Annotation of mandoc/roff.c, Revision 1.304
1.304 ! schwarze 1: /* $Id: roff.c,v 1.303 2017/06/04 00:13:15 schwarze Exp $ */
1.1 kristaps 2: /*
1.267 schwarze 3: * Copyright (c) 2008-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.287 schwarze 4: * Copyright (c) 2010-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
1.1 kristaps 5: *
6: * Permission to use, copy, modify, and distribute this software for any
1.66 kristaps 7: * purpose with or without fee is hereby granted, provided that the above
8: * copyright notice and this permission notice appear in all copies.
1.1 kristaps 9: *
1.106 kristaps 10: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.66 kristaps 11: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.106 kristaps 12: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.66 kristaps 13: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
1.1 kristaps 17: */
1.66 kristaps 18: #include "config.h"
1.225 schwarze 19:
20: #include <sys/types.h>
1.30 kristaps 21:
1.67 kristaps 22: #include <assert.h>
1.85 kristaps 23: #include <ctype.h>
1.245 schwarze 24: #include <limits.h>
1.295 schwarze 25: #include <stddef.h>
26: #include <stdint.h>
1.178 schwarze 27: #include <stdio.h>
1.1 kristaps 28: #include <stdlib.h>
1.67 kristaps 29: #include <string.h>
1.1 kristaps 30:
1.67 kristaps 31: #include "mandoc.h"
1.201 schwarze 32: #include "mandoc_aux.h"
1.295 schwarze 33: #include "mandoc_ohash.h"
1.265 schwarze 34: #include "roff.h"
1.226 schwarze 35: #include "libmandoc.h"
1.266 schwarze 36: #include "roff_int.h"
1.109 kristaps 37: #include "libroff.h"
1.82 kristaps 38:
1.170 schwarze 39: /* Maximum number of string expansions per line, to break infinite loops. */
40: #define EXPAND_LIMIT 1000
41:
1.266 schwarze 42: /* --- data types --------------------------------------------------------- */
43:
1.147 kristaps 44: /*
1.167 kristaps 45: * An incredibly-simple string buffer.
46: */
1.94 kristaps 47: struct roffstr {
1.167 kristaps 48: char *p; /* nil-terminated buffer */
49: size_t sz; /* saved strlen(p) */
1.166 kristaps 50: };
51:
52: /*
1.167 kristaps 53: * A key-value roffstr pair as part of a singly-linked list.
1.166 kristaps 54: */
55: struct roffkv {
56: struct roffstr key;
57: struct roffstr val;
58: struct roffkv *next; /* next in list */
1.94 kristaps 59: };
60:
1.180 schwarze 61: /*
62: * A single number register as part of a singly-linked list.
63: */
64: struct roffreg {
65: struct roffstr key;
1.181 schwarze 66: int val;
1.180 schwarze 67: struct roffreg *next;
68: };
69:
1.295 schwarze 70: /*
71: * Association of request and macro names with token IDs.
72: */
73: struct roffreq {
74: enum roff_tok tok;
75: char name[];
76: };
77:
1.67 kristaps 78: struct roff {
1.128 kristaps 79: struct mparse *parse; /* parse point */
1.296 schwarze 80: struct roff_man *man; /* mdoc or man parser */
1.67 kristaps 81: struct roffnode *last; /* leaf of stack */
1.223 schwarze 82: int *rstack; /* stack of inverted `ie' values */
1.295 schwarze 83: struct ohash *reqtab; /* request lookup table */
1.180 schwarze 84: struct roffreg *regtab; /* number registers */
1.166 kristaps 85: struct roffkv *strtab; /* user-defined strings & macros */
1.167 kristaps 86: struct roffkv *xmbtab; /* multi-byte trans table (`tr') */
87: struct roffstr *xtab; /* single-byte trans table (`tr') */
1.106 kristaps 88: const char *current_string; /* value of last called user macro */
1.118 kristaps 89: struct tbl_node *first_tbl; /* first table parsed */
90: struct tbl_node *last_tbl; /* last table parsed */
91: struct tbl_node *tbl; /* current table being parsed */
1.125 kristaps 92: struct eqn_node *last_eqn; /* last equation parsed */
93: struct eqn_node *first_eqn; /* first equation parsed */
94: struct eqn_node *eqn; /* current equation being parsed */
1.230 schwarze 95: int eqn_inline; /* current equation is inline */
1.223 schwarze 96: int options; /* parse options */
97: int rstacksz; /* current size limit of rstack */
98: int rstackpos; /* position in rstack */
1.227 schwarze 99: int format; /* current file in mdoc or man format */
1.273 schwarze 100: int argc; /* number of args of the last macro */
1.223 schwarze 101: char control; /* control character */
1.303 schwarze 102: char escape; /* escape character */
1.79 kristaps 103: };
104:
1.67 kristaps 105: struct roffnode {
1.294 schwarze 106: enum roff_tok tok; /* type of node */
1.67 kristaps 107: struct roffnode *parent; /* up one in stack */
108: int line; /* parse line */
109: int col; /* parse col */
1.106 kristaps 110: char *name; /* node name, e.g. macro name */
1.79 kristaps 111: char *end; /* end-rules: custom token */
112: int endspan; /* end-rules: next-line or infty */
1.198 schwarze 113: int rule; /* current evaluation rule */
1.67 kristaps 114: };
115:
116: #define ROFF_ARGS struct roff *r, /* parse ctx */ \
1.294 schwarze 117: enum roff_tok tok, /* tok of macro */ \
1.238 schwarze 118: struct buf *buf, /* input buffer */ \
1.67 kristaps 119: int ln, /* parse line */ \
1.75 kristaps 120: int ppos, /* original pos in buffer */ \
121: int pos, /* current pos in buffer */ \
1.74 kristaps 122: int *offs /* reset offset of buffer data */
1.67 kristaps 123:
124: typedef enum rofferr (*roffproc)(ROFF_ARGS);
125:
126: struct roffmac {
1.79 kristaps 127: roffproc proc; /* process new macro */
128: roffproc text; /* process as child text of macro */
129: roffproc sub; /* process as child of macro */
130: int flags;
131: #define ROFFMAC_STRUCT (1 << 0) /* always interpret */
1.67 kristaps 132: };
133:
1.141 kristaps 134: struct predef {
135: const char *name; /* predefined input name */
136: const char *str; /* replacement symbol */
137: };
138:
139: #define PREDEF(__name, __str) \
140: { (__name), (__str) },
141:
1.266 schwarze 142: /* --- function prototypes ------------------------------------------------ */
143:
1.155 kristaps 144: static void roffnode_cleanscope(struct roff *);
145: static void roffnode_pop(struct roff *);
1.294 schwarze 146: static void roffnode_push(struct roff *, enum roff_tok,
1.155 kristaps 147: const char *, int, int);
1.80 kristaps 148: static enum rofferr roff_block(ROFF_ARGS);
149: static enum rofferr roff_block_text(ROFF_ARGS);
150: static enum rofferr roff_block_sub(ROFF_ARGS);
1.296 schwarze 151: static enum rofferr roff_br(ROFF_ARGS);
1.80 kristaps 152: static enum rofferr roff_cblock(ROFF_ARGS);
1.174 kristaps 153: static enum rofferr roff_cc(ROFF_ARGS);
1.195 schwarze 154: static void roff_ccond(struct roff *, int, int);
1.82 kristaps 155: static enum rofferr roff_cond(ROFF_ARGS);
156: static enum rofferr roff_cond_text(ROFF_ARGS);
157: static enum rofferr roff_cond_sub(ROFF_ARGS);
1.92 schwarze 158: static enum rofferr roff_ds(ROFF_ARGS);
1.303 schwarze 159: static enum rofferr roff_ec(ROFF_ARGS);
160: static enum rofferr roff_eo(ROFF_ARGS);
1.238 schwarze 161: static enum rofferr roff_eqndelim(struct roff *, struct buf *, int);
1.271 schwarze 162: static int roff_evalcond(struct roff *r, int, char *, int *);
1.234 kristaps 163: static int roff_evalnum(struct roff *, int,
164: const char *, int *, int *, int);
165: static int roff_evalpar(struct roff *, int,
1.261 schwarze 166: const char *, int *, int *, int);
1.198 schwarze 167: static int roff_evalstrcond(const char *, int *);
1.155 kristaps 168: static void roff_free1(struct roff *);
1.180 schwarze 169: static void roff_freereg(struct roffreg *);
1.167 kristaps 170: static void roff_freestr(struct roffkv *);
1.212 schwarze 171: static size_t roff_getname(struct roff *, char **, int, int);
1.261 schwarze 172: static int roff_getnum(const char *, int *, int *, int);
1.184 schwarze 173: static int roff_getop(const char *, int *, char *);
1.181 schwarze 174: static int roff_getregn(const struct roff *,
175: const char *, size_t);
1.273 schwarze 176: static int roff_getregro(const struct roff *,
177: const char *name);
1.207 schwarze 178: static const char *roff_getstrn(const struct roff *,
1.94 kristaps 179: const char *, size_t);
1.271 schwarze 180: static int roff_hasregn(const struct roff *,
181: const char *, size_t);
1.251 schwarze 182: static enum rofferr roff_insec(ROFF_ARGS);
1.178 schwarze 183: static enum rofferr roff_it(ROFF_ARGS);
1.103 kristaps 184: static enum rofferr roff_line_ignore(ROFF_ARGS);
1.265 schwarze 185: static void roff_man_alloc1(struct roff_man *);
186: static void roff_man_free1(struct roff_man *);
1.300 schwarze 187: static enum rofferr roff_manyarg(ROFF_ARGS);
1.89 kristaps 188: static enum rofferr roff_nr(ROFF_ARGS);
1.297 schwarze 189: static enum rofferr roff_onearg(ROFF_ARGS);
1.294 schwarze 190: static enum roff_tok roff_parse(struct roff *, char *, int *,
1.214 schwarze 191: int, int);
1.238 schwarze 192: static enum rofferr roff_parsetext(struct buf *, int, int *);
193: static enum rofferr roff_res(struct roff *, struct buf *, int, int);
1.122 schwarze 194: static enum rofferr roff_rm(ROFF_ARGS);
1.203 schwarze 195: static enum rofferr roff_rr(ROFF_ARGS);
1.94 kristaps 196: static void roff_setstr(struct roff *,
1.106 kristaps 197: const char *, const char *, int);
1.207 schwarze 198: static void roff_setstrn(struct roffkv **, const char *,
1.164 kristaps 199: size_t, const char *, size_t, int);
1.105 kristaps 200: static enum rofferr roff_so(ROFF_ARGS);
1.164 kristaps 201: static enum rofferr roff_tr(ROFF_ARGS);
1.175 schwarze 202: static enum rofferr roff_Dd(ROFF_ARGS);
203: static enum rofferr roff_TH(ROFF_ARGS);
1.109 kristaps 204: static enum rofferr roff_TE(ROFF_ARGS);
205: static enum rofferr roff_TS(ROFF_ARGS);
1.125 kristaps 206: static enum rofferr roff_EQ(ROFF_ARGS);
207: static enum rofferr roff_EN(ROFF_ARGS);
1.112 kristaps 208: static enum rofferr roff_T_(ROFF_ARGS);
1.251 schwarze 209: static enum rofferr roff_unsupp(ROFF_ARGS);
1.106 kristaps 210: static enum rofferr roff_userdef(ROFF_ARGS);
1.67 kristaps 211:
1.266 schwarze 212: /* --- constant data ------------------------------------------------------ */
213:
1.261 schwarze 214: #define ROFFNUM_SCALE (1 << 0) /* Honour scaling in roff_getnum(). */
215: #define ROFFNUM_WHITE (1 << 1) /* Skip whitespace in roff_evalnum(). */
216:
1.294 schwarze 217: const char *__roff_name[MAN_MAX + 1] = {
1.304 ! schwarze 218: "br", "ft", "ll", "mc",
