULL, 0 }, /* substring */
543: { roff_line_ignore, NULL, NULL, 0 }, /* sv */
544: { roff_insec, NULL, NULL, 0 }, /* sy */
545: { roff_T_, NULL, NULL, 0 }, /* T& */
546: { roff_unsupp, NULL, NULL, 0 }, /* tc */
547: { roff_TE, NULL, NULL, 0 }, /* TE */
1.315 schwarze 548: { roff_Dd, NULL, NULL, 0 }, /* TH */
1.295 schwarze 549: { roff_line_ignore, NULL, NULL, 0 }, /* tkf */
550: { roff_unsupp, NULL, NULL, 0 }, /* tl */
551: { roff_line_ignore, NULL, NULL, 0 }, /* tm */
552: { roff_line_ignore, NULL, NULL, 0 }, /* tm1 */
553: { roff_line_ignore, NULL, NULL, 0 }, /* tmc */
554: { roff_tr, NULL, NULL, 0 }, /* tr */
555: { roff_line_ignore, NULL, NULL, 0 }, /* track */
556: { roff_line_ignore, NULL, NULL, 0 }, /* transchar */
557: { roff_insec, NULL, NULL, 0 }, /* trf */
558: { roff_line_ignore, NULL, NULL, 0 }, /* trimat */
559: { roff_unsupp, NULL, NULL, 0 }, /* trin */
560: { roff_unsupp, NULL, NULL, 0 }, /* trnt */
561: { roff_line_ignore, NULL, NULL, 0 }, /* troff */
562: { roff_TS, NULL, NULL, 0 }, /* TS */
563: { roff_line_ignore, NULL, NULL, 0 }, /* uf */
564: { roff_line_ignore, NULL, NULL, 0 }, /* ul */
565: { roff_unsupp, NULL, NULL, 0 }, /* unformat */
566: { roff_line_ignore, NULL, NULL, 0 }, /* unwatch */
567: { roff_line_ignore, NULL, NULL, 0 }, /* unwatchn */
568: { roff_line_ignore, NULL, NULL, 0 }, /* vpt */
569: { roff_line_ignore, NULL, NULL, 0 }, /* vs */
570: { roff_line_ignore, NULL, NULL, 0 }, /* warn */
571: { roff_line_ignore, NULL, NULL, 0 }, /* warnscale */
572: { roff_line_ignore, NULL, NULL, 0 }, /* watch */
573: { roff_line_ignore, NULL, NULL, 0 }, /* watchlength */
574: { roff_line_ignore, NULL, NULL, 0 }, /* watchn */
575: { roff_unsupp, NULL, NULL, 0 }, /* wh */
576: { roff_unsupp, NULL, NULL, 0 }, /* while */
577: { roff_insec, NULL, NULL, 0 }, /* write */
578: { roff_insec, NULL, NULL, 0 }, /* writec */
579: { roff_insec, NULL, NULL, 0 }, /* writem */
580: { roff_line_ignore, NULL, NULL, 0 }, /* xflag */
581: { roff_cblock, NULL, NULL, 0 }, /* . */
1.306 schwarze 582: { roff_renamed, NULL, NULL, 0 },
1.295 schwarze 583: { roff_userdef, NULL, NULL, 0 }
1.67 kristaps 584: };
585:
1.141 kristaps 586: /* Array of injected predefined strings. */
587: #define PREDEFS_MAX 38
588: static const struct predef predefs[PREDEFS_MAX] = {
589: #include "predefs.in"
590: };
591:
1.305 schwarze 592: static int roffce_lines; /* number of input lines to center */
593: static struct roff_node *roffce_node; /* active request */
1.178 schwarze 594: static int roffit_lines; /* number of lines to delay */
595: static char *roffit_macro; /* nil-terminated macro line */
596:
1.207 schwarze 597:
1.266 schwarze 598: /* --- request table ------------------------------------------------------ */
599:
1.295 schwarze 600: struct ohash *
601: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85 kristaps 602: {
1.295 schwarze 603: struct ohash *htab;
604: struct roffreq *req;
605: enum roff_tok tok;
606: size_t sz;
607: unsigned int slot;
1.85 kristaps 608:
1.295 schwarze 609: htab = mandoc_malloc(sizeof(*htab));
610: mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85 kristaps 611:
1.295 schwarze 612: for (tok = mintok; tok < maxtok; tok++) {
1.296 schwarze 613: if (roff_name[tok] == NULL)
614: continue;
1.295 schwarze 615: sz = strlen(roff_name[tok]);
616: req = mandoc_malloc(sizeof(*req) + sz + 1);
617: req->tok = tok;
618: memcpy(req->name, roff_name[tok], sz + 1);
619: slot = ohash_qlookup(htab, req->name);
620: ohash_insert(htab, slot, req);
1.85 kristaps 621: }
1.295 schwarze 622: return htab;
1.85 kristaps 623: }
624:
1.295 schwarze 625: void
626: roffhash_free(struct ohash *htab)
1.67 kristaps 627: {
1.295 schwarze 628: struct roffreq *req;
629: unsigned int slot;
1.67 kristaps 630:
1.295 schwarze 631: if (htab == NULL)
632: return;
633: for (req = ohash_first(htab, &slot); req != NULL;
634: req = ohash_next(htab, &slot))
635: free(req);
636: ohash_delete(htab);
637: free(htab);
638: }
639:
640: enum roff_tok
641: roffhash_find(struct ohash *htab, const char *name, size_t sz)
642: {
643: struct roffreq *req;
644: const char *end;
645:
646: if (sz) {
647: end = name + sz;
648: req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
649: } else
650: req = ohash_find(htab, ohash_qlookup(htab, name));
651: return req == NULL ? TOKEN_NONE : req->tok;
1.67 kristaps 652: }
653:
1.266 schwarze 654: /* --- stack of request blocks -------------------------------------------- */
655:
1.67 kristaps 656: /*
657: * Pop the current node off of the stack of roff instructions currently
658: * pending.
659: */
660: static void
661: roffnode_pop(struct roff *r)
662: {
663: struct roffnode *p;
664:
1.75 kristaps 665: assert(r->last);
1.207 schwarze 666: p = r->last;
1.82 kristaps 667:
1.75 kristaps 668: r->last = r->last->parent;
1.106 kristaps 669: free(p->name);
670: free(p->end);
1.67 kristaps 671: free(p);
672: }
673:
674: /*
675: * Push a roff node onto the instruction stack. This must later be
676: * removed with roffnode_pop().
677: */
1.98 schwarze 678: static void
1.294 schwarze 679: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106 kristaps 680: int line, int col)
1.67 kristaps 681: {
682: struct roffnode *p;
683:
1.98 schwarze 684: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 685: p->tok = tok;
1.106 kristaps 686: if (name)
687: p->name = mandoc_strdup(name);
1.67 kristaps 688: p->parent = r->last;
689: p->line = line;
690: p->col = col;
1.198 schwarze 691: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 692:
693: r->last = p;
694: }
695:
1.266 schwarze 696: /* --- roff parser state data management ---------------------------------- */
697:
1.67 kristaps 698: static void
699: roff_free1(struct roff *r)
700: {
1.176 schwarze 701: struct tbl_node *tbl;
1.167 kristaps 702: int i;
1.67 kristaps 703:
1.176 schwarze 704: while (NULL != (tbl = r->first_tbl)) {
705: r->first_tbl = tbl->next;
706: tbl_free(tbl);
1.109 kristaps 707: }
1.113 kristaps 708: r->first_tbl = r->last_tbl = r->tbl = NULL;
709:
1.319 schwarze 710: if (r->last_eqn != NULL)
711: eqn_free(r->last_eqn);
712: r->last_eqn = r->eqn = NULL;
1.125 kristaps 713:
1.67 kristaps 714: while (r->last)
715: roffnode_pop(r);
1.109 kristaps 716:
1.223 schwarze 717: free (r->rstack);
718: r->rstack = NULL;
719: r->rstacksz = 0;
720: r->rstackpos = -1;
721:
722: roff_freereg(r->regtab);
723: r->regtab = NULL;
724:
1.167 kristaps 725: roff_freestr(r->strtab);
1.306 schwarze 726: roff_freestr(r->rentab);
1.167 kristaps 727: roff_freestr(r->xmbtab);
1.306 schwarze 728: r->strtab = r->rentab = r->xmbtab = NULL;
1.167 kristaps 729:
730: if (r->xtab)
731: for (i = 0; i < 128; i++)
732: free(r->xtab[i].p);
733: free(r->xtab);
734: r->xtab = NULL;
1.67 kristaps 735: }
736:
737: void
738: roff_reset(struct roff *r)
739: {
740: roff_free1(r);
1.227 schwarze 741: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.303 schwarze 742: r->control = '\0';
743: r->escape = '\\';
1.307 schwarze 744: roffce_lines = 0;
745: roffce_node = NULL;
746: roffit_lines = 0;
747: roffit_macro = NULL;
1.67 kristaps 748: }
749:
750: void
751: roff_free(struct roff *r)
752: {
753: roff_free1(r);
1.295 schwarze 754: roffhash_free(r->reqtab);
1.67 kristaps 755: free(r);
756: }
757:
758: struct roff *
1.279 schwarze 759: roff_alloc(struct mparse *parse, int options)
1.67 kristaps 760: {
761: struct roff *r;
762:
1.98 schwarze 763: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 764: r->parse = parse;
1.295 schwarze 765: r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.199 schwarze 766: r->options = options;
1.227 schwarze 767: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82 kristaps 768: r->rstackpos = -1;
1.303 schwarze 769: r->escape = '\\';
1.277 schwarze 770: return r;
1.265 schwarze 771: }
772:
1.266 schwarze 773: /* --- syntax tree state data management ---------------------------------- */
774:
1.265 schwarze 775: static void
776: roff_man_free1(struct roff_man *man)
777: {
778:
1.266 schwarze 779: if (man->first != NULL)
780: roff_node_delete(man, man->first);
1.265 schwarze 781: free(man->meta.msec);
782: free(man->meta.vol);
783: free(man->meta.os);
784: free(man->meta.arch);
785: free(man->meta.title);
786: free(man->meta.name);
787: free(man->meta.date);
788: }
789:
790: static void
791: roff_man_alloc1(struct roff_man *man)
792: {
793:
794: memset(&man->meta, 0, sizeof(man->meta));
795: man->first = mandoc_calloc(1, sizeof(*man->first));
796: man->first->type = ROFFT_ROOT;
797: man->last = man->first;
798: man->last_es = NULL;
799: man->flags = 0;
800: man->macroset = MACROSET_NONE;
801: man->lastsec = man->lastnamed = SEC_NONE;
802: man->next = ROFF_NEXT_CHILD;
803: }
804:
805: void
806: roff_man_reset(struct roff_man *man)
807: {
808:
809: roff_man_free1(man);
810: roff_man_alloc1(man);
811: }
812:
813: void
814: roff_man_free(struct roff_man *man)
815: {
816:
817: roff_man_free1(man);
818: free(man);
819: }
820:
821: struct roff_man *
822: roff_man_alloc(struct roff *roff, struct mparse *parse,
1.316 schwarze 823: const char *os_s, int quick)
1.265 schwarze 824: {
825: struct roff_man *man;
826:
827: man = mandoc_calloc(1, sizeof(*man));
828: man->parse = parse;
829: man->roff = roff;
1.316 schwarze 830: man->os_s = os_s;
1.265 schwarze 831: man->quick = quick;
832: roff_man_alloc1(man);
1.296 schwarze 833: roff->man = man;
1.277 schwarze 834: return man;
1.67 kristaps 835: }
836:
1.266 schwarze 837: /* --- syntax tree handling ----------------------------------------------- */
838:
839: struct roff_node *
840: roff_node_alloc(struct roff_man *man, int line, int pos,
841: enum roff_type type, int tok)
842: {
843: struct roff_node *n;
844:
845: n = mandoc_calloc(1, sizeof(*n));
846: n->line = line;
847: n->pos = pos;
848: n->tok = tok;
849: n->type = type;
850: n->sec = man->lastsec;
851:
852: if (man->flags & MDOC_SYNOPSIS)
1.285 schwarze 853: n->flags |= NODE_SYNPRETTY;
1.266 schwarze 854: else
1.285 schwarze 855: n->flags &= ~NODE_SYNPRETTY;
1.266 schwarze 856: if (man->flags & MDOC_NEWLINE)
1.285 schwarze 857: n->flags |= NODE_LINE;
1.266 schwarze 858: man->flags &= ~MDOC_NEWLINE;
859:
1.277 schwarze 860: return n;
1.266 schwarze 861: }
862:
863: void
864: roff_node_append(struct roff_man *man, struct roff_node *n)
865: {
866:
867: switch (man->next) {
868: case ROFF_NEXT_SIBLING:
1.281 schwarze 869: if (man->last->next != NULL) {
870: n->next = man->last->next;
871: man->last->next->prev = n;
872: } else
873: man->last->parent->last = n;
1.266 schwarze 874: man->last->next = n;
875: n->prev = man->last;
876: n->parent = man->last->parent;
877: break;
878: case ROFF_NEXT_CHILD:
1.287 schwarze 879: if (man->last->child != NULL) {
880: n->next = man->last->child;
881: man->last->child->prev = n;
882: } else
883: man->last->last = n;
1.266 schwarze 884: man->last->child = n;
885: n->parent = man->last;
886: break;
887: default:
888: abort();
889: }
1.282 schwarze 890: man->last = n;
1.266 schwarze 891:
892: switch (n->type) {
893: case ROFFT_HEAD:
894: n->parent->head = n;
895: break;
896: case ROFFT_BODY:
1.282 schwarze 897: if (n->end != ENDBODY_NOT)
898: return;
1.266 schwarze 899: n->parent->body = n;
900: break;
901: case ROFFT_TAIL:
902: n->parent->tail = n;
903: break;
904: default:
1.282 schwarze 905: return;
1.266 schwarze 906: }
1.282 schwarze 907:
908: /*
909: * Copy over the normalised-data pointer of our parent. Not
910: * everybody has one, but copying a null pointer is fine.
