Annotation of mandoc/roff.c, Revision 1.344
1.344 ! schwarze 1: /* $Id: roff.c,v 1.343 2018/11/26 17:44:34 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:
1.339 schwarze 88: /*
89: * A macro processing context.
90: * More than one is needed when macro calls are nested.
91: */
92: struct mctx {
93: char **argv;
94: int argc;
95: int argsz;
96: };
97:
1.67 kristaps 98: struct roff {
1.128 kristaps 99: struct mparse *parse; /* parse point */
1.296 schwarze 100: struct roff_man *man; /* mdoc or man parser */
1.67 kristaps 101: struct roffnode *last; /* leaf of stack */
1.339 schwarze 102: struct mctx *mstack; /* stack of macro contexts */
1.223 schwarze 103: int *rstack; /* stack of inverted `ie' values */
1.295 schwarze 104: struct ohash *reqtab; /* request lookup table */
1.180 schwarze 105: struct roffreg *regtab; /* number registers */
1.166 kristaps 106: struct roffkv *strtab; /* user-defined strings & macros */
1.306 schwarze 107: struct roffkv *rentab; /* renamed strings & macros */
1.167 kristaps 108: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
109: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.106 kristaps 110: const char *current_string; /* value of last called user macro */
1.118 kristaps 111: struct tbl_node *first_tbl; /* first table parsed */
112: struct tbl_node *last_tbl; /* last table parsed */
113: struct tbl_node *tbl; /* current table being parsed */
1.319 schwarze 114: struct eqn_node *last_eqn; /* equation parser */
115: struct eqn_node *eqn; /* active equation parser */
1.230 schwarze 116: int eqn_inline; /* current equation is inline */
1.223 schwarze 117: int options; /* parse options */
1.339 schwarze 118: int mstacksz; /* current size of mstack */
119: int mstackpos; /* position in mstack */
1.223 schwarze 120: int rstacksz; /* current size limit of rstack */
121: int rstackpos; /* position in rstack */
1.227 schwarze 122: int format; /* current file in mdoc or man format */
1.223 schwarze 123: char control; /* control character */
1.303 schwarze 124: char escape; /* escape character */
1.79 kristaps 125: };
126:
1.67 kristaps 127: struct roffnode {
1.294 schwarze 128: enum roff_tok tok; /* type of node */
1.67 kristaps 129: struct roffnode *parent; /* up one in stack */
130: int line; /* parse line */
131: int col; /* parse col */
1.106 kristaps 132: char *name; /* node name, e.g. macro name */
1.79 kristaps 133: char *end; /* end-rules: custom token */
134: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 135: int rule; /* current evaluation rule */
1.67 kristaps 136: };
137:
138: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.294 schwarze 139: enum roff_tok tok, /* tok of macro */ \
1.238 schwarze 140: struct buf *buf, /* input buffer */ \
1.67 kristaps 141: int ln, /* parse line */ \
1.75 kristaps 142: int ppos, /* original pos in buffer */ \
143: int pos, /* current pos in buffer */ \
1.74 kristaps 144: int *offs /* reset offset of buffer data */
1.67 kristaps 145:
1.340 schwarze 146: typedef int (*roffproc)(ROFF_ARGS);
1.67 kristaps 147:
148: struct roffmac {
1.79 kristaps 149: roffproc proc; /* process new macro */
150: roffproc text; /* process as child text of macro */
151: roffproc sub; /* process as child of macro */
152: int flags;
153: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.67 kristaps 154: };
155:
1.141 kristaps 156: struct predef {
157: const char *name; /* predefined input name */
158: const char *str; /* replacement symbol */
159: };
160:
161: #define PREDEF(__name, __str) \
162: { (__name), (__str) },
163:
1.266 schwarze 164: /* --- function prototypes ------------------------------------------------ */
165:
1.340 schwarze 166: static int roffnode_cleanscope(struct roff *);
167: static int roffnode_pop(struct roff *);
1.294 schwarze 168: static void roffnode_push(struct roff *, enum roff_tok,
1.155 kristaps 169: const char *, int, int);
1.321 schwarze 170: static void roff_addtbl(struct roff_man *, struct tbl_node *);
1.340 schwarze 171: static int roff_als(ROFF_ARGS);
172: static int roff_block(ROFF_ARGS);
173: static int roff_block_text(ROFF_ARGS);
174: static int roff_block_sub(ROFF_ARGS);
175: static int roff_br(ROFF_ARGS);
176: static int roff_cblock(ROFF_ARGS);
177: static int roff_cc(ROFF_ARGS);
178: static int roff_ccond(struct roff *, int, int);
1.341 schwarze 179: static int roff_char(ROFF_ARGS);
1.340 schwarze 180: static int roff_cond(ROFF_ARGS);
181: static int roff_cond_text(ROFF_ARGS);
182: static int roff_cond_sub(ROFF_ARGS);
183: static int roff_ds(ROFF_ARGS);
184: static int roff_ec(ROFF_ARGS);
185: static int roff_eo(ROFF_ARGS);
186: static int roff_eqndelim(struct roff *, struct buf *, int);
1.271 schwarze 187: static int roff_evalcond(struct roff *r, int, char *, int *);
1.234 kristaps 188: static int roff_evalnum(struct roff *, int,
189: const char *, int *, int *, int);
190: static int roff_evalpar(struct roff *, int,
1.261 schwarze 191: const char *, int *, int *, int);
1.198 schwarze 192: static int roff_evalstrcond(const char *, int *);
1.155 kristaps 193: static void roff_free1(struct roff *);
1.180 schwarze 194: static void roff_freereg(struct roffreg *);
1.167 kristaps 195: static void roff_freestr(struct roffkv *);
1.212 schwarze 196: static size_t roff_getname(struct roff *, char **, int, int);
1.261 schwarze 197: static int roff_getnum(const char *, int *, int *, int);
1.184 schwarze 198: static int roff_getop(const char *, int *, char *);
1.327 schwarze 199: static int roff_getregn(struct roff *,
200: const char *, size_t, char);
1.273 schwarze 201: static int roff_getregro(const struct roff *,
202: const char *name);
1.325 schwarze 203: static const char *roff_getstrn(struct roff *,
1.315 schwarze 204: const char *, size_t, int *);
1.271 schwarze 205: static int roff_hasregn(const struct roff *,
206: const char *, size_t);
1.340 schwarze 207: static int roff_insec(ROFF_ARGS);
208: static int roff_it(ROFF_ARGS);
209: static int roff_line_ignore(ROFF_ARGS);
1.265 schwarze 210: static void roff_man_alloc1(struct roff_man *);
211: static void roff_man_free1(struct roff_man *);
1.340 schwarze 212: static int roff_manyarg(ROFF_ARGS);
213: static int roff_nop(ROFF_ARGS);
214: static int roff_nr(ROFF_ARGS);
215: static int roff_onearg(ROFF_ARGS);
1.294 schwarze 216: static enum roff_tok roff_parse(struct roff *, char *, int *,
1.214 schwarze 217: int, int);
1.340 schwarze 218: static int roff_parsetext(struct roff *, struct buf *,
1.305 schwarze 219: int, int *);
1.340 schwarze 220: static int roff_renamed(ROFF_ARGS);
221: static int roff_res(struct roff *, struct buf *, int, int);
222: static int roff_return(ROFF_ARGS);
223: static int roff_rm(ROFF_ARGS);
224: static int roff_rn(ROFF_ARGS);
225: static int roff_rr(ROFF_ARGS);
1.326 schwarze 226: static void roff_setregn(struct roff *, const char *,
1.327 schwarze 227: size_t, int, char, int);
1.94 kristaps 228: static void roff_setstr(struct roff *,
1.106 kristaps 229: const char *, const char *, int);
1.207 schwarze 230: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 231: size_t, const char *, size_t, int);
1.340 schwarze 232: static int roff_shift(ROFF_ARGS);
233: static int roff_so(ROFF_ARGS);
234: static int roff_tr(ROFF_ARGS);
235: static int roff_Dd(ROFF_ARGS);
236: static int roff_TE(ROFF_ARGS);
237: static int roff_TS(ROFF_ARGS);
238: static int roff_EQ(ROFF_ARGS);
239: static int roff_EN(ROFF_ARGS);
240: static int roff_T_(ROFF_ARGS);
241: static int roff_unsupp(ROFF_ARGS);
242: static int roff_userdef(ROFF_ARGS);
1.67 kristaps 243:
1.266 schwarze 244: /* --- constant data ------------------------------------------------------ */
245:
1.261 schwarze 246: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
247: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
248:
1.294 schwarze 249: const char *__roff_name[MAN_MAX + 1] = {
1.305 schwarze 250: "br", "ce", "ft", "ll",
1.312 schwarze 251: "mc", "po", "rj", "sp",
252: "ta", "ti", NULL,
1.294 schwarze 253: "ab", "ad", "af", "aln",
254: "als", "am", "am1", "ami",
255: "ami1", "as", "as1", "asciify",
256: "backtrace", "bd", "bleedat", "blm",
257: "box", "boxa", "bp", "BP",
258: "break", "breakchar", "brnl", "brp",
1.305 schwarze 259: "brpnl", "c2", "cc",
1.294 schwarze 260: "cf", "cflags", "ch", "char",
261: "chop", "class", "close", "CL",
262: "color", "composite", "continue", "cp",
263: "cropat", "cs", "cu", "da",
264: "dch", "Dd", "de", "de1",
265: "defcolor", "dei", "dei1", "device",
266: "devicem", "di", "do", "ds",
267: "ds1", "dwh", "dt", "ec",
268: "ecr", "ecs", "el", "em",
269: "EN", "eo", "EP", "EQ",
270: "errprint", "ev", "evc", "ex",
271: "fallback", "fam", "fc", "fchar",
272: "fcolor", "fdeferlig", "feature", "fkern",
273: "fl", "flig", "fp", "fps",
274: "fschar", "fspacewidth", "fspecial", "ftr",
275: "fzoom", "gcolor", "hc", "hcode",
276: "hidechar", "hla", "hlm", "hpf",
277: "hpfa", "hpfcode", "hw", "hy",
278: "hylang", "hylen", "hym", "hypp",
279: "hys", "ie", "if", "ig",
280: "index", "it", "itc", "IX",
281: "kern", "kernafter", "kernbefore", "kernpair",
282: "lc", "lc_ctype", "lds", "length",
283: "letadj", "lf", "lg", "lhang",
284: "linetabs", "lnr", "lnrf", "lpfx",
1.304 schwarze 285: "ls", "lsm", "lt",
1.294 schwarze 286: "mediasize", "minss", "mk", "mso",
287: "na", "ne", "nh", "nhychar",
288: "nm", "nn", "nop", "nr",
289: "nrf", "nroff", "ns", "nx",
290: "open", "opena", "os", "output",
291: "padj", "papersize", "pc", "pev",
292: "pi", "PI", "pl", "pm",
1.312 schwarze 293: "pn", "pnr", "ps",
1.294 schwarze 294: "psbb", "pshape", "pso", "ptr",
295: "pvs", "rchar", "rd", "recursionlimit",
1.309 schwarze 296: "return", "rfschar", "rhang",
1.294 schwarze 297: "rm", "rn", "rnn", "rr",
298: "rs", "rt", "schar", "sentchar",
299: "shc", "shift", "sizes", "so",
300: "spacewidth", "special", "spreadwarn", "ss",
301: "sty", "substring", "sv", "sy",
1.300 schwarze 302: "T&", "tc", "TE",
1.301 schwarze 303: "TH", "tkf", "tl",
1.294 schwarze 304: "tm", "tm1", "tmc", "tr",
305: "track", "transchar", "trf", "trimat",
306: "trin", "trnt", "troff", "TS",
307: "uf", "ul", "unformat", "unwatch",
308: "unwatchn", "vpt", "vs", "warn",
309: "warnscale", "watch", "watchlength", "watchn",
310: "wh", "while", "write", "writec",
311: "writem", "xflag", ".", NULL,
1.306 schwarze 312: NULL, "text",
1.294 schwarze 313: "Dd", "Dt", "Os", "Sh",
314: "Ss", "Pp", "D1", "Dl",
315: "Bd", "Ed", "Bl", "El",
316: "It", "Ad", "An", "Ap",
317: "Ar", "Cd", "Cm", "Dv",
318: "Er", "Ev", "Ex", "Fa",
319: "Fd", "Fl", "Fn", "Ft",
320: "Ic", "In", "Li", "Nd",
321: "Nm", "Op", "Ot", "Pa",
322: "Rv", "St", "Va", "Vt",
323: "Xr", "%A", "%B", "%D",
324: "%I", "%J", "%N", "%O",
325: "%P", "%R", "%T", "%V",
326: "Ac", "Ao", "Aq", "At",
327: "Bc", "Bf", "Bo", "Bq",
328: "Bsx", "Bx", "Db", "Dc",
329: "Do", "Dq", "Ec", "Ef",
330: "Em", "Eo", "Fx", "Ms",
331: "No", "Ns", "Nx", "Ox",
332: "Pc", "Pf", "Po", "Pq",
333: "Qc", "Ql", "Qo", "Qq",
334: "Re", "Rs", "Sc", "So",
335: "Sq", "Sm", "Sx", "Sy",
336: "Tn", "Ux", "Xc", "Xo",
337: "Fo", "Fc", "Oo", "Oc",
338: "Bk", "Ek", "Bt", "Hf",
339: "Fr", "Ud", "Lb", "Lp",
340: "Lk", "Mt", "Brq", "Bro",
341: "Brc", "%C", "Es", "En",
1.299 schwarze 342: "Dx", "%Q", "%U", "Ta",
343: NULL,
1.294 schwarze 344: "TH", "SH", "SS", "TP",
1.332 schwarze 345: "TQ",
1.294 schwarze 346: "LP", "PP", "P", "IP",
347: "HP", "SM", "SB", "BI",
348: "IB", "BR", "RB", "R",
349: "B", "I", "IR", "RI",
1.299 schwarze 350: "nf", "fi",
1.294 schwarze 351: "RE", "RS", "DT", "UC",
1.297 schwarze 352: "PD", "AT", "in",
1.333 schwarze 353: "SY", "YS", "OP",
354: "EX", "EE", "UR",
1.317 schwarze 355: "UE", "MT", "ME", NULL
1.294 schwarze 356: };
357: const char *const *roff_name = __roff_name;
358:
359: static struct roffmac roffs[TOKEN_NONE] = {
1.296 schwarze 360: { roff_br, NULL, NULL, 0 }, /* br */
1.305 schwarze 361: { roff_onearg, NULL, NULL, 0 }, /* ce */
1.297 schwarze 362: { roff_onearg, NULL, NULL, 0 }, /* ft */
1.298 schwarze 363: { roff_onearg, NULL, NULL, 0 }, /* ll */
1.304 schwarze 364: { roff_onearg, NULL, NULL, 0 }, /* mc */
1.312 schwarze 365: { roff_onearg, NULL, NULL, 0 }, /* po */
1.309 schwarze 366: { roff_onearg, NULL, NULL, 0 }, /* rj */
1.299 schwarze 367: { roff_onearg, NULL, NULL, 0 }, /* sp */
1.300 schwarze 368: { roff_manyarg, NULL, NULL, 0 }, /* ta */
1.301 schwarze 369: { roff_onearg, NULL, NULL, 0 }, /* ti */
1.296 schwarze 370: { NULL, NULL, NULL, 0 }, /* ROFF_MAX */
1.295 schwarze 371: { roff_unsupp, NULL, NULL, 0 }, /* ab */
372: { roff_line_ignore, NULL, NULL, 0 }, /* ad */
373: { roff_line_ignore, NULL, NULL, 0 }, /* af */
374: { roff_unsupp, NULL, NULL, 0 }, /* aln */
1.311 schwarze 375: { roff_als, NULL, NULL, 0 }, /* als */
1.295 schwarze 376: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am */
377: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am1 */
378: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami */
379: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami1 */
380: { roff_ds, NULL, NULL, 0 }, /* as */
381: { roff_ds, NULL, NULL, 0 }, /* as1 */
382: { roff_unsupp, NULL, NULL, 0 }, /* asciify */
383: { roff_line_ignore, NULL, NULL, 0 }, /* backtrace */
384: { roff_line_ignore, NULL, NULL, 0 }, /* bd */
385: { roff_line_ignore, NULL, NULL, 0 }, /* bleedat */
386: { roff_unsupp, NULL, NULL, 0 }, /* blm */
387: { roff_unsupp, NULL, NULL, 0 }, /* box */
388: { roff_unsupp, NULL, NULL, 0 }, /* boxa */
389: { roff_line_ignore, NULL, NULL, 0 }, /* bp */
390: { roff_unsupp, NULL, NULL, 0 }, /* BP */
391: { roff_unsupp, NULL, NULL, 0 }, /* break */
392: { roff_line_ignore, NULL, NULL, 0 }, /* breakchar */
393: { roff_line_ignore, NULL, NULL, 0 }, /* brnl */
1.296 schwarze 394: { roff_br, NULL, NULL, 0 }, /* brp */
1.295 schwarze 395: { roff_line_ignore, NULL, NULL, 0 }, /* brpnl */
396: { roff_unsupp, NULL, NULL, 0 }, /* c2 */
397: { roff_cc, NULL, NULL, 0 }, /* cc */
398: { roff_insec, NULL, NULL, 0 }, /* cf */
399: { roff_line_ignore, NULL, NULL, 0 }, /* cflags */
400: { roff_line_ignore, NULL, NULL, 0 }, /* ch */
1.341 schwarze 401: { roff_char, NULL, NULL, 0 }, /* char */
1.295 schwarze 402: { roff_unsupp, NULL, NULL, 0 }, /* chop */
403: { roff_line_ignore, NULL, NULL, 0 }, /* class */
404: { roff_insec, NULL, NULL, 0 }, /* close */
405: { roff_unsupp, NULL, NULL, 0 }, /* CL */
406: { roff_line_ignore, NULL, NULL, 0 }, /* color */
407: { roff_unsupp, NULL, NULL, 0 }, /* composite */
408: { roff_unsupp, NULL, NULL, 0 }, /* continue */
409: { roff_line_ignore, NULL, NULL, 0 }, /* cp */
410: { roff_line_ignore, NULL, NULL, 0 }, /* cropat */
411: { roff_line_ignore, NULL, NULL, 0 }, /* cs */
412: { roff_line_ignore, NULL, NULL, 0 }, /* cu */
413: { roff_unsupp, NULL, NULL, 0 }, /* da */
414: { roff_unsupp, NULL, NULL, 0 }, /* dch */
415: { roff_Dd, NULL, NULL, 0 }, /* Dd */
416: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de */
417: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de1 */
418: { roff_line_ignore, NULL, NULL, 0 }, /* defcolor */
419: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei */
420: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei1 */
421: { roff_unsupp, NULL, NULL, 0 }, /* device */
422: { roff_unsupp, NULL, NULL, 0 }, /* devicem */
423: { roff_unsupp, NULL, NULL, 0 }, /* di */
424: { roff_unsupp, NULL, NULL, 0 }, /* do */
425: { roff_ds, NULL, NULL, 0 }, /* ds */
426: { roff_ds, NULL, NULL, 0 }, /* ds1 */
427: { roff_unsupp, NULL, NULL, 0 }, /* dwh */
428: { roff_unsupp, NULL, NULL, 0 }, /* dt */
1.303 schwarze 429: { roff_ec, NULL, NULL, 0 }, /* ec */
1.295 schwarze 430: { roff_unsupp, NULL, NULL, 0 }, /* ecr */
431: { roff_unsupp, NULL, NULL, 0 }, /* ecs */
432: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* el */
433: { roff_unsupp, NULL, NULL, 0 }, /* em */
434: { roff_EN, NULL, NULL, 0 }, /* EN */
1.303 schwarze 435: { roff_eo, NULL, NULL, 0 }, /* eo */
1.295 schwarze 436: { roff_unsupp, NULL, NULL, 0 }, /* EP */
437: { roff_EQ, NULL, NULL, 0 }, /* EQ */
438: { roff_line_ignore, NULL, NULL, 0 }, /* errprint */
439: { roff_unsupp, NULL, NULL, 0 }, /* ev */
440: { roff_unsupp, NULL, NULL, 0 }, /* evc */
441: { roff_unsupp, NULL, NULL, 0 }, /* ex */
442: { roff_line_ignore, NULL, NULL, 0 }, /* fallback */
443: { roff_line_ignore, NULL, NULL, 0 }, /* fam */
444: { roff_unsupp, NULL, NULL, 0 }, /* fc */
445: { roff_unsupp, NULL, NULL, 0 }, /* fchar */
446: { roff_line_ignore, NULL, NULL, 0 }, /* fcolor */
447: { roff_line_ignore, NULL, NULL, 0 }, /* fdeferlig */
448: { roff_line_ignore, NULL, NULL, 0 }, /* feature */
449: { roff_line_ignore, NULL, NULL, 0 }, /* fkern */
450: { roff_line_ignore, NULL, NULL, 0 }, /* fl */
451: { roff_line_ignore, NULL, NULL, 0 }, /* flig */
452: { roff_line_ignore, NULL, NULL, 0 }, /* fp */
453: { roff_line_ignore, NULL, NULL, 0 }, /* fps */
454: { roff_unsupp, NULL, NULL, 0 }, /* fschar */
455: { roff_line_ignore, NULL, NULL, 0 }, /* fspacewidth */
456: { roff_line_ignore, NULL, NULL, 0 }, /* fspecial */
457: { roff_line_ignore, NULL, NULL, 0 }, /* ftr */
458: { roff_line_ignore, NULL, NULL, 0 }, /* fzoom */
459: { roff_line_ignore, NULL, NULL, 0 }, /* gcolor */