! 219: "sp", "ta", "ti", NULL,
1.294 schwarze 220: "ab", "ad", "af", "aln",
221: "als", "am", "am1", "ami",
222: "ami1", "as", "as1", "asciify",
223: "backtrace", "bd", "bleedat", "blm",
224: "box", "boxa", "bp", "BP",
225: "break", "breakchar", "brnl", "brp",
226: "brpnl", "c2", "cc", "ce",
227: "cf", "cflags", "ch", "char",
228: "chop", "class", "close", "CL",
229: "color", "composite", "continue", "cp",
230: "cropat", "cs", "cu", "da",
231: "dch", "Dd", "de", "de1",
232: "defcolor", "dei", "dei1", "device",
233: "devicem", "di", "do", "ds",
234: "ds1", "dwh", "dt", "ec",
235: "ecr", "ecs", "el", "em",
236: "EN", "eo", "EP", "EQ",
237: "errprint", "ev", "evc", "ex",
238: "fallback", "fam", "fc", "fchar",
239: "fcolor", "fdeferlig", "feature", "fkern",
240: "fl", "flig", "fp", "fps",
241: "fschar", "fspacewidth", "fspecial", "ftr",
242: "fzoom", "gcolor", "hc", "hcode",
243: "hidechar", "hla", "hlm", "hpf",
244: "hpfa", "hpfcode", "hw", "hy",
245: "hylang", "hylen", "hym", "hypp",
246: "hys", "ie", "if", "ig",
247: "index", "it", "itc", "IX",
248: "kern", "kernafter", "kernbefore", "kernpair",
249: "lc", "lc_ctype", "lds", "length",
250: "letadj", "lf", "lg", "lhang",
251: "linetabs", "lnr", "lnrf", "lpfx",
1.304 ! schwarze 252: "ls", "lsm", "lt",
1.294 schwarze 253: "mediasize", "minss", "mk", "mso",
254: "na", "ne", "nh", "nhychar",
255: "nm", "nn", "nop", "nr",
256: "nrf", "nroff", "ns", "nx",
257: "open", "opena", "os", "output",
258: "padj", "papersize", "pc", "pev",
259: "pi", "PI", "pl", "pm",
260: "pn", "pnr", "po", "ps",
261: "psbb", "pshape", "pso", "ptr",
262: "pvs", "rchar", "rd", "recursionlimit",
263: "return", "rfschar", "rhang", "rj",
264: "rm", "rn", "rnn", "rr",
265: "rs", "rt", "schar", "sentchar",
266: "shc", "shift", "sizes", "so",
267: "spacewidth", "special", "spreadwarn", "ss",
268: "sty", "substring", "sv", "sy",
1.300 schwarze 269: "T&", "tc", "TE",
1.301 schwarze 270: "TH", "tkf", "tl",
1.294 schwarze 271: "tm", "tm1", "tmc", "tr",
272: "track", "transchar", "trf", "trimat",
273: "trin", "trnt", "troff", "TS",
274: "uf", "ul", "unformat", "unwatch",
275: "unwatchn", "vpt", "vs", "warn",
276: "warnscale", "watch", "watchlength", "watchn",
277: "wh", "while", "write", "writec",
278: "writem", "xflag", ".", NULL,
279: "text",
280: "Dd", "Dt", "Os", "Sh",
281: "Ss", "Pp", "D1", "Dl",
282: "Bd", "Ed", "Bl", "El",
283: "It", "Ad", "An", "Ap",
284: "Ar", "Cd", "Cm", "Dv",
285: "Er", "Ev", "Ex", "Fa",
286: "Fd", "Fl", "Fn", "Ft",
287: "Ic", "In", "Li", "Nd",
288: "Nm", "Op", "Ot", "Pa",
289: "Rv", "St", "Va", "Vt",
290: "Xr", "%A", "%B", "%D",
291: "%I", "%J", "%N", "%O",
292: "%P", "%R", "%T", "%V",
293: "Ac", "Ao", "Aq", "At",
294: "Bc", "Bf", "Bo", "Bq",
295: "Bsx", "Bx", "Db", "Dc",
296: "Do", "Dq", "Ec", "Ef",
297: "Em", "Eo", "Fx", "Ms",
298: "No", "Ns", "Nx", "Ox",
299: "Pc", "Pf", "Po", "Pq",
300: "Qc", "Ql", "Qo", "Qq",
301: "Re", "Rs", "Sc", "So",
302: "Sq", "Sm", "Sx", "Sy",
303: "Tn", "Ux", "Xc", "Xo",
304: "Fo", "Fc", "Oo", "Oc",
305: "Bk", "Ek", "Bt", "Hf",
306: "Fr", "Ud", "Lb", "Lp",
307: "Lk", "Mt", "Brq", "Bro",
308: "Brc", "%C", "Es", "En",
1.299 schwarze 309: "Dx", "%Q", "%U", "Ta",
310: NULL,
1.294 schwarze 311: "TH", "SH", "SS", "TP",
312: "LP", "PP", "P", "IP",
313: "HP", "SM", "SB", "BI",
314: "IB", "BR", "RB", "R",
315: "B", "I", "IR", "RI",
1.299 schwarze 316: "nf", "fi",
1.294 schwarze 317: "RE", "RS", "DT", "UC",
1.297 schwarze 318: "PD", "AT", "in",
1.294 schwarze 319: "OP", "EX", "EE", "UR",
1.298 schwarze 320: "UE", NULL
1.294 schwarze 321: };
322: const char *const *roff_name = __roff_name;
323:
324: static struct roffmac roffs[TOKEN_NONE] = {
1.296 schwarze 325: { roff_br, NULL, NULL, 0 }, /* br */
1.297 schwarze 326: { roff_onearg, NULL, NULL, 0 }, /* ft */
1.298 schwarze 327: { roff_onearg, NULL, NULL, 0 }, /* ll */
1.304 ! schwarze 328: { roff_onearg, NULL, NULL, 0 }, /* mc */
1.299 schwarze 329: { roff_onearg, NULL, NULL, 0 }, /* sp */
1.300 schwarze 330: { roff_manyarg, NULL, NULL, 0 }, /* ta */
1.301 schwarze 331: { roff_onearg, NULL, NULL, 0 }, /* ti */
1.296 schwarze 332: { NULL, NULL, NULL, 0 }, /* ROFF_MAX */
1.295 schwarze 333: { roff_unsupp, NULL, NULL, 0 }, /* ab */
334: { roff_line_ignore, NULL, NULL, 0 }, /* ad */
335: { roff_line_ignore, NULL, NULL, 0 }, /* af */
336: { roff_unsupp, NULL, NULL, 0 }, /* aln */
337: { roff_unsupp, NULL, NULL, 0 }, /* als */
338: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am */
339: { roff_block, roff_block_text, roff_block_sub, 0 }, /* am1 */
340: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami */
341: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ami1 */
342: { roff_ds, NULL, NULL, 0 }, /* as */
343: { roff_ds, NULL, NULL, 0 }, /* as1 */
344: { roff_unsupp, NULL, NULL, 0 }, /* asciify */
345: { roff_line_ignore, NULL, NULL, 0 }, /* backtrace */
346: { roff_line_ignore, NULL, NULL, 0 }, /* bd */
347: { roff_line_ignore, NULL, NULL, 0 }, /* bleedat */
348: { roff_unsupp, NULL, NULL, 0 }, /* blm */
349: { roff_unsupp, NULL, NULL, 0 }, /* box */
350: { roff_unsupp, NULL, NULL, 0 }, /* boxa */
351: { roff_line_ignore, NULL, NULL, 0 }, /* bp */
352: { roff_unsupp, NULL, NULL, 0 }, /* BP */
353: { roff_unsupp, NULL, NULL, 0 }, /* break */
354: { roff_line_ignore, NULL, NULL, 0 }, /* breakchar */
355: { roff_line_ignore, NULL, NULL, 0 }, /* brnl */
1.296 schwarze 356: { roff_br, NULL, NULL, 0 }, /* brp */
1.295 schwarze 357: { roff_line_ignore, NULL, NULL, 0 }, /* brpnl */
358: { roff_unsupp, NULL, NULL, 0 }, /* c2 */
359: { roff_cc, NULL, NULL, 0 }, /* cc */
360: { roff_line_ignore, NULL, NULL, 0 }, /* ce */
361: { roff_insec, NULL, NULL, 0 }, /* cf */
362: { roff_line_ignore, NULL, NULL, 0 }, /* cflags */
363: { roff_line_ignore, NULL, NULL, 0 }, /* ch */
364: { roff_unsupp, NULL, NULL, 0 }, /* char */
365: { roff_unsupp, NULL, NULL, 0 }, /* chop */
366: { roff_line_ignore, NULL, NULL, 0 }, /* class */
367: { roff_insec, NULL, NULL, 0 }, /* close */
368: { roff_unsupp, NULL, NULL, 0 }, /* CL */
369: { roff_line_ignore, NULL, NULL, 0 }, /* color */
370: { roff_unsupp, NULL, NULL, 0 }, /* composite */
371: { roff_unsupp, NULL, NULL, 0 }, /* continue */
372: { roff_line_ignore, NULL, NULL, 0 }, /* cp */
373: { roff_line_ignore, NULL, NULL, 0 }, /* cropat */
374: { roff_line_ignore, NULL, NULL, 0 }, /* cs */
375: { roff_line_ignore, NULL, NULL, 0 }, /* cu */
376: { roff_unsupp, NULL, NULL, 0 }, /* da */
377: { roff_unsupp, NULL, NULL, 0 }, /* dch */
378: { roff_Dd, NULL, NULL, 0 }, /* Dd */
379: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de */
380: { roff_block, roff_block_text, roff_block_sub, 0 }, /* de1 */
381: { roff_line_ignore, NULL, NULL, 0 }, /* defcolor */
382: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei */
383: { roff_block, roff_block_text, roff_block_sub, 0 }, /* dei1 */
384: { roff_unsupp, NULL, NULL, 0 }, /* device */
385: { roff_unsupp, NULL, NULL, 0 }, /* devicem */
386: { roff_unsupp, NULL, NULL, 0 }, /* di */
387: { roff_unsupp, NULL, NULL, 0 }, /* do */
388: { roff_ds, NULL, NULL, 0 }, /* ds */
389: { roff_ds, NULL, NULL, 0 }, /* ds1 */
390: { roff_unsupp, NULL, NULL, 0 }, /* dwh */
391: { roff_unsupp, NULL, NULL, 0 }, /* dt */
1.303 schwarze 392: { roff_ec, NULL, NULL, 0 }, /* ec */
1.295 schwarze 393: { roff_unsupp, NULL, NULL, 0 }, /* ecr */
394: { roff_unsupp, NULL, NULL, 0 }, /* ecs */
395: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* el */
396: { roff_unsupp, NULL, NULL, 0 }, /* em */
397: { roff_EN, NULL, NULL, 0 }, /* EN */
1.303 schwarze 398: { roff_eo, NULL, NULL, 0 }, /* eo */
1.295 schwarze 399: { roff_unsupp, NULL, NULL, 0 }, /* EP */
400: { roff_EQ, NULL, NULL, 0 }, /* EQ */
401: { roff_line_ignore, NULL, NULL, 0 }, /* errprint */
402: { roff_unsupp, NULL, NULL, 0 }, /* ev */
403: { roff_unsupp, NULL, NULL, 0 }, /* evc */
404: { roff_unsupp, NULL, NULL, 0 }, /* ex */
405: { roff_line_ignore, NULL, NULL, 0 }, /* fallback */
406: { roff_line_ignore, NULL, NULL, 0 }, /* fam */
407: { roff_unsupp, NULL, NULL, 0 }, /* fc */
408: { roff_unsupp, NULL, NULL, 0 }, /* fchar */
409: { roff_line_ignore, NULL, NULL, 0 }, /* fcolor */
410: { roff_line_ignore, NULL, NULL, 0 }, /* fdeferlig */
411: { roff_line_ignore, NULL, NULL, 0 }, /* feature */
412: { roff_line_ignore, NULL, NULL, 0 }, /* fkern */
413: { roff_line_ignore, NULL, NULL, 0 }, /* fl */
414: { roff_line_ignore, NULL, NULL, 0 }, /* flig */
415: { roff_line_ignore, NULL, NULL, 0 }, /* fp */
416: { roff_line_ignore, NULL, NULL, 0 }, /* fps */
417: { roff_unsupp, NULL, NULL, 0 }, /* fschar */
418: { roff_line_ignore, NULL, NULL, 0 }, /* fspacewidth */
419: { roff_line_ignore, NULL, NULL, 0 }, /* fspecial */
420: { roff_line_ignore, NULL, NULL, 0 }, /* ftr */
421: { roff_line_ignore, NULL, NULL, 0 }, /* fzoom */
422: { roff_line_ignore, NULL, NULL, 0 }, /* gcolor */
423: { roff_line_ignore, NULL, NULL, 0 }, /* hc */
424: { roff_line_ignore, NULL, NULL, 0 }, /* hcode */
425: { roff_line_ignore, NULL, NULL, 0 }, /* hidechar */
426: { roff_line_ignore, NULL, NULL, 0 }, /* hla */
427: { roff_line_ignore, NULL, NULL, 0 }, /* hlm */
428: { roff_line_ignore, NULL, NULL, 0 }, /* hpf */
429: { roff_line_ignore, NULL, NULL, 0 }, /* hpfa */
430: { roff_line_ignore, NULL, NULL, 0 }, /* hpfcode */
431: { roff_line_ignore, NULL, NULL, 0 }, /* hw */
432: { roff_line_ignore, NULL, NULL, 0 }, /* hy */
433: { roff_line_ignore, NULL, NULL, 0 }, /* hylang */
434: { roff_line_ignore, NULL, NULL, 0 }, /* hylen */
435: { roff_line_ignore, NULL, NULL, 0 }, /* hym */
436: { roff_line_ignore, NULL, NULL, 0 }, /* hypp */
437: { roff_line_ignore, NULL, NULL, 0 }, /* hys */
438: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* ie */
439: { roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT }, /* if */
440: { roff_block, roff_block_text, roff_block_sub, 0 }, /* ig */
441: { roff_unsupp, NULL, NULL, 0 }, /* index */
442: { roff_it, NULL, NULL, 0 }, /* it */
443: { roff_unsupp, NULL, NULL, 0 }, /* itc */
444: { roff_line_ignore, NULL, NULL, 0 }, /* IX */
445: { roff_line_ignore, NULL, NULL, 0 }, /* kern */
446: { roff_line_ignore, NULL, NULL, 0 }, /* kernafter */
447: { roff_line_ignore, NULL, NULL, 0 }, /* kernbefore */
448: { roff_line_ignore, NULL, NULL, 0 }, /* kernpair */