911: */
912:
913: n->norm = n->parent->norm;
914: assert(n->parent->type == ROFFT_BLOCK);
1.266 schwarze 915: }
916:
1.267 schwarze 917: void
918: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
919: {
920: struct roff_node *n;
921:
922: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
923: n->string = roff_strdup(man->roff, word);
924: roff_node_append(man, n);
1.286 schwarze 925: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 926: man->next = ROFF_NEXT_SIBLING;
927: }
928:
929: void
930: roff_word_append(struct roff_man *man, const char *word)
931: {
932: struct roff_node *n;
933: char *addstr, *newstr;
934:
935: n = man->last;
936: addstr = roff_strdup(man->roff, word);
937: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
938: free(addstr);
939: free(n->string);
940: n->string = newstr;
941: man->next = ROFF_NEXT_SIBLING;
1.268 schwarze 942: }
943:
944: void
945: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
946: {
947: struct roff_node *n;
948:
949: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
950: roff_node_append(man, n);
951: man->next = ROFF_NEXT_CHILD;
952: }
953:
954: struct roff_node *
955: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
956: {
957: struct roff_node *n;
958:
959: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
960: roff_node_append(man, n);
961: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 962: return n;
1.267 schwarze 963: }
964:
1.266 schwarze 965: struct roff_node *
966: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
967: {
968: struct roff_node *n;
969:
970: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
971: roff_node_append(man, n);
972: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 973: return n;
1.266 schwarze 974: }
975:
976: struct roff_node *
977: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
978: {
979: struct roff_node *n;
980:
981: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
982: roff_node_append(man, n);
983: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 984: return n;
1.267 schwarze 985: }
986:
1.321 schwarze 987: static void
988: roff_addtbl(struct roff_man *man, struct tbl_node *tbl)
1.267 schwarze 989: {
990: struct roff_node *n;
1.321 schwarze 991: const struct tbl_span *span;
1.267 schwarze 992:
993: if (man->macroset == MACROSET_MAN)
1.302 schwarze 994: man_breakscope(man, ROFF_TS);
1.321 schwarze 995: while ((span = tbl_span(tbl)) != NULL) {
996: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
997: n->span = span;
998: roff_node_append(man, n);
999: n->flags |= NODE_VALID | NODE_ENDED;
1000: man->next = ROFF_NEXT_SIBLING;
1001: }
1.266 schwarze 1002: }
1003:
1004: void
1005: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1006: {
1007:
1008: /* Adjust siblings. */
1009:
1010: if (n->prev)
1011: n->prev->next = n->next;
1012: if (n->next)
1013: n->next->prev = n->prev;
1014:
1015: /* Adjust parent. */
1016:
1017: if (n->parent != NULL) {
1018: if (n->parent->child == n)
1019: n->parent->child = n->next;
1020: if (n->parent->last == n)
1021: n->parent->last = n->prev;
1022: }
1023:
1024: /* Adjust parse point. */
1025:
1026: if (man == NULL)
1027: return;
1028: if (man->last == n) {
1029: if (n->prev == NULL) {
1030: man->last = n->parent;
1031: man->next = ROFF_NEXT_CHILD;
1032: } else {
1033: man->last = n->prev;
1034: man->next = ROFF_NEXT_SIBLING;
1035: }
1036: }
1037: if (man->first == n)
1038: man->first = NULL;
1039: }
1040:
1041: void
1042: roff_node_free(struct roff_node *n)
1043: {
1044:
1045: if (n->args != NULL)
1046: mdoc_argv_free(n->args);
1047: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1048: free(n->norm);
1.319 schwarze 1049: if (n->eqn != NULL)
1050: eqn_box_free(n->eqn);
1.266 schwarze 1051: free(n->string);
1052: free(n);
1053: }
1054:
1055: void
1056: roff_node_delete(struct roff_man *man, struct roff_node *n)
1057: {
1058:
1059: while (n->child != NULL)
1060: roff_node_delete(man, n->child);
1061: roff_node_unlink(man, n);
1062: roff_node_free(n);
1.269 schwarze 1063: }
1064:
1065: void
1066: deroff(char **dest, const struct roff_node *n)
1067: {
1068: char *cp;
1069: size_t sz;
1070:
1071: if (n->type != ROFFT_TEXT) {
1072: for (n = n->child; n != NULL; n = n->next)
1073: deroff(dest, n);
1074: return;
1075: }
1076:
1.288 schwarze 1077: /* Skip leading whitespace. */
1.269 schwarze 1078:
1.288 schwarze 1079: for (cp = n->string; *cp != '\0'; cp++) {
1.289 schwarze 1080: if (cp[0] == '\\' && cp[1] != '\0' &&
1081: strchr(" %&0^|~", cp[1]) != NULL)
1.269 schwarze 1082: cp++;
1.288 schwarze 1083: else if ( ! isspace((unsigned char)*cp))
1.269 schwarze 1084: break;
1085: }
1086:
1.289 schwarze 1087: /* Skip trailing backslash. */
1088:
1089: sz = strlen(cp);
1.290 schwarze 1090: if (sz > 0 && cp[sz - 1] == '\\')
1.289 schwarze 1091: sz--;
1092:
1.269 schwarze 1093: /* Skip trailing whitespace. */
1094:
1.289 schwarze 1095: for (; sz; sz--)
1.269 schwarze 1096: if ( ! isspace((unsigned char)cp[sz-1]))
1097: break;
1098:
1099: /* Skip empty strings. */
1100:
1101: if (sz == 0)
1102: return;
1103:
1104: if (*dest == NULL) {
1105: *dest = mandoc_strndup(cp, sz);
1106: return;
1107: }
1108:
1109: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1110: free(*dest);
1111: *dest = cp;
1.266 schwarze 1112: }
1113:
1114: /* --- main functions of the roff parser ---------------------------------- */
1115:
1.94 kristaps 1116: /*
1.206 schwarze 1117: * In the current line, expand escape sequences that tend to get
1118: * used in numerical expressions and conditional requests.
1119: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 1120: */
1.172 schwarze 1121: static enum rofferr
1.238 schwarze 1122: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 1123: {
1.208 schwarze 1124: char ubuf[24]; /* buffer to print the number */
1.205 schwarze 1125: const char *start; /* start of the string to process */
1.209 schwarze 1126: char *stesc; /* start of an escape sequence ('\\') */
1.108 schwarze 1127: const char *stnam; /* start of the name, after "[(*" */
1128: const char *cp; /* end of the name, e.g. before ']' */
1129: const char *res; /* the string to be substituted */
1.238 schwarze 1130: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 1131: size_t maxl; /* expected length of the escape name */
1132: size_t naml; /* actual length of the escape name */
1.237 schwarze 1133: enum mandoc_esc esc; /* type of the escape sequence */
1134: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 1135: int expand_count; /* to avoid infinite loops */
1.206 schwarze 1136: int npos; /* position in numeric expression */
1.218 schwarze 1137: int arg_complete; /* argument not interrupted by eol */
1.303 schwarze 1138: int done; /* no more input available */
1.315 schwarze 1139: int deftype; /* type of definition to paste */
1.323 schwarze 1140: int rcsid; /* kind of RCS id seen */
1.327 ! schwarze 1141: char sign; /* increment number register */
1.206 schwarze 1142: char term; /* character terminating the escape */
1.94 kristaps 1143:
1.303 schwarze 1144: /* Search forward for comments. */
1145:
1146: done = 0;
1147: start = buf->buf + pos;
1148: for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
1149: if (stesc[0] != r->escape || stesc[1] == '\0')
1150: continue;
1151: stesc++;
1152: if (*stesc != '"' && *stesc != '#')
1153: continue;
1.313 schwarze 1154:
1155: /* Comment found, look for RCS id. */
1156:
1.323 schwarze 1157: rcsid = 0;
1.313 schwarze 1158: if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.323 schwarze 1159: rcsid = 1 << MANDOC_OS_OPENBSD;
1.313 schwarze 1160: cp += 8;
1161: } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.323 schwarze 1162: rcsid = 1 << MANDOC_OS_NETBSD;
1.313 schwarze 1163: cp += 7;
1164: }
1165: if (cp != NULL &&
1166: isalnum((unsigned char)*cp) == 0 &&
1.314 schwarze 1167: strchr(cp, '$') != NULL) {
1.323 schwarze 1168: if (r->man->meta.rcsids & rcsid)
1.314 schwarze 1169: mandoc_msg(MANDOCERR_RCS_REP, r->parse,
1170: ln, stesc + 1 - buf->buf, stesc + 1);
1.323 schwarze 1171: r->man->meta.rcsids |= rcsid;
1.314 schwarze 1172: }
1.313 schwarze 1173:
1174: /* Handle trailing whitespace. */
1175:
1.303 schwarze 1176: cp = strchr(stesc--, '\0') - 1;
1177: if (*cp == '\n') {
1178: done = 1;
1179: cp--;
1180: }
1181: if (*cp == ' ' || *cp == '\t')
1182: mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
1183: ln, cp - buf->buf, NULL);
1184: while (stesc > start && stesc[-1] == ' ')
1185: stesc--;
1186: *stesc = '\0';
1187: break;
1188: }
1189: if (stesc == start)
1190: return ROFF_CONT;
1191: stesc--;
1192:
1193: /* Notice the end of the input. */
1194:
1195: if (*stesc == '\n') {
1196: *stesc-- = '\0';
1197: done = 1;
1198: }
1199:
1.170 schwarze 1200: expand_count = 0;
1.303 schwarze 1201: while (stesc >= start) {
1.170 schwarze 1202:
1.205 schwarze 1203: /* Search backwards for the next backslash. */
1204:
1.303 schwarze 1205: if (*stesc != r->escape) {
1206: if (*stesc == '\\') {
1207: *stesc = '\0';
1208: buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
1209: buf->buf, stesc + 1) + 1;
1210: start = nbuf + pos;
1211: stesc = nbuf + (stesc - buf->buf);
1212: free(buf->buf);
1213: buf->buf = nbuf;
1214: }
1215: stesc--;
1.205 schwarze 1216: continue;
1.303 schwarze 1217: }
1.205 schwarze 1218:
1219: /* If it is escaped, skip it. */
1220:
1221: for (cp = stesc - 1; cp >= start; cp--)
1.303 schwarze 1222: if (*cp != r->escape)
1.205 schwarze 1223: break;
1224:
1.238 schwarze 1225: if ((stesc - cp) % 2 == 0) {
1.303 schwarze 1226: while (stesc > cp)
1227: *stesc-- = '\\';
1.205 schwarze 1228: continue;
1.303 schwarze 1229: } else if (stesc[1] != '\0') {
1230: *stesc = '\\';
1231: } else {
1232: *stesc-- = '\0';
1233: if (done)
1234: continue;
1235: else
1236: return ROFF_APPEND;
1.205 schwarze 1237: }
1.108 schwarze 1238:
1.206 schwarze 1239: /* Decide whether to expand or to check only. */
1.108 schwarze 1240:
1.206 schwarze 1241: term = '\0';
1.205 schwarze 1242: cp = stesc + 1;
1.181 schwarze 1243: switch (*cp) {
1.207 schwarze 1244: case '*':
1.181 schwarze 1245: res = NULL;
1246: break;
1.207 schwarze 1247: case 'B':
1248: case 'w':
1.206 schwarze 1249: term = cp[1];
1250: /* FALLTHROUGH */
1.207 schwarze 1251: case 'n':
1.327 ! schwarze 1252: sign = cp[1];
! 1253: if (sign == '+' || sign == '-')
! 1254: cp++;
1.181 schwarze 1255: res = ubuf;
1256: break;
1257: default:
1.237 schwarze 1258: esc = mandoc_escape(&cp, &stnam, &inaml);
1259: if (esc == ESCAPE_ERROR ||
1260: (esc == ESCAPE_SPECIAL &&
1.279 schwarze 1261: mchars_spec2cp(stnam, inaml) < 0))
1.219 schwarze 1262: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1263: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1264: "%.*s", (int)(cp - stesc), stesc);
1.303 schwarze 1265: stesc--;
1.205 schwarze 1266: continue;
1.152 kristaps 1267: }
1268:
1.205 schwarze 1269: if (EXPAND_LIMIT < ++expand_count) {
1270: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1271: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1272: return ROFF_IGN;
1.205 schwarze 1273: }
1.108 schwarze 1274:
1275: /*
1276: * The third character decides the length
1.181 schwarze 1277: * of the name of the string or register.
1.108 schwarze 1278: * Save a pointer to the name.
1279: */
1280:
1.238 schwarze 1281: if (term == '\0') {
1.206 schwarze 1282: switch (*++cp) {
1.207 schwarze 1283: case '\0':
1.206 schwarze 1284: maxl = 0;
1285: break;
1.207 schwarze 1286: case '(':
1.206 schwarze 1287: cp++;
1288: maxl = 2;
1289: break;
1.207 schwarze 1290: case '[':
1.206 schwarze 1291: cp++;
1292: term = ']';
1293: maxl = 0;
1294: break;
1295: default:
1296: maxl = 1;
1297: break;
1298: }
1299: } else {
1300: cp += 2;
1.94 kristaps 1301: maxl = 0;
1302: }
1.108 schwarze 1303: stnam = cp;
1.94 kristaps 1304:
1.108 schwarze 1305: /* Advance to the end of the name. */
1.94 kristaps 1306:
1.253 schwarze 1307: naml = 0;
1.218 schwarze 1308: arg_complete = 1;
1.253 schwarze 1309: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1310: if (*cp == '\0') {
1.219 schwarze 1311: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1312: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1313: arg_complete = 0;
1.206 schwarze 1314: break;
1.153 kristaps 1315: }
1.238 schwarze 1316: if (maxl == 0 && *cp == term) {
1.206 schwarze 1317: cp++;
1.253 schwarze 1318: break;
1319: }
1320: if (*cp++ != '\\' || stesc[1] != 'w') {
1321: naml++;
1322: continue;
1323: }
1324: switch (mandoc_escape(&cp, NULL, NULL)) {
1325: case ESCAPE_SPECIAL:
1326: case ESCAPE_UNICODE:
1327: case ESCAPE_NUMBERED:
1328: case ESCAPE_OVERSTRIKE:
1329: naml++;
1330: break;
1331: default:
1.94 kristaps 1332: break;
1.206 schwarze 1333: }
1.94 kristaps 1334: }
1335:
1.108 schwarze 1336: /*
1337: * Retrieve the replacement string; if it is
1338: * undefined, resume searching for escapes.