460: { roff_line_ignore, NULL, NULL, 0 }, /* hc */
461: { roff_line_ignore, NULL, NULL, 0 }, /* hcode */
462: { roff_line_ignore, NULL, NULL, 0 }, /* hidechar */
463: { roff_line_ignore, NULL, NULL, 0 }, /* hla */
464: { roff_line_ignore, NULL, NULL, 0 }, /* hlm */
465: { roff_line_ignore, NULL, NULL, 0 }, /* hpf */
466: { roff_line_ignore, NULL, NULL, 0 }, /* hpfa */
467: { roff_line_ignore, NULL, NULL, 0 }, /* hpfcode */
468: { roff_line_ignore, NULL, NULL, 0 }, /* hw */
469: { roff_line_ignore, NULL, NULL, 0 }, /* hy */
470: { roff_line_ignore, NULL, NULL, 0 }, /* hylang */
471: { roff_line_ignore, NULL, NULL, 0 }, /* hylen */
472: { roff_line_ignore, NULL, NULL, 0 }, /* hym */
473: { roff_line_ignore, NULL, NULL, 0 }, /* hypp */
474: { roff_line_ignore, NULL, NULL, 0 }, /* hys */
475: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* ie */
476: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* if */
477: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ig */
478: { roff_unsupp, NULL, NULL, 0 }, /* index */
479: { roff_it, NULL, NULL, 0 }, /* it */
480: { roff_unsupp, NULL, NULL, 0 }, /* itc */
481: { roff_line_ignore, NULL, NULL, 0 }, /* IX */
482: { roff_line_ignore, NULL, NULL, 0 }, /* kern */
483: { roff_line_ignore, NULL, NULL, 0 }, /* kernafter */
484: { roff_line_ignore, NULL, NULL, 0 }, /* kernbefore */
485: { roff_line_ignore, NULL, NULL, 0 }, /* kernpair */
486: { roff_unsupp, NULL, NULL, 0 }, /* lc */
487: { roff_unsupp, NULL, NULL, 0 }, /* lc_ctype */
488: { roff_unsupp, NULL, NULL, 0 }, /* lds */
489: { roff_unsupp, NULL, NULL, 0 }, /* length */
490: { roff_line_ignore, NULL, NULL, 0 }, /* letadj */
491: { roff_insec, NULL, NULL, 0 }, /* lf */
492: { roff_line_ignore, NULL, NULL, 0 }, /* lg */
493: { roff_line_ignore, NULL, NULL, 0 }, /* lhang */
494: { roff_unsupp, NULL, NULL, 0 }, /* linetabs */
495: { roff_unsupp, NULL, NULL, 0 }, /* lnr */
496: { roff_unsupp, NULL, NULL, 0 }, /* lnrf */
497: { roff_unsupp, NULL, NULL, 0 }, /* lpfx */
498: { roff_line_ignore, NULL, NULL, 0 }, /* ls */
499: { roff_unsupp, NULL, NULL, 0 }, /* lsm */
500: { roff_line_ignore, NULL, NULL, 0 }, /* lt */
501: { roff_line_ignore, NULL, NULL, 0 }, /* mediasize */
502: { roff_line_ignore, NULL, NULL, 0 }, /* minss */
503: { roff_line_ignore, NULL, NULL, 0 }, /* mk */
504: { roff_insec, NULL, NULL, 0 }, /* mso */
505: { roff_line_ignore, NULL, NULL, 0 }, /* na */
506: { roff_line_ignore, NULL, NULL, 0 }, /* ne */
507: { roff_line_ignore, NULL, NULL, 0 }, /* nh */
508: { roff_line_ignore, NULL, NULL, 0 }, /* nhychar */
509: { roff_unsupp, NULL, NULL, 0 }, /* nm */
510: { roff_unsupp, NULL, NULL, 0 }, /* nn */
1.330 schwarze 511: { roff_nop, NULL, NULL, 0 }, /* nop */
1.295 schwarze 512: { roff_nr, NULL, NULL, 0 }, /* nr */
513: { roff_unsupp, NULL, NULL, 0 }, /* nrf */
514: { roff_line_ignore, NULL, NULL, 0 }, /* nroff */
515: { roff_line_ignore, NULL, NULL, 0 }, /* ns */
516: { roff_insec, NULL, NULL, 0 }, /* nx */
517: { roff_insec, NULL, NULL, 0 }, /* open */
518: { roff_insec, NULL, NULL, 0 }, /* opena */
519: { roff_line_ignore, NULL, NULL, 0 }, /* os */
520: { roff_unsupp, NULL, NULL, 0 }, /* output */
521: { roff_line_ignore, NULL, NULL, 0 }, /* padj */
522: { roff_line_ignore, NULL, NULL, 0 }, /* papersize */
523: { roff_line_ignore, NULL, NULL, 0 }, /* pc */
524: { roff_line_ignore, NULL, NULL, 0 }, /* pev */
525: { roff_insec, NULL, NULL, 0 }, /* pi */
526: { roff_unsupp, NULL, NULL, 0 }, /* PI */
527: { roff_line_ignore, NULL, NULL, 0 }, /* pl */
528: { roff_line_ignore, NULL, NULL, 0 }, /* pm */
529: { roff_line_ignore, NULL, NULL, 0 }, /* pn */
530: { roff_line_ignore, NULL, NULL, 0 }, /* pnr */
531: { roff_line_ignore, NULL, NULL, 0 }, /* ps */
532: { roff_unsupp, NULL, NULL, 0 }, /* psbb */
533: { roff_unsupp, NULL, NULL, 0 }, /* pshape */
534: { roff_insec, NULL, NULL, 0 }, /* pso */
535: { roff_line_ignore, NULL, NULL, 0 }, /* ptr */
536: { roff_line_ignore, NULL, NULL, 0 }, /* pvs */
537: { roff_unsupp, NULL, NULL, 0 }, /* rchar */
538: { roff_line_ignore, NULL, NULL, 0 }, /* rd */
539: { roff_line_ignore, NULL, NULL, 0 }, /* recursionlimit */
1.339 schwarze 540: { roff_return, NULL, NULL, 0 }, /* return */
1.295 schwarze 541: { roff_unsupp, NULL, NULL, 0 }, /* rfschar */
542: { roff_line_ignore, NULL, NULL, 0 }, /* rhang */
543: { roff_rm, NULL, NULL, 0 }, /* rm */
1.306 schwarze 544: { roff_rn, NULL, NULL, 0 }, /* rn */
1.295 schwarze 545: { roff_unsupp, NULL, NULL, 0 }, /* rnn */
546: { roff_rr, NULL, NULL, 0 }, /* rr */
547: { roff_line_ignore, NULL, NULL, 0 }, /* rs */
548: { roff_line_ignore, NULL, NULL, 0 }, /* rt */
549: { roff_unsupp, NULL, NULL, 0 }, /* schar */
550: { roff_line_ignore, NULL, NULL, 0 }, /* sentchar */
551: { roff_line_ignore, NULL, NULL, 0 }, /* shc */
1.339 schwarze 552: { roff_shift, NULL, NULL, 0 }, /* shift */
1.295 schwarze 553: { roff_line_ignore, NULL, NULL, 0 }, /* sizes */
554: { roff_so, NULL, NULL, 0 }, /* so */
555: { roff_line_ignore, NULL, NULL, 0 }, /* spacewidth */
556: { roff_line_ignore, NULL, NULL, 0 }, /* special */
557: { roff_line_ignore, NULL, NULL, 0 }, /* spreadwarn */
558: { roff_line_ignore, NULL, NULL, 0 }, /* ss */
559: { roff_line_ignore, NULL, NULL, 0 }, /* sty */
560: { roff_unsupp, NULL, NULL, 0 }, /* substring */
561: { roff_line_ignore, NULL, NULL, 0 }, /* sv */
562: { roff_insec, NULL, NULL, 0 }, /* sy */
563: { roff_T_, NULL, NULL, 0 }, /* T& */
564: { roff_unsupp, NULL, NULL, 0 }, /* tc */
565: { roff_TE, NULL, NULL, 0 }, /* TE */
1.315 schwarze 566: { roff_Dd, NULL, NULL, 0 }, /* TH */
1.295 schwarze 567: { roff_line_ignore, NULL, NULL, 0 }, /* tkf */
568: { roff_unsupp, NULL, NULL, 0 }, /* tl */
569: { roff_line_ignore, NULL, NULL, 0 }, /* tm */
570: { roff_line_ignore, NULL, NULL, 0 }, /* tm1 */
571: { roff_line_ignore, NULL, NULL, 0 }, /* tmc */
572: { roff_tr, NULL, NULL, 0 }, /* tr */
573: { roff_line_ignore, NULL, NULL, 0 }, /* track */
574: { roff_line_ignore, NULL, NULL, 0 }, /* transchar */
575: { roff_insec, NULL, NULL, 0 }, /* trf */
576: { roff_line_ignore, NULL, NULL, 0 }, /* trimat */
577: { roff_unsupp, NULL, NULL, 0 }, /* trin */
578: { roff_unsupp, NULL, NULL, 0 }, /* trnt */
579: { roff_line_ignore, NULL, NULL, 0 }, /* troff */
580: { roff_TS, NULL, NULL, 0 }, /* TS */
581: { roff_line_ignore, NULL, NULL, 0 }, /* uf */
582: { roff_line_ignore, NULL, NULL, 0 }, /* ul */
583: { roff_unsupp, NULL, NULL, 0 }, /* unformat */
584: { roff_line_ignore, NULL, NULL, 0 }, /* unwatch */
585: { roff_line_ignore, NULL, NULL, 0 }, /* unwatchn */
586: { roff_line_ignore, NULL, NULL, 0 }, /* vpt */
587: { roff_line_ignore, NULL, NULL, 0 }, /* vs */
588: { roff_line_ignore, NULL, NULL, 0 }, /* warn */
589: { roff_line_ignore, NULL, NULL, 0 }, /* warnscale */
590: { roff_line_ignore, NULL, NULL, 0 }, /* watch */
591: { roff_line_ignore, NULL, NULL, 0 }, /* watchlength */
592: { roff_line_ignore, NULL, NULL, 0 }, /* watchn */
593: { roff_unsupp, NULL, NULL, 0 }, /* wh */
1.340 schwarze 594: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /*while*/
1.295 schwarze 595: { roff_insec, NULL, NULL, 0 }, /* write */
596: { roff_insec, NULL, NULL, 0 }, /* writec */
597: { roff_insec, NULL, NULL, 0 }, /* writem */
598: { roff_line_ignore, NULL, NULL, 0 }, /* xflag */
599: { roff_cblock, NULL, NULL, 0 }, /* . */
1.306 schwarze 600: { roff_renamed, NULL, NULL, 0 },
1.295 schwarze 601: { roff_userdef, NULL, NULL, 0 }
1.67 kristaps 602: };
603:
1.141 kristaps 604: /* Array of injected predefined strings. */
605: #define PREDEFS_MAX 38
606: static const struct predef predefs[PREDEFS_MAX] = {
607: #include "predefs.in"
608: };
609:
1.305 schwarze 610: static int roffce_lines; /* number of input lines to center */
611: static struct roff_node *roffce_node; /* active request */
1.178 schwarze 612: static int roffit_lines; /* number of lines to delay */
613: static char *roffit_macro; /* nil-terminated macro line */
614:
1.207 schwarze 615:
1.266 schwarze 616: /* --- request table ------------------------------------------------------ */
617:
1.295 schwarze 618: struct ohash *
619: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85 kristaps 620: {
1.295 schwarze 621: struct ohash *htab;
622: struct roffreq *req;
623: enum roff_tok tok;
624: size_t sz;
625: unsigned int slot;
1.85 kristaps 626:
1.295 schwarze 627: htab = mandoc_malloc(sizeof(*htab));
628: mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85 kristaps 629:
1.295 schwarze 630: for (tok = mintok; tok < maxtok; tok++) {
1.296 schwarze 631: if (roff_name[tok] == NULL)
632: continue;
1.295 schwarze 633: sz = strlen(roff_name[tok]);
634: req = mandoc_malloc(sizeof(*req) + sz + 1);
635: req->tok = tok;
636: memcpy(req->name, roff_name[tok], sz + 1);
637: slot = ohash_qlookup(htab, req->name);
638: ohash_insert(htab, slot, req);
1.85 kristaps 639: }
1.295 schwarze 640: return htab;
1.85 kristaps 641: }
642:
1.295 schwarze 643: void
644: roffhash_free(struct ohash *htab)
1.67 kristaps 645: {
1.295 schwarze 646: struct roffreq *req;
647: unsigned int slot;
1.67 kristaps 648:
1.295 schwarze 649: if (htab == NULL)
650: return;
651: for (req = ohash_first(htab, &slot); req != NULL;
652: req = ohash_next(htab, &slot))
653: free(req);
654: ohash_delete(htab);
655: free(htab);
656: }
657:
658: enum roff_tok
659: roffhash_find(struct ohash *htab, const char *name, size_t sz)
660: {
661: struct roffreq *req;
662: const char *end;
663:
664: if (sz) {
665: end = name + sz;
666: req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
667: } else
668: req = ohash_find(htab, ohash_qlookup(htab, name));
669: return req == NULL ? TOKEN_NONE : req->tok;
1.67 kristaps 670: }
671:
1.266 schwarze 672: /* --- stack of request blocks -------------------------------------------- */
673:
1.67 kristaps 674: /*
675: * Pop the current node off of the stack of roff instructions currently
676: * pending.
677: */
1.340 schwarze 678: static int
1.67 kristaps 679: roffnode_pop(struct roff *r)
680: {
681: struct roffnode *p;
1.340 schwarze 682: int inloop;
1.67 kristaps 683:
1.207 schwarze 684: p = r->last;
1.340 schwarze 685: inloop = p->tok == ROFF_while;
686: r->last = p->parent;
1.106 kristaps 687: free(p->name);
688: free(p->end);
1.67 kristaps 689: free(p);
1.340 schwarze 690: return inloop;
1.67 kristaps 691: }
692:
693: /*
694: * Push a roff node onto the instruction stack. This must later be
695: * removed with roffnode_pop().
696: */
1.98 schwarze 697: static void
1.294 schwarze 698: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106 kristaps 699: int line, int col)
1.67 kristaps 700: {
701: struct roffnode *p;
702:
1.98 schwarze 703: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 704: p->tok = tok;
1.106 kristaps 705: if (name)
706: p->name = mandoc_strdup(name);
1.67 kristaps 707: p->parent = r->last;
708: p->line = line;
709: p->col = col;
1.198 schwarze 710: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 711:
712: r->last = p;
713: }
714:
1.266 schwarze 715: /* --- roff parser state data management ---------------------------------- */
716:
1.67 kristaps 717: static void
718: roff_free1(struct roff *r)
719: {
1.176 schwarze 720: struct tbl_node *tbl;
1.167 kristaps 721: int i;
1.67 kristaps 722:
1.176 schwarze 723: while (NULL != (tbl = r->first_tbl)) {
724: r->first_tbl = tbl->next;
725: tbl_free(tbl);
1.109 kristaps 726: }
1.113 kristaps 727: r->first_tbl = r->last_tbl = r->tbl = NULL;
728:
1.319 schwarze 729: if (r->last_eqn != NULL)
730: eqn_free(r->last_eqn);
731: r->last_eqn = r->eqn = NULL;
1.125 kristaps 732:
1.339 schwarze 733: while (r->mstackpos >= 0)
734: roff_userret(r);
735:
1.67 kristaps 736: while (r->last)
737: roffnode_pop(r);
1.109 kristaps 738:
1.223 schwarze 739: free (r->rstack);
740: r->rstack = NULL;
741: r->rstacksz = 0;
742: r->rstackpos = -1;
743:
744: roff_freereg(r->regtab);
745: r->regtab = NULL;
746:
1.167 kristaps 747: roff_freestr(r->strtab);
1.306 schwarze 748: roff_freestr(r->rentab);
1.167 kristaps 749: roff_freestr(r->xmbtab);
1.306 schwarze 750: r->strtab = r->rentab = r->xmbtab = NULL;
1.167 kristaps 751:
752: if (r->xtab)
753: for (i = 0; i < 128; i++)
754: free(r->xtab[i].p);
755: free(r->xtab);
756: r->xtab = NULL;
1.67 kristaps 757: }
758:
759: void
760: roff_reset(struct roff *r)
761: {
762: roff_free1(r);
1.227 schwarze 763: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.303 schwarze 764: r->control = '\0';
765: r->escape = '\\';
1.307 schwarze 766: roffce_lines = 0;
767: roffce_node = NULL;
768: roffit_lines = 0;
769: roffit_macro = NULL;
1.67 kristaps 770: }
771:
772: void
773: roff_free(struct roff *r)
774: {
1.339 schwarze 775: int i;
776:
1.67 kristaps 777: roff_free1(r);
1.339 schwarze 778: for (i = 0; i < r->mstacksz; i++)
779: free(r->mstack[i].argv);
780: free(r->mstack);
1.295 schwarze 781: roffhash_free(r->reqtab);
1.67 kristaps 782: free(r);
783: }
784:
785: struct roff *
1.279 schwarze 786: roff_alloc(struct mparse *parse, int options)
1.67 kristaps 787: {
788: struct roff *r;
789:
1.98 schwarze 790: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 791: r->parse = parse;
1.328 schwarze 792: r->reqtab = roffhash_alloc(0, ROFF_RENAMED);
1.199 schwarze 793: r->options = options;
1.227 schwarze 794: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.339 schwarze 795: r->mstackpos = -1;
1.82 kristaps 796: r->rstackpos = -1;
1.303 schwarze 797: r->escape = '\\';
1.277 schwarze 798: return r;
1.265 schwarze 799: }
800:
1.266 schwarze 801: /* --- syntax tree state data management ---------------------------------- */
802:
1.265 schwarze 803: static void
804: roff_man_free1(struct roff_man *man)
805: {
806:
1.266 schwarze 807: if (man->first != NULL)
808: roff_node_delete(man, man->first);
1.265 schwarze 809: free(man->meta.msec);
810: free(man->meta.vol);
811: free(man->meta.os);
812: free(man->meta.arch);
813: free(man->meta.title);
814: free(man->meta.name);
815: free(man->meta.date);
816: }
817:
818: static void
819: roff_man_alloc1(struct roff_man *man)
820: {
821:
822: memset(&man->meta, 0, sizeof(man->meta));
823: man->first = mandoc_calloc(1, sizeof(*man->first));
824: man->first->type = ROFFT_ROOT;
825: man->last = man->first;
826: man->last_es = NULL;
827: man->flags = 0;
828: man->macroset = MACROSET_NONE;
829: man->lastsec = man->lastnamed = SEC_NONE;
830: man->next = ROFF_NEXT_CHILD;
831: }
832:
833: void
834: roff_man_reset(struct roff_man *man)
835: {
836:
837: roff_man_free1(man);
838: roff_man_alloc1(man);
839: }
840:
841: void
842: roff_man_free(struct roff_man *man)
843: {
844:
845: roff_man_free1(man);
846: free(man);
847: }
848:
849: struct roff_man *
850: roff_man_alloc(struct roff *roff, struct mparse *parse,
1.316 schwarze 851: const char *os_s, int quick)
1.265 schwarze 852: {
853: struct roff_man *man;
854:
855: man = mandoc_calloc(1, sizeof(*man));
856: man->parse = parse;
857: man->roff = roff;
1.316 schwarze 858: man->os_s = os_s;
1.265 schwarze 859: man->quick = quick;
860: roff_man_alloc1(man);
1.296 schwarze 861: roff->man = man;
1.277 schwarze 862: return man;
1.67 kristaps 863: }
864:
1.266 schwarze 865: /* --- syntax tree handling ----------------------------------------------- */
866:
867: struct roff_node *
868: roff_node_alloc(struct roff_man *man, int line, int pos,
869: enum roff_type type, int tok)
870: {
871: struct roff_node *n;
872:
873: n = mandoc_calloc(1, sizeof(*n));
874: n->line = line;
875: n->pos = pos;
876: n->tok = tok;
877: n->type = type;
878: n->sec = man->lastsec;
879:
880: if (man->flags & MDOC_SYNOPSIS)
1.285 schwarze 881: n->flags |= NODE_SYNPRETTY;
1.266 schwarze 882: else
1.285 schwarze 883: n->flags &= ~NODE_SYNPRETTY;
1.266 schwarze 884: if (man->flags & MDOC_NEWLINE)
1.285 schwarze 885: n->flags |= NODE_LINE;
1.266 schwarze 886: man->flags &= ~MDOC_NEWLINE;
887:
1.277 schwarze 888: return n;
1.266 schwarze 889: }
890:
891: void
892: roff_node_append(struct roff_man *man, struct roff_node *n)
893: {
894:
895: switch (man->next) {
896: case ROFF_NEXT_SIBLING:
1.281 schwarze 897: if (man->last->next != NULL) {
898: n->next = man->last->next;
899: man->last->next->prev = n;
900: } else
901: man->last->parent->last = n;
1.266 schwarze 902: man->last->next = n;
903: n->prev = man->last;
904: n->parent = man->last->parent;
905: break;
906: case ROFF_NEXT_CHILD:
1.287 schwarze 907: if (man->last->child != NULL) {
908: n->next = man->last->child;
909: man->last->child->prev = n;
910: } else
911: man->last->last = n;
1.266 schwarze 912: man->last->child = n;
913: n->parent = man->last;
914: break;
915: default:
916: abort();
917: }
1.282 schwarze 918: man->last = n;
1.266 schwarze 919:
920: switch (n->type) {
921: case ROFFT_HEAD:
922: n->parent->head = n;
923: break;
924: case ROFFT_BODY:
1.282 schwarze 925: if (n->end != ENDBODY_NOT)
926: return;
1.266 schwarze 927: n->parent->body = n;
928: break;
929: case ROFFT_TAIL:
930: n->parent->tail = n;
931: break;
932: default:
1.282 schwarze 933: return;
1.266 schwarze 934: }
1.282 schwarze 935:
936: /*
937: * Copy over the normalised-data pointer of our parent. Not
938: * everybody has one, but copying a null pointer is fine.