449: { roff_unsupp, NULL, NULL, 0 }, /* lc */
450: { roff_unsupp, NULL, NULL, 0 }, /* lc_ctype */
451: { roff_unsupp, NULL, NULL, 0 }, /* lds */
452: { roff_unsupp, NULL, NULL, 0 }, /* length */
453: { roff_line_ignore, NULL, NULL, 0 }, /* letadj */
454: { roff_insec, NULL, NULL, 0 }, /* lf */
455: { roff_line_ignore, NULL, NULL, 0 }, /* lg */
456: { roff_line_ignore, NULL, NULL, 0 }, /* lhang */
457: { roff_unsupp, NULL, NULL, 0 }, /* linetabs */
458: { roff_unsupp, NULL, NULL, 0 }, /* lnr */
459: { roff_unsupp, NULL, NULL, 0 }, /* lnrf */
460: { roff_unsupp, NULL, NULL, 0 }, /* lpfx */
461: { roff_line_ignore, NULL, NULL, 0 }, /* ls */
462: { roff_unsupp, NULL, NULL, 0 }, /* lsm */
463: { roff_line_ignore, NULL, NULL, 0 }, /* lt */
464: { roff_line_ignore, NULL, NULL, 0 }, /* mediasize */
465: { roff_line_ignore, NULL, NULL, 0 }, /* minss */
466: { roff_line_ignore, NULL, NULL, 0 }, /* mk */
467: { roff_insec, NULL, NULL, 0 }, /* mso */
468: { roff_line_ignore, NULL, NULL, 0 }, /* na */
469: { roff_line_ignore, NULL, NULL, 0 }, /* ne */
470: { roff_line_ignore, NULL, NULL, 0 }, /* nh */
471: { roff_line_ignore, NULL, NULL, 0 }, /* nhychar */
472: { roff_unsupp, NULL, NULL, 0 }, /* nm */
473: { roff_unsupp, NULL, NULL, 0 }, /* nn */
474: { roff_unsupp, NULL, NULL, 0 }, /* nop */
475: { roff_nr, NULL, NULL, 0 }, /* nr */
476: { roff_unsupp, NULL, NULL, 0 }, /* nrf */
477: { roff_line_ignore, NULL, NULL, 0 }, /* nroff */
478: { roff_line_ignore, NULL, NULL, 0 }, /* ns */
479: { roff_insec, NULL, NULL, 0 }, /* nx */
480: { roff_insec, NULL, NULL, 0 }, /* open */
481: { roff_insec, NULL, NULL, 0 }, /* opena */
482: { roff_line_ignore, NULL, NULL, 0 }, /* os */
483: { roff_unsupp, NULL, NULL, 0 }, /* output */
484: { roff_line_ignore, NULL, NULL, 0 }, /* padj */
485: { roff_line_ignore, NULL, NULL, 0 }, /* papersize */
486: { roff_line_ignore, NULL, NULL, 0 }, /* pc */
487: { roff_line_ignore, NULL, NULL, 0 }, /* pev */
488: { roff_insec, NULL, NULL, 0 }, /* pi */
489: { roff_unsupp, NULL, NULL, 0 }, /* PI */
490: { roff_line_ignore, NULL, NULL, 0 }, /* pl */
491: { roff_line_ignore, NULL, NULL, 0 }, /* pm */
492: { roff_line_ignore, NULL, NULL, 0 }, /* pn */
493: { roff_line_ignore, NULL, NULL, 0 }, /* pnr */
494: { roff_line_ignore, NULL, NULL, 0 }, /* po */
495: { roff_line_ignore, NULL, NULL, 0 }, /* ps */
496: { roff_unsupp, NULL, NULL, 0 }, /* psbb */
497: { roff_unsupp, NULL, NULL, 0 }, /* pshape */
498: { roff_insec, NULL, NULL, 0 }, /* pso */
499: { roff_line_ignore, NULL, NULL, 0 }, /* ptr */
500: { roff_line_ignore, NULL, NULL, 0 }, /* pvs */
501: { roff_unsupp, NULL, NULL, 0 }, /* rchar */
502: { roff_line_ignore, NULL, NULL, 0 }, /* rd */
503: { roff_line_ignore, NULL, NULL, 0 }, /* recursionlimit */
504: { roff_unsupp, NULL, NULL, 0 }, /* return */
505: { roff_unsupp, NULL, NULL, 0 }, /* rfschar */
506: { roff_line_ignore, NULL, NULL, 0 }, /* rhang */
507: { roff_line_ignore, NULL, NULL, 0 }, /* rj */
508: { roff_rm, NULL, NULL, 0 }, /* rm */
509: { roff_unsupp, NULL, NULL, 0 }, /* rn */
510: { roff_unsupp, NULL, NULL, 0 }, /* rnn */
511: { roff_rr, NULL, NULL, 0 }, /* rr */
512: { roff_line_ignore, NULL, NULL, 0 }, /* rs */
513: { roff_line_ignore, NULL, NULL, 0 }, /* rt */
514: { roff_unsupp, NULL, NULL, 0 }, /* schar */
515: { roff_line_ignore, NULL, NULL, 0 }, /* sentchar */
516: { roff_line_ignore, NULL, NULL, 0 }, /* shc */
517: { roff_unsupp, NULL, NULL, 0 }, /* shift */
518: { roff_line_ignore, NULL, NULL, 0 }, /* sizes */
519: { roff_so, NULL, NULL, 0 }, /* so */
520: { roff_line_ignore, NULL, NULL, 0 }, /* spacewidth */
521: { roff_line_ignore, NULL, NULL, 0 }, /* special */
522: { roff_line_ignore, NULL, NULL, 0 }, /* spreadwarn */
523: { roff_line_ignore, NULL, NULL, 0 }, /* ss */
524: { roff_line_ignore, NULL, NULL, 0 }, /* sty */
525: { roff_unsupp, NULL, NULL, 0 }, /* substring */
526: { roff_line_ignore, NULL, NULL, 0 }, /* sv */
527: { roff_insec, NULL, NULL, 0 }, /* sy */
528: { roff_T_, NULL, NULL, 0 }, /* T& */
529: { roff_unsupp, NULL, NULL, 0 }, /* tc */
530: { roff_TE, NULL, NULL, 0 }, /* TE */
531: { roff_TH, NULL, NULL, 0 }, /* TH */
532: { roff_line_ignore, NULL, NULL, 0 }, /* tkf */
533: { roff_unsupp, NULL, NULL, 0 }, /* tl */
534: { roff_line_ignore, NULL, NULL, 0 }, /* tm */
535: { roff_line_ignore, NULL, NULL, 0 }, /* tm1 */
536: { roff_line_ignore, NULL, NULL, 0 }, /* tmc */
537: { roff_tr, NULL, NULL, 0 }, /* tr */
538: { roff_line_ignore, NULL, NULL, 0 }, /* track */
539: { roff_line_ignore, NULL, NULL, 0 }, /* transchar */
540: { roff_insec, NULL, NULL, 0 }, /* trf */
541: { roff_line_ignore, NULL, NULL, 0 }, /* trimat */
542: { roff_unsupp, NULL, NULL, 0 }, /* trin */
543: { roff_unsupp, NULL, NULL, 0 }, /* trnt */
544: { roff_line_ignore, NULL, NULL, 0 }, /* troff */
545: { roff_TS, NULL, NULL, 0 }, /* TS */
546: { roff_line_ignore, NULL, NULL, 0 }, /* uf */
547: { roff_line_ignore, NULL, NULL, 0 }, /* ul */
548: { roff_unsupp, NULL, NULL, 0 }, /* unformat */
549: { roff_line_ignore, NULL, NULL, 0 }, /* unwatch */
550: { roff_line_ignore, NULL, NULL, 0 }, /* unwatchn */
551: { roff_line_ignore, NULL, NULL, 0 }, /* vpt */
552: { roff_line_ignore, NULL, NULL, 0 }, /* vs */
553: { roff_line_ignore, NULL, NULL, 0 }, /* warn */
554: { roff_line_ignore, NULL, NULL, 0 }, /* warnscale */
555: { roff_line_ignore, NULL, NULL, 0 }, /* watch */
556: { roff_line_ignore, NULL, NULL, 0 }, /* watchlength */
557: { roff_line_ignore, NULL, NULL, 0 }, /* watchn */
558: { roff_unsupp, NULL, NULL, 0 }, /* wh */
559: { roff_unsupp, NULL, NULL, 0 }, /* while */
560: { roff_insec, NULL, NULL, 0 }, /* write */
561: { roff_insec, NULL, NULL, 0 }, /* writec */
562: { roff_insec, NULL, NULL, 0 }, /* writem */
563: { roff_line_ignore, NULL, NULL, 0 }, /* xflag */
564: { roff_cblock, NULL, NULL, 0 }, /* . */
565: { roff_userdef, NULL, NULL, 0 }
1.67 kristaps 566: };
567:
1.200 schwarze 568: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.175 schwarze 569: const char *const __mdoc_reserved[] = {
570: "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
571: "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
572: "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
573: "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.200 schwarze 574: "Dt", "Dv", "Dx", "D1",
575: "Ec", "Ed", "Ef", "Ek", "El", "Em",
576: "En", "Eo", "Er", "Es", "Ev", "Ex",
1.175 schwarze 577: "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.200 schwarze 578: "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
579: "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.175 schwarze 580: "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.200 schwarze 581: "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
582: "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
583: "Sc", "Sh", "Sm", "So", "Sq",
1.175 schwarze 584: "Ss", "St", "Sx", "Sy",
585: "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.200 schwarze 586: "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.175 schwarze 587: "%P", "%Q", "%R", "%T", "%U", "%V",
588: NULL
589: };
590:
1.200 schwarze 591: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.175 schwarze 592: const char *const __man_reserved[] = {
1.200 schwarze 593: "AT", "B", "BI", "BR", "DT",
594: "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
595: "LP", "OP", "P", "PD", "PP",
596: "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
597: "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.175 schwarze 598: NULL
599: };
600:
1.141 kristaps 601: /* Array of injected predefined strings. */
602: #define PREDEFS_MAX 38
603: static const struct predef predefs[PREDEFS_MAX] = {
604: #include "predefs.in"
605: };
606:
1.178 schwarze 607: static int roffit_lines; /* number of lines to delay */
608: static char *roffit_macro; /* nil-terminated macro line */
609:
1.207 schwarze 610:
1.266 schwarze 611: /* --- request table ------------------------------------------------------ */
612:
1.295 schwarze 613: struct ohash *
614: roffhash_alloc(enum roff_tok mintok, enum roff_tok maxtok)
1.85 kristaps 615: {
1.295 schwarze 616: struct ohash *htab;
617: struct roffreq *req;
618: enum roff_tok tok;
619: size_t sz;
620: unsigned int slot;
1.85 kristaps 621:
1.295 schwarze 622: htab = mandoc_malloc(sizeof(*htab));
623: mandoc_ohash_init(htab, 8, offsetof(struct roffreq, name));
1.85 kristaps 624:
1.295 schwarze 625: for (tok = mintok; tok < maxtok; tok++) {
1.296 schwarze 626: if (roff_name[tok] == NULL)
627: continue;
1.295 schwarze 628: sz = strlen(roff_name[tok]);
629: req = mandoc_malloc(sizeof(*req) + sz + 1);
630: req->tok = tok;
631: memcpy(req->name, roff_name[tok], sz + 1);
632: slot = ohash_qlookup(htab, req->name);
633: ohash_insert(htab, slot, req);
1.85 kristaps 634: }
1.295 schwarze 635: return htab;
1.85 kristaps 636: }
637:
1.295 schwarze 638: void
639: roffhash_free(struct ohash *htab)
1.67 kristaps 640: {
1.295 schwarze 641: struct roffreq *req;
642: unsigned int slot;
1.67 kristaps 643:
1.295 schwarze 644: if (htab == NULL)
645: return;
646: for (req = ohash_first(htab, &slot); req != NULL;
647: req = ohash_next(htab, &slot))
648: free(req);
649: ohash_delete(htab);
650: free(htab);
651: }
652:
653: enum roff_tok
654: roffhash_find(struct ohash *htab, const char *name, size_t sz)
655: {
656: struct roffreq *req;
657: const char *end;
658:
659: if (sz) {
660: end = name + sz;
661: req = ohash_find(htab, ohash_qlookupi(htab, name, &end));
662: } else
663: req = ohash_find(htab, ohash_qlookup(htab, name));
664: return req == NULL ? TOKEN_NONE : req->tok;
1.67 kristaps 665: }
666:
1.266 schwarze 667: /* --- stack of request blocks -------------------------------------------- */
668:
1.67 kristaps 669: /*
670: * Pop the current node off of the stack of roff instructions currently
671: * pending.
672: */
673: static void
674: roffnode_pop(struct roff *r)
675: {
676: struct roffnode *p;
677:
1.75 kristaps 678: assert(r->last);
1.207 schwarze 679: p = r->last;
1.82 kristaps 680:
1.75 kristaps 681: r->last = r->last->parent;
1.106 kristaps 682: free(p->name);
683: free(p->end);
1.67 kristaps 684: free(p);
685: }
686:
687: /*
688: * Push a roff node onto the instruction stack. This must later be
689: * removed with roffnode_pop().