1339: */
1340:
1.206 schwarze 1341: switch (stesc[1]) {
1.207 schwarze 1342: case '*':
1.315 schwarze 1343: if (arg_complete) {
1344: deftype = ROFFDEF_USER | ROFFDEF_PRE;
1345: res = roff_getstrn(r, stnam, naml, &deftype);
1346: }
1.206 schwarze 1347: break;
1.207 schwarze 1348: case 'B':
1.206 schwarze 1349: npos = 0;
1.218 schwarze 1350: ubuf[0] = arg_complete &&
1.261 schwarze 1351: roff_evalnum(r, ln, stnam, &npos,
1352: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1353: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1354: ubuf[1] = '\0';
1355: break;
1.207 schwarze 1356: case 'n':
1.218 schwarze 1357: if (arg_complete)
1358: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.327 ! schwarze 1359: roff_getregn(r, stnam, naml, sign));
1.218 schwarze 1360: else
1361: ubuf[0] = '\0';
1.206 schwarze 1362: break;
1.207 schwarze 1363: case 'w':
1.218 schwarze 1364: /* use even incomplete args */
1.208 schwarze 1365: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1366: 24 * (int)naml);
1367: break;
1368: }
1.94 kristaps 1369:
1.238 schwarze 1370: if (res == NULL) {
1.219 schwarze 1371: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1372: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1373: "%.*s", (int)naml, stnam);
1.142 kristaps 1374: res = "";
1.246 schwarze 1375: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1376: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1377: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1378: return ROFF_IGN;
1.94 kristaps 1379: }
1380:
1.108 schwarze 1381: /* Replace the escape sequence by the string. */
1382:
1.209 schwarze 1383: *stesc = '\0';
1.238 schwarze 1384: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1385: buf->buf, res, cp) + 1;
1.94 kristaps 1386:
1.205 schwarze 1387: /* Prepare for the next replacement. */
1.94 kristaps 1388:
1.205 schwarze 1389: start = nbuf + pos;
1.238 schwarze 1390: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1391: free(buf->buf);
1392: buf->buf = nbuf;
1.154 kristaps 1393: }
1.277 schwarze 1394: return ROFF_CONT;
1.154 kristaps 1395: }
1396:
1397: /*
1.275 schwarze 1398: * Process text streams.
1.154 kristaps 1399: */
1400: static enum rofferr
1.305 schwarze 1401: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.154 kristaps 1402: {
1403: size_t sz;
1404: const char *start;
1.178 schwarze 1405: char *p;
1406: int isz;
1.154 kristaps 1407: enum mandoc_esc esc;
1408:
1.275 schwarze 1409: /* Spring the input line trap. */
1410:
1411: if (roffit_lines == 1) {
1412: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1413: free(buf->buf);
1414: buf->buf = p;
1415: buf->sz = isz + 1;
1416: *offs = 0;
1417: free(roffit_macro);
1418: roffit_lines = 0;
1.277 schwarze 1419: return ROFF_REPARSE;
1.275 schwarze 1420: } else if (roffit_lines > 1)
1421: --roffit_lines;
1422:
1.305 schwarze 1423: if (roffce_node != NULL && buf->buf[pos] != '\0') {
1424: if (roffce_lines < 1) {
1425: r->man->last = roffce_node;
1426: r->man->next = ROFF_NEXT_SIBLING;
1427: roffce_lines = 0;
1428: roffce_node = NULL;
1429: } else
1430: roffce_lines--;
1431: }
1432:
1.275 schwarze 1433: /* Convert all breakable hyphens into ASCII_HYPH. */
1434:
1.238 schwarze 1435: start = p = buf->buf + pos;
1.154 kristaps 1436:
1.238 schwarze 1437: while (*p != '\0') {
1.154 kristaps 1438: sz = strcspn(p, "-\\");
1439: p += sz;
1440:
1.238 schwarze 1441: if (*p == '\0')
1.159 kristaps 1442: break;
1443:
1.238 schwarze 1444: if (*p == '\\') {
1.154 kristaps 1445: /* Skip over escapes. */
1446: p++;
1.189 schwarze 1447: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1448: if (esc == ESCAPE_ERROR)
1.154 kristaps 1449: break;
1.264 schwarze 1450: while (*p == '-')
1451: p++;
1.155 kristaps 1452: continue;
1.159 kristaps 1453: } else if (p == start) {
1.158 kristaps 1454: p++;
1.155 kristaps 1455: continue;
1.158 kristaps 1456: }
1.155 kristaps 1457:
1.171 schwarze 1458: if (isalpha((unsigned char)p[-1]) &&
1459: isalpha((unsigned char)p[1]))
1.155 kristaps 1460: *p = ASCII_HYPH;
1461: p++;
1.94 kristaps 1462: }
1.277 schwarze 1463: return ROFF_CONT;
1.94 kristaps 1464: }
1465:
1.67 kristaps 1466: enum rofferr
1.238 schwarze 1467: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1468: {
1.294 schwarze 1469: enum roff_tok t;
1.109 kristaps 1470: enum rofferr e;
1.238 schwarze 1471: int pos; /* parse point */
1.243 schwarze 1472: int spos; /* saved parse point for messages */
1.238 schwarze 1473: int ppos; /* original offset in buf->buf */
1474: int ctl; /* macro line (boolean) */
1475:
1476: ppos = pos = *offs;
1.79 kristaps 1477:
1.230 schwarze 1478: /* Handle in-line equation delimiters. */
1479:
1.236 schwarze 1480: if (r->tbl == NULL &&
1481: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1482: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1483: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1484: if (e == ROFF_REPARSE)
1.277 schwarze 1485: return e;
1.230 schwarze 1486: assert(e == ROFF_CONT);
1487: }
1488:
1489: /* Expand some escape sequences. */
1.94 kristaps 1490:
1.238 schwarze 1491: e = roff_res(r, buf, ln, pos);
1.303 schwarze 1492: if (e == ROFF_IGN || e == ROFF_APPEND)
1.277 schwarze 1493: return e;
1.238 schwarze 1494: assert(e == ROFF_CONT);
1.94 kristaps 1495:
1.238 schwarze 1496: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1497:
1.94 kristaps 1498: /*
1.79 kristaps 1499: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1500: * text through the macro's filter.
1501: * Equations process all content themselves.
1502: * Tables process almost all content themselves, but we want
1503: * to warn about macros before passing it there.
1.79 kristaps 1504: */
1.74 kristaps 1505:
1.252 schwarze 1506: if (r->last != NULL && ! ctl) {
1.78 kristaps 1507: t = r->last->tok;
1.238 schwarze 1508: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.294 schwarze 1509: if (e == ROFF_IGN)
1.277 schwarze 1510: return e;
1.294 schwarze 1511: assert(e == ROFF_CONT);
1.182 schwarze 1512: }
1.319 schwarze 1513: if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
1514: eqn_read(r->eqn, buf->buf + ppos);
1515: return ROFF_IGN;
1516: }
1.321 schwarze 1517: if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
1518: tbl_read(r->tbl, ln, buf->buf, ppos);
1519: roff_addtbl(r->man, r->tbl);
1520: return ROFF_IGN;
1521: }
1.252 schwarze 1522: if ( ! ctl)
1.305 schwarze 1523: return roff_parsetext(r, buf, pos, offs);
1.228 schwarze 1524:
1525: /* Skip empty request lines. */
1526:
1.238 schwarze 1527: if (buf->buf[pos] == '"') {
1.228 schwarze 1528: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1529: ln, pos, NULL);
1.277 schwarze 1530: return ROFF_IGN;
1.238 schwarze 1531: } else if (buf->buf[pos] == '\0')
1.277 schwarze 1532: return ROFF_IGN;
1.67 kristaps 1533:
1.79 kristaps 1534: /*
1535: * If a scope is open, go to the child handler for that macro,
1536: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1537: * Don't do so if an equation is open.
1.79 kristaps 1538: */
1.78 kristaps 1539:
1.79 kristaps 1540: if (r->last) {
1541: t = r->last->tok;
1.277 schwarze 1542: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1543: }
1.78 kristaps 1544:
1.243 schwarze 1545: /* No scope is open. This is a new request or macro. */
1546:
1547: spos = pos;
1548: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1549:
1550: /* Tables ignore most macros. */
1551:
1.308 schwarze 1552: if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.309 schwarze 1553: t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.243 schwarze 1554: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1555: ln, pos, buf->buf + spos);
1.308 schwarze 1556: if (t != TOKEN_NONE)
1.277 schwarze 1557: return ROFF_IGN;
1.257 schwarze 1558: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1559: pos++;
1.291 schwarze 1560: while (buf->buf[pos] == ' ')
1.257 schwarze 1561: pos++;
1.321 schwarze 1562: tbl_read(r->tbl, ln, buf->buf, pos);
1563: roff_addtbl(r->man, r->tbl);
1564: return ROFF_IGN;
1.243 schwarze 1565: }
1566:
1.305 schwarze 1567: /* For now, let high level macros abort .ce mode. */
1568:
1569: if (ctl && roffce_node != NULL &&
1.324 schwarze 1570: (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
1571: t == ROFF_TH || t == ROFF_TS)) {
1.305 schwarze 1572: r->man->last = roffce_node;
1573: r->man->next = ROFF_NEXT_SIBLING;
1574: roffce_lines = 0;
1575: roffce_node = NULL;
1576: }
1577:
1.79 kristaps 1578: /*
1.243 schwarze 1579: * This is neither a roff request nor a user-defined macro.
1580: * Let the standard macro set parsers handle it.
1.79 kristaps 1581: */
1.67 kristaps 1582:
1.294 schwarze 1583: if (t == TOKEN_NONE)
1.277 schwarze 1584: return ROFF_CONT;
1.243 schwarze 1585:
1586: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1587:
1.296 schwarze 1588: return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.74 kristaps 1589: }
1590:
1.117 kristaps 1591: void
1.74 kristaps 1592: roff_endparse(struct roff *r)
1593: {
1.321 schwarze 1594: if (r->last != NULL)
1.221 schwarze 1595: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1596: r->last->line, r->last->col,
1.295 schwarze 1597: roff_name[r->last->tok]);
1.117 kristaps 1598:
1.321 schwarze 1599: if (r->eqn != NULL) {
1.221 schwarze 1600: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1.319 schwarze 1601: r->eqn->node->line, r->eqn->node->pos, "EQ");
1602: eqn_parse(r->eqn);
1603: r->eqn = NULL;
1.125 kristaps 1604: }
1605:
1.321 schwarze 1606: if (r->tbl != NULL) {
1.221 schwarze 1607: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1608: r->tbl->line, r->tbl->pos, "TS");
1.321 schwarze 1609: tbl_end(r->tbl);
1610: r->tbl = NULL;
1.117 kristaps 1611: }
1.67 kristaps 1612: }
1613:
1614: /*
1615: * Parse a roff node's type from the input buffer. This must be in the
1616: * form of ".foo xxx" in the usual way.
1617: */
1.294 schwarze 1618: static enum roff_tok
1.214 schwarze 1619: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1620: {
1.214 schwarze 1621: char *cp;
1.106 kristaps 1622: const char *mac;
1623: size_t maclen;
1.315 schwarze 1624: int deftype;
1.294 schwarze 1625: enum roff_tok t;
1.67 kristaps 1626:
1.214 schwarze 1627: cp = buf + *pos;
1628:
1629: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294 schwarze 1630: return TOKEN_NONE;
1.67 kristaps 1631:
1.214 schwarze 1632: mac = cp;
1633: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1634:
1.315 schwarze 1635: deftype = ROFFDEF_USER | ROFFDEF_REN;
1636: r->current_string = roff_getstrn(r, mac, maclen, &deftype);
1637: switch (deftype) {
1638: case ROFFDEF_USER:
1639: t = ROFF_USERDEF;
1640: break;
1641: case ROFFDEF_REN:
1642: t = ROFF_RENAMED;
1643: break;
1644: default:
1645: t = roffhash_find(r->reqtab, mac, maclen);
1646: break;
1647: }
1.294 schwarze 1648: if (t != TOKEN_NONE)
1.214 schwarze 1649: *pos = cp - buf;
1.325 schwarze 1650: else if (deftype == ROFFDEF_UNDEF) {
1651: /* Using an undefined macro defines it to be empty. */
1652: roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
1653: roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
1654: }
1.277 schwarze 1655: return t;
1.67 kristaps 1656: }
1657:
1.266 schwarze 1658: /* --- handling of request blocks ----------------------------------------- */
1659:
1.67 kristaps 1660: static enum rofferr
1.76 kristaps 1661: roff_cblock(ROFF_ARGS)
1.67 kristaps 1662: {
1663:
1.79 kristaps 1664: /*
1665: * A block-close `..' should only be invoked as a child of an
1666: * ignore macro, otherwise raise a warning and just ignore it.