939: */
940:
941: n->norm = n->parent->norm;
942: assert(n->parent->type == ROFFT_BLOCK);
1.266 schwarze 943: }
944:
1.267 schwarze 945: void
946: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
947: {
948: struct roff_node *n;
949:
950: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
951: n->string = roff_strdup(man->roff, word);
952: roff_node_append(man, n);
1.286 schwarze 953: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 954: man->next = ROFF_NEXT_SIBLING;
955: }
956:
957: void
958: roff_word_append(struct roff_man *man, const char *word)
959: {
960: struct roff_node *n;
961: char *addstr, *newstr;
962:
963: n = man->last;
964: addstr = roff_strdup(man->roff, word);
965: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
966: free(addstr);
967: free(n->string);
968: n->string = newstr;
969: man->next = ROFF_NEXT_SIBLING;
1.268 schwarze 970: }
971:
972: void
973: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
974: {
975: struct roff_node *n;
976:
977: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
978: roff_node_append(man, n);
979: man->next = ROFF_NEXT_CHILD;
980: }
981:
982: struct roff_node *
983: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
984: {
985: struct roff_node *n;
986:
987: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
988: roff_node_append(man, n);
989: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 990: return n;
1.267 schwarze 991: }
992:
1.266 schwarze 993: struct roff_node *
994: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
995: {
996: struct roff_node *n;
997:
998: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
999: roff_node_append(man, n);
1000: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 1001: return n;
1.266 schwarze 1002: }
1003:
1004: struct roff_node *
1005: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
1006: {
1007: struct roff_node *n;
1008:
1009: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
1010: roff_node_append(man, n);
1011: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 1012: return n;
1.267 schwarze 1013: }
1014:
1.321 schwarze 1015: static void
1016: roff_addtbl(struct roff_man *man, struct tbl_node *tbl)
1.267 schwarze 1017: {
1018: struct roff_node *n;
1.321 schwarze 1019: const struct tbl_span *span;
1.267 schwarze 1020:
1021: if (man->macroset == MACROSET_MAN)
1.302 schwarze 1022: man_breakscope(man, ROFF_TS);
1.321 schwarze 1023: while ((span = tbl_span(tbl)) != NULL) {
1024: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
1025: n->span = span;
1026: roff_node_append(man, n);
1027: n->flags |= NODE_VALID | NODE_ENDED;
1028: man->next = ROFF_NEXT_SIBLING;
1029: }
1.266 schwarze 1030: }
1031:
1032: void
1033: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1034: {
1035:
1036: /* Adjust siblings. */
1037:
1038: if (n->prev)
1039: n->prev->next = n->next;
1040: if (n->next)
1041: n->next->prev = n->prev;
1042:
1043: /* Adjust parent. */
1044:
1045: if (n->parent != NULL) {
1046: if (n->parent->child == n)
1047: n->parent->child = n->next;
1048: if (n->parent->last == n)
1049: n->parent->last = n->prev;
1050: }
1051:
1052: /* Adjust parse point. */
1053:
1054: if (man == NULL)
1055: return;
1056: if (man->last == n) {
1057: if (n->prev == NULL) {
1058: man->last = n->parent;
1059: man->next = ROFF_NEXT_CHILD;
1060: } else {
1061: man->last = n->prev;
1062: man->next = ROFF_NEXT_SIBLING;
1063: }
1064: }
1065: if (man->first == n)
1066: man->first = NULL;
1.344 ! schwarze 1067: }
! 1068:
! 1069: void
! 1070: roff_node_relink(struct roff_man *man, struct roff_node *n)
! 1071: {
! 1072: roff_node_unlink(man, n);
! 1073: n->prev = n->next = NULL;
! 1074: roff_node_append(man, n);
1.266 schwarze 1075: }
1076:
1077: void
1078: roff_node_free(struct roff_node *n)
1079: {
1080:
1081: if (n->args != NULL)
1082: mdoc_argv_free(n->args);
1083: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1084: free(n->norm);
1.319 schwarze 1085: if (n->eqn != NULL)
1086: eqn_box_free(n->eqn);
1.266 schwarze 1087: free(n->string);
1088: free(n);
1089: }
1090:
1091: void
1092: roff_node_delete(struct roff_man *man, struct roff_node *n)
1093: {
1094:
1095: while (n->child != NULL)
1096: roff_node_delete(man, n->child);
1097: roff_node_unlink(man, n);
1098: roff_node_free(n);
1.269 schwarze 1099: }
1100:
1101: void
1102: deroff(char **dest, const struct roff_node *n)
1103: {
1104: char *cp;
1105: size_t sz;
1106:
1107: if (n->type != ROFFT_TEXT) {
1108: for (n = n->child; n != NULL; n = n->next)
1109: deroff(dest, n);
1110: return;
1111: }
1112:
1.288 schwarze 1113: /* Skip leading whitespace. */
1.269 schwarze 1114:
1.288 schwarze 1115: for (cp = n->string; *cp != '\0'; cp++) {
1.289 schwarze 1116: if (cp[0] == '\\' && cp[1] != '\0' &&
1117: strchr(" %&0^|~", cp[1]) != NULL)
1.269 schwarze 1118: cp++;
1.288 schwarze 1119: else if ( ! isspace((unsigned char)*cp))
1.269 schwarze 1120: break;
1121: }
1122:
1.289 schwarze 1123: /* Skip trailing backslash. */
1124:
1125: sz = strlen(cp);
1.290 schwarze 1126: if (sz > 0 && cp[sz - 1] == '\\')
1.289 schwarze 1127: sz--;
1128:
1.269 schwarze 1129: /* Skip trailing whitespace. */
1130:
1.289 schwarze 1131: for (; sz; sz--)
1.269 schwarze 1132: if ( ! isspace((unsigned char)cp[sz-1]))
1133: break;
1134:
1135: /* Skip empty strings. */
1136:
1137: if (sz == 0)
1138: return;
1139:
1140: if (*dest == NULL) {
1141: *dest = mandoc_strndup(cp, sz);
1142: return;
1143: }
1144:
1145: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1146: free(*dest);
1147: *dest = cp;
1.266 schwarze 1148: }
1149:
1150: /* --- main functions of the roff parser ---------------------------------- */
1151:
1.94 kristaps 1152: /*
1.206 schwarze 1153: * In the current line, expand escape sequences that tend to get
1154: * used in numerical expressions and conditional requests.
1155: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 1156: */
1.340 schwarze 1157: static int
1.238 schwarze 1158: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 1159: {
1.339 schwarze 1160: struct mctx *ctx; /* current macro call context */
1.208 schwarze 1161: char ubuf[24]; /* buffer to print the number */
1.328 schwarze 1162: struct roff_node *n; /* used for header comments */
1.205 schwarze 1163: const char *start; /* start of the string to process */
1.209 schwarze 1164: char *stesc; /* start of an escape sequence ('\\') */
1.328 schwarze 1165: char *ep; /* end of comment string */
1.108 schwarze 1166: const char *stnam; /* start of the name, after "[(*" */
1167: const char *cp; /* end of the name, e.g. before ']' */
1168: const char *res; /* the string to be substituted */
1.238 schwarze 1169: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 1170: size_t maxl; /* expected length of the escape name */
1171: size_t naml; /* actual length of the escape name */
1.339 schwarze 1172: size_t asz; /* length of the replacement */
1173: size_t rsz; /* length of the rest of the string */
1.237 schwarze 1174: enum mandoc_esc esc; /* type of the escape sequence */
1175: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 1176: int expand_count; /* to avoid infinite loops */
1.206 schwarze 1177: int npos; /* position in numeric expression */
1.218 schwarze 1178: int arg_complete; /* argument not interrupted by eol */
1.339 schwarze 1179: int quote_args; /* true for \\$@, false for \\$* */
1.303 schwarze 1180: int done; /* no more input available */
1.315 schwarze 1181: int deftype; /* type of definition to paste */
1.323 schwarze 1182: int rcsid; /* kind of RCS id seen */
1.327 schwarze 1183: char sign; /* increment number register */
1.206 schwarze 1184: char term; /* character terminating the escape */
1.94 kristaps 1185:
1.303 schwarze 1186: /* Search forward for comments. */
1187:
1188: done = 0;
1189: start = buf->buf + pos;
1190: for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
1191: if (stesc[0] != r->escape || stesc[1] == '\0')
1192: continue;
1193: stesc++;
1194: if (*stesc != '"' && *stesc != '#')
1195: continue;
1.313 schwarze 1196:
1197: /* Comment found, look for RCS id. */
1198:
1.323 schwarze 1199: rcsid = 0;
1.313 schwarze 1200: if ((cp = strstr(stesc, "$" "OpenBSD")) != NULL) {
1.323 schwarze 1201: rcsid = 1 << MANDOC_OS_OPENBSD;
1.313 schwarze 1202: cp += 8;
1203: } else if ((cp = strstr(stesc, "$" "NetBSD")) != NULL) {
1.323 schwarze 1204: rcsid = 1 << MANDOC_OS_NETBSD;
1.313 schwarze 1205: cp += 7;
1206: }
1207: if (cp != NULL &&
1208: isalnum((unsigned char)*cp) == 0 &&
1.314 schwarze 1209: strchr(cp, '$') != NULL) {
1.323 schwarze 1210: if (r->man->meta.rcsids & rcsid)
1.314 schwarze 1211: mandoc_msg(MANDOCERR_RCS_REP, r->parse,
1212: ln, stesc + 1 - buf->buf, stesc + 1);
1.323 schwarze 1213: r->man->meta.rcsids |= rcsid;
1.314 schwarze 1214: }
1.313 schwarze 1215:
1216: /* Handle trailing whitespace. */
1217:
1.328 schwarze 1218: ep = strchr(stesc--, '\0') - 1;
1219: if (*ep == '\n') {
1.303 schwarze 1220: done = 1;
1.328 schwarze 1221: ep--;
1.303 schwarze 1222: }
1.328 schwarze 1223: if (*ep == ' ' || *ep == '\t')
1.303 schwarze 1224: mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
1.328 schwarze 1225: ln, ep - buf->buf, NULL);
1226:
1227: /*
1228: * Save comments preceding the title macro
1229: * in the syntax tree.
1230: */
1231:
1232: if (r->format == 0) {
1233: while (*ep == ' ' || *ep == '\t')
1234: ep--;
1235: ep[1] = '\0';
1236: n = roff_node_alloc(r->man,
1237: ln, stesc + 1 - buf->buf,
1238: ROFFT_COMMENT, TOKEN_NONE);
1239: n->string = mandoc_strdup(stesc + 2);
1240: roff_node_append(r->man, n);
1241: n->flags |= NODE_VALID | NODE_ENDED;
1242: r->man->next = ROFF_NEXT_SIBLING;
1243: }
1244:
1.334 schwarze 1245: /* Line continuation with comment. */
1246:
1247: if (stesc[1] == '#') {
1248: *stesc = '\0';
1.340 schwarze 1249: return ROFF_IGN | ROFF_APPEND;
1.334 schwarze 1250: }
1251:
1252: /* Discard normal comments. */
1.328 schwarze 1253:
1.335 schwarze 1254: while (stesc > start && stesc[-1] == ' ' &&
1255: (stesc == start + 1 || stesc[-2] != '\\'))
1.303 schwarze 1256: stesc--;
1257: *stesc = '\0';
1258: break;
1259: }
1260: if (stesc == start)
1261: return ROFF_CONT;
1262: stesc--;
1263:
1264: /* Notice the end of the input. */
1265:
1266: if (*stesc == '\n') {
1267: *stesc-- = '\0';
1268: done = 1;
1269: }
1270:
1.170 schwarze 1271: expand_count = 0;
1.303 schwarze 1272: while (stesc >= start) {
1.170 schwarze 1273:
1.205 schwarze 1274: /* Search backwards for the next backslash. */
1275:
1.303 schwarze 1276: if (*stesc != r->escape) {
1277: if (*stesc == '\\') {
1278: *stesc = '\0';
1279: buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
1280: buf->buf, stesc + 1) + 1;
1281: start = nbuf + pos;
1282: stesc = nbuf + (stesc - buf->buf);
1283: free(buf->buf);
1284: buf->buf = nbuf;
1285: }
1286: stesc--;
1.205 schwarze 1287: continue;
1.303 schwarze 1288: }
1.205 schwarze 1289:
1290: /* If it is escaped, skip it. */
1291:
1292: for (cp = stesc - 1; cp >= start; cp--)
1.303 schwarze 1293: if (*cp != r->escape)
1.205 schwarze 1294: break;
1295:
1.238 schwarze 1296: if ((stesc - cp) % 2 == 0) {
1.303 schwarze 1297: while (stesc > cp)
1298: *stesc-- = '\\';
1.205 schwarze 1299: continue;
1.303 schwarze 1300: } else if (stesc[1] != '\0') {
1301: *stesc = '\\';
1302: } else {
1303: *stesc-- = '\0';
1304: if (done)
1305: continue;
1306: else
1.340 schwarze 1307: return ROFF_IGN | ROFF_APPEND;
1.205 schwarze 1308: }
1.108 schwarze 1309:
1.206 schwarze 1310: /* Decide whether to expand or to check only. */
1.108 schwarze 1311:
1.206 schwarze 1312: term = '\0';
1.205 schwarze 1313: cp = stesc + 1;
1.181 schwarze 1314: switch (*cp) {
1.207 schwarze 1315: case '*':
1.339 schwarze 1316: case '$':
1.181 schwarze 1317: res = NULL;
1318: break;
1.207 schwarze 1319: case 'B':
1320: case 'w':
1.206 schwarze 1321: term = cp[1];
1322: /* FALLTHROUGH */
1.207 schwarze 1323: case 'n':
1.327 schwarze 1324: sign = cp[1];
1325: if (sign == '+' || sign == '-')
1326: cp++;
1.181 schwarze 1327: res = ubuf;
1328: break;
1329: default:
1.237 schwarze 1330: esc = mandoc_escape(&cp, &stnam, &inaml);
1331: if (esc == ESCAPE_ERROR ||
1332: (esc == ESCAPE_SPECIAL &&
1.279 schwarze 1333: mchars_spec2cp(stnam, inaml) < 0))
1.219 schwarze 1334: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1335: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1336: "%.*s", (int)(cp - stesc), stesc);
1.303 schwarze 1337: stesc--;
1.205 schwarze 1338: continue;
1.152 kristaps 1339: }
1340:
1.205 schwarze 1341: if (EXPAND_LIMIT < ++expand_count) {
1342: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1343: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1344: return ROFF_IGN;
1.205 schwarze 1345: }
1.108 schwarze 1346:
1347: /*
1348: * The third character decides the length
1.181 schwarze 1349: * of the name of the string or register.
1.108 schwarze 1350: * Save a pointer to the name.
1351: */
1352:
1.238 schwarze 1353: if (term == '\0') {
1.206 schwarze 1354: switch (*++cp) {
1.207 schwarze 1355: case '\0':
1.206 schwarze 1356: maxl = 0;
1357: break;
1.207 schwarze 1358: case '(':
1.206 schwarze 1359: cp++;
1360: maxl = 2;
1361: break;
1.207 schwarze 1362: case '[':
1.206 schwarze 1363: cp++;
1364: term = ']';
1365: maxl = 0;
1366: break;
1367: default:
1368: maxl = 1;
1369: break;
1370: }
1371: } else {
1372: cp += 2;
1.94 kristaps 1373: maxl = 0;
1374: }
1.108 schwarze 1375: stnam = cp;
1.94 kristaps 1376:
1.108 schwarze 1377: /* Advance to the end of the name. */
1.94 kristaps 1378:
1.253 schwarze 1379: naml = 0;
1.218 schwarze 1380: arg_complete = 1;
1.253 schwarze 1381: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1382: if (*cp == '\0') {
1.219 schwarze 1383: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1384: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1385: arg_complete = 0;
1.206 schwarze 1386: break;
1.153 kristaps 1387: }
1.238 schwarze 1388: if (maxl == 0 && *cp == term) {
1.206 schwarze 1389: cp++;
1.253 schwarze 1390: break;
1391: }
1392: if (*cp++ != '\\' || stesc[1] != 'w') {
1393: naml++;
1394: continue;
1395: }
1396: switch (mandoc_escape(&cp, NULL, NULL)) {
1397: case ESCAPE_SPECIAL:
1398: case ESCAPE_UNICODE:
1399: case ESCAPE_NUMBERED:
1400: case ESCAPE_OVERSTRIKE:
1401: naml++;
1402: break;
1403: default:
1.94 kristaps 1404: break;
1.206 schwarze 1405: }
1.94 kristaps 1406: }
1407:
1.108 schwarze 1408: /*
1409: * Retrieve the replacement string; if it is
1410: * undefined, resume searching for escapes.
1411: */
1412:
1.206 schwarze 1413: switch (stesc[1]) {
1.207 schwarze 1414: case '*':
1.315 schwarze 1415: if (arg_complete) {
1416: deftype = ROFFDEF_USER | ROFFDEF_PRE;
1417: res = roff_getstrn(r, stnam, naml, &deftype);
1.331 schwarze 1418:
1419: /*
1420: * If not overriden, let \*(.T
1421: * through to the formatters.