690: */
1.98 schwarze 691: static void
1.294 schwarze 692: roffnode_push(struct roff *r, enum roff_tok tok, const char *name,
1.106 kristaps 693: int line, int col)
1.67 kristaps 694: {
695: struct roffnode *p;
696:
1.98 schwarze 697: p = mandoc_calloc(1, sizeof(struct roffnode));
1.67 kristaps 698: p->tok = tok;
1.106 kristaps 699: if (name)
700: p->name = mandoc_strdup(name);
1.67 kristaps 701: p->parent = r->last;
702: p->line = line;
703: p->col = col;
1.198 schwarze 704: p->rule = p->parent ? p->parent->rule : 0;
1.67 kristaps 705:
706: r->last = p;
707: }
708:
1.266 schwarze 709: /* --- roff parser state data management ---------------------------------- */
710:
1.67 kristaps 711: static void
712: roff_free1(struct roff *r)
713: {
1.176 schwarze 714: struct tbl_node *tbl;
1.125 kristaps 715: struct eqn_node *e;
1.167 kristaps 716: int i;
1.67 kristaps 717:
1.176 schwarze 718: while (NULL != (tbl = r->first_tbl)) {
719: r->first_tbl = tbl->next;
720: tbl_free(tbl);
1.109 kristaps 721: }
1.113 kristaps 722: r->first_tbl = r->last_tbl = r->tbl = NULL;
723:
1.125 kristaps 724: while (NULL != (e = r->first_eqn)) {
725: r->first_eqn = e->next;
726: eqn_free(e);
727: }
728: r->first_eqn = r->last_eqn = r->eqn = NULL;
729:
1.67 kristaps 730: while (r->last)
731: roffnode_pop(r);
1.109 kristaps 732:
1.223 schwarze 733: free (r->rstack);
734: r->rstack = NULL;
735: r->rstacksz = 0;
736: r->rstackpos = -1;
737:
738: roff_freereg(r->regtab);
739: r->regtab = NULL;
740:
1.167 kristaps 741: roff_freestr(r->strtab);
742: roff_freestr(r->xmbtab);
743: r->strtab = r->xmbtab = NULL;
744:
745: if (r->xtab)
746: for (i = 0; i < 128; i++)
747: free(r->xtab[i].p);
748: free(r->xtab);
749: r->xtab = NULL;
1.67 kristaps 750: }
751:
752: void
753: roff_reset(struct roff *r)
754: {
755: roff_free1(r);
1.227 schwarze 756: r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.303 schwarze 757: r->control = '\0';
758: r->escape = '\\';
1.67 kristaps 759: }
760:
761: void
762: roff_free(struct roff *r)
763: {
764: roff_free1(r);
1.295 schwarze 765: roffhash_free(r->reqtab);
1.67 kristaps 766: free(r);
767: }
768:
769: struct roff *
1.279 schwarze 770: roff_alloc(struct mparse *parse, int options)
1.67 kristaps 771: {
772: struct roff *r;
773:
1.98 schwarze 774: r = mandoc_calloc(1, sizeof(struct roff));
1.128 kristaps 775: r->parse = parse;
1.295 schwarze 776: r->reqtab = roffhash_alloc(0, ROFF_USERDEF);
1.199 schwarze 777: r->options = options;
1.227 schwarze 778: r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.82 kristaps 779: r->rstackpos = -1;
1.303 schwarze 780: r->escape = '\\';
1.277 schwarze 781: return r;
1.265 schwarze 782: }
783:
1.266 schwarze 784: /* --- syntax tree state data management ---------------------------------- */
785:
1.265 schwarze 786: static void
787: roff_man_free1(struct roff_man *man)
788: {
789:
1.266 schwarze 790: if (man->first != NULL)
791: roff_node_delete(man, man->first);
1.265 schwarze 792: free(man->meta.msec);
793: free(man->meta.vol);
794: free(man->meta.os);
795: free(man->meta.arch);
796: free(man->meta.title);
797: free(man->meta.name);
798: free(man->meta.date);
799: }
800:
801: static void
802: roff_man_alloc1(struct roff_man *man)
803: {
804:
805: memset(&man->meta, 0, sizeof(man->meta));
806: man->first = mandoc_calloc(1, sizeof(*man->first));
807: man->first->type = ROFFT_ROOT;
808: man->last = man->first;
809: man->last_es = NULL;
810: man->flags = 0;
811: man->macroset = MACROSET_NONE;
812: man->lastsec = man->lastnamed = SEC_NONE;
813: man->next = ROFF_NEXT_CHILD;
814: }
815:
816: void
817: roff_man_reset(struct roff_man *man)
818: {
819:
820: roff_man_free1(man);
821: roff_man_alloc1(man);
822: }
823:
824: void
825: roff_man_free(struct roff_man *man)
826: {
827:
828: roff_man_free1(man);
829: free(man);
830: }
831:
832: struct roff_man *
833: roff_man_alloc(struct roff *roff, struct mparse *parse,
834: const char *defos, int quick)
835: {
836: struct roff_man *man;
837:
838: man = mandoc_calloc(1, sizeof(*man));
839: man->parse = parse;
840: man->roff = roff;
841: man->defos = defos;
842: man->quick = quick;
843: roff_man_alloc1(man);
1.296 schwarze 844: roff->man = man;
1.277 schwarze 845: return man;
1.67 kristaps 846: }
847:
1.266 schwarze 848: /* --- syntax tree handling ----------------------------------------------- */
849:
850: struct roff_node *
851: roff_node_alloc(struct roff_man *man, int line, int pos,
852: enum roff_type type, int tok)
853: {
854: struct roff_node *n;
855:
856: n = mandoc_calloc(1, sizeof(*n));
857: n->line = line;
858: n->pos = pos;
859: n->tok = tok;
860: n->type = type;
861: n->sec = man->lastsec;
862:
863: if (man->flags & MDOC_SYNOPSIS)
1.285 schwarze 864: n->flags |= NODE_SYNPRETTY;
1.266 schwarze 865: else
1.285 schwarze 866: n->flags &= ~NODE_SYNPRETTY;
1.266 schwarze 867: if (man->flags & MDOC_NEWLINE)
1.285 schwarze 868: n->flags |= NODE_LINE;
1.266 schwarze 869: man->flags &= ~MDOC_NEWLINE;
870:
1.277 schwarze 871: return n;
1.266 schwarze 872: }
873:
874: void
875: roff_node_append(struct roff_man *man, struct roff_node *n)
876: {
877:
878: switch (man->next) {
879: case ROFF_NEXT_SIBLING:
1.281 schwarze 880: if (man->last->next != NULL) {
881: n->next = man->last->next;
882: man->last->next->prev = n;
883: } else
884: man->last->parent->last = n;
1.266 schwarze 885: man->last->next = n;
886: n->prev = man->last;
887: n->parent = man->last->parent;
888: break;
889: case ROFF_NEXT_CHILD:
1.287 schwarze 890: if (man->last->child != NULL) {
891: n->next = man->last->child;
892: man->last->child->prev = n;
893: } else
894: man->last->last = n;
1.266 schwarze 895: man->last->child = n;
896: n->parent = man->last;
897: break;
898: default:
899: abort();
900: }
1.282 schwarze 901: man->last = n;
1.266 schwarze 902:
903: switch (n->type) {
904: case ROFFT_HEAD:
905: n->parent->head = n;
906: break;
907: case ROFFT_BODY:
1.282 schwarze 908: if (n->end != ENDBODY_NOT)
909: return;
1.266 schwarze 910: n->parent->body = n;
911: break;
912: case ROFFT_TAIL:
913: n->parent->tail = n;
914: break;
915: default:
1.282 schwarze 916: return;
1.266 schwarze 917: }
1.282 schwarze 918:
919: /*
920: * Copy over the normalised-data pointer of our parent. Not
921: * everybody has one, but copying a null pointer is fine.
922: */
923:
924: n->norm = n->parent->norm;
925: assert(n->parent->type == ROFFT_BLOCK);
1.266 schwarze 926: }
927:
1.267 schwarze 928: void
929: roff_word_alloc(struct roff_man *man, int line, int pos, const char *word)
930: {
931: struct roff_node *n;
932:
933: n = roff_node_alloc(man, line, pos, ROFFT_TEXT, TOKEN_NONE);
934: n->string = roff_strdup(man->roff, word);
935: roff_node_append(man, n);
1.286 schwarze 936: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 937: man->next = ROFF_NEXT_SIBLING;
938: }
939:
940: void
941: roff_word_append(struct roff_man *man, const char *word)
942: {
943: struct roff_node *n;
944: char *addstr, *newstr;
945:
946: n = man->last;
947: addstr = roff_strdup(man->roff, word);
948: mandoc_asprintf(&newstr, "%s %s", n->string, addstr);
949: free(addstr);
950: free(n->string);
951: n->string = newstr;
952: man->next = ROFF_NEXT_SIBLING;
1.268 schwarze 953: }
954:
955: void
956: roff_elem_alloc(struct roff_man *man, int line, int pos, int tok)
957: {
958: struct roff_node *n;
959:
960: n = roff_node_alloc(man, line, pos, ROFFT_ELEM, tok);
961: roff_node_append(man, n);
962: man->next = ROFF_NEXT_CHILD;
963: }
964:
965: struct roff_node *
966: roff_block_alloc(struct roff_man *man, int line, int pos, int tok)
967: {
968: struct roff_node *n;
969:
970: n = roff_node_alloc(man, line, pos, ROFFT_BLOCK, tok);
971: roff_node_append(man, n);
972: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 973: return n;
1.267 schwarze 974: }
975:
1.266 schwarze 976: struct roff_node *
977: roff_head_alloc(struct roff_man *man, int line, int pos, int tok)
978: {
979: struct roff_node *n;
980:
981: n = roff_node_alloc(man, line, pos, ROFFT_HEAD, tok);
982: roff_node_append(man, n);
983: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 984: return n;
1.266 schwarze 985: }
986:
987: struct roff_node *
988: roff_body_alloc(struct roff_man *man, int line, int pos, int tok)
989: {
990: struct roff_node *n;
991:
992: n = roff_node_alloc(man, line, pos, ROFFT_BODY, tok);
993: roff_node_append(man, n);
994: man->next = ROFF_NEXT_CHILD;
1.277 schwarze 995: return n;
1.267 schwarze 996: }
997:
998: void
999: roff_addeqn(struct roff_man *man, const struct eqn *eqn)
1000: {
1001: struct roff_node *n;
1002:
1003: n = roff_node_alloc(man, eqn->ln, eqn->pos, ROFFT_EQN, TOKEN_NONE);
1004: n->eqn = eqn;
1005: if (eqn->ln > man->last->line)
1.285 schwarze 1006: n->flags |= NODE_LINE;
1.267 schwarze 1007: roff_node_append(man, n);
1008: man->next = ROFF_NEXT_SIBLING;
1009: }
1010:
1011: void
1012: roff_addtbl(struct roff_man *man, const struct tbl_span *tbl)
1013: {
1014: struct roff_node *n;
1015:
1016: if (man->macroset == MACROSET_MAN)
1.302 schwarze 1017: man_breakscope(man, ROFF_TS);
1.267 schwarze 1018: n = roff_node_alloc(man, tbl->line, 0, ROFFT_TBL, TOKEN_NONE);
1019: n->span = tbl;
1020: roff_node_append(man, n);
1.286 schwarze 1021: n->flags |= NODE_VALID | NODE_ENDED;
1.267 schwarze 1022: man->next = ROFF_NEXT_SIBLING;
1.266 schwarze 1023: }
1024:
1025: void
1026: roff_node_unlink(struct roff_man *man, struct roff_node *n)
1027: {
1028:
1029: /* Adjust siblings. */
1030:
1031: if (n->prev)
1032: n->prev->next = n->next;
1033: if (n->next)
1034: n->next->prev = n->prev;
1035:
1036: /* Adjust parent. */
1037:
1038: if (n->parent != NULL) {
1039: if (n->parent->child == n)
1040: n->parent->child = n->next;
1041: if (n->parent->last == n)
1042: n->parent->last = n->prev;
1043: }
1044:
1045: /* Adjust parse point. */
1046:
1047: if (man == NULL)
1048: return;
1049: if (man->last == n) {
1050: if (n->prev == NULL) {
1051: man->last = n->parent;
1052: man->next = ROFF_NEXT_CHILD;
1053: } else {
1054: man->last = n->prev;
1055: man->next = ROFF_NEXT_SIBLING;
1056: }
1057: }
1058: if (man->first == n)
1059: man->first = NULL;
1060: }
1061:
1062: void
1063: roff_node_free(struct roff_node *n)
1064: {
1065:
1066: if (n->args != NULL)
1067: mdoc_argv_free(n->args);
1068: if (n->type == ROFFT_BLOCK || n->type == ROFFT_ELEM)
1069: free(n->norm);
1070: free(n->string);
1071: free(n);
1072: }
1073:
1074: void
1075: roff_node_delete(struct roff_man *man, struct roff_node *n)
1076: {
1077:
1078: while (n->child != NULL)
1079: roff_node_delete(man, n->child);
1080: roff_node_unlink(man, n);
1081: roff_node_free(n);
1.269 schwarze 1082: }
1083:
1084: void
1085: deroff(char **dest, const struct roff_node *n)
1086: {
1087: char *cp;
1088: size_t sz;
1089:
1090: if (n->type != ROFFT_TEXT) {
1091: for (n = n->child; n != NULL; n = n->next)
1092: deroff(dest, n);
1093: return;
1094: }
1095:
1.288 schwarze 1096: /* Skip leading whitespace. */
1.269 schwarze 1097:
1.288 schwarze 1098: for (cp = n->string; *cp != '\0'; cp++) {
1.289 schwarze 1099: if (cp[0] == '\\' && cp[1] != '\0' &&
1100: strchr(" %&0^|~", cp[1]) != NULL)
1.269 schwarze 1101: cp++;
1.288 schwarze 1102: else if ( ! isspace((unsigned char)*cp))
1.269 schwarze 1103: break;
1104: }
1105:
1.289 schwarze 1106: /* Skip trailing backslash. */
1107:
1108: sz = strlen(cp);
1.290 schwarze 1109: if (sz > 0 && cp[sz - 1] == '\\')
1.289 schwarze 1110: sz--;
1111:
1.269 schwarze 1112: /* Skip trailing whitespace. */
1113:
1.289 schwarze 1114: for (; sz; sz--)
1.269 schwarze 1115: if ( ! isspace((unsigned char)cp[sz-1]))
1116: break;
1117:
1118: /* Skip empty strings. */
1119:
1120: if (sz == 0)
1121: return;
1122:
1123: if (*dest == NULL) {
1124: *dest = mandoc_strndup(cp, sz);
1125: return;
1126: }
1127:
1128: mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp);
1129: free(*dest);
1130: *dest = cp;
1.266 schwarze 1131: }
1132:
1133: /* --- main functions of the roff parser ---------------------------------- */
1134:
1.94 kristaps 1135: /*
1.206 schwarze 1136: * In the current line, expand escape sequences that tend to get
1137: * used in numerical expressions and conditional requests.
1138: * Also check the syntax of the remaining escape sequences.