1667: */
1668:
1.238 schwarze 1669: if (r->last == NULL) {
1.221 schwarze 1670: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1671: ln, ppos, "..");
1.277 schwarze 1672: return ROFF_IGN;
1.76 kristaps 1673: }
1.67 kristaps 1674:
1.81 kristaps 1675: switch (r->last->tok) {
1.207 schwarze 1676: case ROFF_am:
1.220 schwarze 1677: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.207 schwarze 1678: case ROFF_ami:
1679: case ROFF_de:
1.108 schwarze 1680: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.207 schwarze 1681: case ROFF_dei:
1682: case ROFF_ig:
1.81 kristaps 1683: break;
1684: default:
1.221 schwarze 1685: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1686: ln, ppos, "..");
1.277 schwarze 1687: return ROFF_IGN;
1.76 kristaps 1688: }
1.67 kristaps 1689:
1.238 schwarze 1690: if (buf->buf[pos] != '\0')
1.217 schwarze 1691: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1692: ".. %s", buf->buf + pos);
1.71 kristaps 1693:
1694: roffnode_pop(r);
1.76 kristaps 1695: roffnode_cleanscope(r);
1.277 schwarze 1696: return ROFF_IGN;
1.71 kristaps 1697:
1.67 kristaps 1698: }
1699:
1.76 kristaps 1700: static void
1701: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1702: {
1703:
1.76 kristaps 1704: while (r->last) {
1.173 schwarze 1705: if (--r->last->endspan != 0)
1.76 kristaps 1706: break;
1707: roffnode_pop(r);
1708: }
1.67 kristaps 1709: }
1710:
1.195 schwarze 1711: static void
1712: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1713: {
1714:
1.76 kristaps 1715: if (NULL == r->last) {
1.221 schwarze 1716: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1717: ln, ppos, "\\}");
1.195 schwarze 1718: return;
1.76 kristaps 1719: }
1720:
1.82 kristaps 1721: switch (r->last->tok) {
1.207 schwarze 1722: case ROFF_el:
1723: case ROFF_ie:
1724: case ROFF_if:
1.82 kristaps 1725: break;
1726: default:
1.221 schwarze 1727: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1728: ln, ppos, "\\}");
1.195 schwarze 1729: return;
1.75 kristaps 1730: }
1731:
1.76 kristaps 1732: if (r->last->endspan > -1) {
1.221 schwarze 1733: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1734: ln, ppos, "\\}");
1.195 schwarze 1735: return;
1.76 kristaps 1736: }
1737:
1.75 kristaps 1738: roffnode_pop(r);
1.76 kristaps 1739: roffnode_cleanscope(r);
1.195 schwarze 1740: return;
1.76 kristaps 1741: }
1742:
1743: static enum rofferr
1.80 kristaps 1744: roff_block(ROFF_ARGS)
1.76 kristaps 1745: {
1.315 schwarze 1746: const char *name, *value;
1747: char *call, *cp, *iname, *rname;
1748: size_t csz, namesz, rsz;
1749: int deftype;
1.106 kristaps 1750:
1.220 schwarze 1751: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1752:
1.238 schwarze 1753: if (tok == ROFF_de1)
1.220 schwarze 1754: tok = ROFF_de;
1.251 schwarze 1755: else if (tok == ROFF_dei1)
1756: tok = ROFF_dei;
1.238 schwarze 1757: else if (tok == ROFF_am1)
1.220 schwarze 1758: tok = ROFF_am;
1.251 schwarze 1759: else if (tok == ROFF_ami1)
1760: tok = ROFF_ami;
1.220 schwarze 1761:
1762: /* Parse the macro name argument. */
1763:
1.238 schwarze 1764: cp = buf->buf + pos;
1765: if (tok == ROFF_ig) {
1.220 schwarze 1766: iname = NULL;
1767: namesz = 0;
1768: } else {
1769: iname = cp;
1770: namesz = roff_getname(r, &cp, ln, ppos);
1771: iname[namesz] = '\0';
1772: }
1.107 kristaps 1773:
1.220 schwarze 1774: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1775:
1.238 schwarze 1776: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315 schwarze 1777: deftype = ROFFDEF_USER;
1778: name = roff_getstrn(r, iname, namesz, &deftype);
1779: if (name == NULL) {
1.220 schwarze 1780: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1781: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1782: "%.*s", (int)namesz, iname);
1783: namesz = 0;
1784: } else
1785: namesz = strlen(name);
1786: } else
1787: name = iname;
1.107 kristaps 1788:
1.238 schwarze 1789: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1790: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.295 schwarze 1791: ln, ppos, roff_name[tok]);
1.277 schwarze 1792: return ROFF_IGN;
1.220 schwarze 1793: }
1.80 kristaps 1794:
1.106 kristaps 1795: roffnode_push(r, tok, name, ln, ppos);
1796:
1797: /*
1798: * At the beginning of a `de' macro, clear the existing string
1799: * with the same name, if there is one. New content will be
1.193 schwarze 1800: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1801: */
1.107 kristaps 1802:
1.311 schwarze 1803: if (tok == ROFF_de || tok == ROFF_dei) {
1.213 schwarze 1804: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.311 schwarze 1805: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.315 schwarze 1806: } else if (tok == ROFF_am || tok == ROFF_ami) {
1807: deftype = ROFFDEF_ANY;
1808: value = roff_getstrn(r, iname, namesz, &deftype);
1809: switch (deftype) { /* Before appending, ... */
1810: case ROFFDEF_PRE: /* copy predefined to user-defined. */
1811: roff_setstrn(&r->strtab, name, namesz,
1812: value, strlen(value), 0);
1813: break;
1814: case ROFFDEF_REN: /* call original standard macro. */
1815: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1816: (int)strlen(value), value);
1817: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1818: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1819: free(call);
1820: break;
1821: case ROFFDEF_STD: /* rename and call standard macro. */
1822: rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
1823: roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
1824: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1825: (int)rsz, rname);
1826: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1827: free(call);
1828: free(rname);
1829: break;
1830: default:
1831: break;
1832: }
1.311 schwarze 1833: }
1.76 kristaps 1834:
1.238 schwarze 1835: if (*cp == '\0')
1.277 schwarze 1836: return ROFF_IGN;
1.78 kristaps 1837:
1.220 schwarze 1838: /* Get the custom end marker. */
1.107 kristaps 1839:
1.220 schwarze 1840: iname = cp;
1.213 schwarze 1841: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1842:
1843: /* Resolve the end marker if it is indirect. */
1844:
1.238 schwarze 1845: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315 schwarze 1846: deftype = ROFFDEF_USER;
1847: name = roff_getstrn(r, iname, namesz, &deftype);
1848: if (name == NULL) {
1.220 schwarze 1849: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1850: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1851: "%.*s", (int)namesz, iname);
1852: namesz = 0;
1853: } else
1854: namesz = strlen(name);
1855: } else
1856: name = iname;
1857:
1.213 schwarze 1858: if (namesz)
1859: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1860:
1.238 schwarze 1861: if (*cp != '\0')
1.217 schwarze 1862: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.295 schwarze 1863: ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74 kristaps 1864:
1.277 schwarze 1865: return ROFF_IGN;
1.78 kristaps 1866: }
1867:
1868: static enum rofferr
1.80 kristaps 1869: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1870: {
1.294 schwarze 1871: enum roff_tok t;
1.79 kristaps 1872: int i, j;
1873:
1874: /*
1875: * First check whether a custom macro exists at this level. If
1876: * it does, then check against it. This is some of groff's
1877: * stranger behaviours. If we encountered a custom end-scope
1878: * tag and that tag also happens to be a "real" macro, then we
1879: * need to try interpreting it again as a real macro. If it's
1880: * not, then return ignore. Else continue.
1881: */
1882:
1883: if (r->last->end) {
1.130 kristaps 1884: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1885: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1886: break;
1887:
1.238 schwarze 1888: if (r->last->end[j] == '\0' &&
1889: (buf->buf[i] == '\0' ||
1890: buf->buf[i] == ' ' ||
1891: buf->buf[i] == '\t')) {
1.79 kristaps 1892: roffnode_pop(r);
1893: roffnode_cleanscope(r);
1894:
1.238 schwarze 1895: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1896: i++;
1897:
1898: pos = i;
1.238 schwarze 1899: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294 schwarze 1900: TOKEN_NONE)
1.277 schwarze 1901: return ROFF_RERUN;
1902: return ROFF_IGN;
1.79 kristaps 1903: }
1904: }
1905:
1906: /*
1907: * If we have no custom end-query or lookup failed, then try
1908: * pulling it out of the hashtable.
1909: */
1910:
1.238 schwarze 1911: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1912:
1.238 schwarze 1913: if (t != ROFF_cblock) {
1914: if (tok != ROFF_ig)
1915: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.277 schwarze 1916: return ROFF_IGN;
1.106 kristaps 1917: }
1.79 kristaps 1918:
1.277 schwarze 1919: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1920: }
1921:
1922: static enum rofferr
1.80 kristaps 1923: roff_block_text(ROFF_ARGS)
1.78 kristaps 1924: {
1925:
1.238 schwarze 1926: if (tok != ROFF_ig)
1927: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1928:
1.277 schwarze 1929: return ROFF_IGN;
1.78 kristaps 1930: }
1931:
1932: static enum rofferr
1.82 kristaps 1933: roff_cond_sub(ROFF_ARGS)
1934: {
1.294 schwarze 1935: enum roff_tok t;
1.139 kristaps 1936: char *ep;
1.198 schwarze 1937: int rr;
1.82 kristaps 1938:
1939: rr = r->last->rule;
1.139 kristaps 1940: roffnode_cleanscope(r);
1.144 kristaps 1941:
1.196 schwarze 1942: /*
1943: * If `\}' occurs on a macro line without a preceding macro,
1944: * drop the line completely.
1945: */
1946:
1.238 schwarze 1947: ep = buf->buf + pos;
1948: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1949: rr = 0;
1.196 schwarze 1950:
1.177 schwarze 1951: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1952:
1.238 schwarze 1953: while ((ep = strchr(ep, '\\')) != NULL) {
1.318 schwarze 1954: switch (ep[1]) {
1955: case '}':
1956: memmove(ep, ep + 2, strlen(ep + 2) + 1);
1957: roff_ccond(r, ln, ep - buf->buf);
1958: break;
1959: case '\0':
1960: ++ep;
1961: break;
1962: default:
1963: ep += 2;
1964: break;
1.197 schwarze 1965: }
1.177 schwarze 1966: }
1.318 schwarze 1967:
1968: /*
1969: * Fully handle known macros when they are structurally
1970: * required or when the conditional evaluated to true.
1971: */
1972:
1973: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1974: return t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT)
1975: ? (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs) : rr
1976: ? ROFF_CONT : ROFF_IGN;
1.82 kristaps 1977: }
1978:
1979: static enum rofferr
1980: roff_cond_text(ROFF_ARGS)
1.78 kristaps 1981: {
1.140 kristaps 1982: char *ep;
1.198 schwarze 1983: int rr;
1.82 kristaps 1984:
1985: rr = r->last->rule;
1.140 kristaps 1986: roffnode_cleanscope(r);
1.82 kristaps 1987:
1.238 schwarze 1988: ep = buf->buf + pos;
1989: while ((ep = strchr(ep, '\\')) != NULL) {
1990: if (*(++ep) == '}') {
1.197 schwarze 1991: *ep = '&';
1.238 schwarze 1992: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1993: }
1.247 schwarze 1994: if (*ep != '\0')
1995: ++ep;
1.78 kristaps 1996: }
1.277 schwarze 1997: return rr ? ROFF_CONT : ROFF_IGN;
1.74 kristaps 1998: }
1999:
1.266 schwarze 2000: /* --- handling of numeric and conditional expressions -------------------- */
2001:
1.204 schwarze 2002: /*
2003: * Parse a single signed integer number. Stop at the first non-digit.
2004: * If there is at least one digit, return success and advance the
2005: * parse point, else return failure and let the parse point unchanged.
2006: * Ignore overflows, treat them just like the C language.
2007: */
1.184 schwarze 2008: static int
1.261 schwarze 2009: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 2010: {
1.261 schwarze 2011: int myres, scaled, n, p;
1.206 schwarze 2012:
2013: if (NULL == res)
2014: res = &myres;
1.184 schwarze 2015:
2016: p = *pos;
2017: n = v[p] == '-';
1.261 schwarze 2018: if (n || v[p] == '+')
1.184 schwarze 2019: p++;
2020:
1.261 schwarze 2021: if (flags & ROFFNUM_WHITE)
2022: while (isspace((unsigned char)v[p]))
2023: p++;
2024:
1.184 schwarze 2025: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 2026: *res = 10 * *res + v[p] - '0';
1.184 schwarze 2027: if (p == *pos + n)
2028: return 0;
2029:
2030: if (n)
2031: *res = -*res;
2032:
1.254 schwarze 2033: /* Each number may be followed by one optional scaling unit. */
2034:
2035: switch (v[p]) {
2036: case 'f':
1.261 schwarze 2037: scaled = *res * 65536;
1.254 schwarze 2038: break;
2039: case 'i':
1.261 schwarze 2040: scaled = *res * 240;
1.254 schwarze 2041: break;
2042: case 'c':
1.261 schwarze 2043: scaled = *res * 240 / 2.54;
1.254 schwarze 2044: break;
2045: case 'v':
2046: case 'P':
1.261 schwarze 2047: scaled = *res * 40;
1.254 schwarze 2048: break;
2049: case 'm':
2050: case 'n':
1.261 schwarze 2051: scaled = *res * 24;
1.254 schwarze 2052: break;
2053: case 'p':
1.261 schwarze 2054: scaled = *res * 10 / 3;
1.254 schwarze 2055: break;
2056: case 'u':
1.261 schwarze 2057: scaled = *res;
1.254 schwarze 2058: break;
2059: case 'M':
1.261 schwarze 2060: scaled = *res * 6 / 25;
1.254 schwarze 2061: break;
2062: default:
1.261 schwarze 2063: scaled = *res;
1.254 schwarze 2064: p--;
2065: break;
2066: }
1.261 schwarze 2067: if (flags & ROFFNUM_SCALE)
2068: *res = scaled;
1.254 schwarze 2069:
2070: *pos = p + 1;
1.277 schwarze 2071: return 1;
1.184 schwarze 2072: }
2073:
1.198 schwarze 2074: /*
2075: * Evaluate a string comparison condition.
2076: * The first character is the delimiter.
2077: * Succeed if the string up to its second occurrence
2078: * matches the string up to its third occurence.
2079: * Advance the cursor after the third occurrence
2080: * or lacking that, to the end of the line.
2081: */
2082: static int
2083: roff_evalstrcond(const char *v, int *pos)
2084: {
2085: const char *s1, *s2, *s3;
2086: int match;
2087:
2088: match = 0;
2089: s1 = v + *pos; /* initial delimiter */
2090: s2 = s1 + 1; /* for scanning the first string */
2091: s3 = strchr(s2, *s1); /* for scanning the second string */
2092:
2093: if (NULL == s3) /* found no middle delimiter */
2094: goto out;
2095:
2096: while ('\0' != *++s3) {
2097: if (*s2 != *s3) { /* mismatch */
2098: s3 = strchr(s3, *s1);
2099: break;
2100: }
2101: if (*s3 == *s1) { /* found the final delimiter */
2102: match = 1;
2103: break;
2104: }
2105: s2++;
2106: }
2107:
2108: out:
2109: if (NULL == s3)
2110: s3 = strchr(s2, '\0');
1.242 schwarze 2111: else if (*s3 != '\0')
1.198 schwarze 2112: s3++;
2113: *pos = s3 - v;
1.277 schwarze 2114: return match;
1.198 schwarze 2115: }
2116:
1.204 schwarze 2117: /*
2118: * Evaluate an optionally negated single character, numerical,
2119: * or string condition.
2120: */
1.198 schwarze 2121: static int
1.271 schwarze 2122: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88 kristaps 2123: {
1.271 schwarze 2124: char *cp, *name;
2125: size_t sz;
1.315 schwarze 2126: int deftype, number, savepos, istrue, wanttrue;
1.88 kristaps 2127:
1.198 schwarze 2128: if ('!' == v[*pos]) {
2129: wanttrue = 0;
2130: (*pos)++;
2131: } else
2132: wanttrue = 1;
2133:
1.88 kristaps 2134: switch (v[*pos]) {
1.240 schwarze 2135: case '\0':
1.277 schwarze 2136: return 0;
1.207 schwarze 2137: case 'n':
2138: case 'o':
1.88 kristaps 2139: (*pos)++;
1.277 schwarze 2140: return wanttrue;
1.207 schwarze 2141: case 'c':
2142: case 'e':
2143: case 't':
1.239 schwarze 2144: case 'v':
1.88 kristaps 2145: (*pos)++;
1.277 schwarze 2146: return !wanttrue;
1.310 schwarze 2147: case 'd':
1.271 schwarze 2148: case 'r':
1.310 schwarze 2149: cp = v + *pos + 1;
2150: while (*cp == ' ')
2151: cp++;
2152: name = cp;
2153: sz = roff_getname(r, &cp, ln, cp - v);
1.315 schwarze 2154: if (sz == 0)
2155: istrue = 0;
2156: else if (v[*pos] == 'r')
2157: istrue = roff_hasregn(r, name, sz);
2158: else {
2159: deftype = ROFFDEF_ANY;
2160: roff_getstrn(r, name, sz, &deftype);
2161: istrue = !!deftype;
2162: }
1.271 schwarze 2163: *pos = cp - v;
1.310 schwarze 2164: return istrue == wanttrue;
1.88 kristaps 2165: default:
2166: break;
2167: }
2168:
1.241 schwarze 2169: savepos = *pos;
1.261 schwarze 2170: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277 schwarze 2171: return (number > 0) == wanttrue;
1.241 schwarze 2172: else if (*pos == savepos)
1.277 schwarze 2173: return roff_evalstrcond(v, pos) == wanttrue;
1.204 schwarze 2174: else
1.277 schwarze 2175: return 0;
1.88 kristaps 2176: }
2177:
1.74 kristaps 2178: static enum rofferr
1.103 kristaps 2179: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 2180: {
1.123 schwarze 2181:
1.277 schwarze 2182: return ROFF_IGN;
1.89 kristaps 2183: }
2184:
1.104 kristaps 2185: static enum rofferr
1.251 schwarze 2186: roff_insec(ROFF_ARGS)
2187: {
2188:
2189: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.295 schwarze 2190: ln, ppos, roff_name[tok]);
1.277 schwarze 2191: return ROFF_IGN;
1.251 schwarze 2192: }
2193:
2194: static enum rofferr
2195: roff_unsupp(ROFF_ARGS)
2196: {
2197:
2198: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.295 schwarze 2199: ln, ppos, roff_name[tok]);
1.277 schwarze 2200: return ROFF_IGN;
1.251 schwarze 2201: }
2202:
2203: static enum rofferr
1.82 kristaps 2204: roff_cond(ROFF_ARGS)
1.74 kristaps 2205: {
1.173 schwarze 2206:
2207: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 2208:
1.207 schwarze 2209: /*
1.134 kristaps 2210: * An `.el' has no conditional body: it will consume the value
2211: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 2212: * defaults to DENY.