1422: */
1423:
1424: if (res == NULL && naml == 2 &&
1425: stnam[0] == '.' && stnam[1] == 'T') {
1426: roff_setstrn(&r->strtab,
1427: ".T", 2, NULL, 0, 0);
1428: stesc--;
1429: continue;
1430: }
1.315 schwarze 1431: }
1.206 schwarze 1432: break;
1.339 schwarze 1433: case '$':
1434: if (r->mstackpos < 0) {
1435: mandoc_vmsg(MANDOCERR_ARG_UNDEF,
1436: r->parse, ln, (int)(stesc - buf->buf),
1437: "%.3s", stesc);
1438: break;
1439: }
1440: ctx = r->mstack + r->mstackpos;
1441: npos = stesc[2] - '1';
1442: if (npos >= 0 && npos <= 8) {
1443: res = npos < ctx->argc ?
1444: ctx->argv[npos] : "";
1445: break;
1446: }
1447: if (stesc[2] == '*')
1448: quote_args = 0;
1449: else if (stesc[2] == '@')
1450: quote_args = 1;
1451: else {
1452: mandoc_vmsg(MANDOCERR_ARG_NONUM,
1453: r->parse, ln, (int)(stesc - buf->buf),
1454: "%.3s", stesc);
1455: break;
1456: }
1457: asz = 0;
1458: for (npos = 0; npos < ctx->argc; npos++) {
1459: if (npos)
1460: asz++; /* blank */
1461: if (quote_args)
1462: asz += 2; /* quotes */
1463: asz += strlen(ctx->argv[npos]);
1464: }
1465: if (asz != 3) {
1466: rsz = buf->sz - (stesc - buf->buf) - 3;
1467: if (asz < 3)
1468: memmove(stesc + asz, stesc + 3, rsz);
1469: buf->sz += asz - 3;
1470: nbuf = mandoc_realloc(buf->buf, buf->sz);
1471: start = nbuf + pos;
1472: stesc = nbuf + (stesc - buf->buf);
1473: buf->buf = nbuf;
1474: if (asz > 3)
1475: memmove(stesc + asz, stesc + 3, rsz);
1476: }
1477: for (npos = 0; npos < ctx->argc; npos++) {
1478: if (npos)
1479: *stesc++ = ' ';
1480: if (quote_args)
1481: *stesc++ = '"';
1482: cp = ctx->argv[npos];
1483: while (*cp != '\0')
1484: *stesc++ = *cp++;
1485: if (quote_args)
1486: *stesc++ = '"';
1487: }
1488: continue;
1.207 schwarze 1489: case 'B':
1.206 schwarze 1490: npos = 0;
1.218 schwarze 1491: ubuf[0] = arg_complete &&
1.261 schwarze 1492: roff_evalnum(r, ln, stnam, &npos,
1493: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1494: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1495: ubuf[1] = '\0';
1496: break;
1.207 schwarze 1497: case 'n':
1.218 schwarze 1498: if (arg_complete)
1499: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.327 schwarze 1500: roff_getregn(r, stnam, naml, sign));
1.218 schwarze 1501: else
1502: ubuf[0] = '\0';
1.206 schwarze 1503: break;
1.207 schwarze 1504: case 'w':
1.218 schwarze 1505: /* use even incomplete args */
1.208 schwarze 1506: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1507: 24 * (int)naml);
1508: break;
1509: }
1.94 kristaps 1510:
1.238 schwarze 1511: if (res == NULL) {
1.339 schwarze 1512: if (stesc[1] == '*')
1513: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1514: r->parse, ln, (int)(stesc - buf->buf),
1515: "%.*s", (int)naml, stnam);
1.142 kristaps 1516: res = "";
1.246 schwarze 1517: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1518: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1519: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1520: return ROFF_IGN;
1.94 kristaps 1521: }
1522:
1.108 schwarze 1523: /* Replace the escape sequence by the string. */
1524:
1.209 schwarze 1525: *stesc = '\0';
1.238 schwarze 1526: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1527: buf->buf, res, cp) + 1;
1.94 kristaps 1528:
1.205 schwarze 1529: /* Prepare for the next replacement. */
1.94 kristaps 1530:
1.205 schwarze 1531: start = nbuf + pos;
1.238 schwarze 1532: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1533: free(buf->buf);
1534: buf->buf = nbuf;
1.154 kristaps 1535: }
1.277 schwarze 1536: return ROFF_CONT;
1.154 kristaps 1537: }
1538:
1539: /*
1.275 schwarze 1540: * Process text streams.
1.154 kristaps 1541: */
1.340 schwarze 1542: static int
1.305 schwarze 1543: roff_parsetext(struct roff *r, struct buf *buf, int pos, int *offs)
1.154 kristaps 1544: {
1545: size_t sz;
1546: const char *start;
1.178 schwarze 1547: char *p;
1548: int isz;
1.154 kristaps 1549: enum mandoc_esc esc;
1550:
1.275 schwarze 1551: /* Spring the input line trap. */
1552:
1553: if (roffit_lines == 1) {
1554: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1555: free(buf->buf);
1556: buf->buf = p;
1557: buf->sz = isz + 1;
1558: *offs = 0;
1559: free(roffit_macro);
1560: roffit_lines = 0;
1.277 schwarze 1561: return ROFF_REPARSE;
1.275 schwarze 1562: } else if (roffit_lines > 1)
1563: --roffit_lines;
1564:
1.305 schwarze 1565: if (roffce_node != NULL && buf->buf[pos] != '\0') {
1566: if (roffce_lines < 1) {
1567: r->man->last = roffce_node;
1568: r->man->next = ROFF_NEXT_SIBLING;
1569: roffce_lines = 0;
1570: roffce_node = NULL;
1571: } else
1572: roffce_lines--;
1573: }
1574:
1.275 schwarze 1575: /* Convert all breakable hyphens into ASCII_HYPH. */
1576:
1.238 schwarze 1577: start = p = buf->buf + pos;
1.154 kristaps 1578:
1.238 schwarze 1579: while (*p != '\0') {
1.154 kristaps 1580: sz = strcspn(p, "-\\");
1581: p += sz;
1582:
1.238 schwarze 1583: if (*p == '\0')
1.159 kristaps 1584: break;
1585:
1.238 schwarze 1586: if (*p == '\\') {
1.154 kristaps 1587: /* Skip over escapes. */
1588: p++;
1.189 schwarze 1589: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1590: if (esc == ESCAPE_ERROR)
1.154 kristaps 1591: break;
1.264 schwarze 1592: while (*p == '-')
1593: p++;
1.155 kristaps 1594: continue;
1.159 kristaps 1595: } else if (p == start) {
1.158 kristaps 1596: p++;
1.155 kristaps 1597: continue;
1.158 kristaps 1598: }
1.155 kristaps 1599:
1.171 schwarze 1600: if (isalpha((unsigned char)p[-1]) &&
1601: isalpha((unsigned char)p[1]))
1.155 kristaps 1602: *p = ASCII_HYPH;
1603: p++;
1.94 kristaps 1604: }
1.277 schwarze 1605: return ROFF_CONT;
1.94 kristaps 1606: }
1607:
1.340 schwarze 1608: int
1.238 schwarze 1609: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1610: {
1.294 schwarze 1611: enum roff_tok t;
1.340 schwarze 1612: int e;
1.238 schwarze 1613: int pos; /* parse point */
1.243 schwarze 1614: int spos; /* saved parse point for messages */
1.238 schwarze 1615: int ppos; /* original offset in buf->buf */
1616: int ctl; /* macro line (boolean) */
1617:
1618: ppos = pos = *offs;
1.79 kristaps 1619:
1.230 schwarze 1620: /* Handle in-line equation delimiters. */
1621:
1.236 schwarze 1622: if (r->tbl == NULL &&
1623: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1624: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1625: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1626: if (e == ROFF_REPARSE)
1.277 schwarze 1627: return e;
1.230 schwarze 1628: assert(e == ROFF_CONT);
1629: }
1630:
1631: /* Expand some escape sequences. */
1.94 kristaps 1632:
1.238 schwarze 1633: e = roff_res(r, buf, ln, pos);
1.340 schwarze 1634: if ((e & ROFF_MASK) == ROFF_IGN)
1.277 schwarze 1635: return e;
1.238 schwarze 1636: assert(e == ROFF_CONT);
1.94 kristaps 1637:
1.238 schwarze 1638: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1639:
1.94 kristaps 1640: /*
1.79 kristaps 1641: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1642: * text through the macro's filter.
1643: * Equations process all content themselves.
1644: * Tables process almost all content themselves, but we want
1645: * to warn about macros before passing it there.
1.79 kristaps 1646: */
1.74 kristaps 1647:
1.252 schwarze 1648: if (r->last != NULL && ! ctl) {
1.78 kristaps 1649: t = r->last->tok;
1.238 schwarze 1650: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.340 schwarze 1651: if ((e & ROFF_MASK) == ROFF_IGN)
1.277 schwarze 1652: return e;
1.340 schwarze 1653: e &= ~ROFF_MASK;
1654: } else
1655: e = ROFF_IGN;
1.319 schwarze 1656: if (r->eqn != NULL && strncmp(buf->buf + ppos, ".EN", 3)) {
1657: eqn_read(r->eqn, buf->buf + ppos);
1.340 schwarze 1658: return e;
1.319 schwarze 1659: }
1.321 schwarze 1660: if (r->tbl != NULL && (ctl == 0 || buf->buf[pos] == '\0')) {
1661: tbl_read(r->tbl, ln, buf->buf, ppos);
1662: roff_addtbl(r->man, r->tbl);
1.340 schwarze 1663: return e;
1.321 schwarze 1664: }
1.252 schwarze 1665: if ( ! ctl)
1.340 schwarze 1666: return roff_parsetext(r, buf, pos, offs) | e;
1.228 schwarze 1667:
1668: /* Skip empty request lines. */
1669:
1.238 schwarze 1670: if (buf->buf[pos] == '"') {
1.228 schwarze 1671: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1672: ln, pos, NULL);
1.277 schwarze 1673: return ROFF_IGN;
1.238 schwarze 1674: } else if (buf->buf[pos] == '\0')
1.277 schwarze 1675: return ROFF_IGN;
1.67 kristaps 1676:
1.79 kristaps 1677: /*
1678: * If a scope is open, go to the child handler for that macro,
1679: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1680: * Don't do so if an equation is open.
1.79 kristaps 1681: */
1.78 kristaps 1682:
1.79 kristaps 1683: if (r->last) {
1684: t = r->last->tok;
1.277 schwarze 1685: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1686: }
1.78 kristaps 1687:
1.243 schwarze 1688: /* No scope is open. This is a new request or macro. */
1689:
1690: spos = pos;
1691: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1692:
1693: /* Tables ignore most macros. */
1694:
1.308 schwarze 1695: if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS ||
1.309 schwarze 1696: t == ROFF_br || t == ROFF_ce || t == ROFF_rj || t == ROFF_sp)) {
1.243 schwarze 1697: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1698: ln, pos, buf->buf + spos);
1.308 schwarze 1699: if (t != TOKEN_NONE)
1.277 schwarze 1700: return ROFF_IGN;
1.257 schwarze 1701: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1702: pos++;
1.291 schwarze 1703: while (buf->buf[pos] == ' ')
1.257 schwarze 1704: pos++;
1.321 schwarze 1705: tbl_read(r->tbl, ln, buf->buf, pos);
1706: roff_addtbl(r->man, r->tbl);
1707: return ROFF_IGN;
1.243 schwarze 1708: }
1709:
1.305 schwarze 1710: /* For now, let high level macros abort .ce mode. */
1711:
1712: if (ctl && roffce_node != NULL &&
1.324 schwarze 1713: (t == TOKEN_NONE || t == ROFF_Dd || t == ROFF_EQ ||
1714: t == ROFF_TH || t == ROFF_TS)) {
1.305 schwarze 1715: r->man->last = roffce_node;
1716: r->man->next = ROFF_NEXT_SIBLING;
1717: roffce_lines = 0;
1718: roffce_node = NULL;
1719: }
1720:
1.79 kristaps 1721: /*
1.243 schwarze 1722: * This is neither a roff request nor a user-defined macro.
1723: * Let the standard macro set parsers handle it.
1.79 kristaps 1724: */
1.67 kristaps 1725:
1.294 schwarze 1726: if (t == TOKEN_NONE)
1.277 schwarze 1727: return ROFF_CONT;
1.243 schwarze 1728:
1729: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1730:
1.296 schwarze 1731: return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.74 kristaps 1732: }
1733:
1.339 schwarze 1734: /*
1735: * Internal interface function to tell the roff parser that execution
1736: * of the current macro ended. This is required because macro
1737: * definitions usually do not end with a .return request.
1738: */
1739: void
1740: roff_userret(struct roff *r)
1741: {
1742: struct mctx *ctx;
1743: int i;
1744:
1745: assert(r->mstackpos >= 0);
1746: ctx = r->mstack + r->mstackpos;
1747: for (i = 0; i < ctx->argc; i++)
1748: free(ctx->argv[i]);
1749: ctx->argc = 0;
1750: r->mstackpos--;
1751: }
1752:
1.117 kristaps 1753: void
1.74 kristaps 1754: roff_endparse(struct roff *r)
1755: {
1.321 schwarze 1756: if (r->last != NULL)
1.221 schwarze 1757: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1758: r->last->line, r->last->col,
1.295 schwarze 1759: roff_name[r->last->tok]);
1.117 kristaps 1760:
1.321 schwarze 1761: if (r->eqn != NULL) {
1.221 schwarze 1762: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1.319 schwarze 1763: r->eqn->node->line, r->eqn->node->pos, "EQ");
1764: eqn_parse(r->eqn);
1765: r->eqn = NULL;
1.125 kristaps 1766: }
1767:
1.321 schwarze 1768: if (r->tbl != NULL) {
1.221 schwarze 1769: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1770: r->tbl->line, r->tbl->pos, "TS");
1.321 schwarze 1771: tbl_end(r->tbl);
1772: r->tbl = NULL;
1.117 kristaps 1773: }
1.67 kristaps 1774: }
1775:
1776: /*
1777: * Parse a roff node's type from the input buffer. This must be in the
1778: * form of ".foo xxx" in the usual way.
1779: */
1.294 schwarze 1780: static enum roff_tok
1.214 schwarze 1781: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1782: {
1.214 schwarze 1783: char *cp;
1.106 kristaps 1784: const char *mac;
1785: size_t maclen;
1.315 schwarze 1786: int deftype;
1.294 schwarze 1787: enum roff_tok t;
1.67 kristaps 1788:
1.214 schwarze 1789: cp = buf + *pos;
1790:
1791: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294 schwarze 1792: return TOKEN_NONE;
1.67 kristaps 1793:
1.214 schwarze 1794: mac = cp;
1795: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1796:
1.315 schwarze 1797: deftype = ROFFDEF_USER | ROFFDEF_REN;
1798: r->current_string = roff_getstrn(r, mac, maclen, &deftype);
1799: switch (deftype) {
1800: case ROFFDEF_USER:
1801: t = ROFF_USERDEF;
1802: break;
1803: case ROFFDEF_REN:
1804: t = ROFF_RENAMED;
1805: break;
1806: default:
1807: t = roffhash_find(r->reqtab, mac, maclen);
1808: break;
1809: }
1.294 schwarze 1810: if (t != TOKEN_NONE)
1.214 schwarze 1811: *pos = cp - buf;
1.325 schwarze 1812: else if (deftype == ROFFDEF_UNDEF) {
1813: /* Using an undefined macro defines it to be empty. */
1814: roff_setstrn(&r->strtab, mac, maclen, "", 0, 0);
1815: roff_setstrn(&r->rentab, mac, maclen, NULL, 0, 0);
1816: }
1.277 schwarze 1817: return t;
1.67 kristaps 1818: }
1819:
1.266 schwarze 1820: /* --- handling of request blocks ----------------------------------------- */
1821:
1.340 schwarze 1822: static int
1.76 kristaps 1823: roff_cblock(ROFF_ARGS)
1.67 kristaps 1824: {
1825:
1.79 kristaps 1826: /*
1827: * A block-close `..' should only be invoked as a child of an
1828: * ignore macro, otherwise raise a warning and just ignore it.