1.154 kristaps 1139: */
1.172 schwarze 1140: static enum rofferr
1.238 schwarze 1141: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.94 kristaps 1142: {
1.208 schwarze 1143: char ubuf[24]; /* buffer to print the number */
1.205 schwarze 1144: const char *start; /* start of the string to process */
1.209 schwarze 1145: char *stesc; /* start of an escape sequence ('\\') */
1.108 schwarze 1146: const char *stnam; /* start of the name, after "[(*" */
1147: const char *cp; /* end of the name, e.g. before ']' */
1148: const char *res; /* the string to be substituted */
1.238 schwarze 1149: char *nbuf; /* new buffer to copy buf->buf to */
1.181 schwarze 1150: size_t maxl; /* expected length of the escape name */
1151: size_t naml; /* actual length of the escape name */
1.237 schwarze 1152: enum mandoc_esc esc; /* type of the escape sequence */
1153: int inaml; /* length returned from mandoc_escape() */
1.181 schwarze 1154: int expand_count; /* to avoid infinite loops */
1.206 schwarze 1155: int npos; /* position in numeric expression */
1.218 schwarze 1156: int arg_complete; /* argument not interrupted by eol */
1.303 schwarze 1157: int done; /* no more input available */
1.206 schwarze 1158: char term; /* character terminating the escape */
1.94 kristaps 1159:
1.303 schwarze 1160: /* Search forward for comments. */
1161:
1162: done = 0;
1163: start = buf->buf + pos;
1164: for (stesc = buf->buf + pos; *stesc != '\0'; stesc++) {
1165: if (stesc[0] != r->escape || stesc[1] == '\0')
1166: continue;
1167: stesc++;
1168: if (*stesc != '"' && *stesc != '#')
1169: continue;
1170: cp = strchr(stesc--, '\0') - 1;
1171: if (*cp == '\n') {
1172: done = 1;
1173: cp--;
1174: }
1175: if (*cp == ' ' || *cp == '\t')
1176: mandoc_msg(MANDOCERR_SPACE_EOL, r->parse,
1177: ln, cp - buf->buf, NULL);
1178: while (stesc > start && stesc[-1] == ' ')
1179: stesc--;
1180: *stesc = '\0';
1181: break;
1182: }
1183: if (stesc == start)
1184: return ROFF_CONT;
1185: stesc--;
1186:
1187: /* Notice the end of the input. */
1188:
1189: if (*stesc == '\n') {
1190: *stesc-- = '\0';
1191: done = 1;
1192: }
1193:
1.170 schwarze 1194: expand_count = 0;
1.303 schwarze 1195: while (stesc >= start) {
1.170 schwarze 1196:
1.205 schwarze 1197: /* Search backwards for the next backslash. */
1198:
1.303 schwarze 1199: if (*stesc != r->escape) {
1200: if (*stesc == '\\') {
1201: *stesc = '\0';
1202: buf->sz = mandoc_asprintf(&nbuf, "%s\\e%s",
1203: buf->buf, stesc + 1) + 1;
1204: start = nbuf + pos;
1205: stesc = nbuf + (stesc - buf->buf);
1206: free(buf->buf);
1207: buf->buf = nbuf;
1208: }
1209: stesc--;
1.205 schwarze 1210: continue;
1.303 schwarze 1211: }
1.205 schwarze 1212:
1213: /* If it is escaped, skip it. */
1214:
1215: for (cp = stesc - 1; cp >= start; cp--)
1.303 schwarze 1216: if (*cp != r->escape)
1.205 schwarze 1217: break;
1218:
1.238 schwarze 1219: if ((stesc - cp) % 2 == 0) {
1.303 schwarze 1220: while (stesc > cp)
1221: *stesc-- = '\\';
1.205 schwarze 1222: continue;
1.303 schwarze 1223: } else if (stesc[1] != '\0') {
1224: *stesc = '\\';
1225: } else {
1226: *stesc-- = '\0';
1227: if (done)
1228: continue;
1229: else
1230: return ROFF_APPEND;
1.205 schwarze 1231: }
1.108 schwarze 1232:
1.206 schwarze 1233: /* Decide whether to expand or to check only. */
1.108 schwarze 1234:
1.206 schwarze 1235: term = '\0';
1.205 schwarze 1236: cp = stesc + 1;
1.181 schwarze 1237: switch (*cp) {
1.207 schwarze 1238: case '*':
1.181 schwarze 1239: res = NULL;
1240: break;
1.207 schwarze 1241: case 'B':
1242: case 'w':
1.206 schwarze 1243: term = cp[1];
1244: /* FALLTHROUGH */
1.207 schwarze 1245: case 'n':
1.181 schwarze 1246: res = ubuf;
1247: break;
1248: default:
1.237 schwarze 1249: esc = mandoc_escape(&cp, &stnam, &inaml);
1250: if (esc == ESCAPE_ERROR ||
1251: (esc == ESCAPE_SPECIAL &&
1.279 schwarze 1252: mchars_spec2cp(stnam, inaml) < 0))
1.219 schwarze 1253: mandoc_vmsg(MANDOCERR_ESC_BAD,
1.238 schwarze 1254: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1255: "%.*s", (int)(cp - stesc), stesc);
1.303 schwarze 1256: stesc--;
1.205 schwarze 1257: continue;
1.152 kristaps 1258: }
1259:
1.205 schwarze 1260: if (EXPAND_LIMIT < ++expand_count) {
1261: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.238 schwarze 1262: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1263: return ROFF_IGN;
1.205 schwarze 1264: }
1.108 schwarze 1265:
1266: /*
1267: * The third character decides the length
1.181 schwarze 1268: * of the name of the string or register.
1.108 schwarze 1269: * Save a pointer to the name.
1270: */
1271:
1.238 schwarze 1272: if (term == '\0') {
1.206 schwarze 1273: switch (*++cp) {
1.207 schwarze 1274: case '\0':
1.206 schwarze 1275: maxl = 0;
1276: break;
1.207 schwarze 1277: case '(':
1.206 schwarze 1278: cp++;
1279: maxl = 2;
1280: break;
1.207 schwarze 1281: case '[':
1.206 schwarze 1282: cp++;
1283: term = ']';
1284: maxl = 0;
1285: break;
1286: default:
1287: maxl = 1;
1288: break;
1289: }
1290: } else {
1291: cp += 2;
1.94 kristaps 1292: maxl = 0;
1293: }
1.108 schwarze 1294: stnam = cp;
1.94 kristaps 1295:
1.108 schwarze 1296: /* Advance to the end of the name. */
1.94 kristaps 1297:
1.253 schwarze 1298: naml = 0;
1.218 schwarze 1299: arg_complete = 1;
1.253 schwarze 1300: while (maxl == 0 || naml < maxl) {
1.238 schwarze 1301: if (*cp == '\0') {
1.219 schwarze 1302: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 1303: ln, (int)(stesc - buf->buf), stesc);
1.218 schwarze 1304: arg_complete = 0;
1.206 schwarze 1305: break;
1.153 kristaps 1306: }
1.238 schwarze 1307: if (maxl == 0 && *cp == term) {
1.206 schwarze 1308: cp++;
1.253 schwarze 1309: break;
1310: }
1311: if (*cp++ != '\\' || stesc[1] != 'w') {
1312: naml++;
1313: continue;
1314: }
1315: switch (mandoc_escape(&cp, NULL, NULL)) {
1316: case ESCAPE_SPECIAL:
1317: case ESCAPE_UNICODE:
1318: case ESCAPE_NUMBERED:
1319: case ESCAPE_OVERSTRIKE:
1320: naml++;
1321: break;
1322: default:
1.94 kristaps 1323: break;
1.206 schwarze 1324: }
1.94 kristaps 1325: }
1326:
1.108 schwarze 1327: /*
1328: * Retrieve the replacement string; if it is
1329: * undefined, resume searching for escapes.
1330: */
1331:
1.206 schwarze 1332: switch (stesc[1]) {
1.207 schwarze 1333: case '*':
1.218 schwarze 1334: if (arg_complete)
1335: res = roff_getstrn(r, stnam, naml);
1.206 schwarze 1336: break;
1.207 schwarze 1337: case 'B':
1.206 schwarze 1338: npos = 0;
1.218 schwarze 1339: ubuf[0] = arg_complete &&
1.261 schwarze 1340: roff_evalnum(r, ln, stnam, &npos,
1341: NULL, ROFFNUM_SCALE) &&
1.218 schwarze 1342: stnam + npos + 1 == cp ? '1' : '0';
1.206 schwarze 1343: ubuf[1] = '\0';
1344: break;
1.207 schwarze 1345: case 'n':
1.218 schwarze 1346: if (arg_complete)
1347: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1348: roff_getregn(r, stnam, naml));
1349: else
1350: ubuf[0] = '\0';
1.206 schwarze 1351: break;
1.207 schwarze 1352: case 'w':
1.218 schwarze 1353: /* use even incomplete args */
1.208 schwarze 1354: (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.206 schwarze 1355: 24 * (int)naml);
1356: break;
1357: }
1.94 kristaps 1358:
1.238 schwarze 1359: if (res == NULL) {
1.219 schwarze 1360: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1361: r->parse, ln, (int)(stesc - buf->buf),
1.219 schwarze 1362: "%.*s", (int)naml, stnam);
1.142 kristaps 1363: res = "";
1.246 schwarze 1364: } else if (buf->sz + strlen(res) > SHRT_MAX) {
1365: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1366: ln, (int)(stesc - buf->buf), NULL);
1.277 schwarze 1367: return ROFF_IGN;
1.94 kristaps 1368: }
1369:
1.108 schwarze 1370: /* Replace the escape sequence by the string. */
1371:
1.209 schwarze 1372: *stesc = '\0';
1.238 schwarze 1373: buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
1374: buf->buf, res, cp) + 1;
1.94 kristaps 1375:
1.205 schwarze 1376: /* Prepare for the next replacement. */
1.94 kristaps 1377:
1.205 schwarze 1378: start = nbuf + pos;
1.238 schwarze 1379: stesc = nbuf + (stesc - buf->buf) + strlen(res);
1380: free(buf->buf);
1381: buf->buf = nbuf;
1.154 kristaps 1382: }
1.277 schwarze 1383: return ROFF_CONT;
1.154 kristaps 1384: }
1385:
1386: /*
1.275 schwarze 1387: * Process text streams.
1.154 kristaps 1388: */
1389: static enum rofferr
1.238 schwarze 1390: roff_parsetext(struct buf *buf, int pos, int *offs)
1.154 kristaps 1391: {
1392: size_t sz;
1393: const char *start;
1.178 schwarze 1394: char *p;
1395: int isz;
1.154 kristaps 1396: enum mandoc_esc esc;
1397:
1.275 schwarze 1398: /* Spring the input line trap. */
1399:
1400: if (roffit_lines == 1) {
1401: isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
1402: free(buf->buf);
1403: buf->buf = p;
1404: buf->sz = isz + 1;
1405: *offs = 0;
1406: free(roffit_macro);
1407: roffit_lines = 0;
1.277 schwarze 1408: return ROFF_REPARSE;
1.275 schwarze 1409: } else if (roffit_lines > 1)
1410: --roffit_lines;
1411:
1412: /* Convert all breakable hyphens into ASCII_HYPH. */
1413:
1.238 schwarze 1414: start = p = buf->buf + pos;
1.154 kristaps 1415:
1.238 schwarze 1416: while (*p != '\0') {
1.154 kristaps 1417: sz = strcspn(p, "-\\");
1418: p += sz;
1419:
1.238 schwarze 1420: if (*p == '\0')
1.159 kristaps 1421: break;
1422:
1.238 schwarze 1423: if (*p == '\\') {
1.154 kristaps 1424: /* Skip over escapes. */
1425: p++;
1.189 schwarze 1426: esc = mandoc_escape((const char **)&p, NULL, NULL);
1.238 schwarze 1427: if (esc == ESCAPE_ERROR)
1.154 kristaps 1428: break;
1.264 schwarze 1429: while (*p == '-')
1430: p++;
1.155 kristaps 1431: continue;
1.159 kristaps 1432: } else if (p == start) {
1.158 kristaps 1433: p++;
1.155 kristaps 1434: continue;
1.158 kristaps 1435: }
1.155 kristaps 1436:
1.171 schwarze 1437: if (isalpha((unsigned char)p[-1]) &&
1438: isalpha((unsigned char)p[1]))
1.155 kristaps 1439: *p = ASCII_HYPH;
1440: p++;
1.94 kristaps 1441: }
1.277 schwarze 1442: return ROFF_CONT;
1.94 kristaps 1443: }
1444:
1.67 kristaps 1445: enum rofferr
1.238 schwarze 1446: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.67 kristaps 1447: {
1.294 schwarze 1448: enum roff_tok t;
1.109 kristaps 1449: enum rofferr e;
1.238 schwarze 1450: int pos; /* parse point */
1.243 schwarze 1451: int spos; /* saved parse point for messages */
1.238 schwarze 1452: int ppos; /* original offset in buf->buf */
1453: int ctl; /* macro line (boolean) */
1454:
1455: ppos = pos = *offs;
1.79 kristaps 1456:
1.230 schwarze 1457: /* Handle in-line equation delimiters. */
1458:
1.236 schwarze 1459: if (r->tbl == NULL &&
1460: r->last_eqn != NULL && r->last_eqn->delim &&
1.230 schwarze 1461: (r->eqn == NULL || r->eqn_inline)) {
1.238 schwarze 1462: e = roff_eqndelim(r, buf, pos);
1.230 schwarze 1463: if (e == ROFF_REPARSE)
1.277 schwarze 1464: return e;
1.230 schwarze 1465: assert(e == ROFF_CONT);
1466: }
1467:
1468: /* Expand some escape sequences. */
1.94 kristaps 1469:
1.238 schwarze 1470: e = roff_res(r, buf, ln, pos);
1.303 schwarze 1471: if (e == ROFF_IGN || e == ROFF_APPEND)
1.277 schwarze 1472: return e;
1.238 schwarze 1473: assert(e == ROFF_CONT);
1.94 kristaps 1474:
1.238 schwarze 1475: ctl = roff_getcontrol(r, buf->buf, &pos);
1.130 kristaps 1476:
1.94 kristaps 1477: /*
1.79 kristaps 1478: * First, if a scope is open and we're not a macro, pass the
1.252 schwarze 1479: * text through the macro's filter.
1480: * Equations process all content themselves.
1481: * Tables process almost all content themselves, but we want
1482: * to warn about macros before passing it there.
1.79 kristaps 1483: */
1.74 kristaps 1484:
1.252 schwarze 1485: if (r->last != NULL && ! ctl) {
1.78 kristaps 1486: t = r->last->tok;
1.238 schwarze 1487: e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
1.294 schwarze 1488: if (e == ROFF_IGN)
1.277 schwarze 1489: return e;
1.294 schwarze 1490: assert(e == ROFF_CONT);
1.182 schwarze 1491: }
1.252 schwarze 1492: if (r->eqn != NULL)
1.277 schwarze 1493: return eqn_read(&r->eqn, ln, buf->buf, ppos, offs);
1.252 schwarze 1494: if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.277 schwarze 1495: return tbl_read(r->tbl, ln, buf->buf, ppos);
1.252 schwarze 1496: if ( ! ctl)
1.277 schwarze 1497: return roff_parsetext(buf, pos, offs);
1.228 schwarze 1498:
1499: /* Skip empty request lines. */
1500:
1.238 schwarze 1501: if (buf->buf[pos] == '"') {
1.228 schwarze 1502: mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
1503: ln, pos, NULL);
1.277 schwarze 1504: return ROFF_IGN;
1.238 schwarze 1505: } else if (buf->buf[pos] == '\0')
1.277 schwarze 1506: return ROFF_IGN;
1.67 kristaps 1507:
1.79 kristaps 1508: /*
1509: * If a scope is open, go to the child handler for that macro,
1510: * as it may want to preprocess before doing anything with it.