1.134 kristaps 2213: *
2214: * If we're not an `el', however, then evaluate the conditional.
2215: */
1.133 kristaps 2216:
1.238 schwarze 2217: r->last->rule = tok == ROFF_el ?
1.207 schwarze 2218: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 2219: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 2220:
1.134 kristaps 2221: /*
2222: * An if-else will put the NEGATION of the current evaluated
2223: * conditional into the stack of rules.
2224: */
2225:
1.238 schwarze 2226: if (tok == ROFF_ie) {
1.223 schwarze 2227: if (r->rstackpos + 1 == r->rstacksz) {
2228: r->rstacksz += 16;
2229: r->rstack = mandoc_reallocarray(r->rstack,
2230: r->rstacksz, sizeof(int));
1.134 kristaps 2231: }
1.198 schwarze 2232: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 2233: }
1.88 kristaps 2234:
2235: /* If the parent has false as its rule, then so do we. */
2236:
1.198 schwarze 2237: if (r->last->parent && !r->last->parent->rule)
2238: r->last->rule = 0;
1.88 kristaps 2239:
2240: /*
1.173 schwarze 2241: * Determine scope.
2242: * If there is nothing on the line after the conditional,
2243: * not even whitespace, use next-line scope.
1.88 kristaps 2244: */
1.74 kristaps 2245:
1.238 schwarze 2246: if (buf->buf[pos] == '\0') {
1.173 schwarze 2247: r->last->endspan = 2;
2248: goto out;
2249: }
2250:
1.238 schwarze 2251: while (buf->buf[pos] == ' ')
1.173 schwarze 2252: pos++;
2253:
2254: /* An opening brace requests multiline scope. */
1.75 kristaps 2255:
1.238 schwarze 2256: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 2257: r->last->endspan = -1;
2258: pos += 2;
1.272 schwarze 2259: while (buf->buf[pos] == ' ')
2260: pos++;
1.173 schwarze 2261: goto out;
1.207 schwarze 2262: }
1.74 kristaps 2263:
1.77 kristaps 2264: /*
1.173 schwarze 2265: * Anything else following the conditional causes
2266: * single-line scope. Warn if the scope contains
2267: * nothing but trailing whitespace.
1.77 kristaps 2268: */
2269:
1.238 schwarze 2270: if (buf->buf[pos] == '\0')
1.216 schwarze 2271: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.295 schwarze 2272: ln, ppos, roff_name[tok]);
1.77 kristaps 2273:
1.173 schwarze 2274: r->last->endspan = 1;
1.74 kristaps 2275:
1.173 schwarze 2276: out:
1.75 kristaps 2277: *offs = pos;
1.277 schwarze 2278: return ROFF_RERUN;
1.83 schwarze 2279: }
2280:
2281: static enum rofferr
1.92 schwarze 2282: roff_ds(ROFF_ARGS)
2283: {
1.212 schwarze 2284: char *string;
2285: const char *name;
2286: size_t namesz;
1.96 kristaps 2287:
1.251 schwarze 2288: /* Ignore groff compatibility mode for now. */
2289:
2290: if (tok == ROFF_ds1)
2291: tok = ROFF_ds;
2292: else if (tok == ROFF_as1)
2293: tok = ROFF_as;
2294:
1.96 kristaps 2295: /*
1.212 schwarze 2296: * The first word is the name of the string.
2297: * If it is empty or terminated by an escape sequence,
2298: * abort the `ds' request without defining anything.
1.96 kristaps 2299: */
1.92 schwarze 2300:
1.238 schwarze 2301: name = string = buf->buf + pos;
2302: if (*name == '\0')
1.277 schwarze 2303: return ROFF_IGN;
1.92 schwarze 2304:
1.212 schwarze 2305: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 2306: if (name[namesz] == '\\')
1.277 schwarze 2307: return ROFF_IGN;
1.212 schwarze 2308:
2309: /* Read past the initial double-quote, if any. */
1.238 schwarze 2310: if (*string == '"')
1.92 schwarze 2311: string++;
2312:
1.96 kristaps 2313: /* The rest is the value. */
1.212 schwarze 2314: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2315: ROFF_as == tok);
1.311 schwarze 2316: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.277 schwarze 2317: return ROFF_IGN;
1.92 schwarze 2318: }
2319:
1.204 schwarze 2320: /*
2321: * Parse a single operator, one or two characters long.
2322: * If the operator is recognized, return success and advance the
2323: * parse point, else return failure and let the parse point unchanged.
2324: */
2325: static int
2326: roff_getop(const char *v, int *pos, char *res)
2327: {
2328:
2329: *res = v[*pos];
2330:
2331: switch (*res) {
1.207 schwarze 2332: case '+':
2333: case '-':
2334: case '*':
2335: case '/':
2336: case '%':
2337: case '&':
2338: case ':':
1.204 schwarze 2339: break;
2340: case '<':
2341: switch (v[*pos + 1]) {
1.207 schwarze 2342: case '=':
1.204 schwarze 2343: *res = 'l';
2344: (*pos)++;
2345: break;
1.207 schwarze 2346: case '>':
1.204 schwarze 2347: *res = '!';
2348: (*pos)++;
2349: break;
1.207 schwarze 2350: case '?':
1.204 schwarze 2351: *res = 'i';
2352: (*pos)++;
2353: break;
2354: default:
2355: break;
2356: }
2357: break;
2358: case '>':
2359: switch (v[*pos + 1]) {
1.207 schwarze 2360: case '=':
1.204 schwarze 2361: *res = 'g';
2362: (*pos)++;
2363: break;
1.207 schwarze 2364: case '?':
1.204 schwarze 2365: *res = 'a';
2366: (*pos)++;
2367: break;
2368: default:
2369: break;
2370: }
2371: break;
2372: case '=':
2373: if ('=' == v[*pos + 1])
2374: (*pos)++;
2375: break;
2376: default:
1.277 schwarze 2377: return 0;
1.204 schwarze 2378: }
2379: (*pos)++;
2380:
1.277 schwarze 2381: return *res;
1.204 schwarze 2382: }
2383:
2384: /*
2385: * Evaluate either a parenthesized numeric expression
2386: * or a single signed integer number.
2387: */
2388: static int
1.235 schwarze 2389: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2390: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2391: {
2392:
2393: if ('(' != v[*pos])
1.277 schwarze 2394: return roff_getnum(v, pos, res, flags);
1.204 schwarze 2395:
2396: (*pos)++;
1.261 schwarze 2397: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277 schwarze 2398: return 0;
1.204 schwarze 2399:
1.206 schwarze 2400: /*
2401: * Omission of the closing parenthesis
2402: * is an error in validation mode,
2403: * but ignored in evaluation mode.
2404: */
2405:
1.204 schwarze 2406: if (')' == v[*pos])
2407: (*pos)++;
1.206 schwarze 2408: else if (NULL == res)
1.277 schwarze 2409: return 0;
1.204 schwarze 2410:
1.277 schwarze 2411: return 1;
1.204 schwarze 2412: }
2413:
2414: /*
2415: * Evaluate a complete numeric expression.
2416: * Proceed left to right, there is no concept of precedence.
2417: */
2418: static int
1.235 schwarze 2419: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2420: int *pos, int *res, int flags)
1.204 schwarze 2421: {
2422: int mypos, operand2;
2423: char operator;
2424:
2425: if (NULL == pos) {
2426: mypos = 0;
2427: pos = &mypos;
2428: }
2429:
1.261 schwarze 2430: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2431: while (isspace((unsigned char)v[*pos]))
2432: (*pos)++;
2433:
1.261 schwarze 2434: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277 schwarze 2435: return 0;
1.204 schwarze 2436:
2437: while (1) {
1.261 schwarze 2438: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2439: while (isspace((unsigned char)v[*pos]))
2440: (*pos)++;
2441:
2442: if ( ! roff_getop(v, pos, &operator))
2443: break;
2444:
1.261 schwarze 2445: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2446: while (isspace((unsigned char)v[*pos]))
2447: (*pos)++;
2448:
1.261 schwarze 2449: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277 schwarze 2450: return 0;
1.204 schwarze 2451:
1.261 schwarze 2452: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2453: while (isspace((unsigned char)v[*pos]))
2454: (*pos)++;
1.206 schwarze 2455:
2456: if (NULL == res)
2457: continue;
1.204 schwarze 2458:
2459: switch (operator) {
1.207 schwarze 2460: case '+':
1.204 schwarze 2461: *res += operand2;
2462: break;
1.207 schwarze 2463: case '-':
1.204 schwarze 2464: *res -= operand2;
2465: break;
1.207 schwarze 2466: case '*':
1.204 schwarze 2467: *res *= operand2;
2468: break;
1.207 schwarze 2469: case '/':
1.244 schwarze 2470: if (operand2 == 0) {
1.235 schwarze 2471: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2472: r->parse, ln, *pos, v);
2473: *res = 0;
2474: break;
2475: }
1.204 schwarze 2476: *res /= operand2;
2477: break;
1.207 schwarze 2478: case '%':
1.244 schwarze 2479: if (operand2 == 0) {
2480: mandoc_msg(MANDOCERR_DIVZERO,
2481: r->parse, ln, *pos, v);
2482: *res = 0;
2483: break;
2484: }
1.204 schwarze 2485: *res %= operand2;
2486: break;
1.207 schwarze 2487: case '<':
1.204 schwarze 2488: *res = *res < operand2;
2489: break;
1.207 schwarze 2490: case '>':
1.204 schwarze 2491: *res = *res > operand2;
2492: break;
1.207 schwarze 2493: case 'l':
1.204 schwarze 2494: *res = *res <= operand2;
2495: break;
1.207 schwarze 2496: case 'g':
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 '&':
1.204 schwarze 2506: *res = *res && operand2;
2507: break;
1.207 schwarze 2508: case ':':
1.204 schwarze 2509: *res = *res || operand2;
2510: break;
1.207 schwarze 2511: case 'i':
1.204 schwarze 2512: if (operand2 < *res)
2513: *res = operand2;
2514: break;
1.207 schwarze 2515: case 'a':
1.204 schwarze 2516: if (operand2 > *res)
2517: *res = operand2;
2518: break;
2519: default:
2520: abort();
2521: }
2522: }
1.277 schwarze 2523: return 1;
1.204 schwarze 2524: }
2525:
1.266 schwarze 2526: /* --- register management ------------------------------------------------ */
2527:
1.180 schwarze 2528: void
1.187 schwarze 2529: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2530: {
1.327 ! schwarze 2531: roff_setregn(r, name, strlen(name), val, sign, INT_MIN);
1.326 schwarze 2532: }
2533:
2534: static void
2535: roff_setregn(struct roff *r, const char *name, size_t len,
1.327 ! schwarze 2536: int val, char sign, int step)
1.326 schwarze 2537: {
1.180 schwarze 2538: struct roffreg *reg;
2539:
2540: /* Search for an existing register with the same name. */
2541: reg = r->regtab;
2542:
1.326 schwarze 2543: while (reg != NULL && (reg->key.sz != len ||
2544: strncmp(reg->key.p, name, len) != 0))
1.180 schwarze 2545: reg = reg->next;
1.147 kristaps 2546:
1.180 schwarze 2547: if (NULL == reg) {
2548: /* Create a new register. */
2549: reg = mandoc_malloc(sizeof(struct roffreg));
1.326 schwarze 2550: reg->key.p = mandoc_strndup(name, len);
2551: reg->key.sz = len;
1.187 schwarze 2552: reg->val = 0;
1.327 ! schwarze 2553: reg->step = 0;
1.180 schwarze 2554: reg->next = r->regtab;
2555: r->regtab = reg;
2556: }
2557:
1.187 schwarze 2558: if ('+' == sign)
2559: reg->val += val;
2560: else if ('-' == sign)
2561: reg->val -= val;
2562: else
2563: reg->val = val;
1.327 ! schwarze 2564: if (step != INT_MIN)
! 2565: reg->step = step;
1.147 kristaps 2566: }
2567:
1.192 schwarze 2568: /*
2569: * Handle some predefined read-only number registers.
2570: * For now, return -1 if the requested register is not predefined;
2571: * in case a predefined read-only register having the value -1
2572: * were to turn up, another special value would have to be chosen.