1829: */
1830:
1.238 schwarze 1831: if (r->last == NULL) {
1.221 schwarze 1832: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1833: ln, ppos, "..");
1.277 schwarze 1834: return ROFF_IGN;
1.76 kristaps 1835: }
1.67 kristaps 1836:
1.81 kristaps 1837: switch (r->last->tok) {
1.207 schwarze 1838: case ROFF_am:
1.220 schwarze 1839: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.207 schwarze 1840: case ROFF_ami:
1841: case ROFF_de:
1.108 schwarze 1842: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.207 schwarze 1843: case ROFF_dei:
1844: case ROFF_ig:
1.81 kristaps 1845: break;
1846: default:
1.221 schwarze 1847: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1848: ln, ppos, "..");
1.277 schwarze 1849: return ROFF_IGN;
1.76 kristaps 1850: }
1.67 kristaps 1851:
1.238 schwarze 1852: if (buf->buf[pos] != '\0')
1.217 schwarze 1853: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1854: ".. %s", buf->buf + pos);
1.71 kristaps 1855:
1856: roffnode_pop(r);
1.76 kristaps 1857: roffnode_cleanscope(r);
1.277 schwarze 1858: return ROFF_IGN;
1.71 kristaps 1859:
1.67 kristaps 1860: }
1861:
1.340 schwarze 1862: static int
1.76 kristaps 1863: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1864: {
1.340 schwarze 1865: int inloop;
1.67 kristaps 1866:
1.340 schwarze 1867: inloop = 0;
1868: while (r->last != NULL) {
1.173 schwarze 1869: if (--r->last->endspan != 0)
1.76 kristaps 1870: break;
1.340 schwarze 1871: inloop += roffnode_pop(r);
1.76 kristaps 1872: }
1.340 schwarze 1873: return inloop;
1.67 kristaps 1874: }
1875:
1.340 schwarze 1876: static int
1.195 schwarze 1877: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1878: {
1.76 kristaps 1879: if (NULL == r->last) {
1.221 schwarze 1880: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1881: ln, ppos, "\\}");
1.340 schwarze 1882: return 0;
1.76 kristaps 1883: }
1884:
1.82 kristaps 1885: switch (r->last->tok) {
1.207 schwarze 1886: case ROFF_el:
1887: case ROFF_ie:
1888: case ROFF_if:
1.340 schwarze 1889: case ROFF_while:
1.82 kristaps 1890: break;
1891: default:
1.221 schwarze 1892: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1893: ln, ppos, "\\}");
1.340 schwarze 1894: return 0;
1.75 kristaps 1895: }
1896:
1.76 kristaps 1897: if (r->last->endspan > -1) {
1.221 schwarze 1898: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1899: ln, ppos, "\\}");
1.340 schwarze 1900: return 0;
1.76 kristaps 1901: }
1902:
1.340 schwarze 1903: return roffnode_pop(r) + roffnode_cleanscope(r);
1.76 kristaps 1904: }
1905:
1.340 schwarze 1906: static int
1.80 kristaps 1907: roff_block(ROFF_ARGS)
1.76 kristaps 1908: {
1.315 schwarze 1909: const char *name, *value;
1910: char *call, *cp, *iname, *rname;
1911: size_t csz, namesz, rsz;
1912: int deftype;
1.106 kristaps 1913:
1.220 schwarze 1914: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1915:
1.238 schwarze 1916: if (tok == ROFF_de1)
1.220 schwarze 1917: tok = ROFF_de;
1.251 schwarze 1918: else if (tok == ROFF_dei1)
1919: tok = ROFF_dei;
1.238 schwarze 1920: else if (tok == ROFF_am1)
1.220 schwarze 1921: tok = ROFF_am;
1.251 schwarze 1922: else if (tok == ROFF_ami1)
1923: tok = ROFF_ami;
1.220 schwarze 1924:
1925: /* Parse the macro name argument. */
1926:
1.238 schwarze 1927: cp = buf->buf + pos;
1928: if (tok == ROFF_ig) {
1.220 schwarze 1929: iname = NULL;
1930: namesz = 0;
1931: } else {
1932: iname = cp;
1933: namesz = roff_getname(r, &cp, ln, ppos);
1934: iname[namesz] = '\0';
1935: }
1.107 kristaps 1936:
1.220 schwarze 1937: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1938:
1.238 schwarze 1939: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315 schwarze 1940: deftype = ROFFDEF_USER;
1941: name = roff_getstrn(r, iname, namesz, &deftype);
1942: if (name == NULL) {
1.220 schwarze 1943: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1944: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1945: "%.*s", (int)namesz, iname);
1946: namesz = 0;
1947: } else
1948: namesz = strlen(name);
1949: } else
1950: name = iname;
1.107 kristaps 1951:
1.238 schwarze 1952: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1953: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.295 schwarze 1954: ln, ppos, roff_name[tok]);
1.277 schwarze 1955: return ROFF_IGN;
1.220 schwarze 1956: }
1.80 kristaps 1957:
1.106 kristaps 1958: roffnode_push(r, tok, name, ln, ppos);
1959:
1960: /*
1961: * At the beginning of a `de' macro, clear the existing string
1962: * with the same name, if there is one. New content will be
1.193 schwarze 1963: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1964: */
1.107 kristaps 1965:
1.311 schwarze 1966: if (tok == ROFF_de || tok == ROFF_dei) {
1.213 schwarze 1967: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.311 schwarze 1968: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.315 schwarze 1969: } else if (tok == ROFF_am || tok == ROFF_ami) {
1970: deftype = ROFFDEF_ANY;
1971: value = roff_getstrn(r, iname, namesz, &deftype);
1972: switch (deftype) { /* Before appending, ... */
1973: case ROFFDEF_PRE: /* copy predefined to user-defined. */
1974: roff_setstrn(&r->strtab, name, namesz,
1975: value, strlen(value), 0);
1976: break;
1977: case ROFFDEF_REN: /* call original standard macro. */
1978: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1979: (int)strlen(value), value);
1980: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1981: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1982: free(call);
1983: break;
1984: case ROFFDEF_STD: /* rename and call standard macro. */
1985: rsz = mandoc_asprintf(&rname, "__%s_renamed", name);
1986: roff_setstrn(&r->rentab, rname, rsz, name, namesz, 0);
1987: csz = mandoc_asprintf(&call, ".%.*s \\$* \\\"\n",
1988: (int)rsz, rname);
1989: roff_setstrn(&r->strtab, name, namesz, call, csz, 0);
1990: free(call);
1991: free(rname);
1992: break;
1993: default:
1994: break;
1995: }
1.311 schwarze 1996: }
1.76 kristaps 1997:
1.238 schwarze 1998: if (*cp == '\0')
1.277 schwarze 1999: return ROFF_IGN;
1.78 kristaps 2000:
1.220 schwarze 2001: /* Get the custom end marker. */
1.107 kristaps 2002:
1.220 schwarze 2003: iname = cp;
1.213 schwarze 2004: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 2005:
2006: /* Resolve the end marker if it is indirect. */
2007:
1.238 schwarze 2008: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1.315 schwarze 2009: deftype = ROFFDEF_USER;
2010: name = roff_getstrn(r, iname, namesz, &deftype);
2011: if (name == NULL) {
1.220 schwarze 2012: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 2013: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 2014: "%.*s", (int)namesz, iname);
2015: namesz = 0;
2016: } else
2017: namesz = strlen(name);
2018: } else
2019: name = iname;
2020:
1.213 schwarze 2021: if (namesz)
2022: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 2023:
1.238 schwarze 2024: if (*cp != '\0')
1.217 schwarze 2025: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.295 schwarze 2026: ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74 kristaps 2027:
1.277 schwarze 2028: return ROFF_IGN;
1.78 kristaps 2029: }
2030:
1.340 schwarze 2031: static int
1.80 kristaps 2032: roff_block_sub(ROFF_ARGS)
1.79 kristaps 2033: {
1.294 schwarze 2034: enum roff_tok t;
1.79 kristaps 2035: int i, j;
2036:
2037: /*
2038: * First check whether a custom macro exists at this level. If
2039: * it does, then check against it. This is some of groff's
2040: * stranger behaviours. If we encountered a custom end-scope
2041: * tag and that tag also happens to be a "real" macro, then we
2042: * need to try interpreting it again as a real macro. If it's
2043: * not, then return ignore. Else continue.
2044: */
2045:
2046: if (r->last->end) {
1.130 kristaps 2047: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 2048: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 2049: break;
2050:
1.238 schwarze 2051: if (r->last->end[j] == '\0' &&
2052: (buf->buf[i] == '\0' ||
2053: buf->buf[i] == ' ' ||
2054: buf->buf[i] == '\t')) {
1.79 kristaps 2055: roffnode_pop(r);
2056: roffnode_cleanscope(r);
2057:
1.238 schwarze 2058: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 2059: i++;
2060:
2061: pos = i;
1.238 schwarze 2062: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294 schwarze 2063: TOKEN_NONE)
1.277 schwarze 2064: return ROFF_RERUN;
2065: return ROFF_IGN;
1.79 kristaps 2066: }
2067: }
2068:
2069: /*
2070: * If we have no custom end-query or lookup failed, then try
2071: * pulling it out of the hashtable.
2072: */
2073:
1.238 schwarze 2074: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 2075:
1.238 schwarze 2076: if (t != ROFF_cblock) {
2077: if (tok != ROFF_ig)
2078: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.277 schwarze 2079: return ROFF_IGN;
1.106 kristaps 2080: }
1.79 kristaps 2081:
1.277 schwarze 2082: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 2083: }
2084:
1.340 schwarze 2085: static int
1.80 kristaps 2086: roff_block_text(ROFF_ARGS)
1.78 kristaps 2087: {
2088:
1.238 schwarze 2089: if (tok != ROFF_ig)
2090: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 2091:
1.277 schwarze 2092: return ROFF_IGN;
1.78 kristaps 2093: }
2094:
1.340 schwarze 2095: static int
1.82 kristaps 2096: roff_cond_sub(ROFF_ARGS)
2097: {
1.340 schwarze 2098: char *ep;
2099: int endloop, irc, rr;
1.294 schwarze 2100: enum roff_tok t;
1.82 kristaps 2101:
1.340 schwarze 2102: irc = ROFF_IGN;
1.82 kristaps 2103: rr = r->last->rule;
1.340 schwarze 2104: endloop = tok != ROFF_while ? ROFF_IGN :
2105: rr ? ROFF_LOOPCONT : ROFF_LOOPEXIT;
2106: if (roffnode_cleanscope(r))
2107: irc |= endloop;
1.144 kristaps 2108:
1.196 schwarze 2109: /*
2110: * If `\}' occurs on a macro line without a preceding macro,
2111: * drop the line completely.
2112: */
2113:
1.238 schwarze 2114: ep = buf->buf + pos;
2115: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 2116: rr = 0;
1.196 schwarze 2117:
1.343 schwarze 2118: /*
2119: * The closing delimiter `\}' rewinds the conditional scope
2120: * but is otherwise ignored when interpreting the line.
2121: */
1.144 kristaps 2122:
1.238 schwarze 2123: while ((ep = strchr(ep, '\\')) != NULL) {
1.318 schwarze 2124: switch (ep[1]) {
2125: case '}':
2126: memmove(ep, ep + 2, strlen(ep + 2) + 1);
1.340 schwarze 2127: if (roff_ccond(r, ln, ep - buf->buf))
2128: irc |= endloop;
1.318 schwarze 2129: break;
2130: case '\0':
2131: ++ep;
2132: break;
2133: default:
2134: ep += 2;
2135: break;
1.197 schwarze 2136: }
1.177 schwarze 2137: }
1.318 schwarze 2138:
2139: /*
2140: * Fully handle known macros when they are structurally
2141: * required or when the conditional evaluated to true.
2142: */
2143:
2144: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.340 schwarze 2145: irc |= t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT) ?
2146: (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs) :
2147: rr ? ROFF_CONT : ROFF_IGN;
2148: return irc;
1.82 kristaps 2149: }
2150:
1.340 schwarze 2151: static int
1.82 kristaps 2152: roff_cond_text(ROFF_ARGS)
1.78 kristaps 2153: {
1.140 kristaps 2154: char *ep;
1.340 schwarze 2155: int endloop, irc, rr;
1.82 kristaps 2156:
1.340 schwarze 2157: irc = ROFF_IGN;
1.82 kristaps 2158: rr = r->last->rule;
1.340 schwarze 2159: endloop = tok != ROFF_while ? ROFF_IGN :
2160: rr ? ROFF_LOOPCONT : ROFF_LOOPEXIT;
2161: if (roffnode_cleanscope(r))
2162: irc |= endloop;
1.82 kristaps 2163:
1.343 schwarze 2164: /*
2165: * If `\}' occurs on a text line with neither preceding
2166: * nor following characters, drop the line completely.
2167: */
2168:
1.238 schwarze 2169: ep = buf->buf + pos;
1.343 schwarze 2170: if (strcmp(ep, "\\}") == 0)
2171: rr = 0;
2172:
2173: /*
2174: * The closing delimiter `\}' rewinds the conditional scope
2175: * but is otherwise ignored when interpreting the line.
2176: */
2177:
1.238 schwarze 2178: while ((ep = strchr(ep, '\\')) != NULL) {
1.343 schwarze 2179: switch (ep[1]) {
2180: case '}':
2181: memmove(ep, ep + 2, strlen(ep + 2) + 1);
2182: if (roff_ccond(r, ln, ep - buf->buf))
1.340 schwarze 2183: irc |= endloop;
1.343 schwarze 2184: break;
2185: case '\0':
2186: ++ep;
2187: break;
2188: default:
2189: ep += 2;
2190: break;
1.197 schwarze 2191: }
1.78 kristaps 2192: }
1.340 schwarze 2193: if (rr)
2194: irc |= ROFF_CONT;
2195: return irc;
1.74 kristaps 2196: }
2197:
1.266 schwarze 2198: /* --- handling of numeric and conditional expressions -------------------- */
2199:
1.204 schwarze 2200: /*
2201: * Parse a single signed integer number. Stop at the first non-digit.
2202: * If there is at least one digit, return success and advance the
2203: * parse point, else return failure and let the parse point unchanged.
2204: * Ignore overflows, treat them just like the C language.
2205: */
1.184 schwarze 2206: static int
1.261 schwarze 2207: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 2208: {
1.261 schwarze 2209: int myres, scaled, n, p;
1.206 schwarze 2210:
2211: if (NULL == res)
2212: res = &myres;
1.184 schwarze 2213:
2214: p = *pos;
2215: n = v[p] == '-';
1.261 schwarze 2216: if (n || v[p] == '+')
1.184 schwarze 2217: p++;
2218:
1.261 schwarze 2219: if (flags & ROFFNUM_WHITE)
2220: while (isspace((unsigned char)v[p]))
2221: p++;
2222:
1.184 schwarze 2223: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 2224: *res = 10 * *res + v[p] - '0';
1.184 schwarze 2225: if (p == *pos + n)
2226: return 0;
2227:
2228: if (n)
2229: *res = -*res;
2230:
1.254 schwarze 2231: /* Each number may be followed by one optional scaling unit. */
2232:
2233: switch (v[p]) {
2234: case 'f':
1.261 schwarze 2235: scaled = *res * 65536;
1.254 schwarze 2236: break;
2237: case 'i':
1.261 schwarze 2238: scaled = *res * 240;
1.254 schwarze 2239: break;
2240: case 'c':
1.261 schwarze 2241: scaled = *res * 240 / 2.54;
1.254 schwarze 2242: break;
2243: case 'v':
2244: case 'P':
1.261 schwarze 2245: scaled = *res * 40;
1.254 schwarze 2246: break;
2247: case 'm':
2248: case 'n':
1.261 schwarze 2249: scaled = *res * 24;
1.254 schwarze 2250: break;
2251: case 'p':
1.261 schwarze 2252: scaled = *res * 10 / 3;
1.254 schwarze 2253: break;
2254: case 'u':
1.261 schwarze 2255: scaled = *res;
1.254 schwarze 2256: break;
2257: case 'M':
1.261 schwarze 2258: scaled = *res * 6 / 25;
1.254 schwarze 2259: break;
2260: default:
1.261 schwarze 2261: scaled = *res;
1.254 schwarze 2262: p--;
2263: break;
2264: }
1.261 schwarze 2265: if (flags & ROFFNUM_SCALE)
2266: *res = scaled;
1.254 schwarze 2267:
2268: *pos = p + 1;
1.277 schwarze 2269: return 1;
1.184 schwarze 2270: }
2271:
1.198 schwarze 2272: /*
2273: * Evaluate a string comparison condition.
2274: * The first character is the delimiter.
2275: * Succeed if the string up to its second occurrence
2276: * matches the string up to its third occurence.
2277: * Advance the cursor after the third occurrence
2278: * or lacking that, to the end of the line.
2279: */
2280: static int
2281: roff_evalstrcond(const char *v, int *pos)
2282: {
2283: const char *s1, *s2, *s3;
2284: int match;
2285:
2286: match = 0;
2287: s1 = v + *pos; /* initial delimiter */
2288: s2 = s1 + 1; /* for scanning the first string */
2289: s3 = strchr(s2, *s1); /* for scanning the second string */
2290:
2291: if (NULL == s3) /* found no middle delimiter */
2292: goto out;
2293:
2294: while ('\0' != *++s3) {
2295: if (*s2 != *s3) { /* mismatch */
2296: s3 = strchr(s3, *s1);
2297: break;
2298: }
2299: if (*s3 == *s1) { /* found the final delimiter */
2300: match = 1;
2301: break;
2302: }
2303: s2++;
2304: }
2305:
2306: out:
2307: if (NULL == s3)
2308: s3 = strchr(s2, '\0');
1.242 schwarze 2309: else if (*s3 != '\0')
1.198 schwarze 2310: s3++;
2311: *pos = s3 - v;
1.277 schwarze 2312: return match;
1.198 schwarze 2313: }
2314:
1.204 schwarze 2315: /*
2316: * Evaluate an optionally negated single character, numerical,
2317: * or string condition.
2318: */
1.198 schwarze 2319: static int
1.271 schwarze 2320: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88 kristaps 2321: {
1.336 schwarze 2322: const char *start, *end;
2323: char *cp, *name;
2324: size_t sz;
2325: int deftype, len, number, savepos, istrue, wanttrue;
1.88 kristaps 2326:
1.198 schwarze 2327: if ('!' == v[*pos]) {
2328: wanttrue = 0;
2329: (*pos)++;
2330: } else
2331: wanttrue = 1;
2332:
1.88 kristaps 2333: switch (v[*pos]) {
1.240 schwarze 2334: case '\0':
1.277 schwarze 2335: return 0;
1.207 schwarze 2336: case 'n':
2337: case 'o':
1.88 kristaps 2338: (*pos)++;
1.277 schwarze 2339: return wanttrue;
1.207 schwarze 2340: case 'e':
2341: case 't':
1.239 schwarze 2342: case 'v':
1.88 kristaps 2343: (*pos)++;
1.277 schwarze 2344: return !wanttrue;
1.336 schwarze 2345: case 'c':
2346: do {
2347: (*pos)++;
2348: } while (v[*pos] == ' ');
2349:
2350: /*
2351: * Quirk for groff compatibility:
2352: * The horizontal tab is neither available nor unavailable.
2353: */
2354:
2355: if (v[*pos] == '\t') {
2356: (*pos)++;
2357: return 0;
2358: }
2359:
2360: /* Printable ASCII characters are available. */
2361:
2362: if (v[*pos] != '\\') {
2363: (*pos)++;
2364: return wanttrue;
2365: }
2366:
2367: end = v + ++*pos;
2368: switch (mandoc_escape(&end, &start, &len)) {
2369: case ESCAPE_SPECIAL:
2370: istrue = mchars_spec2cp(start, len) != -1;
2371: break;
2372: case ESCAPE_UNICODE:
2373: istrue = 1;
2374: break;
2375: case ESCAPE_NUMBERED:
2376: istrue = mchars_num2char(start, len) != -1;
2377: break;
2378: default:
2379: istrue = !wanttrue;
2380: break;
2381: }
2382: *pos = end - v;
2383: return istrue == wanttrue;
1.310 schwarze 2384: case 'd':
1.271 schwarze 2385: case 'r':
1.310 schwarze 2386: cp = v + *pos + 1;
2387: while (*cp == ' ')
2388: cp++;
2389: name = cp;
2390: sz = roff_getname(r, &cp, ln, cp - v);
1.315 schwarze 2391: if (sz == 0)
2392: istrue = 0;
2393: else if (v[*pos] == 'r')
2394: istrue = roff_hasregn(r, name, sz);
2395: else {
2396: deftype = ROFFDEF_ANY;
2397: roff_getstrn(r, name, sz, &deftype);
2398: istrue = !!deftype;
2399: }
1.271 schwarze 2400: *pos = cp - v;
1.310 schwarze 2401: return istrue == wanttrue;
1.88 kristaps 2402: default:
2403: break;
2404: }
2405:
1.241 schwarze 2406: savepos = *pos;
1.261 schwarze 2407: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277 schwarze 2408: return (number > 0) == wanttrue;
1.241 schwarze 2409: else if (*pos == savepos)
1.277 schwarze 2410: return roff_evalstrcond(v, pos) == wanttrue;
1.204 schwarze 2411: else
1.277 schwarze 2412: return 0;
1.88 kristaps 2413: }
2414:
1.340 schwarze 2415: static int
1.103 kristaps 2416: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 2417: {
1.123 schwarze 2418:
1.277 schwarze 2419: return ROFF_IGN;
1.89 kristaps 2420: }
2421:
1.340 schwarze 2422: static int
1.251 schwarze 2423: roff_insec(ROFF_ARGS)
2424: {
2425:
2426: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.295 schwarze 2427: ln, ppos, roff_name[tok]);
1.277 schwarze 2428: return ROFF_IGN;
1.251 schwarze 2429: }
2430:
1.340 schwarze 2431: static int
1.251 schwarze 2432: roff_unsupp(ROFF_ARGS)
2433: {
2434:
2435: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.295 schwarze 2436: ln, ppos, roff_name[tok]);
1.277 schwarze 2437: return ROFF_IGN;
1.251 schwarze 2438: }
2439:
1.340 schwarze 2440: static int
1.82 kristaps 2441: roff_cond(ROFF_ARGS)
1.74 kristaps 2442: {
1.340 schwarze 2443: int irc;
1.173 schwarze 2444:
2445: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 2446:
1.207 schwarze 2447: /*
1.134 kristaps 2448: * An `.el' has no conditional body: it will consume the value
2449: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 2450: * defaults to DENY.
1.134 kristaps 2451: *
2452: * If we're not an `el', however, then evaluate the conditional.
2453: */
1.133 kristaps 2454:
1.238 schwarze 2455: r->last->rule = tok == ROFF_el ?
1.207 schwarze 2456: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 2457: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 2458:
1.134 kristaps 2459: /*
2460: * An if-else will put the NEGATION of the current evaluated
2461: * conditional into the stack of rules.
2462: */
2463:
1.238 schwarze 2464: if (tok == ROFF_ie) {
1.223 schwarze 2465: if (r->rstackpos + 1 == r->rstacksz) {
2466: r->rstacksz += 16;
2467: r->rstack = mandoc_reallocarray(r->rstack,
2468: r->rstacksz, sizeof(int));
1.134 kristaps 2469: }
1.198 schwarze 2470: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 2471: }
1.88 kristaps 2472:
2473: /* If the parent has false as its rule, then so do we. */
2474:
1.198 schwarze 2475: if (r->last->parent && !r->last->parent->rule)
2476: r->last->rule = 0;
1.88 kristaps 2477:
2478: /*
1.173 schwarze 2479: * Determine scope.
2480: * If there is nothing on the line after the conditional,
2481: * not even whitespace, use next-line scope.
1.340 schwarze 2482: * Except that .while does not support next-line scope.