1.125 kristaps 1511: * Don't do so if an equation is open.
1.79 kristaps 1512: */
1.78 kristaps 1513:
1.79 kristaps 1514: if (r->last) {
1515: t = r->last->tok;
1.277 schwarze 1516: return (*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1517: }
1.78 kristaps 1518:
1.243 schwarze 1519: /* No scope is open. This is a new request or macro. */
1520:
1521: spos = pos;
1522: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1523:
1524: /* Tables ignore most macros. */
1525:
1.294 schwarze 1526: if (r->tbl != NULL && (t == TOKEN_NONE || t == ROFF_TS)) {
1.243 schwarze 1527: mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
1528: ln, pos, buf->buf + spos);
1.257 schwarze 1529: if (t == ROFF_TS)
1.277 schwarze 1530: return ROFF_IGN;
1.257 schwarze 1531: while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
1532: pos++;
1.291 schwarze 1533: while (buf->buf[pos] == ' ')
1.257 schwarze 1534: pos++;
1.277 schwarze 1535: return tbl_read(r->tbl, ln, buf->buf, pos);
1.243 schwarze 1536: }
1537:
1.79 kristaps 1538: /*
1.243 schwarze 1539: * This is neither a roff request nor a user-defined macro.
1540: * Let the standard macro set parsers handle it.
1.79 kristaps 1541: */
1.67 kristaps 1542:
1.294 schwarze 1543: if (t == TOKEN_NONE)
1.277 schwarze 1544: return ROFF_CONT;
1.243 schwarze 1545:
1546: /* Execute a roff request or a user defined macro. */
1.67 kristaps 1547:
1.296 schwarze 1548: return (*roffs[t].proc)(r, t, buf, ln, spos, pos, offs);
1.74 kristaps 1549: }
1550:
1.117 kristaps 1551: void
1.74 kristaps 1552: roff_endparse(struct roff *r)
1553: {
1554:
1.110 kristaps 1555: if (r->last)
1.221 schwarze 1556: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1557: r->last->line, r->last->col,
1.295 schwarze 1558: roff_name[r->last->tok]);
1.117 kristaps 1559:
1.125 kristaps 1560: if (r->eqn) {
1.221 schwarze 1561: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1562: r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.151 kristaps 1563: eqn_end(&r->eqn);
1.125 kristaps 1564: }
1565:
1.117 kristaps 1566: if (r->tbl) {
1.221 schwarze 1567: mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
1568: r->tbl->line, r->tbl->pos, "TS");
1.151 kristaps 1569: tbl_end(&r->tbl);
1.117 kristaps 1570: }
1.67 kristaps 1571: }
1572:
1573: /*
1574: * Parse a roff node's type from the input buffer. This must be in the
1575: * form of ".foo xxx" in the usual way.
1576: */
1.294 schwarze 1577: static enum roff_tok
1.214 schwarze 1578: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.67 kristaps 1579: {
1.214 schwarze 1580: char *cp;
1.106 kristaps 1581: const char *mac;
1582: size_t maclen;
1.294 schwarze 1583: enum roff_tok t;
1.67 kristaps 1584:
1.214 schwarze 1585: cp = buf + *pos;
1586:
1587: if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.294 schwarze 1588: return TOKEN_NONE;
1.67 kristaps 1589:
1.214 schwarze 1590: mac = cp;
1591: maclen = roff_getname(r, &cp, ln, ppos);
1.67 kristaps 1592:
1.106 kristaps 1593: t = (r->current_string = roff_getstrn(r, mac, maclen))
1.295 schwarze 1594: ? ROFF_USERDEF : roffhash_find(r->reqtab, mac, maclen);
1.67 kristaps 1595:
1.294 schwarze 1596: if (t != TOKEN_NONE)
1.214 schwarze 1597: *pos = cp - buf;
1.67 kristaps 1598:
1.277 schwarze 1599: return t;
1.67 kristaps 1600: }
1601:
1.266 schwarze 1602: /* --- handling of request blocks ----------------------------------------- */
1603:
1.67 kristaps 1604: static enum rofferr
1.76 kristaps 1605: roff_cblock(ROFF_ARGS)
1.67 kristaps 1606: {
1607:
1.79 kristaps 1608: /*
1609: * A block-close `..' should only be invoked as a child of an
1610: * ignore macro, otherwise raise a warning and just ignore it.
1611: */
1612:
1.238 schwarze 1613: if (r->last == NULL) {
1.221 schwarze 1614: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1615: ln, ppos, "..");
1.277 schwarze 1616: return ROFF_IGN;
1.76 kristaps 1617: }
1.67 kristaps 1618:
1.81 kristaps 1619: switch (r->last->tok) {
1.207 schwarze 1620: case ROFF_am:
1.220 schwarze 1621: /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.207 schwarze 1622: case ROFF_ami:
1623: case ROFF_de:
1.108 schwarze 1624: /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.207 schwarze 1625: case ROFF_dei:
1626: case ROFF_ig:
1.81 kristaps 1627: break;
1628: default:
1.221 schwarze 1629: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1630: ln, ppos, "..");
1.277 schwarze 1631: return ROFF_IGN;
1.76 kristaps 1632: }
1.67 kristaps 1633:
1.238 schwarze 1634: if (buf->buf[pos] != '\0')
1.217 schwarze 1635: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 1636: ".. %s", buf->buf + pos);
1.71 kristaps 1637:
1638: roffnode_pop(r);
1.76 kristaps 1639: roffnode_cleanscope(r);
1.277 schwarze 1640: return ROFF_IGN;
1.71 kristaps 1641:
1.67 kristaps 1642: }
1643:
1.76 kristaps 1644: static void
1645: roffnode_cleanscope(struct roff *r)
1.67 kristaps 1646: {
1647:
1.76 kristaps 1648: while (r->last) {
1.173 schwarze 1649: if (--r->last->endspan != 0)
1.76 kristaps 1650: break;
1651: roffnode_pop(r);
1652: }
1.67 kristaps 1653: }
1654:
1.195 schwarze 1655: static void
1656: roff_ccond(struct roff *r, int ln, int ppos)
1.74 kristaps 1657: {
1658:
1.76 kristaps 1659: if (NULL == r->last) {
1.221 schwarze 1660: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1661: ln, ppos, "\\}");
1.195 schwarze 1662: return;
1.76 kristaps 1663: }
1664:
1.82 kristaps 1665: switch (r->last->tok) {
1.207 schwarze 1666: case ROFF_el:
1667: case ROFF_ie:
1668: case ROFF_if:
1.82 kristaps 1669: break;
1670: default:
1.221 schwarze 1671: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1672: ln, ppos, "\\}");
1.195 schwarze 1673: return;
1.75 kristaps 1674: }
1675:
1.76 kristaps 1676: if (r->last->endspan > -1) {
1.221 schwarze 1677: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
1678: ln, ppos, "\\}");
1.195 schwarze 1679: return;
1.76 kristaps 1680: }
1681:
1.75 kristaps 1682: roffnode_pop(r);
1.76 kristaps 1683: roffnode_cleanscope(r);
1.195 schwarze 1684: return;
1.76 kristaps 1685: }
1686:
1687: static enum rofferr
1.80 kristaps 1688: roff_block(ROFF_ARGS)
1.76 kristaps 1689: {
1.220 schwarze 1690: const char *name;
1691: char *iname, *cp;
1.213 schwarze 1692: size_t namesz;
1.106 kristaps 1693:
1.220 schwarze 1694: /* Ignore groff compatibility mode for now. */
1.76 kristaps 1695:
1.238 schwarze 1696: if (tok == ROFF_de1)
1.220 schwarze 1697: tok = ROFF_de;
1.251 schwarze 1698: else if (tok == ROFF_dei1)
1699: tok = ROFF_dei;
1.238 schwarze 1700: else if (tok == ROFF_am1)
1.220 schwarze 1701: tok = ROFF_am;
1.251 schwarze 1702: else if (tok == ROFF_ami1)
1703: tok = ROFF_ami;
1.220 schwarze 1704:
1705: /* Parse the macro name argument. */
1706:
1.238 schwarze 1707: cp = buf->buf + pos;
1708: if (tok == ROFF_ig) {
1.220 schwarze 1709: iname = NULL;
1710: namesz = 0;
1711: } else {
1712: iname = cp;
1713: namesz = roff_getname(r, &cp, ln, ppos);
1714: iname[namesz] = '\0';
1715: }
1.107 kristaps 1716:
1.220 schwarze 1717: /* Resolve the macro name argument if it is indirect. */
1.107 kristaps 1718:
1.238 schwarze 1719: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1720: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1721: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1722: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1723: "%.*s", (int)namesz, iname);
1724: namesz = 0;
1725: } else
1726: namesz = strlen(name);
1727: } else
1728: name = iname;
1.107 kristaps 1729:
1.238 schwarze 1730: if (namesz == 0 && tok != ROFF_ig) {
1.220 schwarze 1731: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
1.295 schwarze 1732: ln, ppos, roff_name[tok]);
1.277 schwarze 1733: return ROFF_IGN;
1.220 schwarze 1734: }
1.80 kristaps 1735:
1.106 kristaps 1736: roffnode_push(r, tok, name, ln, ppos);
1737:
1738: /*
1739: * At the beginning of a `de' macro, clear the existing string
1740: * with the same name, if there is one. New content will be
1.193 schwarze 1741: * appended from roff_block_text() in multiline mode.
1.106 kristaps 1742: */
1.107 kristaps 1743:
1.238 schwarze 1744: if (tok == ROFF_de || tok == ROFF_dei)
1.213 schwarze 1745: roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.76 kristaps 1746:
1.238 schwarze 1747: if (*cp == '\0')
1.277 schwarze 1748: return ROFF_IGN;
1.78 kristaps 1749:
1.220 schwarze 1750: /* Get the custom end marker. */
1.107 kristaps 1751:
1.220 schwarze 1752: iname = cp;
1.213 schwarze 1753: namesz = roff_getname(r, &cp, ln, ppos);
1.220 schwarze 1754:
1755: /* Resolve the end marker if it is indirect. */
1756:
1.238 schwarze 1757: if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
1758: if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.220 schwarze 1759: mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.238 schwarze 1760: r->parse, ln, (int)(iname - buf->buf),
1.220 schwarze 1761: "%.*s", (int)namesz, iname);
1762: namesz = 0;
1763: } else
1764: namesz = strlen(name);
1765: } else
1766: name = iname;
1767:
1.213 schwarze 1768: if (namesz)
1769: r->last->end = mandoc_strndup(name, namesz);
1.78 kristaps 1770:
1.238 schwarze 1771: if (*cp != '\0')
1.217 schwarze 1772: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
1.295 schwarze 1773: ln, pos, ".%s ... %s", roff_name[tok], cp);
1.74 kristaps 1774:
1.277 schwarze 1775: return ROFF_IGN;
1.78 kristaps 1776: }
1777:
1778: static enum rofferr
1.80 kristaps 1779: roff_block_sub(ROFF_ARGS)
1.79 kristaps 1780: {
1.294 schwarze 1781: enum roff_tok t;
1.79 kristaps 1782: int i, j;
1783:
1784: /*
1785: * First check whether a custom macro exists at this level. If
1786: * it does, then check against it. This is some of groff's
1787: * stranger behaviours. If we encountered a custom end-scope
1788: * tag and that tag also happens to be a "real" macro, then we
1789: * need to try interpreting it again as a real macro. If it's
1790: * not, then return ignore. Else continue.
1791: */
1792:
1793: if (r->last->end) {
1.130 kristaps 1794: for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.238 schwarze 1795: if (buf->buf[i] != r->last->end[j])
1.79 kristaps 1796: break;
1797:
1.238 schwarze 1798: if (r->last->end[j] == '\0' &&
1799: (buf->buf[i] == '\0' ||
1800: buf->buf[i] == ' ' ||
1801: buf->buf[i] == '\t')) {
1.79 kristaps 1802: roffnode_pop(r);
1803: roffnode_cleanscope(r);
1804:
1.238 schwarze 1805: while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.130 kristaps 1806: i++;
1807:
1808: pos = i;
1.238 schwarze 1809: if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
1.294 schwarze 1810: TOKEN_NONE)
1.277 schwarze 1811: return ROFF_RERUN;
1812: return ROFF_IGN;
1.79 kristaps 1813: }
1814: }
1815:
1816: /*
1817: * If we have no custom end-query or lookup failed, then try
1818: * pulling it out of the hashtable.
1819: */
1820:
1.238 schwarze 1821: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.79 kristaps 1822:
1.238 schwarze 1823: if (t != ROFF_cblock) {
1824: if (tok != ROFF_ig)
1825: roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.277 schwarze 1826: return ROFF_IGN;
1.106 kristaps 1827: }
1.79 kristaps 1828:
1.277 schwarze 1829: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.79 kristaps 1830: }
1831:
1832: static enum rofferr
1.80 kristaps 1833: roff_block_text(ROFF_ARGS)
1.78 kristaps 1834: {
1835:
1.238 schwarze 1836: if (tok != ROFF_ig)
1837: roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.106 kristaps 1838:
1.277 schwarze 1839: return ROFF_IGN;
1.78 kristaps 1840: }
1841:
1842: static enum rofferr
1.82 kristaps 1843: roff_cond_sub(ROFF_ARGS)
1844: {
1.294 schwarze 1845: enum roff_tok t;
1.139 kristaps 1846: char *ep;
1.198 schwarze 1847: int rr;
1.82 kristaps 1848:
1849: rr = r->last->rule;
1.139 kristaps 1850: roffnode_cleanscope(r);
1.238 schwarze 1851: t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.82 kristaps 1852:
1.139 kristaps 1853: /*
1.177 schwarze 1854: * Fully handle known macros when they are structurally
1855: * required or when the conditional evaluated to true.