2573: */
2574: static int
1.273 schwarze 2575: roff_getregro(const struct roff *r, const char *name)
1.192 schwarze 2576: {
2577:
2578: switch (*name) {
1.273 schwarze 2579: case '$': /* Number of arguments of the last macro evaluated. */
1.277 schwarze 2580: return r->argc;
1.207 schwarze 2581: case 'A': /* ASCII approximation mode is always off. */
1.277 schwarze 2582: return 0;
1.207 schwarze 2583: case 'g': /* Groff compatibility mode is always on. */
1.277 schwarze 2584: return 1;
1.207 schwarze 2585: case 'H': /* Fixed horizontal resolution. */
1.277 schwarze 2586: return 24;
1.207 schwarze 2587: case 'j': /* Always adjust left margin only. */
1.277 schwarze 2588: return 0;
1.207 schwarze 2589: case 'T': /* Some output device is always defined. */
1.277 schwarze 2590: return 1;
1.207 schwarze 2591: case 'V': /* Fixed vertical resolution. */
1.277 schwarze 2592: return 40;
1.192 schwarze 2593: default:
1.277 schwarze 2594: return -1;
1.192 schwarze 2595: }
2596: }
2597:
1.181 schwarze 2598: int
1.326 schwarze 2599: roff_getreg(struct roff *r, const char *name)
1.147 kristaps 2600: {
1.327 ! schwarze 2601: return roff_getregn(r, name, strlen(name), '\0');
1.181 schwarze 2602: }
2603:
2604: static int
1.327 ! schwarze 2605: roff_getregn(struct roff *r, const char *name, size_t len, char sign)
1.181 schwarze 2606: {
2607: struct roffreg *reg;
1.192 schwarze 2608: int val;
2609:
2610: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2611: val = roff_getregro(r, name + 1);
1.192 schwarze 2612: if (-1 != val)
1.277 schwarze 2613: return val;
1.192 schwarze 2614: }
1.181 schwarze 2615:
1.327 ! schwarze 2616: for (reg = r->regtab; reg; reg = reg->next) {
1.181 schwarze 2617: if (len == reg->key.sz &&
1.327 ! schwarze 2618: 0 == strncmp(name, reg->key.p, len)) {
! 2619: switch (sign) {
! 2620: case '+':
! 2621: reg->val += reg->step;
! 2622: break;
! 2623: case '-':
! 2624: reg->val -= reg->step;
! 2625: break;
! 2626: default:
! 2627: break;
! 2628: }
1.277 schwarze 2629: return reg->val;
1.327 ! schwarze 2630: }
! 2631: }
1.271 schwarze 2632:
1.327 ! schwarze 2633: roff_setregn(r, name, len, 0, '\0', INT_MIN);
1.277 schwarze 2634: return 0;
1.271 schwarze 2635: }
2636:
2637: static int
2638: roff_hasregn(const struct roff *r, const char *name, size_t len)
2639: {
2640: struct roffreg *reg;
2641: int val;
2642:
2643: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2644: val = roff_getregro(r, name + 1);
1.271 schwarze 2645: if (-1 != val)
1.277 schwarze 2646: return 1;
1.271 schwarze 2647: }
2648:
2649: for (reg = r->regtab; reg; reg = reg->next)
2650: if (len == reg->key.sz &&
2651: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2652: return 1;
1.147 kristaps 2653:
1.277 schwarze 2654: return 0;
1.147 kristaps 2655: }
2656:
1.180 schwarze 2657: static void
2658: roff_freereg(struct roffreg *reg)
1.147 kristaps 2659: {
1.180 schwarze 2660: struct roffreg *old_reg;
1.147 kristaps 2661:
1.180 schwarze 2662: while (NULL != reg) {
2663: free(reg->key.p);
2664: old_reg = reg;
2665: reg = reg->next;
2666: free(old_reg);
2667: }
1.147 kristaps 2668: }
1.92 schwarze 2669:
2670: static enum rofferr
1.89 kristaps 2671: roff_nr(ROFF_ARGS)
1.83 schwarze 2672: {
1.327 ! schwarze 2673: char *key, *val, *step;
1.212 schwarze 2674: size_t keysz;
1.327 ! schwarze 2675: int iv, is, len;
1.187 schwarze 2676: char sign;
1.89 kristaps 2677:
1.238 schwarze 2678: key = val = buf->buf + pos;
2679: if (*key == '\0')
1.277 schwarze 2680: return ROFF_IGN;
1.212 schwarze 2681:
2682: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2683: if (key[keysz] == '\\')
1.277 schwarze 2684: return ROFF_IGN;
1.89 kristaps 2685:
1.187 schwarze 2686: sign = *val;
1.238 schwarze 2687: if (sign == '+' || sign == '-')
1.187 schwarze 2688: val++;
2689:
1.327 ! schwarze 2690: len = 0;
! 2691: if (roff_evalnum(r, ln, val, &len, &iv, ROFFNUM_SCALE) == 0)
! 2692: return ROFF_IGN;
! 2693:
! 2694: step = val + len;
! 2695: while (isspace((unsigned char)*step))
! 2696: step++;
! 2697: if (roff_evalnum(r, ln, step, NULL, &is, 0) == 0)
! 2698: is = INT_MIN;
1.109 kristaps 2699:
1.327 ! schwarze 2700: roff_setregn(r, key, keysz, iv, sign, is);
1.277 schwarze 2701: return ROFF_IGN;
1.203 schwarze 2702: }
2703:
2704: static enum rofferr
2705: roff_rr(ROFF_ARGS)
2706: {
2707: struct roffreg *reg, **prev;
1.212 schwarze 2708: char *name, *cp;
2709: size_t namesz;
1.203 schwarze 2710:
1.238 schwarze 2711: name = cp = buf->buf + pos;
2712: if (*name == '\0')
1.277 schwarze 2713: return ROFF_IGN;
1.212 schwarze 2714: namesz = roff_getname(r, &cp, ln, pos);
2715: name[namesz] = '\0';
1.203 schwarze 2716:
2717: prev = &r->regtab;
2718: while (1) {
2719: reg = *prev;
1.238 schwarze 2720: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2721: break;
2722: prev = ®->next;
2723: }
1.238 schwarze 2724: if (reg != NULL) {
1.203 schwarze 2725: *prev = reg->next;
2726: free(reg->key.p);
2727: free(reg);
2728: }
1.277 schwarze 2729: return ROFF_IGN;
1.122 schwarze 2730: }
2731:
1.266 schwarze 2732: /* --- handler functions for roff requests -------------------------------- */
2733:
1.122 schwarze 2734: static enum rofferr
2735: roff_rm(ROFF_ARGS)
2736: {
2737: const char *name;
2738: char *cp;
1.212 schwarze 2739: size_t namesz;
1.122 schwarze 2740:
1.238 schwarze 2741: cp = buf->buf + pos;
2742: while (*cp != '\0') {
1.212 schwarze 2743: name = cp;
1.238 schwarze 2744: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2745: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.311 schwarze 2746: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.238 schwarze 2747: if (name[namesz] == '\\')
1.212 schwarze 2748: break;
1.122 schwarze 2749: }
1.277 schwarze 2750: return ROFF_IGN;
1.178 schwarze 2751: }
2752:
2753: static enum rofferr
2754: roff_it(ROFF_ARGS)
2755: {
2756: int iv;
2757:
2758: /* Parse the number of lines. */
1.261 schwarze 2759:
2760: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 2761: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2762: ln, ppos, buf->buf + 1);
1.277 schwarze 2763: return ROFF_IGN;
1.178 schwarze 2764: }
2765:
1.262 schwarze 2766: while (isspace((unsigned char)buf->buf[pos]))
2767: pos++;
2768:
2769: /*
2770: * Arm the input line trap.
2771: * Special-casing "an-trap" is an ugly workaround to cope
2772: * with DocBook stupidly fiddling with man(7) internals.
2773: */
1.261 schwarze 2774:
1.178 schwarze 2775: roffit_lines = iv;
1.262 schwarze 2776: roffit_macro = mandoc_strdup(iv != 1 ||
2777: strcmp(buf->buf + pos, "an-trap") ?
2778: buf->buf + pos : "br");
1.277 schwarze 2779: return ROFF_IGN;
1.175 schwarze 2780: }
2781:
2782: static enum rofferr
2783: roff_Dd(ROFF_ARGS)
2784: {
1.315 schwarze 2785: int mask;
2786: enum roff_tok t, te;
1.227 schwarze 2787:
1.315 schwarze 2788: switch (tok) {
2789: case ROFF_Dd:
2790: tok = MDOC_Dd;
2791: te = MDOC_MAX;
2792: if (r->format == 0)
2793: r->format = MPARSE_MDOC;
2794: mask = MPARSE_MDOC | MPARSE_QUICK;
2795: break;
2796: case ROFF_TH:
2797: tok = MAN_TH;
2798: te = MAN_MAX;
2799: if (r->format == 0)
2800: r->format = MPARSE_MAN;
2801: mask = MPARSE_QUICK;
2802: break;
2803: default:
2804: abort();
2805: }
2806: if ((r->options & mask) == 0)
2807: for (t = tok; t < te; t++)
2808: roff_setstr(r, roff_name[t], NULL, 0);
1.277 schwarze 2809: return ROFF_CONT;
1.109 kristaps 2810: }
2811:
2812: static enum rofferr
2813: roff_TE(ROFF_ARGS)
2814: {
1.321 schwarze 2815: if (r->tbl == NULL) {
1.221 schwarze 2816: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2817: ln, ppos, "TE");
1.321 schwarze 2818: return ROFF_IGN;
2819: }
2820: if (tbl_end(r->tbl) == 0) {
2821: r->tbl = NULL;
1.258 schwarze 2822: free(buf->buf);
2823: buf->buf = mandoc_strdup(".sp");
2824: buf->sz = 4;
1.277 schwarze 2825: return ROFF_REPARSE;
1.258 schwarze 2826: }
1.321 schwarze 2827: r->tbl = NULL;
1.277 schwarze 2828: return ROFF_IGN;
1.112 kristaps 2829: }
2830:
2831: static enum rofferr
2832: roff_T_(ROFF_ARGS)
2833: {
2834:
2835: if (NULL == r->tbl)
1.221 schwarze 2836: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2837: ln, ppos, "T&");
1.112 kristaps 2838: else
1.296 schwarze 2839: tbl_restart(ln, ppos, r->tbl);
1.112 kristaps 2840:
1.277 schwarze 2841: return ROFF_IGN;
1.109 kristaps 2842: }
2843:
1.230 schwarze 2844: /*
2845: * Handle in-line equation delimiters.
2846: */
2847: static enum rofferr
1.238 schwarze 2848: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2849: {
1.233 schwarze 2850: char *cp1, *cp2;
2851: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2852:
1.230 schwarze 2853: /*
2854: * Outside equations, look for an opening delimiter.
2855: * If we are inside an equation, we already know it is
2856: * in-line, or this function wouldn't have been called;
2857: * so look for a closing delimiter.
2858: */
2859:
1.238 schwarze 2860: cp1 = buf->buf + pos;
1.230 schwarze 2861: cp2 = strchr(cp1, r->eqn == NULL ?
2862: r->last_eqn->odelim : r->last_eqn->cdelim);
2863: if (cp2 == NULL)
1.277 schwarze 2864: return ROFF_CONT;
1.230 schwarze 2865:
1.233 schwarze 2866: *cp2++ = '\0';
2867: bef_pr = bef_nl = aft_nl = aft_pr = "";
2868:
2869: /* Handle preceding text, protecting whitespace. */
2870:
1.238 schwarze 2871: if (*buf->buf != '\0') {
1.233 schwarze 2872: if (r->eqn == NULL)
2873: bef_pr = "\\&";
2874: bef_nl = "\n";
2875: }
2876:
2877: /*
2878: * Prepare replacing the delimiter with an equation macro
2879: * and drop leading white space from the equation.