1.88 kristaps 2483: */
1.74 kristaps 2484:
1.340 schwarze 2485: if (buf->buf[pos] == '\0' && tok != ROFF_while) {
1.173 schwarze 2486: r->last->endspan = 2;
2487: goto out;
2488: }
2489:
1.238 schwarze 2490: while (buf->buf[pos] == ' ')
1.173 schwarze 2491: pos++;
2492:
2493: /* An opening brace requests multiline scope. */
1.75 kristaps 2494:
1.238 schwarze 2495: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 2496: r->last->endspan = -1;
2497: pos += 2;
1.272 schwarze 2498: while (buf->buf[pos] == ' ')
2499: pos++;
1.173 schwarze 2500: goto out;
1.207 schwarze 2501: }
1.74 kristaps 2502:
1.77 kristaps 2503: /*
1.173 schwarze 2504: * Anything else following the conditional causes
2505: * single-line scope. Warn if the scope contains
2506: * nothing but trailing whitespace.
1.77 kristaps 2507: */
2508:
1.238 schwarze 2509: if (buf->buf[pos] == '\0')
1.216 schwarze 2510: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.295 schwarze 2511: ln, ppos, roff_name[tok]);
1.77 kristaps 2512:
1.173 schwarze 2513: r->last->endspan = 1;
1.74 kristaps 2514:
1.173 schwarze 2515: out:
1.75 kristaps 2516: *offs = pos;
1.340 schwarze 2517: irc = ROFF_RERUN;
2518: if (tok == ROFF_while)
2519: irc |= ROFF_WHILE;
2520: return irc;
1.83 schwarze 2521: }
2522:
1.340 schwarze 2523: static int
1.92 schwarze 2524: roff_ds(ROFF_ARGS)
2525: {
1.212 schwarze 2526: char *string;
2527: const char *name;
2528: size_t namesz;
1.96 kristaps 2529:
1.251 schwarze 2530: /* Ignore groff compatibility mode for now. */
2531:
2532: if (tok == ROFF_ds1)
2533: tok = ROFF_ds;
2534: else if (tok == ROFF_as1)
2535: tok = ROFF_as;
2536:
1.96 kristaps 2537: /*
1.212 schwarze 2538: * The first word is the name of the string.
2539: * If it is empty or terminated by an escape sequence,
2540: * abort the `ds' request without defining anything.
1.96 kristaps 2541: */
1.92 schwarze 2542:
1.238 schwarze 2543: name = string = buf->buf + pos;
2544: if (*name == '\0')
1.277 schwarze 2545: return ROFF_IGN;
1.92 schwarze 2546:
1.212 schwarze 2547: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 2548: if (name[namesz] == '\\')
1.277 schwarze 2549: return ROFF_IGN;
1.212 schwarze 2550:
2551: /* Read past the initial double-quote, if any. */
1.238 schwarze 2552: if (*string == '"')
1.92 schwarze 2553: string++;
2554:
1.96 kristaps 2555: /* The rest is the value. */
1.212 schwarze 2556: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2557: ROFF_as == tok);
1.311 schwarze 2558: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.277 schwarze 2559: return ROFF_IGN;
1.92 schwarze 2560: }
2561:
1.204 schwarze 2562: /*
2563: * Parse a single operator, one or two characters long.
2564: * If the operator is recognized, return success and advance the
2565: * parse point, else return failure and let the parse point unchanged.
2566: */
2567: static int
2568: roff_getop(const char *v, int *pos, char *res)
2569: {
2570:
2571: *res = v[*pos];
2572:
2573: switch (*res) {
1.207 schwarze 2574: case '+':
2575: case '-':
2576: case '*':
2577: case '/':
2578: case '%':
2579: case '&':
2580: case ':':
1.204 schwarze 2581: break;
2582: case '<':
2583: switch (v[*pos + 1]) {
1.207 schwarze 2584: case '=':
1.204 schwarze 2585: *res = 'l';
2586: (*pos)++;
2587: break;
1.207 schwarze 2588: case '>':
1.204 schwarze 2589: *res = '!';
2590: (*pos)++;
2591: break;
1.207 schwarze 2592: case '?':
1.204 schwarze 2593: *res = 'i';
2594: (*pos)++;
2595: break;
2596: default:
2597: break;
2598: }
2599: break;
2600: case '>':
2601: switch (v[*pos + 1]) {
1.207 schwarze 2602: case '=':
1.204 schwarze 2603: *res = 'g';
2604: (*pos)++;
2605: break;
1.207 schwarze 2606: case '?':
1.204 schwarze 2607: *res = 'a';
2608: (*pos)++;
2609: break;
2610: default:
2611: break;
2612: }
2613: break;
2614: case '=':
2615: if ('=' == v[*pos + 1])
2616: (*pos)++;
2617: break;
2618: default:
1.277 schwarze 2619: return 0;
1.204 schwarze 2620: }
2621: (*pos)++;
2622:
1.277 schwarze 2623: return *res;
1.204 schwarze 2624: }
2625:
2626: /*
2627: * Evaluate either a parenthesized numeric expression
2628: * or a single signed integer number.
2629: */
2630: static int
1.235 schwarze 2631: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2632: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2633: {
2634:
2635: if ('(' != v[*pos])
1.277 schwarze 2636: return roff_getnum(v, pos, res, flags);
1.204 schwarze 2637:
2638: (*pos)++;
1.261 schwarze 2639: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277 schwarze 2640: return 0;
1.204 schwarze 2641:
1.206 schwarze 2642: /*
2643: * Omission of the closing parenthesis
2644: * is an error in validation mode,
2645: * but ignored in evaluation mode.
2646: */
2647:
1.204 schwarze 2648: if (')' == v[*pos])
2649: (*pos)++;
1.206 schwarze 2650: else if (NULL == res)
1.277 schwarze 2651: return 0;
1.204 schwarze 2652:
1.277 schwarze 2653: return 1;
1.204 schwarze 2654: }
2655:
2656: /*
2657: * Evaluate a complete numeric expression.
2658: * Proceed left to right, there is no concept of precedence.
2659: */
2660: static int
1.235 schwarze 2661: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2662: int *pos, int *res, int flags)
1.204 schwarze 2663: {
2664: int mypos, operand2;
2665: char operator;
2666:
2667: if (NULL == pos) {
2668: mypos = 0;
2669: pos = &mypos;
2670: }
2671:
1.261 schwarze 2672: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2673: while (isspace((unsigned char)v[*pos]))
2674: (*pos)++;
2675:
1.261 schwarze 2676: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277 schwarze 2677: return 0;
1.204 schwarze 2678:
2679: while (1) {
1.261 schwarze 2680: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2681: while (isspace((unsigned char)v[*pos]))
2682: (*pos)++;
2683:
2684: if ( ! roff_getop(v, pos, &operator))
2685: break;
2686:
1.261 schwarze 2687: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2688: while (isspace((unsigned char)v[*pos]))
2689: (*pos)++;
2690:
1.261 schwarze 2691: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277 schwarze 2692: return 0;
1.204 schwarze 2693:
1.261 schwarze 2694: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2695: while (isspace((unsigned char)v[*pos]))
2696: (*pos)++;
1.206 schwarze 2697:
2698: if (NULL == res)
2699: continue;
1.204 schwarze 2700:
2701: switch (operator) {
1.207 schwarze 2702: case '+':
1.204 schwarze 2703: *res += operand2;
2704: break;
1.207 schwarze 2705: case '-':
1.204 schwarze 2706: *res -= operand2;
2707: break;
1.207 schwarze 2708: case '*':
1.204 schwarze 2709: *res *= operand2;
2710: break;
1.207 schwarze 2711: case '/':
1.244 schwarze 2712: if (operand2 == 0) {
1.235 schwarze 2713: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2714: r->parse, ln, *pos, v);
2715: *res = 0;
2716: break;
2717: }
1.204 schwarze 2718: *res /= operand2;
2719: break;
1.207 schwarze 2720: case '%':
1.244 schwarze 2721: if (operand2 == 0) {
2722: mandoc_msg(MANDOCERR_DIVZERO,
2723: r->parse, ln, *pos, v);
2724: *res = 0;
2725: break;
2726: }
1.204 schwarze 2727: *res %= operand2;
2728: break;
1.207 schwarze 2729: case '<':
1.204 schwarze 2730: *res = *res < operand2;
2731: break;
1.207 schwarze 2732: case '>':
1.204 schwarze 2733: *res = *res > operand2;
2734: break;
1.207 schwarze 2735: case 'l':
1.204 schwarze 2736: *res = *res <= operand2;
2737: break;
1.207 schwarze 2738: case 'g':
1.204 schwarze 2739: *res = *res >= operand2;
2740: break;
1.207 schwarze 2741: case '=':
1.204 schwarze 2742: *res = *res == operand2;
2743: break;
1.207 schwarze 2744: case '!':
1.204 schwarze 2745: *res = *res != operand2;
2746: break;
1.207 schwarze 2747: case '&':
1.204 schwarze 2748: *res = *res && operand2;
2749: break;
1.207 schwarze 2750: case ':':
1.204 schwarze 2751: *res = *res || operand2;
2752: break;
1.207 schwarze 2753: case 'i':
1.204 schwarze 2754: if (operand2 < *res)
2755: *res = operand2;
2756: break;
1.207 schwarze 2757: case 'a':
1.204 schwarze 2758: if (operand2 > *res)
2759: *res = operand2;
2760: break;
2761: default:
2762: abort();
2763: }
2764: }
1.277 schwarze 2765: return 1;
1.204 schwarze 2766: }
2767:
1.266 schwarze 2768: /* --- register management ------------------------------------------------ */
2769:
1.180 schwarze 2770: void
1.187 schwarze 2771: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2772: {
1.327 schwarze 2773: roff_setregn(r, name, strlen(name), val, sign, INT_MIN);
1.326 schwarze 2774: }
2775:
2776: static void
2777: roff_setregn(struct roff *r, const char *name, size_t len,
1.327 schwarze 2778: int val, char sign, int step)
1.326 schwarze 2779: {
1.180 schwarze 2780: struct roffreg *reg;
2781:
2782: /* Search for an existing register with the same name. */
2783: reg = r->regtab;
2784:
1.326 schwarze 2785: while (reg != NULL && (reg->key.sz != len ||
2786: strncmp(reg->key.p, name, len) != 0))
1.180 schwarze 2787: reg = reg->next;
1.147 kristaps 2788:
1.180 schwarze 2789: if (NULL == reg) {
2790: /* Create a new register. */
2791: reg = mandoc_malloc(sizeof(struct roffreg));
1.326 schwarze 2792: reg->key.p = mandoc_strndup(name, len);
2793: reg->key.sz = len;
1.187 schwarze 2794: reg->val = 0;
1.327 schwarze 2795: reg->step = 0;
1.180 schwarze 2796: reg->next = r->regtab;
2797: r->regtab = reg;
2798: }
2799:
1.187 schwarze 2800: if ('+' == sign)
2801: reg->val += val;
2802: else if ('-' == sign)
2803: reg->val -= val;
2804: else
2805: reg->val = val;
1.327 schwarze 2806: if (step != INT_MIN)
2807: reg->step = step;
1.147 kristaps 2808: }
2809:
1.192 schwarze 2810: /*
2811: * Handle some predefined read-only number registers.
2812: * For now, return -1 if the requested register is not predefined;
2813: * in case a predefined read-only register having the value -1
2814: * were to turn up, another special value would have to be chosen.
2815: */
2816: static int
1.273 schwarze 2817: roff_getregro(const struct roff *r, const char *name)
1.192 schwarze 2818: {
2819:
2820: switch (*name) {
1.273 schwarze 2821: case '$': /* Number of arguments of the last macro evaluated. */
1.339 schwarze 2822: return r->mstackpos < 0 ? 0 : r->mstack[r->mstackpos].argc;
1.207 schwarze 2823: case 'A': /* ASCII approximation mode is always off. */
1.277 schwarze 2824: return 0;
1.207 schwarze 2825: case 'g': /* Groff compatibility mode is always on. */
1.277 schwarze 2826: return 1;
1.207 schwarze 2827: case 'H': /* Fixed horizontal resolution. */
1.277 schwarze 2828: return 24;
1.207 schwarze 2829: case 'j': /* Always adjust left margin only. */
1.277 schwarze 2830: return 0;
1.207 schwarze 2831: case 'T': /* Some output device is always defined. */
1.277 schwarze 2832: return 1;
1.207 schwarze 2833: case 'V': /* Fixed vertical resolution. */
1.277 schwarze 2834: return 40;
1.192 schwarze 2835: default:
1.277 schwarze 2836: return -1;
1.192 schwarze 2837: }
2838: }
2839:
1.181 schwarze 2840: int
1.326 schwarze 2841: roff_getreg(struct roff *r, const char *name)
1.147 kristaps 2842: {
1.327 schwarze 2843: return roff_getregn(r, name, strlen(name), '\0');
1.181 schwarze 2844: }
2845:
2846: static int
1.327 schwarze 2847: roff_getregn(struct roff *r, const char *name, size_t len, char sign)
1.181 schwarze 2848: {
2849: struct roffreg *reg;
1.192 schwarze 2850: int val;
2851:
2852: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2853: val = roff_getregro(r, name + 1);
1.192 schwarze 2854: if (-1 != val)
1.277 schwarze 2855: return val;
1.192 schwarze 2856: }
1.181 schwarze 2857:
1.327 schwarze 2858: for (reg = r->regtab; reg; reg = reg->next) {
1.181 schwarze 2859: if (len == reg->key.sz &&
1.327 schwarze 2860: 0 == strncmp(name, reg->key.p, len)) {
2861: switch (sign) {
2862: case '+':
2863: reg->val += reg->step;
2864: break;
2865: case '-':
2866: reg->val -= reg->step;
2867: break;
2868: default:
2869: break;
2870: }
1.277 schwarze 2871: return reg->val;
1.327 schwarze 2872: }
2873: }
1.271 schwarze 2874:
1.327 schwarze 2875: roff_setregn(r, name, len, 0, '\0', INT_MIN);
1.277 schwarze 2876: return 0;
1.271 schwarze 2877: }
2878:
2879: static int
2880: roff_hasregn(const struct roff *r, const char *name, size_t len)
2881: {
2882: struct roffreg *reg;
2883: int val;
2884:
2885: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2886: val = roff_getregro(r, name + 1);
1.271 schwarze 2887: if (-1 != val)
1.277 schwarze 2888: return 1;
1.271 schwarze 2889: }
2890:
2891: for (reg = r->regtab; reg; reg = reg->next)
2892: if (len == reg->key.sz &&
2893: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2894: return 1;
1.147 kristaps 2895:
1.277 schwarze 2896: return 0;
1.147 kristaps 2897: }
2898:
1.180 schwarze 2899: static void
2900: roff_freereg(struct roffreg *reg)
1.147 kristaps 2901: {
1.180 schwarze 2902: struct roffreg *old_reg;
1.147 kristaps 2903:
1.180 schwarze 2904: while (NULL != reg) {
2905: free(reg->key.p);
2906: old_reg = reg;
2907: reg = reg->next;
2908: free(old_reg);
2909: }
1.147 kristaps 2910: }
1.92 schwarze 2911:
1.340 schwarze 2912: static int
1.89 kristaps 2913: roff_nr(ROFF_ARGS)
1.83 schwarze 2914: {
1.327 schwarze 2915: char *key, *val, *step;
1.212 schwarze 2916: size_t keysz;
1.327 schwarze 2917: int iv, is, len;
1.187 schwarze 2918: char sign;
1.89 kristaps 2919:
1.238 schwarze 2920: key = val = buf->buf + pos;
2921: if (*key == '\0')
1.277 schwarze 2922: return ROFF_IGN;
1.212 schwarze 2923:
2924: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2925: if (key[keysz] == '\\')
1.277 schwarze 2926: return ROFF_IGN;
1.89 kristaps 2927:
1.187 schwarze 2928: sign = *val;
1.238 schwarze 2929: if (sign == '+' || sign == '-')
1.187 schwarze 2930: val++;
2931:
1.327 schwarze 2932: len = 0;
2933: if (roff_evalnum(r, ln, val, &len, &iv, ROFFNUM_SCALE) == 0)
2934: return ROFF_IGN;
2935:
2936: step = val + len;
2937: while (isspace((unsigned char)*step))
2938: step++;
2939: if (roff_evalnum(r, ln, step, NULL, &is, 0) == 0)
2940: is = INT_MIN;
1.109 kristaps 2941:
1.327 schwarze 2942: roff_setregn(r, key, keysz, iv, sign, is);
1.277 schwarze 2943: return ROFF_IGN;
1.203 schwarze 2944: }
2945:
1.340 schwarze 2946: static int
1.203 schwarze 2947: roff_rr(ROFF_ARGS)
2948: {
2949: struct roffreg *reg, **prev;
1.212 schwarze 2950: char *name, *cp;
2951: size_t namesz;
1.203 schwarze 2952:
1.238 schwarze 2953: name = cp = buf->buf + pos;
2954: if (*name == '\0')
1.277 schwarze 2955: return ROFF_IGN;
1.212 schwarze 2956: namesz = roff_getname(r, &cp, ln, pos);
2957: name[namesz] = '\0';
1.203 schwarze 2958:
2959: prev = &r->regtab;
2960: while (1) {
2961: reg = *prev;
1.238 schwarze 2962: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2963: break;
2964: prev = ®->next;
2965: }
1.238 schwarze 2966: if (reg != NULL) {
1.203 schwarze 2967: *prev = reg->next;
2968: free(reg->key.p);
2969: free(reg);
2970: }
1.277 schwarze 2971: return ROFF_IGN;
1.122 schwarze 2972: }
2973:
1.266 schwarze 2974: /* --- handler functions for roff requests -------------------------------- */
2975:
1.340 schwarze 2976: static int
1.122 schwarze 2977: roff_rm(ROFF_ARGS)
2978: {
2979: const char *name;
2980: char *cp;
1.212 schwarze 2981: size_t namesz;
1.122 schwarze 2982:
1.238 schwarze 2983: cp = buf->buf + pos;
2984: while (*cp != '\0') {
1.212 schwarze 2985: name = cp;
1.238 schwarze 2986: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2987: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.311 schwarze 2988: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.238 schwarze 2989: if (name[namesz] == '\\')
1.212 schwarze 2990: break;
1.122 schwarze 2991: }
1.277 schwarze 2992: return ROFF_IGN;
1.178 schwarze 2993: }
2994:
1.340 schwarze 2995: static int
1.178 schwarze 2996: roff_it(ROFF_ARGS)
2997: {
2998: int iv;
2999:
3000: /* Parse the number of lines. */
1.261 schwarze 3001:
3002: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 3003: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 3004: ln, ppos, buf->buf + 1);
1.277 schwarze 3005: return ROFF_IGN;
1.178 schwarze 3006: }
3007:
1.262 schwarze 3008: while (isspace((unsigned char)buf->buf[pos]))
3009: pos++;
3010:
3011: /*
3012: * Arm the input line trap.
3013: * Special-casing "an-trap" is an ugly workaround to cope
3014: * with DocBook stupidly fiddling with man(7) internals.
3015: */
1.261 schwarze 3016:
1.178 schwarze 3017: roffit_lines = iv;
1.262 schwarze 3018: roffit_macro = mandoc_strdup(iv != 1 ||
3019: strcmp(buf->buf + pos, "an-trap") ?