1.87 kristaps 1856: */
1857:
1.294 schwarze 1858: if (t != TOKEN_NONE && (rr || roffs[t].flags & ROFFMAC_STRUCT))
1.277 schwarze 1859: return (*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs);
1.144 kristaps 1860:
1.196 schwarze 1861: /*
1862: * If `\}' occurs on a macro line without a preceding macro,
1863: * drop the line completely.
1864: */
1865:
1.238 schwarze 1866: ep = buf->buf + pos;
1867: if (ep[0] == '\\' && ep[1] == '}')
1.198 schwarze 1868: rr = 0;
1.196 schwarze 1869:
1.177 schwarze 1870: /* Always check for the closing delimiter `\}'. */
1.144 kristaps 1871:
1.238 schwarze 1872: while ((ep = strchr(ep, '\\')) != NULL) {
1873: if (*(++ep) == '}') {
1.197 schwarze 1874: *ep = '&';
1.238 schwarze 1875: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1876: }
1.247 schwarze 1877: if (*ep != '\0')
1878: ++ep;
1.177 schwarze 1879: }
1.277 schwarze 1880: return rr ? ROFF_CONT : ROFF_IGN;
1.82 kristaps 1881: }
1882:
1883: static enum rofferr
1884: roff_cond_text(ROFF_ARGS)
1.78 kristaps 1885: {
1.140 kristaps 1886: char *ep;
1.198 schwarze 1887: int rr;
1.82 kristaps 1888:
1889: rr = r->last->rule;
1.140 kristaps 1890: roffnode_cleanscope(r);
1.82 kristaps 1891:
1.238 schwarze 1892: ep = buf->buf + pos;
1893: while ((ep = strchr(ep, '\\')) != NULL) {
1894: if (*(++ep) == '}') {
1.197 schwarze 1895: *ep = '&';
1.238 schwarze 1896: roff_ccond(r, ln, ep - buf->buf - 1);
1.197 schwarze 1897: }
1.247 schwarze 1898: if (*ep != '\0')
1899: ++ep;
1.78 kristaps 1900: }
1.277 schwarze 1901: return rr ? ROFF_CONT : ROFF_IGN;
1.74 kristaps 1902: }
1903:
1.266 schwarze 1904: /* --- handling of numeric and conditional expressions -------------------- */
1905:
1.204 schwarze 1906: /*
1907: * Parse a single signed integer number. Stop at the first non-digit.
1908: * If there is at least one digit, return success and advance the
1909: * parse point, else return failure and let the parse point unchanged.
1910: * Ignore overflows, treat them just like the C language.
1911: */
1.184 schwarze 1912: static int
1.261 schwarze 1913: roff_getnum(const char *v, int *pos, int *res, int flags)
1.184 schwarze 1914: {
1.261 schwarze 1915: int myres, scaled, n, p;
1.206 schwarze 1916:
1917: if (NULL == res)
1918: res = &myres;
1.184 schwarze 1919:
1920: p = *pos;
1921: n = v[p] == '-';
1.261 schwarze 1922: if (n || v[p] == '+')
1.184 schwarze 1923: p++;
1924:
1.261 schwarze 1925: if (flags & ROFFNUM_WHITE)
1926: while (isspace((unsigned char)v[p]))
1927: p++;
1928:
1.184 schwarze 1929: for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.204 schwarze 1930: *res = 10 * *res + v[p] - '0';
1.184 schwarze 1931: if (p == *pos + n)
1932: return 0;
1933:
1934: if (n)
1935: *res = -*res;
1936:
1.254 schwarze 1937: /* Each number may be followed by one optional scaling unit. */
1938:
1939: switch (v[p]) {
1940: case 'f':
1.261 schwarze 1941: scaled = *res * 65536;
1.254 schwarze 1942: break;
1943: case 'i':
1.261 schwarze 1944: scaled = *res * 240;
1.254 schwarze 1945: break;
1946: case 'c':
1.261 schwarze 1947: scaled = *res * 240 / 2.54;
1.254 schwarze 1948: break;
1949: case 'v':
1950: case 'P':
1.261 schwarze 1951: scaled = *res * 40;
1.254 schwarze 1952: break;
1953: case 'm':
1954: case 'n':
1.261 schwarze 1955: scaled = *res * 24;
1.254 schwarze 1956: break;
1957: case 'p':
1.261 schwarze 1958: scaled = *res * 10 / 3;
1.254 schwarze 1959: break;
1960: case 'u':
1.261 schwarze 1961: scaled = *res;
1.254 schwarze 1962: break;
1963: case 'M':
1.261 schwarze 1964: scaled = *res * 6 / 25;
1.254 schwarze 1965: break;
1966: default:
1.261 schwarze 1967: scaled = *res;
1.254 schwarze 1968: p--;
1969: break;
1970: }
1.261 schwarze 1971: if (flags & ROFFNUM_SCALE)
1972: *res = scaled;
1.254 schwarze 1973:
1974: *pos = p + 1;
1.277 schwarze 1975: return 1;
1.184 schwarze 1976: }
1977:
1.198 schwarze 1978: /*
1979: * Evaluate a string comparison condition.
1980: * The first character is the delimiter.
1981: * Succeed if the string up to its second occurrence
1982: * matches the string up to its third occurence.
1983: * Advance the cursor after the third occurrence
1984: * or lacking that, to the end of the line.
1985: */
1986: static int
1987: roff_evalstrcond(const char *v, int *pos)
1988: {
1989: const char *s1, *s2, *s3;
1990: int match;
1991:
1992: match = 0;
1993: s1 = v + *pos; /* initial delimiter */
1994: s2 = s1 + 1; /* for scanning the first string */
1995: s3 = strchr(s2, *s1); /* for scanning the second string */
1996:
1997: if (NULL == s3) /* found no middle delimiter */
1998: goto out;
1999:
2000: while ('\0' != *++s3) {
2001: if (*s2 != *s3) { /* mismatch */
2002: s3 = strchr(s3, *s1);
2003: break;
2004: }
2005: if (*s3 == *s1) { /* found the final delimiter */
2006: match = 1;
2007: break;
2008: }
2009: s2++;
2010: }
2011:
2012: out:
2013: if (NULL == s3)
2014: s3 = strchr(s2, '\0');
1.242 schwarze 2015: else if (*s3 != '\0')
1.198 schwarze 2016: s3++;
2017: *pos = s3 - v;
1.277 schwarze 2018: return match;
1.198 schwarze 2019: }
2020:
1.204 schwarze 2021: /*
2022: * Evaluate an optionally negated single character, numerical,
2023: * or string condition.
2024: */
1.198 schwarze 2025: static int
1.271 schwarze 2026: roff_evalcond(struct roff *r, int ln, char *v, int *pos)
1.88 kristaps 2027: {
1.271 schwarze 2028: char *cp, *name;
2029: size_t sz;
1.241 schwarze 2030: int number, savepos, wanttrue;
1.88 kristaps 2031:
1.198 schwarze 2032: if ('!' == v[*pos]) {
2033: wanttrue = 0;
2034: (*pos)++;
2035: } else
2036: wanttrue = 1;
2037:
1.88 kristaps 2038: switch (v[*pos]) {
1.240 schwarze 2039: case '\0':
1.277 schwarze 2040: return 0;
1.207 schwarze 2041: case 'n':
2042: case 'o':
1.88 kristaps 2043: (*pos)++;
1.277 schwarze 2044: return wanttrue;
1.207 schwarze 2045: case 'c':
2046: case 'd':
2047: case 'e':
2048: case 't':
1.239 schwarze 2049: case 'v':
1.88 kristaps 2050: (*pos)++;
1.277 schwarze 2051: return !wanttrue;
1.271 schwarze 2052: case 'r':
2053: cp = name = v + ++*pos;
2054: sz = roff_getname(r, &cp, ln, *pos);
2055: *pos = cp - v;
1.277 schwarze 2056: return (sz && roff_hasregn(r, name, sz)) == wanttrue;
1.88 kristaps 2057: default:
2058: break;
2059: }
2060:
1.241 schwarze 2061: savepos = *pos;
1.261 schwarze 2062: if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.277 schwarze 2063: return (number > 0) == wanttrue;
1.241 schwarze 2064: else if (*pos == savepos)
1.277 schwarze 2065: return roff_evalstrcond(v, pos) == wanttrue;
1.204 schwarze 2066: else
1.277 schwarze 2067: return 0;
1.88 kristaps 2068: }
2069:
1.74 kristaps 2070: static enum rofferr
1.103 kristaps 2071: roff_line_ignore(ROFF_ARGS)
1.89 kristaps 2072: {
1.123 schwarze 2073:
1.277 schwarze 2074: return ROFF_IGN;
1.89 kristaps 2075: }
2076:
1.104 kristaps 2077: static enum rofferr
1.251 schwarze 2078: roff_insec(ROFF_ARGS)
2079: {
2080:
2081: mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
1.295 schwarze 2082: ln, ppos, roff_name[tok]);
1.277 schwarze 2083: return ROFF_IGN;
1.251 schwarze 2084: }
2085:
2086: static enum rofferr
2087: roff_unsupp(ROFF_ARGS)
2088: {
2089:
2090: mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
1.295 schwarze 2091: ln, ppos, roff_name[tok]);
1.277 schwarze 2092: return ROFF_IGN;
1.251 schwarze 2093: }
2094:
2095: static enum rofferr
1.82 kristaps 2096: roff_cond(ROFF_ARGS)
1.74 kristaps 2097: {
1.173 schwarze 2098:
2099: roffnode_push(r, tok, NULL, ln, ppos);
1.74 kristaps 2100:
1.207 schwarze 2101: /*
1.134 kristaps 2102: * An `.el' has no conditional body: it will consume the value
2103: * of the current rstack entry set in prior `ie' calls or
1.207 schwarze 2104: * defaults to DENY.
1.134 kristaps 2105: *
2106: * If we're not an `el', however, then evaluate the conditional.
2107: */
1.133 kristaps 2108:
1.238 schwarze 2109: r->last->rule = tok == ROFF_el ?
1.207 schwarze 2110: (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.238 schwarze 2111: roff_evalcond(r, ln, buf->buf, &pos);
1.77 kristaps 2112:
1.134 kristaps 2113: /*
2114: * An if-else will put the NEGATION of the current evaluated
2115: * conditional into the stack of rules.
2116: */
2117:
1.238 schwarze 2118: if (tok == ROFF_ie) {
1.223 schwarze 2119: if (r->rstackpos + 1 == r->rstacksz) {
2120: r->rstacksz += 16;
2121: r->rstack = mandoc_reallocarray(r->rstack,
2122: r->rstacksz, sizeof(int));
1.134 kristaps 2123: }
1.198 schwarze 2124: r->rstack[++r->rstackpos] = !r->last->rule;
1.82 kristaps 2125: }
1.88 kristaps 2126:
2127: /* If the parent has false as its rule, then so do we. */
2128:
1.198 schwarze 2129: if (r->last->parent && !r->last->parent->rule)
2130: r->last->rule = 0;
1.88 kristaps 2131:
2132: /*
1.173 schwarze 2133: * Determine scope.
2134: * If there is nothing on the line after the conditional,
2135: * not even whitespace, use next-line scope.
1.88 kristaps 2136: */
1.74 kristaps 2137:
1.238 schwarze 2138: if (buf->buf[pos] == '\0') {
1.173 schwarze 2139: r->last->endspan = 2;
2140: goto out;
2141: }
2142:
1.238 schwarze 2143: while (buf->buf[pos] == ' ')
1.173 schwarze 2144: pos++;
2145:
2146: /* An opening brace requests multiline scope. */
1.75 kristaps 2147:
1.238 schwarze 2148: if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.75 kristaps 2149: r->last->endspan = -1;
2150: pos += 2;
1.272 schwarze 2151: while (buf->buf[pos] == ' ')
2152: pos++;
1.173 schwarze 2153: goto out;
1.207 schwarze 2154: }
1.74 kristaps 2155:
1.77 kristaps 2156: /*
1.173 schwarze 2157: * Anything else following the conditional causes
2158: * single-line scope. Warn if the scope contains
2159: * nothing but trailing whitespace.
1.77 kristaps 2160: */
2161:
1.238 schwarze 2162: if (buf->buf[pos] == '\0')
1.216 schwarze 2163: mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
1.295 schwarze 2164: ln, ppos, roff_name[tok]);
1.77 kristaps 2165:
1.173 schwarze 2166: r->last->endspan = 1;
1.74 kristaps 2167:
1.173 schwarze 2168: out:
1.75 kristaps 2169: *offs = pos;
1.277 schwarze 2170: return ROFF_RERUN;
1.83 schwarze 2171: }
2172:
2173: static enum rofferr
1.92 schwarze 2174: roff_ds(ROFF_ARGS)
2175: {
1.212 schwarze 2176: char *string;
2177: const char *name;
2178: size_t namesz;
1.96 kristaps 2179:
1.251 schwarze 2180: /* Ignore groff compatibility mode for now. */
2181:
2182: if (tok == ROFF_ds1)
2183: tok = ROFF_ds;
2184: else if (tok == ROFF_as1)
2185: tok = ROFF_as;
2186:
1.96 kristaps 2187: /*
1.212 schwarze 2188: * The first word is the name of the string.
2189: * If it is empty or terminated by an escape sequence,
2190: * abort the `ds' request without defining anything.
1.96 kristaps 2191: */
1.92 schwarze 2192:
1.238 schwarze 2193: name = string = buf->buf + pos;
2194: if (*name == '\0')
1.277 schwarze 2195: return ROFF_IGN;
1.92 schwarze 2196:
1.212 schwarze 2197: namesz = roff_getname(r, &string, ln, pos);
1.238 schwarze 2198: if (name[namesz] == '\\')
1.277 schwarze 2199: return ROFF_IGN;
1.212 schwarze 2200:
2201: /* Read past the initial double-quote, if any. */
1.238 schwarze 2202: if (*string == '"')
1.92 schwarze 2203: string++;
2204:
1.96 kristaps 2205: /* The rest is the value. */
1.212 schwarze 2206: roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
2207: ROFF_as == tok);
1.277 schwarze 2208: return ROFF_IGN;
1.92 schwarze 2209: }
2210:
1.204 schwarze 2211: /*
2212: * Parse a single operator, one or two characters long.