2880: */
1.230 schwarze 2881:
1.233 schwarze 2882: if (r->eqn == NULL) {
2883: while (*cp2 == ' ')
2884: cp2++;
2885: mac = ".EQ";
2886: } else
2887: mac = ".EN";
2888:
2889: /* Handle following text, protecting whitespace. */
2890:
2891: if (*cp2 != '\0') {
2892: aft_nl = "\n";
2893: if (r->eqn != NULL)
2894: aft_pr = "\\&";
2895: }
2896:
2897: /* Do the actual replacement. */
2898:
1.238 schwarze 2899: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2900: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2901: free(buf->buf);
2902: buf->buf = cp1;
1.230 schwarze 2903:
2904: /* Toggle the in-line state of the eqn subsystem. */
2905:
2906: r->eqn_inline = r->eqn == NULL;
1.277 schwarze 2907: return ROFF_REPARSE;
1.151 kristaps 2908: }
2909:
1.235 schwarze 2910: static enum rofferr
2911: roff_EQ(ROFF_ARGS)
1.125 kristaps 2912: {
1.319 schwarze 2913: struct roff_node *n;
2914:
1.322 schwarze 2915: if (r->man->macroset == MACROSET_MAN)
2916: man_breakscope(r->man, ROFF_EQ);
1.319 schwarze 2917: n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
2918: if (ln > r->man->last->line)
2919: n->flags |= NODE_LINE;
2920: n->eqn = mandoc_calloc(1, sizeof(*n->eqn));
2921: n->eqn->expectargs = UINT_MAX;
2922: roff_node_append(r->man, n);
2923: r->man->next = ROFF_NEXT_SIBLING;
1.125 kristaps 2924:
1.238 schwarze 2925: assert(r->eqn == NULL);
1.319 schwarze 2926: if (r->last_eqn == NULL)
2927: r->last_eqn = eqn_alloc(r->parse);
2928: else
2929: eqn_reset(r->last_eqn);
2930: r->eqn = r->last_eqn;
2931: r->eqn->node = n;
1.151 kristaps 2932:
1.238 schwarze 2933: if (buf->buf[pos] != '\0')
1.235 schwarze 2934: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 2935: ".EQ %s", buf->buf + pos);
1.151 kristaps 2936:
1.277 schwarze 2937: return ROFF_IGN;
1.125 kristaps 2938: }
2939:
2940: static enum rofferr
2941: roff_EN(ROFF_ARGS)
2942: {
1.319 schwarze 2943: if (r->eqn != NULL) {
2944: eqn_parse(r->eqn);
2945: r->eqn = NULL;
2946: } else
2947: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
2948: if (buf->buf[pos] != '\0')
2949: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
2950: "EN %s", buf->buf + pos);
1.277 schwarze 2951: return ROFF_IGN;
1.125 kristaps 2952: }
2953:
2954: static enum rofferr
1.109 kristaps 2955: roff_TS(ROFF_ARGS)
2956: {
1.321 schwarze 2957: if (r->tbl != NULL) {
1.221 schwarze 2958: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2959: ln, ppos, "TS breaks TS");
1.321 schwarze 2960: tbl_end(r->tbl);
1.115 kristaps 2961: }
1.321 schwarze 2962: r->tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 2963: if (r->last_tbl)
1.321 schwarze 2964: r->last_tbl->next = r->tbl;
1.113 kristaps 2965: else
1.321 schwarze 2966: r->first_tbl = r->tbl;
2967: r->last_tbl = r->tbl;
1.297 schwarze 2968: return ROFF_IGN;
2969: }
2970:
2971: static enum rofferr
2972: roff_onearg(ROFF_ARGS)
2973: {
2974: struct roff_node *n;
2975: char *cp;
1.305 schwarze 2976: int npos;
1.297 schwarze 2977:
1.302 schwarze 2978: if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.320 schwarze 2979: (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
2980: tok == ROFF_ti))
1.302 schwarze 2981: man_breakscope(r->man, tok);
2982:
1.309 schwarze 2983: if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.305 schwarze 2984: r->man->last = roffce_node;
2985: r->man->next = ROFF_NEXT_SIBLING;
2986: }
2987:
1.297 schwarze 2988: roff_elem_alloc(r->man, ln, ppos, tok);
2989: n = r->man->last;
2990:
2991: cp = buf->buf + pos;
2992: if (*cp != '\0') {
2993: while (*cp != '\0' && *cp != ' ')
2994: cp++;
2995: while (*cp == ' ')
2996: *cp++ = '\0';
2997: if (*cp != '\0')
2998: mandoc_vmsg(MANDOCERR_ARG_EXCESS,
2999: r->parse, ln, cp - buf->buf,
3000: "%s ... %s", roff_name[tok], cp);
3001: roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.300 schwarze 3002: }
3003:
1.309 schwarze 3004: if (tok == ROFF_ce || tok == ROFF_rj) {
3005: if (r->man->last->type == ROFFT_ELEM) {
1.305 schwarze 3006: roff_word_alloc(r->man, ln, pos, "1");
3007: r->man->last->flags |= NODE_NOSRC;
3008: }
3009: npos = 0;
3010: if (roff_evalnum(r, ln, r->man->last->string, &npos,
3011: &roffce_lines, 0) == 0) {
3012: mandoc_vmsg(MANDOCERR_CE_NONUM,
3013: r->parse, ln, pos, "ce %s", buf->buf + pos);
3014: roffce_lines = 1;
3015: }
3016: if (roffce_lines < 1) {
3017: r->man->last = r->man->last->parent;
3018: roffce_node = NULL;
3019: roffce_lines = 0;
3020: } else
3021: roffce_node = r->man->last->parent;
3022: } else {
3023: n->flags |= NODE_VALID | NODE_ENDED;
3024: r->man->last = n;
3025: }
3026: n->flags |= NODE_LINE;
1.300 schwarze 3027: r->man->next = ROFF_NEXT_SIBLING;
3028: return ROFF_IGN;
3029: }
3030:
3031: static enum rofferr
3032: roff_manyarg(ROFF_ARGS)
3033: {
3034: struct roff_node *n;
3035: char *sp, *ep;
3036:
3037: roff_elem_alloc(r->man, ln, ppos, tok);
3038: n = r->man->last;
3039:
3040: for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
3041: while (*ep != '\0' && *ep != ' ')
3042: ep++;
3043: while (*ep == ' ')
3044: *ep++ = '\0';
3045: roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.297 schwarze 3046: }
3047:
3048: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3049: r->man->last = n;
3050: r->man->next = ROFF_NEXT_SIBLING;
1.277 schwarze 3051: return ROFF_IGN;
1.251 schwarze 3052: }
3053:
3054: static enum rofferr
1.311 schwarze 3055: roff_als(ROFF_ARGS)
3056: {
3057: char *oldn, *newn, *end, *value;
3058: size_t oldsz, newsz, valsz;
3059:
3060: newn = oldn = buf->buf + pos;
3061: if (*newn == '\0')
3062: return ROFF_IGN;
3063:
3064: newsz = roff_getname(r, &oldn, ln, pos);
3065: if (newn[newsz] == '\\' || *oldn == '\0')
3066: return ROFF_IGN;
3067:
3068: end = oldn;
3069: oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
3070: if (oldsz == 0)
3071: return ROFF_IGN;
3072:
1.315 schwarze 3073: valsz = mandoc_asprintf(&value, ".%.*s \\$*\\\"\n",
3074: (int)oldsz, oldn);
1.311 schwarze 3075: roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
3076: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3077: free(value);
3078: return ROFF_IGN;
3079: }
3080:
3081: static enum rofferr
1.296 schwarze 3082: roff_br(ROFF_ARGS)
1.251 schwarze 3083: {
1.302 schwarze 3084: if (r->man->flags & (MAN_BLINE | MAN_ELINE))
3085: man_breakscope(r->man, ROFF_br);
1.296 schwarze 3086: roff_elem_alloc(r->man, ln, ppos, ROFF_br);
3087: if (buf->buf[pos] != '\0')
3088: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
3089: "%s %s", roff_name[tok], buf->buf + pos);
3090: r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3091: r->man->next = ROFF_NEXT_SIBLING;
3092: return ROFF_IGN;
1.92 schwarze 3093: }
3094:
1.105 kristaps 3095: static enum rofferr
1.174 kristaps 3096: roff_cc(ROFF_ARGS)
3097: {
3098: const char *p;
3099:
1.238 schwarze 3100: p = buf->buf + pos;
1.174 kristaps 3101:
1.238 schwarze 3102: if (*p == '\0' || (r->control = *p++) == '.')
1.303 schwarze 3103: r->control = '\0';
1.174 kristaps 3104:
1.238 schwarze 3105: if (*p != '\0')
1.260 schwarze 3106: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3107: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 3108:
1.277 schwarze 3109: return ROFF_IGN;
1.174 kristaps 3110: }
3111:
3112: static enum rofferr
1.303 schwarze 3113: roff_ec(ROFF_ARGS)
3114: {
3115: const char *p;
3116:
3117: p = buf->buf + pos;
3118: if (*p == '\0')
3119: r->escape = '\\';
3120: else {
3121: r->escape = *p;
3122: if (*++p != '\0')
3123: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3124: ln, p - buf->buf, "ec ... %s", p);
3125: }
3126: return ROFF_IGN;
3127: }
3128:
3129: static enum rofferr
3130: roff_eo(ROFF_ARGS)
3131: {
3132: r->escape = '\0';
3133: if (buf->buf[pos] != '\0')
3134: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
3135: ln, pos, "eo %s", buf->buf + pos);
3136: return ROFF_IGN;
3137: }
3138:
3139: static enum rofferr
1.164 kristaps 3140: roff_tr(ROFF_ARGS)
3141: {
3142: const char *p, *first, *second;
3143: size_t fsz, ssz;
3144: enum mandoc_esc esc;
3145:
1.238 schwarze 3146: p = buf->buf + pos;
1.164 kristaps 3147:
1.238 schwarze 3148: if (*p == '\0') {
1.260 schwarze 3149: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.277 schwarze 3150: return ROFF_IGN;
1.164 kristaps 3151: }
3152:
1.238 schwarze 3153: while (*p != '\0') {
1.164 kristaps 3154: fsz = ssz = 1;
3155:
3156: first = p++;
1.238 schwarze 3157: if (*first == '\\') {
1.164 kristaps 3158: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3159: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3160: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3161: ln, (int)(p - buf->buf), first);
1.277 schwarze 3162: return ROFF_IGN;
1.164 kristaps 3163: }
3164: fsz = (size_t)(p - first);
3165: }
3166:
3167: second = p++;
1.238 schwarze 3168: if (*second == '\\') {
1.164 kristaps 3169: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3170: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3171: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3172: ln, (int)(p - buf->buf), second);
1.277 schwarze 3173: return ROFF_IGN;
1.164 kristaps 3174: }
3175: ssz = (size_t)(p - second);
1.238 schwarze 3176: } else if (*second == '\0') {
1.260 schwarze 3177: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
3178: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 3179: second = " ";
1.165 kristaps 3180: p--;
1.164 kristaps 3181: }
3182:
1.167 kristaps 3183: if (fsz > 1) {
1.207 schwarze 3184: roff_setstrn(&r->xmbtab, first, fsz,
3185: second, ssz, 0);
1.167 kristaps 3186: continue;
3187: }
3188:
1.238 schwarze 3189: if (r->xtab == NULL)
1.207 schwarze 3190: r->xtab = mandoc_calloc(128,
3191: sizeof(struct roffstr));
1.167 kristaps 3192:
3193: free(r->xtab[(int)*first].p);
3194: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
3195: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 3196: }
3197:
1.277 schwarze 3198: return ROFF_IGN;
1.164 kristaps 3199: }
3200:
3201: static enum rofferr
1.306 schwarze 3202: roff_rn(ROFF_ARGS)
3203: {
3204: const char *value;
3205: char *oldn, *newn, *end;
3206: size_t oldsz, newsz;
1.315 schwarze 3207: int deftype;
1.306 schwarze 3208:
3209: oldn = newn = buf->buf + pos;
3210: if (*oldn == '\0')
3211: return ROFF_IGN;
3212:
3213: oldsz = roff_getname(r, &newn, ln, pos);
3214: if (oldn[oldsz] == '\\' || *newn == '\0')
3215: return ROFF_IGN;
3216:
3217: end = newn;
3218: newsz = roff_getname(r, &end, ln, newn - buf->buf);
3219: if (newsz == 0)
3220: return ROFF_IGN;
3221:
1.315 schwarze 3222: deftype = ROFFDEF_ANY;
3223: value = roff_getstrn(r, oldn, oldsz, &deftype);
3224: switch (deftype) {
3225: case ROFFDEF_USER:
1.306 schwarze 3226: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3227: roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
3228: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.315 schwarze 3229: break;
3230: case ROFFDEF_PRE:
3231: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3232: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3233: break;
3234: case ROFFDEF_REN:
1.306 schwarze 3235: roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
3236: roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.315 schwarze 3237: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3238: break;
3239: case ROFFDEF_STD:
1.306 schwarze 3240: roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.315 schwarze 3241: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3242: break;
3243: default:
3244: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3245: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3246: break;
3247: }
1.306 schwarze 3248: return ROFF_IGN;
3249: }
3250:
3251: static enum rofferr
1.105 kristaps 3252: roff_so(ROFF_ARGS)
3253: {
1.249 schwarze 3254: char *name, *cp;
1.105 kristaps 3255:
1.238 schwarze 3256: name = buf->buf + pos;
1.224 schwarze 3257: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 3258:
3259: /*
3260: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3261: * opening anything that's not in our cwd or anything beneath
3262: * it. Thus, explicitly disallow traversing up the file-system
3263: * or using absolute paths.
3264: */
3265:
1.238 schwarze 3266: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 3267: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3268: ".so %s", name);
1.249 schwarze 3269: buf->sz = mandoc_asprintf(&cp,
3270: ".sp\nSee the file %s.\n.sp", name) + 1;
3271: free(buf->buf);
3272: buf->buf = cp;
3273: *offs = 0;
1.277 schwarze 3274: return ROFF_REPARSE;
1.105 kristaps 3275: }
3276:
3277: *offs = pos;
1.277 schwarze 3278: return ROFF_SO;
1.105 kristaps 3279: }
1.92 schwarze 3280:
1.266 schwarze 3281: /* --- user defined strings and macros ------------------------------------ */
3282:
1.106 kristaps 3283: static enum rofferr
3284: roff_userdef(ROFF_ARGS)
1.99 kristaps 3285: {
1.311 schwarze 3286: const char *arg[16], *ap;
1.106 kristaps 3287: char *cp, *n1, *n2;
1.292 schwarze 3288: int expand_count, i, ib, ie;
1.263 schwarze 3289: size_t asz, rsz;
1.106 kristaps 3290:
3291: /*
3292: * Collect pointers to macro argument strings
1.188 schwarze 3293: * and NUL-terminate them.
1.106 kristaps 3294: */
1.263 schwarze 3295:
1.273 schwarze 3296: r->argc = 0;
1.238 schwarze 3297: cp = buf->buf + pos;
1.311 schwarze 3298: for (i = 0; i < 16; i++) {
1.273 schwarze 3299: if (*cp == '\0')
3300: arg[i] = "";
3301: else {
3302: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
3303: r->argc = i + 1;
3304: }
3305: }
1.99 kristaps 3306:
1.106 kristaps 3307: /*
3308: * Expand macro arguments.
1.99 kristaps 3309: */
1.263 schwarze 3310:
3311: buf->sz = strlen(r->current_string) + 1;
1.292 schwarze 3312: n1 = n2 = cp = mandoc_malloc(buf->sz);
1.263 schwarze 3313: memcpy(n1, r->current_string, buf->sz);
1.292 schwarze 3314: expand_count = 0;
1.263 schwarze 3315: while (*cp != '\0') {
3316:
3317: /* Scan ahead for the next argument invocation. */
3318:
3319: if (*cp++ != '\\')
3320: continue;
3321: if (*cp++ != '$')
3322: continue;
1.274 schwarze 3323: if (*cp == '*') { /* \\$* inserts all arguments */
3324: ib = 0;
3325: ie = r->argc - 1;
3326: } else { /* \\$1 .. \\$9 insert one argument */
3327: ib = ie = *cp - '1';
3328: if (ib < 0 || ib > 8)
3329: continue;
3330: }
1.263 schwarze 3331: cp -= 2;
1.292 schwarze 3332:
3333: /*
3334: * Prevent infinite recursion.
3335: */
3336:
3337: if (cp >= n2)
3338: expand_count = 1;
3339: else if (++expand_count > EXPAND_LIMIT) {
3340: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
3341: ln, (int)(cp - n1), NULL);
1.293 schwarze 3342: free(buf->buf);
3343: buf->buf = n1;
1.292 schwarze 3344: return ROFF_IGN;
3345: }
1.263 schwarze 3346:
3347: /*
3348: * Determine the size of the expanded argument,
3349: * taking escaping of quotes into account.
3350: */
3351:
1.274 schwarze 3352: asz = ie > ib ? ie - ib : 0; /* for blanks */
3353: for (i = ib; i <= ie; i++) {
3354: for (ap = arg[i]; *ap != '\0'; ap++) {
3355: asz++;
3356: if (*ap == '"')
3357: asz += 3;
3358: }
1.263 schwarze 3359: }
3360: if (asz != 3) {
3361:
3362: /*
3363: * Determine the size of the rest of the
3364: * unexpanded macro, including the NUL.
3365: */
3366:
3367: rsz = buf->sz - (cp - n1) - 3;
3368:
3369: /*
3370: * When shrinking, move before
3371: * releasing the storage.
3372: */
3373:
3374: if (asz < 3)
3375: memmove(cp + asz, cp + 3, rsz);
3376:
3377: /*
3378: * Resize the storage for the macro
3379: * and readjust the parse pointer.
3380: */
3381:
3382: buf->sz += asz - 3;
3383: n2 = mandoc_realloc(n1, buf->sz);
3384: cp = n2 + (cp - n1);
3385: n1 = n2;
3386:
3387: /*
3388: * When growing, make room
3389: * for the expanded argument.
3390: */
3391:
3392: if (asz > 3)
3393: memmove(cp + asz, cp + 3, rsz);
3394: }
3395:
3396: /* Copy the expanded argument, escaping quotes. */
3397:
3398: n2 = cp;
1.274 schwarze 3399: for (i = ib; i <= ie; i++) {
3400: for (ap = arg[i]; *ap != '\0'; ap++) {
3401: if (*ap == '"') {
3402: memcpy(n2, "\\(dq", 4);
3403: n2 += 4;
3404: } else
3405: *n2++ = *ap;
3406: }
3407: if (i < ie)
3408: *n2++ = ' ';
1.106 kristaps 3409: }
1.99 kristaps 3410: }
3411:
1.106 kristaps 3412: /*
3413: * Replace the macro invocation
3414: * by the expanded macro.