3020: buf->buf + pos : "br");
1.277 schwarze 3021: return ROFF_IGN;
1.175 schwarze 3022: }
3023:
1.340 schwarze 3024: static int
1.175 schwarze 3025: roff_Dd(ROFF_ARGS)
3026: {
1.315 schwarze 3027: int mask;
3028: enum roff_tok t, te;
1.227 schwarze 3029:
1.315 schwarze 3030: switch (tok) {
3031: case ROFF_Dd:
3032: tok = MDOC_Dd;
3033: te = MDOC_MAX;
3034: if (r->format == 0)
3035: r->format = MPARSE_MDOC;
3036: mask = MPARSE_MDOC | MPARSE_QUICK;
3037: break;
3038: case ROFF_TH:
3039: tok = MAN_TH;
3040: te = MAN_MAX;
3041: if (r->format == 0)
3042: r->format = MPARSE_MAN;
3043: mask = MPARSE_QUICK;
3044: break;
3045: default:
3046: abort();
3047: }
3048: if ((r->options & mask) == 0)
3049: for (t = tok; t < te; t++)
3050: roff_setstr(r, roff_name[t], NULL, 0);
1.277 schwarze 3051: return ROFF_CONT;
1.109 kristaps 3052: }
3053:
1.340 schwarze 3054: static int
1.109 kristaps 3055: roff_TE(ROFF_ARGS)
3056: {
1.321 schwarze 3057: if (r->tbl == NULL) {
1.221 schwarze 3058: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
3059: ln, ppos, "TE");
1.321 schwarze 3060: return ROFF_IGN;
3061: }
3062: if (tbl_end(r->tbl) == 0) {
3063: r->tbl = NULL;
1.258 schwarze 3064: free(buf->buf);
3065: buf->buf = mandoc_strdup(".sp");
3066: buf->sz = 4;
1.329 schwarze 3067: *offs = 0;
1.277 schwarze 3068: return ROFF_REPARSE;
1.258 schwarze 3069: }
1.321 schwarze 3070: r->tbl = NULL;
1.277 schwarze 3071: return ROFF_IGN;
1.112 kristaps 3072: }
3073:
1.340 schwarze 3074: static int
1.112 kristaps 3075: roff_T_(ROFF_ARGS)
3076: {
3077:
3078: if (NULL == r->tbl)
1.221 schwarze 3079: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
3080: ln, ppos, "T&");
1.112 kristaps 3081: else
1.296 schwarze 3082: tbl_restart(ln, ppos, r->tbl);
1.112 kristaps 3083:
1.277 schwarze 3084: return ROFF_IGN;
1.109 kristaps 3085: }
3086:
1.230 schwarze 3087: /*
3088: * Handle in-line equation delimiters.
3089: */
1.340 schwarze 3090: static int
1.238 schwarze 3091: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 3092: {
1.233 schwarze 3093: char *cp1, *cp2;
3094: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 3095:
1.230 schwarze 3096: /*
3097: * Outside equations, look for an opening delimiter.
3098: * If we are inside an equation, we already know it is
3099: * in-line, or this function wouldn't have been called;
3100: * so look for a closing delimiter.
3101: */
3102:
1.238 schwarze 3103: cp1 = buf->buf + pos;
1.230 schwarze 3104: cp2 = strchr(cp1, r->eqn == NULL ?
3105: r->last_eqn->odelim : r->last_eqn->cdelim);
3106: if (cp2 == NULL)
1.277 schwarze 3107: return ROFF_CONT;
1.230 schwarze 3108:
1.233 schwarze 3109: *cp2++ = '\0';
3110: bef_pr = bef_nl = aft_nl = aft_pr = "";
3111:
3112: /* Handle preceding text, protecting whitespace. */
3113:
1.238 schwarze 3114: if (*buf->buf != '\0') {
1.233 schwarze 3115: if (r->eqn == NULL)
3116: bef_pr = "\\&";
3117: bef_nl = "\n";
3118: }
3119:
3120: /*
3121: * Prepare replacing the delimiter with an equation macro
3122: * and drop leading white space from the equation.
3123: */
1.230 schwarze 3124:
1.233 schwarze 3125: if (r->eqn == NULL) {
3126: while (*cp2 == ' ')
3127: cp2++;
3128: mac = ".EQ";
3129: } else
3130: mac = ".EN";
3131:
3132: /* Handle following text, protecting whitespace. */
3133:
3134: if (*cp2 != '\0') {
3135: aft_nl = "\n";
3136: if (r->eqn != NULL)
3137: aft_pr = "\\&";
3138: }
3139:
3140: /* Do the actual replacement. */
3141:
1.238 schwarze 3142: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 3143: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 3144: free(buf->buf);
3145: buf->buf = cp1;
1.230 schwarze 3146:
3147: /* Toggle the in-line state of the eqn subsystem. */
3148:
3149: r->eqn_inline = r->eqn == NULL;
1.277 schwarze 3150: return ROFF_REPARSE;
1.151 kristaps 3151: }
3152:
1.340 schwarze 3153: static int
1.235 schwarze 3154: roff_EQ(ROFF_ARGS)
1.125 kristaps 3155: {
1.319 schwarze 3156: struct roff_node *n;
3157:
1.322 schwarze 3158: if (r->man->macroset == MACROSET_MAN)
3159: man_breakscope(r->man, ROFF_EQ);
1.319 schwarze 3160: n = roff_node_alloc(r->man, ln, ppos, ROFFT_EQN, TOKEN_NONE);
3161: if (ln > r->man->last->line)
3162: n->flags |= NODE_LINE;
3163: n->eqn = mandoc_calloc(1, sizeof(*n->eqn));
3164: n->eqn->expectargs = UINT_MAX;
3165: roff_node_append(r->man, n);
3166: r->man->next = ROFF_NEXT_SIBLING;
1.125 kristaps 3167:
1.238 schwarze 3168: assert(r->eqn == NULL);
1.319 schwarze 3169: if (r->last_eqn == NULL)
3170: r->last_eqn = eqn_alloc(r->parse);
3171: else
3172: eqn_reset(r->last_eqn);
3173: r->eqn = r->last_eqn;
3174: r->eqn->node = n;
1.151 kristaps 3175:
1.238 schwarze 3176: if (buf->buf[pos] != '\0')
1.235 schwarze 3177: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 3178: ".EQ %s", buf->buf + pos);
1.151 kristaps 3179:
1.277 schwarze 3180: return ROFF_IGN;
1.125 kristaps 3181: }
3182:
1.340 schwarze 3183: static int
1.125 kristaps 3184: roff_EN(ROFF_ARGS)
3185: {
1.319 schwarze 3186: if (r->eqn != NULL) {
3187: eqn_parse(r->eqn);
3188: r->eqn = NULL;
3189: } else
3190: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
3191: if (buf->buf[pos] != '\0')
3192: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
3193: "EN %s", buf->buf + pos);
1.277 schwarze 3194: return ROFF_IGN;
1.125 kristaps 3195: }
3196:
1.340 schwarze 3197: static int
1.109 kristaps 3198: roff_TS(ROFF_ARGS)
3199: {
1.321 schwarze 3200: if (r->tbl != NULL) {
1.221 schwarze 3201: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
3202: ln, ppos, "TS breaks TS");
1.321 schwarze 3203: tbl_end(r->tbl);
1.115 kristaps 3204: }
1.321 schwarze 3205: r->tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 3206: if (r->last_tbl)
1.321 schwarze 3207: r->last_tbl->next = r->tbl;
1.113 kristaps 3208: else
1.321 schwarze 3209: r->first_tbl = r->tbl;
3210: r->last_tbl = r->tbl;
1.297 schwarze 3211: return ROFF_IGN;
3212: }
3213:
1.340 schwarze 3214: static int
1.297 schwarze 3215: roff_onearg(ROFF_ARGS)
3216: {
3217: struct roff_node *n;
3218: char *cp;
1.305 schwarze 3219: int npos;
1.297 schwarze 3220:
1.302 schwarze 3221: if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
1.320 schwarze 3222: (tok == ROFF_ce || tok == ROFF_rj || tok == ROFF_sp ||
3223: tok == ROFF_ti))
1.302 schwarze 3224: man_breakscope(r->man, tok);
3225:
1.309 schwarze 3226: if (roffce_node != NULL && (tok == ROFF_ce || tok == ROFF_rj)) {
1.305 schwarze 3227: r->man->last = roffce_node;
3228: r->man->next = ROFF_NEXT_SIBLING;
3229: }
3230:
1.297 schwarze 3231: roff_elem_alloc(r->man, ln, ppos, tok);
3232: n = r->man->last;
3233:
3234: cp = buf->buf + pos;
3235: if (*cp != '\0') {
3236: while (*cp != '\0' && *cp != ' ')
3237: cp++;
3238: while (*cp == ' ')
3239: *cp++ = '\0';
3240: if (*cp != '\0')
3241: mandoc_vmsg(MANDOCERR_ARG_EXCESS,
3242: r->parse, ln, cp - buf->buf,
3243: "%s ... %s", roff_name[tok], cp);
3244: roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.300 schwarze 3245: }
3246:
1.309 schwarze 3247: if (tok == ROFF_ce || tok == ROFF_rj) {
3248: if (r->man->last->type == ROFFT_ELEM) {
1.305 schwarze 3249: roff_word_alloc(r->man, ln, pos, "1");
3250: r->man->last->flags |= NODE_NOSRC;
3251: }
3252: npos = 0;
3253: if (roff_evalnum(r, ln, r->man->last->string, &npos,
3254: &roffce_lines, 0) == 0) {
3255: mandoc_vmsg(MANDOCERR_CE_NONUM,
3256: r->parse, ln, pos, "ce %s", buf->buf + pos);
3257: roffce_lines = 1;
3258: }
3259: if (roffce_lines < 1) {
3260: r->man->last = r->man->last->parent;
3261: roffce_node = NULL;
3262: roffce_lines = 0;
3263: } else
3264: roffce_node = r->man->last->parent;
3265: } else {
3266: n->flags |= NODE_VALID | NODE_ENDED;
3267: r->man->last = n;
3268: }
3269: n->flags |= NODE_LINE;
1.300 schwarze 3270: r->man->next = ROFF_NEXT_SIBLING;
3271: return ROFF_IGN;
3272: }
3273:
1.340 schwarze 3274: static int
1.300 schwarze 3275: roff_manyarg(ROFF_ARGS)
3276: {
3277: struct roff_node *n;
3278: char *sp, *ep;
3279:
3280: roff_elem_alloc(r->man, ln, ppos, tok);
3281: n = r->man->last;
3282:
3283: for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
3284: while (*ep != '\0' && *ep != ' ')
3285: ep++;
3286: while (*ep == ' ')
3287: *ep++ = '\0';
3288: roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.297 schwarze 3289: }
3290:
3291: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3292: r->man->last = n;
3293: r->man->next = ROFF_NEXT_SIBLING;
1.277 schwarze 3294: return ROFF_IGN;
1.251 schwarze 3295: }
3296:
1.340 schwarze 3297: static int
1.311 schwarze 3298: roff_als(ROFF_ARGS)
3299: {
3300: char *oldn, *newn, *end, *value;
3301: size_t oldsz, newsz, valsz;
3302:
3303: newn = oldn = buf->buf + pos;
3304: if (*newn == '\0')
3305: return ROFF_IGN;
3306:
3307: newsz = roff_getname(r, &oldn, ln, pos);
3308: if (newn[newsz] == '\\' || *oldn == '\0')
3309: return ROFF_IGN;
3310:
3311: end = oldn;
3312: oldsz = roff_getname(r, &end, ln, oldn - buf->buf);
3313: if (oldsz == 0)
3314: return ROFF_IGN;
3315:
1.338 schwarze 3316: valsz = mandoc_asprintf(&value, ".%.*s \\$@\\\"\n",
1.315 schwarze 3317: (int)oldsz, oldn);
1.311 schwarze 3318: roff_setstrn(&r->strtab, newn, newsz, value, valsz, 0);
3319: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3320: free(value);
3321: return ROFF_IGN;
3322: }
3323:
1.340 schwarze 3324: static int
1.296 schwarze 3325: roff_br(ROFF_ARGS)
1.251 schwarze 3326: {
1.302 schwarze 3327: if (r->man->flags & (MAN_BLINE | MAN_ELINE))
3328: man_breakscope(r->man, ROFF_br);
1.296 schwarze 3329: roff_elem_alloc(r->man, ln, ppos, ROFF_br);
3330: if (buf->buf[pos] != '\0')
3331: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
3332: "%s %s", roff_name[tok], buf->buf + pos);
3333: r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
3334: r->man->next = ROFF_NEXT_SIBLING;
3335: return ROFF_IGN;
1.92 schwarze 3336: }
3337:
1.340 schwarze 3338: static int
1.174 kristaps 3339: roff_cc(ROFF_ARGS)
3340: {
3341: const char *p;
3342:
1.238 schwarze 3343: p = buf->buf + pos;
1.174 kristaps 3344:
1.238 schwarze 3345: if (*p == '\0' || (r->control = *p++) == '.')
1.303 schwarze 3346: r->control = '\0';
1.174 kristaps 3347:
1.238 schwarze 3348: if (*p != '\0')
1.260 schwarze 3349: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3350: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 3351:
1.341 schwarze 3352: return ROFF_IGN;
3353: }
3354:
3355: static int
3356: roff_char(ROFF_ARGS)
3357: {
3358: const char *p, *kp, *vp;
3359: size_t ksz, vsz;
3360: int font;
3361:
3362: /* Parse the character to be replaced. */
3363:
3364: kp = buf->buf + pos;
3365: p = kp + 1;
3366: if (*kp == '\0' || (*kp == '\\' &&
3367: mandoc_escape(&p, NULL, NULL) != ESCAPE_SPECIAL) ||
3368: (*p != ' ' && *p != '\0')) {
3369: mandoc_vmsg(MANDOCERR_CHAR_ARG, r->parse,
3370: ln, pos, "char %s", kp);
3371: return ROFF_IGN;
3372: }
3373: ksz = p - kp;
3374: while (*p == ' ')
3375: p++;
3376:
3377: /*
3378: * If the replacement string contains a font escape sequence,
3379: * we have to restore the font at the end.
3380: */
3381:
3382: vp = p;
3383: vsz = strlen(p);
3384: font = 0;
3385: while (*p != '\0') {
3386: if (*p++ != '\\')
3387: continue;
3388: switch (mandoc_escape(&p, NULL, NULL)) {
3389: case ESCAPE_FONT:
3390: case ESCAPE_FONTROMAN:
3391: case ESCAPE_FONTITALIC:
3392: case ESCAPE_FONTBOLD:
3393: case ESCAPE_FONTBI:
1.342 schwarze 3394: case ESCAPE_FONTCW:
1.341 schwarze 3395: case ESCAPE_FONTPREV:
3396: font++;
3397: break;
3398: default:
3399: break;
3400: }
3401: }
3402: if (font > 1)
3403: mandoc_msg(MANDOCERR_CHAR_FONT, r->parse,
3404: ln, vp - buf->buf, vp);
3405:
3406: /*
3407: * Approximate the effect of .char using the .tr tables.
3408: * XXX In groff, .char and .tr interact differently.
3409: */
3410:
3411: if (ksz == 1) {
3412: if (r->xtab == NULL)
3413: r->xtab = mandoc_calloc(128, sizeof(*r->xtab));
3414: assert((unsigned int)*kp < 128);
3415: free(r->xtab[(int)*kp].p);
3416: r->xtab[(int)*kp].sz = mandoc_asprintf(&r->xtab[(int)*kp].p,
3417: "%s%s", vp, font ? "\fP" : "");
3418: } else {
3419: roff_setstrn(&r->xmbtab, kp, ksz, vp, vsz, 0);
3420: if (font)
3421: roff_setstrn(&r->xmbtab, kp, ksz, "\\fP", 3, 1);
3422: }
1.277 schwarze 3423: return ROFF_IGN;
1.174 kristaps 3424: }
3425:
1.340 schwarze 3426: static int
1.303 schwarze 3427: roff_ec(ROFF_ARGS)
3428: {
3429: const char *p;
3430:
3431: p = buf->buf + pos;
3432: if (*p == '\0')
3433: r->escape = '\\';
3434: else {
3435: r->escape = *p;
3436: if (*++p != '\0')
3437: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
3438: ln, p - buf->buf, "ec ... %s", p);
3439: }
3440: return ROFF_IGN;
3441: }
3442:
1.340 schwarze 3443: static int
1.303 schwarze 3444: roff_eo(ROFF_ARGS)
3445: {
3446: r->escape = '\0';
3447: if (buf->buf[pos] != '\0')
3448: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
3449: ln, pos, "eo %s", buf->buf + pos);
3450: return ROFF_IGN;
1.330 schwarze 3451: }
3452:
1.340 schwarze 3453: static int
1.330 schwarze 3454: roff_nop(ROFF_ARGS)
3455: {
3456: while (buf->buf[pos] == ' ')
3457: pos++;
3458: *offs = pos;
3459: return ROFF_RERUN;
1.303 schwarze 3460: }
3461:
1.340 schwarze 3462: static int
1.164 kristaps 3463: roff_tr(ROFF_ARGS)
3464: {
3465: const char *p, *first, *second;
3466: size_t fsz, ssz;
3467: enum mandoc_esc esc;
3468:
1.238 schwarze 3469: p = buf->buf + pos;
1.164 kristaps 3470:
1.238 schwarze 3471: if (*p == '\0') {
1.260 schwarze 3472: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.277 schwarze 3473: return ROFF_IGN;
1.164 kristaps 3474: }
3475:
1.238 schwarze 3476: while (*p != '\0') {
1.164 kristaps 3477: fsz = ssz = 1;
3478:
3479: first = p++;
1.238 schwarze 3480: if (*first == '\\') {
1.164 kristaps 3481: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3482: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3483: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3484: ln, (int)(p - buf->buf), first);
1.277 schwarze 3485: return ROFF_IGN;
1.164 kristaps 3486: }
3487: fsz = (size_t)(p - first);
3488: }
3489:
3490: second = p++;
1.238 schwarze 3491: if (*second == '\\') {
1.164 kristaps 3492: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 3493: if (esc == ESCAPE_ERROR) {
1.219 schwarze 3494: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 3495: ln, (int)(p - buf->buf), second);
1.277 schwarze 3496: return ROFF_IGN;
1.164 kristaps 3497: }
3498: ssz = (size_t)(p - second);
1.238 schwarze 3499: } else if (*second == '\0') {
1.260 schwarze 3500: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
3501: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 3502: second = " ";
1.165 kristaps 3503: p--;
1.164 kristaps 3504: }
3505:
1.167 kristaps 3506: if (fsz > 1) {
1.207 schwarze 3507: roff_setstrn(&r->xmbtab, first, fsz,
3508: second, ssz, 0);
1.167 kristaps 3509: continue;
3510: }
3511:
1.238 schwarze 3512: if (r->xtab == NULL)
1.207 schwarze 3513: r->xtab = mandoc_calloc(128,
3514: sizeof(struct roffstr));
1.167 kristaps 3515:
3516: free(r->xtab[(int)*first].p);
3517: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
3518: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 3519: }
3520:
1.277 schwarze 3521: return ROFF_IGN;
1.164 kristaps 3522: }
3523:
1.339 schwarze 3524: /*
3525: * Implementation of the .return request.
3526: * There is no need to call roff_userret() from here.
3527: * The read module will call that after rewinding the reader stack
3528: * to the place from where the current macro was called.