3415: */
1.263 schwarze 3416:
1.238 schwarze 3417: free(buf->buf);
3418: buf->buf = n1;
1.248 schwarze 3419: *offs = 0;
1.106 kristaps 3420:
1.277 schwarze 3421: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
3422: ROFF_REPARSE : ROFF_APPEND;
1.99 kristaps 3423: }
1.121 schwarze 3424:
1.306 schwarze 3425: /*
3426: * Calling a high-level macro that was renamed with .rn.
3427: * r->current_string has already been set up by roff_parse().
3428: */
3429: static enum rofferr
3430: roff_renamed(ROFF_ARGS)
3431: {
3432: char *nbuf;
3433:
1.315 schwarze 3434: buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
3435: buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.306 schwarze 3436: free(buf->buf);
3437: buf->buf = nbuf;
3438: return ROFF_CONT;
3439: }
3440:
1.212 schwarze 3441: static size_t
1.121 schwarze 3442: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3443: {
3444: char *name, *cp;
1.212 schwarze 3445: size_t namesz;
1.121 schwarze 3446:
3447: name = *cpp;
3448: if ('\0' == *name)
1.277 schwarze 3449: return 0;
1.121 schwarze 3450:
1.212 schwarze 3451: /* Read until end of name and terminate it with NUL. */
3452: for (cp = name; 1; cp++) {
3453: if ('\0' == *cp || ' ' == *cp) {
3454: namesz = cp - name;
3455: break;
3456: }
1.121 schwarze 3457: if ('\\' != *cp)
3458: continue;
1.215 schwarze 3459: namesz = cp - name;
3460: if ('{' == cp[1] || '}' == cp[1])
3461: break;
1.121 schwarze 3462: cp++;
3463: if ('\\' == *cp)
3464: continue;
1.224 schwarze 3465: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3466: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 3467: mandoc_escape((const char **)&cp, NULL, NULL);
3468: break;
1.121 schwarze 3469: }
3470:
3471: /* Read past spaces. */
3472: while (' ' == *cp)
3473: cp++;
3474:
3475: *cpp = cp;
1.277 schwarze 3476: return namesz;
1.121 schwarze 3477: }
3478:
1.106 kristaps 3479: /*
3480: * Store *string into the user-defined string called *name.
3481: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 3482: * append == 0: replace mode
3483: * append == 1: single-line append mode
3484: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 3485: */
1.94 kristaps 3486: static void
1.106 kristaps 3487: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 3488: int append)
1.92 schwarze 3489: {
1.315 schwarze 3490: size_t namesz;
1.164 kristaps 3491:
1.315 schwarze 3492: namesz = strlen(name);
3493: roff_setstrn(&r->strtab, name, namesz, string,
1.207 schwarze 3494: string ? strlen(string) : 0, append);
1.315 schwarze 3495: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.164 kristaps 3496: }
3497:
3498: static void
1.166 kristaps 3499: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 3500: const char *string, size_t stringsz, int append)
1.164 kristaps 3501: {
1.166 kristaps 3502: struct roffkv *n;
1.164 kristaps 3503: char *c;
3504: int i;
3505: size_t oldch, newch;
1.92 schwarze 3506:
1.106 kristaps 3507: /* Search for an existing string with the same name. */
1.164 kristaps 3508: n = *r;
3509:
1.211 schwarze 3510: while (n && (namesz != n->key.sz ||
3511: strncmp(n->key.p, name, namesz)))
1.92 schwarze 3512: n = n->next;
1.94 kristaps 3513:
3514: if (NULL == n) {
1.106 kristaps 3515: /* Create a new string table entry. */
1.166 kristaps 3516: n = mandoc_malloc(sizeof(struct roffkv));
3517: n->key.p = mandoc_strndup(name, namesz);
3518: n->key.sz = namesz;
3519: n->val.p = NULL;
3520: n->val.sz = 0;
1.164 kristaps 3521: n->next = *r;
3522: *r = n;
1.193 schwarze 3523: } else if (0 == append) {
1.166 kristaps 3524: free(n->val.p);
3525: n->val.p = NULL;
3526: n->val.sz = 0;
1.106 kristaps 3527: }
3528:
3529: if (NULL == string)
3530: return;
3531:
3532: /*
3533: * One additional byte for the '\n' in multiline mode,
3534: * and one for the terminating '\0'.
3535: */
1.193 schwarze 3536: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 3537:
1.166 kristaps 3538: if (NULL == n->val.p) {
3539: n->val.p = mandoc_malloc(newch);
3540: *n->val.p = '\0';
1.106 kristaps 3541: oldch = 0;
3542: } else {
1.166 kristaps 3543: oldch = n->val.sz;
3544: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 3545: }
3546:
3547: /* Skip existing content in the destination buffer. */
1.166 kristaps 3548: c = n->val.p + (int)oldch;
1.106 kristaps 3549:
3550: /* Append new content to the destination buffer. */
1.164 kristaps 3551: i = 0;
3552: while (i < (int)stringsz) {
1.106 kristaps 3553: /*
3554: * Rudimentary roff copy mode:
3555: * Handle escaped backslashes.
3556: */
1.164 kristaps 3557: if ('\\' == string[i] && '\\' == string[i + 1])
3558: i++;
3559: *c++ = string[i++];
1.106 kristaps 3560: }
1.94 kristaps 3561:
1.106 kristaps 3562: /* Append terminating bytes. */
1.193 schwarze 3563: if (1 < append)
1.106 kristaps 3564: *c++ = '\n';
1.163 kristaps 3565:
1.106 kristaps 3566: *c = '\0';
1.166 kristaps 3567: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 3568: }
3569:
1.94 kristaps 3570: static const char *
1.325 schwarze 3571: roff_getstrn(struct roff *r, const char *name, size_t len,
1.315 schwarze 3572: int *deftype)
1.92 schwarze 3573: {
1.315 schwarze 3574: const struct roffkv *n;
1.325 schwarze 3575: int found, i;
1.315 schwarze 3576: enum roff_tok tok;
3577:
1.325 schwarze 3578: found = 0;
3579: for (n = r->strtab; n != NULL; n = n->next) {
3580: if (strncmp(name, n->key.p, len) != 0 ||
3581: n->key.p[len] != '\0' || n->val.p == NULL)
3582: continue;
3583: if (*deftype & ROFFDEF_USER) {
3584: *deftype = ROFFDEF_USER;
3585: return n->val.p;
3586: } else {
3587: found = 1;
3588: break;
3589: }
3590: }
3591: for (n = r->rentab; n != NULL; n = n->next) {
3592: if (strncmp(name, n->key.p, len) != 0 ||
3593: n->key.p[len] != '\0' || n->val.p == NULL)
3594: continue;
3595: if (*deftype & ROFFDEF_REN) {
3596: *deftype = ROFFDEF_REN;
3597: return n->val.p;
3598: } else {
3599: found = 1;
3600: break;
1.315 schwarze 3601: }
3602: }
1.325 schwarze 3603: for (i = 0; i < PREDEFS_MAX; i++) {
3604: if (strncmp(name, predefs[i].name, len) != 0 ||
3605: predefs[i].name[len] != '\0')
3606: continue;
3607: if (*deftype & ROFFDEF_PRE) {
3608: *deftype = ROFFDEF_PRE;
3609: return predefs[i].str;
3610: } else {
3611: found = 1;
3612: break;
1.315 schwarze 3613: }
3614: }
1.325 schwarze 3615: if (r->man->macroset != MACROSET_MAN) {
3616: for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
3617: if (strncmp(name, roff_name[tok], len) != 0 ||
3618: roff_name[tok][len] != '\0')
3619: continue;
3620: if (*deftype & ROFFDEF_STD) {
3621: *deftype = ROFFDEF_STD;
3622: return NULL;
3623: } else {
3624: found = 1;
3625: break;
1.315 schwarze 3626: }
3627: }
3628: }
1.325 schwarze 3629: if (r->man->macroset != MACROSET_MDOC) {
3630: for (tok = MAN_TH; tok < MAN_MAX; tok++) {
3631: if (strncmp(name, roff_name[tok], len) != 0 ||
3632: roff_name[tok][len] != '\0')
3633: continue;
3634: if (*deftype & ROFFDEF_STD) {
3635: *deftype = ROFFDEF_STD;
3636: return NULL;
3637: } else {
3638: found = 1;
3639: break;
1.315 schwarze 3640: }
3641: }
1.325 schwarze 3642: }
3643:
3644: if (found == 0 && *deftype != ROFFDEF_ANY) {
3645: if (*deftype & ROFFDEF_REN) {
3646: /*
3647: * This might still be a request,
3648: * so do not treat it as undefined yet.
3649: */
3650: *deftype = ROFFDEF_UNDEF;
3651: return NULL;
1.315 schwarze 3652: }
1.325 schwarze 3653:
3654: /* Using an undefined string defines it to be empty. */
3655:
3656: roff_setstrn(&r->strtab, name, len, "", 0, 0);
3657: roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.315 schwarze 3658: }
1.325 schwarze 3659:
1.315 schwarze 3660: *deftype = 0;
1.277 schwarze 3661: return NULL;
1.92 schwarze 3662: }
3663:
1.94 kristaps 3664: static void
1.167 kristaps 3665: roff_freestr(struct roffkv *r)
1.92 schwarze 3666: {
1.166 kristaps 3667: struct roffkv *n, *nn;
1.92 schwarze 3668:
1.167 kristaps 3669: for (n = r; n; n = nn) {
1.166 kristaps 3670: free(n->key.p);
3671: free(n->val.p);
1.92 schwarze 3672: nn = n->next;
3673: free(n);
3674: }
1.114 kristaps 3675: }
1.266 schwarze 3676:
3677: /* --- accessors and utility functions ------------------------------------ */
1.164 kristaps 3678:
3679: /*
3680: * Duplicate an input string, making the appropriate character
3681: * conversations (as stipulated by `tr') along the way.
3682: * Returns a heap-allocated string with all the replacements made.
3683: */
3684: char *
3685: roff_strdup(const struct roff *r, const char *p)
3686: {
1.166 kristaps 3687: const struct roffkv *cp;
1.164 kristaps 3688: char *res;
3689: const char *pp;
3690: size_t ssz, sz;
3691: enum mandoc_esc esc;
3692:
1.167 kristaps 3693: if (NULL == r->xmbtab && NULL == r->xtab)
1.277 schwarze 3694: return mandoc_strdup(p);
1.164 kristaps 3695: else if ('\0' == *p)
1.277 schwarze 3696: return mandoc_strdup("");
1.164 kristaps 3697:
3698: /*
3699: * Step through each character looking for term matches
3700: * (remember that a `tr' can be invoked with an escape, which is
3701: * a glyph but the escape is multi-character).
3702: * We only do this if the character hash has been initialised
3703: * and the string is >0 length.
3704: */
3705:
3706: res = NULL;
3707: ssz = 0;
3708:
3709: while ('\0' != *p) {
1.289 schwarze 3710: assert((unsigned int)*p < 128);
3711: if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167 kristaps 3712: sz = r->xtab[(int)*p].sz;
3713: res = mandoc_realloc(res, ssz + sz + 1);
3714: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3715: ssz += sz;
3716: p++;
3717: continue;
3718: } else if ('\\' != *p) {
3719: res = mandoc_realloc(res, ssz + 2);
3720: res[ssz++] = *p++;
3721: continue;
3722: }
3723:
1.164 kristaps 3724: /* Search for term matches. */
1.167 kristaps 3725: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 3726: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 3727: break;
3728:
3729: if (NULL != cp) {
3730: /*
3731: * A match has been found.
3732: * Append the match to the array and move
3733: * forward by its keysize.
3734: */
1.207 schwarze 3735: res = mandoc_realloc(res,
3736: ssz + cp->val.sz + 1);
1.166 kristaps 3737: memcpy(res + ssz, cp->val.p, cp->val.sz);
3738: ssz += cp->val.sz;
3739: p += (int)cp->key.sz;
1.164 kristaps 3740: continue;
3741: }
3742:
1.167 kristaps 3743: /*
3744: * Handle escapes carefully: we need to copy
3745: * over just the escape itself, or else we might
3746: * do replacements within the escape itself.
3747: * Make sure to pass along the bogus string.
3748: */
3749: pp = p++;
3750: esc = mandoc_escape(&p, NULL, NULL);
3751: if (ESCAPE_ERROR == esc) {
3752: sz = strlen(pp);
1.164 kristaps 3753: res = mandoc_realloc(res, ssz + sz + 1);
3754: memcpy(res + ssz, pp, sz);
1.167 kristaps 3755: break;
1.164 kristaps 3756: }
1.207 schwarze 3757: /*
3758: * We bail out on bad escapes.
1.167 kristaps 3759: * No need to warn: we already did so when
3760: * roff_res() was called.
3761: */
3762: sz = (int)(p - pp);
3763: res = mandoc_realloc(res, ssz + sz + 1);
3764: memcpy(res + ssz, pp, sz);
3765: ssz += sz;
1.164 kristaps 3766: }
3767:
3768: res[(int)ssz] = '\0';
1.277 schwarze 3769: return res;
1.227 schwarze 3770: }
3771:
3772: int
3773: roff_getformat(const struct roff *r)
3774: {
3775:
1.277 schwarze 3776: return r->format;
1.174 kristaps 3777: }
3778:
3779: /*
1.207 schwarze 3780: * Find out whether a line is a macro line or not.
1.174 kristaps 3781: * If it is, adjust the current position and return one; if it isn't,
3782: * return zero and don't change the current position.
3783: * If the control character has been set with `.cc', then let that grain
3784: * precedence.
3785: * This is slighly contrary to groff, where using the non-breaking
3786: * control character when `cc' has been invoked will cause the
3787: * non-breaking macro contents to be printed verbatim.
3788: */
3789: int
3790: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3791: {
3792: int pos;
3793:
3794: pos = *ppos;
3795:
1.303 schwarze 3796: if (r->control != '\0' && cp[pos] == r->control)
1.174 kristaps 3797: pos++;
1.303 schwarze 3798: else if (r->control != '\0')
1.277 schwarze 3799: return 0;
1.174 kristaps 3800: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3801: pos += 2;
3802: else if ('.' == cp[pos] || '\'' == cp[pos])
3803: pos++;
3804: else
1.277 schwarze 3805: return 0;
1.174 kristaps 3806:
3807: while (' ' == cp[pos] || '\t' == cp[pos])
3808: pos++;
3809:
3810: *ppos = pos;
1.277 schwarze 3811: return 1;
1.74 kristaps 3812: }
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