3529: */
1.340 schwarze 3530: static int
1.339 schwarze 3531: roff_return(ROFF_ARGS)
3532: {
3533: if (r->mstackpos >= 0)
1.340 schwarze 3534: return ROFF_IGN | ROFF_USERRET;
1.339 schwarze 3535:
3536: mandoc_msg(MANDOCERR_REQ_NOMAC, r->parse, ln, ppos, "return");
3537: return ROFF_IGN;
3538: }
3539:
1.340 schwarze 3540: static int
1.306 schwarze 3541: roff_rn(ROFF_ARGS)
3542: {
3543: const char *value;
3544: char *oldn, *newn, *end;
3545: size_t oldsz, newsz;
1.315 schwarze 3546: int deftype;
1.306 schwarze 3547:
3548: oldn = newn = buf->buf + pos;
3549: if (*oldn == '\0')
3550: return ROFF_IGN;
3551:
3552: oldsz = roff_getname(r, &newn, ln, pos);
3553: if (oldn[oldsz] == '\\' || *newn == '\0')
3554: return ROFF_IGN;
3555:
3556: end = newn;
3557: newsz = roff_getname(r, &end, ln, newn - buf->buf);
3558: if (newsz == 0)
3559: return ROFF_IGN;
3560:
1.315 schwarze 3561: deftype = ROFFDEF_ANY;
3562: value = roff_getstrn(r, oldn, oldsz, &deftype);
3563: switch (deftype) {
3564: case ROFFDEF_USER:
1.306 schwarze 3565: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3566: roff_setstrn(&r->strtab, oldn, oldsz, NULL, 0, 0);
3567: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
1.315 schwarze 3568: break;
3569: case ROFFDEF_PRE:
3570: roff_setstrn(&r->strtab, newn, newsz, value, strlen(value), 0);
3571: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3572: break;
3573: case ROFFDEF_REN:
1.306 schwarze 3574: roff_setstrn(&r->rentab, newn, newsz, value, strlen(value), 0);
3575: roff_setstrn(&r->rentab, oldn, oldsz, NULL, 0, 0);
1.315 schwarze 3576: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3577: break;
3578: case ROFFDEF_STD:
1.306 schwarze 3579: roff_setstrn(&r->rentab, newn, newsz, oldn, oldsz, 0);
1.315 schwarze 3580: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3581: break;
3582: default:
3583: roff_setstrn(&r->strtab, newn, newsz, NULL, 0, 0);
3584: roff_setstrn(&r->rentab, newn, newsz, NULL, 0, 0);
3585: break;
3586: }
1.306 schwarze 3587: return ROFF_IGN;
3588: }
3589:
1.340 schwarze 3590: static int
1.339 schwarze 3591: roff_shift(ROFF_ARGS)
3592: {
3593: struct mctx *ctx;
3594: int levels, i;
3595:
3596: levels = 1;
3597: if (buf->buf[pos] != '\0' &&
3598: roff_evalnum(r, ln, buf->buf, &pos, &levels, 0) == 0) {
3599: mandoc_vmsg(MANDOCERR_CE_NONUM, r->parse,
3600: ln, pos, "shift %s", buf->buf + pos);
3601: levels = 1;
3602: }
3603: if (r->mstackpos < 0) {
3604: mandoc_msg(MANDOCERR_REQ_NOMAC, r->parse, ln, ppos, "shift");
3605: return ROFF_IGN;
3606: }
3607: ctx = r->mstack + r->mstackpos;
3608: if (levels > ctx->argc) {
3609: mandoc_vmsg(MANDOCERR_SHIFT, r->parse,
3610: ln, pos, "%d, but max is %d", levels, ctx->argc);
3611: levels = ctx->argc;
3612: }
3613: if (levels == 0)
3614: return ROFF_IGN;
3615: for (i = 0; i < levels; i++)
3616: free(ctx->argv[i]);
3617: ctx->argc -= levels;
3618: for (i = 0; i < ctx->argc; i++)
3619: ctx->argv[i] = ctx->argv[i + levels];
3620: return ROFF_IGN;
3621: }
3622:
1.340 schwarze 3623: static int
1.105 kristaps 3624: roff_so(ROFF_ARGS)
3625: {
1.249 schwarze 3626: char *name, *cp;
1.105 kristaps 3627:
1.238 schwarze 3628: name = buf->buf + pos;
1.224 schwarze 3629: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 3630:
3631: /*
3632: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3633: * opening anything that's not in our cwd or anything beneath
3634: * it. Thus, explicitly disallow traversing up the file-system
3635: * or using absolute paths.
3636: */
3637:
1.238 schwarze 3638: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 3639: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3640: ".so %s", name);
1.249 schwarze 3641: buf->sz = mandoc_asprintf(&cp,
3642: ".sp\nSee the file %s.\n.sp", name) + 1;
3643: free(buf->buf);
3644: buf->buf = cp;
3645: *offs = 0;
1.277 schwarze 3646: return ROFF_REPARSE;
1.105 kristaps 3647: }
3648:
3649: *offs = pos;
1.277 schwarze 3650: return ROFF_SO;
1.105 kristaps 3651: }
1.92 schwarze 3652:
1.266 schwarze 3653: /* --- user defined strings and macros ------------------------------------ */
3654:
1.340 schwarze 3655: static int
1.106 kristaps 3656: roff_userdef(ROFF_ARGS)
1.99 kristaps 3657: {
1.339 schwarze 3658: struct mctx *ctx;
3659: char *arg, *ap, *dst, *src;
3660: size_t sz;
3661:
3662: /* Initialize a new macro stack context. */
3663:
3664: if (++r->mstackpos == r->mstacksz) {
3665: r->mstack = mandoc_recallocarray(r->mstack,
3666: r->mstacksz, r->mstacksz + 8, sizeof(*r->mstack));
3667: r->mstacksz += 8;
3668: }
3669: ctx = r->mstack + r->mstackpos;
3670: ctx->argsz = 0;
3671: ctx->argc = 0;
3672: ctx->argv = NULL;
3673:
3674: /*
3675: * Collect pointers to macro argument strings,
3676: * NUL-terminating them and escaping quotes.
3677: */
3678:
3679: src = buf->buf + pos;
3680: while (*src != '\0') {
3681: if (ctx->argc == ctx->argsz) {
3682: ctx->argsz += 8;
3683: ctx->argv = mandoc_reallocarray(ctx->argv,
3684: ctx->argsz, sizeof(*ctx->argv));
3685: }
3686: arg = mandoc_getarg(r->parse, &src, ln, &pos);
3687: sz = 1; /* For the terminating NUL. */
3688: for (ap = arg; *ap != '\0'; ap++)
3689: sz += *ap == '"' ? 4 : 1;
3690: ctx->argv[ctx->argc++] = dst = mandoc_malloc(sz);
3691: for (ap = arg; *ap != '\0'; ap++) {
3692: if (*ap == '"') {
3693: memcpy(dst, "\\(dq", 4);
3694: dst += 4;
3695: } else
3696: *dst++ = *ap;
1.106 kristaps 3697: }
1.339 schwarze 3698: *dst = '\0';
1.337 schwarze 3699: }
3700:
1.339 schwarze 3701: /* Replace the macro invocation by the macro definition. */
1.263 schwarze 3702:
1.238 schwarze 3703: free(buf->buf);
1.339 schwarze 3704: buf->buf = mandoc_strdup(r->current_string);
3705: buf->sz = strlen(buf->buf) + 1;
1.248 schwarze 3706: *offs = 0;
1.106 kristaps 3707:
1.277 schwarze 3708: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.340 schwarze 3709: ROFF_REPARSE | ROFF_USERCALL : ROFF_IGN | ROFF_APPEND;
1.99 kristaps 3710: }
1.121 schwarze 3711:
1.306 schwarze 3712: /*
3713: * Calling a high-level macro that was renamed with .rn.
3714: * r->current_string has already been set up by roff_parse().
3715: */
1.340 schwarze 3716: static int
1.306 schwarze 3717: roff_renamed(ROFF_ARGS)
3718: {
3719: char *nbuf;
3720:
1.315 schwarze 3721: buf->sz = mandoc_asprintf(&nbuf, ".%s%s%s", r->current_string,
3722: buf->buf[pos] == '\0' ? "" : " ", buf->buf + pos) + 1;
1.306 schwarze 3723: free(buf->buf);
3724: buf->buf = nbuf;
1.329 schwarze 3725: *offs = 0;
1.306 schwarze 3726: return ROFF_CONT;
3727: }
3728:
1.212 schwarze 3729: static size_t
1.121 schwarze 3730: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3731: {
3732: char *name, *cp;
1.212 schwarze 3733: size_t namesz;
1.121 schwarze 3734:
3735: name = *cpp;
3736: if ('\0' == *name)
1.277 schwarze 3737: return 0;
1.121 schwarze 3738:
1.212 schwarze 3739: /* Read until end of name and terminate it with NUL. */
3740: for (cp = name; 1; cp++) {
3741: if ('\0' == *cp || ' ' == *cp) {
3742: namesz = cp - name;
3743: break;
3744: }
1.121 schwarze 3745: if ('\\' != *cp)
3746: continue;
1.215 schwarze 3747: namesz = cp - name;
3748: if ('{' == cp[1] || '}' == cp[1])
3749: break;
1.121 schwarze 3750: cp++;
3751: if ('\\' == *cp)
3752: continue;
1.224 schwarze 3753: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3754: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 3755: mandoc_escape((const char **)&cp, NULL, NULL);
3756: break;
1.121 schwarze 3757: }
3758:
3759: /* Read past spaces. */
3760: while (' ' == *cp)
3761: cp++;
3762:
3763: *cpp = cp;
1.277 schwarze 3764: return namesz;
1.121 schwarze 3765: }
3766:
1.106 kristaps 3767: /*
3768: * Store *string into the user-defined string called *name.
3769: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 3770: * append == 0: replace mode
3771: * append == 1: single-line append mode
3772: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 3773: */
1.94 kristaps 3774: static void
1.106 kristaps 3775: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 3776: int append)
1.92 schwarze 3777: {
1.315 schwarze 3778: size_t namesz;
1.164 kristaps 3779:
1.315 schwarze 3780: namesz = strlen(name);
3781: roff_setstrn(&r->strtab, name, namesz, string,
1.207 schwarze 3782: string ? strlen(string) : 0, append);
1.315 schwarze 3783: roff_setstrn(&r->rentab, name, namesz, NULL, 0, 0);
1.164 kristaps 3784: }
3785:
3786: static void
1.166 kristaps 3787: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 3788: const char *string, size_t stringsz, int append)
1.164 kristaps 3789: {
1.166 kristaps 3790: struct roffkv *n;
1.164 kristaps 3791: char *c;
3792: int i;
3793: size_t oldch, newch;
1.92 schwarze 3794:
1.106 kristaps 3795: /* Search for an existing string with the same name. */
1.164 kristaps 3796: n = *r;
3797:
1.211 schwarze 3798: while (n && (namesz != n->key.sz ||
3799: strncmp(n->key.p, name, namesz)))
1.92 schwarze 3800: n = n->next;
1.94 kristaps 3801:
3802: if (NULL == n) {
1.106 kristaps 3803: /* Create a new string table entry. */
1.166 kristaps 3804: n = mandoc_malloc(sizeof(struct roffkv));
3805: n->key.p = mandoc_strndup(name, namesz);
3806: n->key.sz = namesz;
3807: n->val.p = NULL;
3808: n->val.sz = 0;
1.164 kristaps 3809: n->next = *r;
3810: *r = n;
1.193 schwarze 3811: } else if (0 == append) {
1.166 kristaps 3812: free(n->val.p);
3813: n->val.p = NULL;
3814: n->val.sz = 0;
1.106 kristaps 3815: }
3816:
3817: if (NULL == string)
3818: return;
3819:
3820: /*
3821: * One additional byte for the '\n' in multiline mode,
3822: * and one for the terminating '\0'.
3823: */
1.193 schwarze 3824: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 3825:
1.166 kristaps 3826: if (NULL == n->val.p) {
3827: n->val.p = mandoc_malloc(newch);
3828: *n->val.p = '\0';
1.106 kristaps 3829: oldch = 0;
3830: } else {
1.166 kristaps 3831: oldch = n->val.sz;
3832: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 3833: }
3834:
3835: /* Skip existing content in the destination buffer. */
1.166 kristaps 3836: c = n->val.p + (int)oldch;
1.106 kristaps 3837:
3838: /* Append new content to the destination buffer. */
1.164 kristaps 3839: i = 0;
3840: while (i < (int)stringsz) {
1.106 kristaps 3841: /*
3842: * Rudimentary roff copy mode:
3843: * Handle escaped backslashes.
3844: */
1.164 kristaps 3845: if ('\\' == string[i] && '\\' == string[i + 1])
3846: i++;
3847: *c++ = string[i++];
1.106 kristaps 3848: }
1.94 kristaps 3849:
1.106 kristaps 3850: /* Append terminating bytes. */
1.193 schwarze 3851: if (1 < append)
1.106 kristaps 3852: *c++ = '\n';
1.163 kristaps 3853:
1.106 kristaps 3854: *c = '\0';
1.166 kristaps 3855: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 3856: }
3857:
1.94 kristaps 3858: static const char *
1.325 schwarze 3859: roff_getstrn(struct roff *r, const char *name, size_t len,
1.315 schwarze 3860: int *deftype)
1.92 schwarze 3861: {
1.315 schwarze 3862: const struct roffkv *n;
1.325 schwarze 3863: int found, i;
1.315 schwarze 3864: enum roff_tok tok;
3865:
1.325 schwarze 3866: found = 0;
3867: for (n = r->strtab; n != NULL; n = n->next) {
3868: if (strncmp(name, n->key.p, len) != 0 ||
3869: n->key.p[len] != '\0' || n->val.p == NULL)
3870: continue;
3871: if (*deftype & ROFFDEF_USER) {
3872: *deftype = ROFFDEF_USER;
3873: return n->val.p;
3874: } else {
3875: found = 1;
3876: break;
3877: }
3878: }
3879: for (n = r->rentab; n != NULL; n = n->next) {
3880: if (strncmp(name, n->key.p, len) != 0 ||
3881: n->key.p[len] != '\0' || n->val.p == NULL)
3882: continue;
3883: if (*deftype & ROFFDEF_REN) {
3884: *deftype = ROFFDEF_REN;
3885: return n->val.p;
3886: } else {
3887: found = 1;
3888: break;
1.315 schwarze 3889: }
3890: }
1.325 schwarze 3891: for (i = 0; i < PREDEFS_MAX; i++) {
3892: if (strncmp(name, predefs[i].name, len) != 0 ||
3893: predefs[i].name[len] != '\0')
3894: continue;
3895: if (*deftype & ROFFDEF_PRE) {
3896: *deftype = ROFFDEF_PRE;
3897: return predefs[i].str;
3898: } else {
3899: found = 1;
3900: break;
1.315 schwarze 3901: }
3902: }
1.325 schwarze 3903: if (r->man->macroset != MACROSET_MAN) {
3904: for (tok = MDOC_Dd; tok < MDOC_MAX; tok++) {
3905: if (strncmp(name, roff_name[tok], len) != 0 ||
3906: roff_name[tok][len] != '\0')
3907: continue;
3908: if (*deftype & ROFFDEF_STD) {
3909: *deftype = ROFFDEF_STD;
3910: return NULL;
3911: } else {
3912: found = 1;
3913: break;
1.315 schwarze 3914: }
3915: }
3916: }
1.325 schwarze 3917: if (r->man->macroset != MACROSET_MDOC) {
3918: for (tok = MAN_TH; tok < MAN_MAX; tok++) {
3919: if (strncmp(name, roff_name[tok], len) != 0 ||
3920: roff_name[tok][len] != '\0')
3921: continue;
3922: if (*deftype & ROFFDEF_STD) {
3923: *deftype = ROFFDEF_STD;
3924: return NULL;
3925: } else {
3926: found = 1;
3927: break;
1.315 schwarze 3928: }
3929: }
1.325 schwarze 3930: }
3931:
3932: if (found == 0 && *deftype != ROFFDEF_ANY) {
3933: if (*deftype & ROFFDEF_REN) {
3934: /*
3935: * This might still be a request,
3936: * so do not treat it as undefined yet.
3937: */
3938: *deftype = ROFFDEF_UNDEF;
3939: return NULL;
1.315 schwarze 3940: }
1.325 schwarze 3941:
3942: /* Using an undefined string defines it to be empty. */
3943:
3944: roff_setstrn(&r->strtab, name, len, "", 0, 0);
3945: roff_setstrn(&r->rentab, name, len, NULL, 0, 0);
1.315 schwarze 3946: }
1.325 schwarze 3947:
1.315 schwarze 3948: *deftype = 0;
1.277 schwarze 3949: return NULL;
1.92 schwarze 3950: }
3951:
1.94 kristaps 3952: static void
1.167 kristaps 3953: roff_freestr(struct roffkv *r)
1.92 schwarze 3954: {
1.166 kristaps 3955: struct roffkv *n, *nn;
1.92 schwarze 3956:
1.167 kristaps 3957: for (n = r; n; n = nn) {
1.166 kristaps 3958: free(n->key.p);
3959: free(n->val.p);
1.92 schwarze 3960: nn = n->next;
3961: free(n);
3962: }
1.114 kristaps 3963: }
1.266 schwarze 3964:
3965: /* --- accessors and utility functions ------------------------------------ */
1.164 kristaps 3966:
3967: /*
3968: * Duplicate an input string, making the appropriate character
3969: * conversations (as stipulated by `tr') along the way.
3970: * Returns a heap-allocated string with all the replacements made.
3971: */
3972: char *
3973: roff_strdup(const struct roff *r, const char *p)
3974: {
1.166 kristaps 3975: const struct roffkv *cp;
1.164 kristaps 3976: char *res;
3977: const char *pp;
3978: size_t ssz, sz;
3979: enum mandoc_esc esc;
3980:
1.167 kristaps 3981: if (NULL == r->xmbtab && NULL == r->xtab)
1.277 schwarze 3982: return mandoc_strdup(p);
1.164 kristaps 3983: else if ('\0' == *p)
1.277 schwarze 3984: return mandoc_strdup("");
1.164 kristaps 3985:
3986: /*
3987: * Step through each character looking for term matches
3988: * (remember that a `tr' can be invoked with an escape, which is
3989: * a glyph but the escape is multi-character).
3990: * We only do this if the character hash has been initialised
3991: * and the string is >0 length.
3992: */
3993:
3994: res = NULL;
3995: ssz = 0;
3996:
3997: while ('\0' != *p) {
1.289 schwarze 3998: assert((unsigned int)*p < 128);
3999: if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167 kristaps 4000: sz = r->xtab[(int)*p].sz;
4001: res = mandoc_realloc(res, ssz + sz + 1);
4002: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
4003: ssz += sz;
4004: p++;
4005: continue;
4006: } else if ('\\' != *p) {
4007: res = mandoc_realloc(res, ssz + 2);
4008: res[ssz++] = *p++;
4009: continue;
4010: }
4011:
1.164 kristaps 4012: /* Search for term matches. */
1.167 kristaps 4013: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 4014: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 4015: break;
4016:
4017: if (NULL != cp) {
4018: /*
4019: * A match has been found.
4020: * Append the match to the array and move
4021: * forward by its keysize.
4022: */
1.207 schwarze 4023: res = mandoc_realloc(res,
4024: ssz + cp->val.sz + 1);
1.166 kristaps 4025: memcpy(res + ssz, cp->val.p, cp->val.sz);
4026: ssz += cp->val.sz;
4027: p += (int)cp->key.sz;
1.164 kristaps 4028: continue;
4029: }
4030:
1.167 kristaps 4031: /*
4032: * Handle escapes carefully: we need to copy
4033: * over just the escape itself, or else we might
4034: * do replacements within the escape itself.
4035: * Make sure to pass along the bogus string.
4036: */
4037: pp = p++;
4038: esc = mandoc_escape(&p, NULL, NULL);
4039: if (ESCAPE_ERROR == esc) {
4040: sz = strlen(pp);
1.164 kristaps 4041: res = mandoc_realloc(res, ssz + sz + 1);
4042: memcpy(res + ssz, pp, sz);
1.167 kristaps 4043: break;
1.164 kristaps 4044: }
1.207 schwarze 4045: /*
4046: * We bail out on bad escapes.
1.167 kristaps 4047: * No need to warn: we already did so when
4048: * roff_res() was called.
4049: */
4050: sz = (int)(p - pp);
4051: res = mandoc_realloc(res, ssz + sz + 1);
4052: memcpy(res + ssz, pp, sz);
4053: ssz += sz;
1.164 kristaps 4054: }
4055:
4056: res[(int)ssz] = '\0';
1.277 schwarze 4057: return res;
1.227 schwarze 4058: }
4059:
4060: int
4061: roff_getformat(const struct roff *r)
4062: {
4063:
1.277 schwarze 4064: return r->format;
1.174 kristaps 4065: }
4066:
4067: /*
1.207 schwarze 4068: * Find out whether a line is a macro line or not.
1.174 kristaps 4069: * If it is, adjust the current position and return one; if it isn't,
4070: * return zero and don't change the current position.
4071: * If the control character has been set with `.cc', then let that grain
4072: * precedence.
4073: * This is slighly contrary to groff, where using the non-breaking
4074: * control character when `cc' has been invoked will cause the
4075: * non-breaking macro contents to be printed verbatim.
4076: */
4077: int
4078: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
4079: {
4080: int pos;
4081:
4082: pos = *ppos;
4083:
1.303 schwarze 4084: if (r->control != '\0' && cp[pos] == r->control)
1.174 kristaps 4085: pos++;
1.303 schwarze 4086: else if (r->control != '\0')
1.277 schwarze 4087: return 0;
1.174 kristaps 4088: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
4089: pos += 2;
4090: else if ('.' == cp[pos] || '\'' == cp[pos])
4091: pos++;
4092: else
1.277 schwarze 4093: return 0;
1.174 kristaps 4094:
4095: while (' ' == cp[pos] || '\t' == cp[pos])
4096: pos++;
4097:
4098: *ppos = pos;
1.277 schwarze 4099: return 1;
1.74 kristaps 4100: }
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