2213: * If the operator is recognized, return success and advance the
2214: * parse point, else return failure and let the parse point unchanged.
2215: */
2216: static int
2217: roff_getop(const char *v, int *pos, char *res)
2218: {
2219:
2220: *res = v[*pos];
2221:
2222: switch (*res) {
1.207 schwarze 2223: case '+':
2224: case '-':
2225: case '*':
2226: case '/':
2227: case '%':
2228: case '&':
2229: case ':':
1.204 schwarze 2230: break;
2231: case '<':
2232: switch (v[*pos + 1]) {
1.207 schwarze 2233: case '=':
1.204 schwarze 2234: *res = 'l';
2235: (*pos)++;
2236: break;
1.207 schwarze 2237: case '>':
1.204 schwarze 2238: *res = '!';
2239: (*pos)++;
2240: break;
1.207 schwarze 2241: case '?':
1.204 schwarze 2242: *res = 'i';
2243: (*pos)++;
2244: break;
2245: default:
2246: break;
2247: }
2248: break;
2249: case '>':
2250: switch (v[*pos + 1]) {
1.207 schwarze 2251: case '=':
1.204 schwarze 2252: *res = 'g';
2253: (*pos)++;
2254: break;
1.207 schwarze 2255: case '?':
1.204 schwarze 2256: *res = 'a';
2257: (*pos)++;
2258: break;
2259: default:
2260: break;
2261: }
2262: break;
2263: case '=':
2264: if ('=' == v[*pos + 1])
2265: (*pos)++;
2266: break;
2267: default:
1.277 schwarze 2268: return 0;
1.204 schwarze 2269: }
2270: (*pos)++;
2271:
1.277 schwarze 2272: return *res;
1.204 schwarze 2273: }
2274:
2275: /*
2276: * Evaluate either a parenthesized numeric expression
2277: * or a single signed integer number.
2278: */
2279: static int
1.235 schwarze 2280: roff_evalpar(struct roff *r, int ln,
1.261 schwarze 2281: const char *v, int *pos, int *res, int flags)
1.204 schwarze 2282: {
2283:
2284: if ('(' != v[*pos])
1.277 schwarze 2285: return roff_getnum(v, pos, res, flags);
1.204 schwarze 2286:
2287: (*pos)++;
1.261 schwarze 2288: if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.277 schwarze 2289: return 0;
1.204 schwarze 2290:
1.206 schwarze 2291: /*
2292: * Omission of the closing parenthesis
2293: * is an error in validation mode,
2294: * but ignored in evaluation mode.
2295: */
2296:
1.204 schwarze 2297: if (')' == v[*pos])
2298: (*pos)++;
1.206 schwarze 2299: else if (NULL == res)
1.277 schwarze 2300: return 0;
1.204 schwarze 2301:
1.277 schwarze 2302: return 1;
1.204 schwarze 2303: }
2304:
2305: /*
2306: * Evaluate a complete numeric expression.
2307: * Proceed left to right, there is no concept of precedence.
2308: */
2309: static int
1.235 schwarze 2310: roff_evalnum(struct roff *r, int ln, const char *v,
1.261 schwarze 2311: int *pos, int *res, int flags)
1.204 schwarze 2312: {
2313: int mypos, operand2;
2314: char operator;
2315:
2316: if (NULL == pos) {
2317: mypos = 0;
2318: pos = &mypos;
2319: }
2320:
1.261 schwarze 2321: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2322: while (isspace((unsigned char)v[*pos]))
2323: (*pos)++;
2324:
1.261 schwarze 2325: if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.277 schwarze 2326: return 0;
1.204 schwarze 2327:
2328: while (1) {
1.261 schwarze 2329: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2330: while (isspace((unsigned char)v[*pos]))
2331: (*pos)++;
2332:
2333: if ( ! roff_getop(v, pos, &operator))
2334: break;
2335:
1.261 schwarze 2336: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2337: while (isspace((unsigned char)v[*pos]))
2338: (*pos)++;
2339:
1.261 schwarze 2340: if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.277 schwarze 2341: return 0;
1.204 schwarze 2342:
1.261 schwarze 2343: if (flags & ROFFNUM_WHITE)
1.204 schwarze 2344: while (isspace((unsigned char)v[*pos]))
2345: (*pos)++;
1.206 schwarze 2346:
2347: if (NULL == res)
2348: continue;
1.204 schwarze 2349:
2350: switch (operator) {
1.207 schwarze 2351: case '+':
1.204 schwarze 2352: *res += operand2;
2353: break;
1.207 schwarze 2354: case '-':
1.204 schwarze 2355: *res -= operand2;
2356: break;
1.207 schwarze 2357: case '*':
1.204 schwarze 2358: *res *= operand2;
2359: break;
1.207 schwarze 2360: case '/':
1.244 schwarze 2361: if (operand2 == 0) {
1.235 schwarze 2362: mandoc_msg(MANDOCERR_DIVZERO,
1.234 kristaps 2363: r->parse, ln, *pos, v);
2364: *res = 0;
2365: break;
2366: }
1.204 schwarze 2367: *res /= operand2;
2368: break;
1.207 schwarze 2369: case '%':
1.244 schwarze 2370: if (operand2 == 0) {
2371: mandoc_msg(MANDOCERR_DIVZERO,
2372: r->parse, ln, *pos, v);
2373: *res = 0;
2374: break;
2375: }
1.204 schwarze 2376: *res %= operand2;
2377: break;
1.207 schwarze 2378: case '<':
1.204 schwarze 2379: *res = *res < operand2;
2380: break;
1.207 schwarze 2381: case '>':
1.204 schwarze 2382: *res = *res > operand2;
2383: break;
1.207 schwarze 2384: case 'l':
1.204 schwarze 2385: *res = *res <= operand2;
2386: break;
1.207 schwarze 2387: case 'g':
1.204 schwarze 2388: *res = *res >= operand2;
2389: break;
1.207 schwarze 2390: case '=':
1.204 schwarze 2391: *res = *res == operand2;
2392: break;
1.207 schwarze 2393: case '!':
1.204 schwarze 2394: *res = *res != operand2;
2395: break;
1.207 schwarze 2396: case '&':
1.204 schwarze 2397: *res = *res && operand2;
2398: break;
1.207 schwarze 2399: case ':':
1.204 schwarze 2400: *res = *res || operand2;
2401: break;
1.207 schwarze 2402: case 'i':
1.204 schwarze 2403: if (operand2 < *res)
2404: *res = operand2;
2405: break;
1.207 schwarze 2406: case 'a':
1.204 schwarze 2407: if (operand2 > *res)
2408: *res = operand2;
2409: break;
2410: default:
2411: abort();
2412: }
2413: }
1.277 schwarze 2414: return 1;
1.204 schwarze 2415: }
2416:
1.266 schwarze 2417: /* --- register management ------------------------------------------------ */
2418:
1.180 schwarze 2419: void
1.187 schwarze 2420: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.147 kristaps 2421: {
1.180 schwarze 2422: struct roffreg *reg;
2423:
2424: /* Search for an existing register with the same name. */
2425: reg = r->regtab;
2426:
2427: while (reg && strcmp(name, reg->key.p))
2428: reg = reg->next;
1.147 kristaps 2429:
1.180 schwarze 2430: if (NULL == reg) {
2431: /* Create a new register. */
2432: reg = mandoc_malloc(sizeof(struct roffreg));
2433: reg->key.p = mandoc_strdup(name);
2434: reg->key.sz = strlen(name);
1.187 schwarze 2435: reg->val = 0;
1.180 schwarze 2436: reg->next = r->regtab;
2437: r->regtab = reg;
2438: }
2439:
1.187 schwarze 2440: if ('+' == sign)
2441: reg->val += val;
2442: else if ('-' == sign)
2443: reg->val -= val;
2444: else
2445: reg->val = val;
1.147 kristaps 2446: }
2447:
1.192 schwarze 2448: /*
2449: * Handle some predefined read-only number registers.
2450: * For now, return -1 if the requested register is not predefined;
2451: * in case a predefined read-only register having the value -1
2452: * were to turn up, another special value would have to be chosen.
2453: */
2454: static int
1.273 schwarze 2455: roff_getregro(const struct roff *r, const char *name)
1.192 schwarze 2456: {
2457:
2458: switch (*name) {
1.273 schwarze 2459: case '$': /* Number of arguments of the last macro evaluated. */
1.277 schwarze 2460: return r->argc;
1.207 schwarze 2461: case 'A': /* ASCII approximation mode is always off. */
1.277 schwarze 2462: return 0;
1.207 schwarze 2463: case 'g': /* Groff compatibility mode is always on. */
1.277 schwarze 2464: return 1;
1.207 schwarze 2465: case 'H': /* Fixed horizontal resolution. */
1.277 schwarze 2466: return 24;
1.207 schwarze 2467: case 'j': /* Always adjust left margin only. */
1.277 schwarze 2468: return 0;
1.207 schwarze 2469: case 'T': /* Some output device is always defined. */
1.277 schwarze 2470: return 1;
1.207 schwarze 2471: case 'V': /* Fixed vertical resolution. */
1.277 schwarze 2472: return 40;
1.192 schwarze 2473: default:
1.277 schwarze 2474: return -1;
1.192 schwarze 2475: }
2476: }
2477:
1.181 schwarze 2478: int
1.180 schwarze 2479: roff_getreg(const struct roff *r, const char *name)
1.147 kristaps 2480: {
1.180 schwarze 2481: struct roffreg *reg;
1.192 schwarze 2482: int val;
2483:
2484: if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
1.273 schwarze 2485: val = roff_getregro(r, name + 1);
1.192 schwarze 2486: if (-1 != val)
1.277 schwarze 2487: return val;
1.192 schwarze 2488: }
1.180 schwarze 2489:
2490: for (reg = r->regtab; reg; reg = reg->next)
2491: if (0 == strcmp(name, reg->key.p))
1.277 schwarze 2492: return reg->val;
1.181 schwarze 2493:
1.277 schwarze 2494: return 0;
1.181 schwarze 2495: }
2496:
2497: static int
2498: roff_getregn(const struct roff *r, const char *name, size_t len)
2499: {
2500: struct roffreg *reg;
1.192 schwarze 2501: int val;
2502:
2503: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2504: val = roff_getregro(r, name + 1);
1.192 schwarze 2505: if (-1 != val)
1.277 schwarze 2506: return val;
1.192 schwarze 2507: }
1.181 schwarze 2508:
2509: for (reg = r->regtab; reg; reg = reg->next)
2510: if (len == reg->key.sz &&
2511: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2512: return reg->val;
1.271 schwarze 2513:
1.277 schwarze 2514: return 0;
1.271 schwarze 2515: }
2516:
2517: static int
2518: roff_hasregn(const struct roff *r, const char *name, size_t len)
2519: {
2520: struct roffreg *reg;
2521: int val;
2522:
2523: if ('.' == name[0] && 2 == len) {
1.273 schwarze 2524: val = roff_getregro(r, name + 1);
1.271 schwarze 2525: if (-1 != val)
1.277 schwarze 2526: return 1;
1.271 schwarze 2527: }
2528:
2529: for (reg = r->regtab; reg; reg = reg->next)
2530: if (len == reg->key.sz &&
2531: 0 == strncmp(name, reg->key.p, len))
1.277 schwarze 2532: return 1;
1.147 kristaps 2533:
1.277 schwarze 2534: return 0;
1.147 kristaps 2535: }
2536:
1.180 schwarze 2537: static void
2538: roff_freereg(struct roffreg *reg)
1.147 kristaps 2539: {
1.180 schwarze 2540: struct roffreg *old_reg;
1.147 kristaps 2541:
1.180 schwarze 2542: while (NULL != reg) {
2543: free(reg->key.p);
2544: old_reg = reg;
2545: reg = reg->next;
2546: free(old_reg);
2547: }
1.147 kristaps 2548: }
1.92 schwarze 2549:
2550: static enum rofferr
1.89 kristaps 2551: roff_nr(ROFF_ARGS)
1.83 schwarze 2552: {
1.212 schwarze 2553: char *key, *val;
2554: size_t keysz;
1.138 kristaps 2555: int iv;
1.187 schwarze 2556: char sign;
1.89 kristaps 2557:
1.238 schwarze 2558: key = val = buf->buf + pos;
2559: if (*key == '\0')
1.277 schwarze 2560: return ROFF_IGN;
1.212 schwarze 2561:
2562: keysz = roff_getname(r, &val, ln, pos);
1.238 schwarze 2563: if (key[keysz] == '\\')
1.277 schwarze 2564: return ROFF_IGN;
1.212 schwarze 2565: key[keysz] = '\0';
1.89 kristaps 2566:
1.187 schwarze 2567: sign = *val;
1.238 schwarze 2568: if (sign == '+' || sign == '-')
1.187 schwarze 2569: val++;
2570:
1.261 schwarze 2571: if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.204 schwarze 2572: roff_setreg(r, key, iv, sign);
1.109 kristaps 2573:
1.277 schwarze 2574: return ROFF_IGN;
1.203 schwarze 2575: }
2576:
2577: static enum rofferr
2578: roff_rr(ROFF_ARGS)
2579: {
2580: struct roffreg *reg, **prev;
1.212 schwarze 2581: char *name, *cp;
2582: size_t namesz;
1.203 schwarze 2583:
1.238 schwarze 2584: name = cp = buf->buf + pos;
2585: if (*name == '\0')
1.277 schwarze 2586: return ROFF_IGN;
1.212 schwarze 2587: namesz = roff_getname(r, &cp, ln, pos);
2588: name[namesz] = '\0';
1.203 schwarze 2589:
2590: prev = &r->regtab;
2591: while (1) {
2592: reg = *prev;
1.238 schwarze 2593: if (reg == NULL || !strcmp(name, reg->key.p))
1.203 schwarze 2594: break;
2595: prev = ®->next;
2596: }
1.238 schwarze 2597: if (reg != NULL) {
1.203 schwarze 2598: *prev = reg->next;
2599: free(reg->key.p);
2600: free(reg);
2601: }
1.277 schwarze 2602: return ROFF_IGN;
1.122 schwarze 2603: }
2604:
1.266 schwarze 2605: /* --- handler functions for roff requests -------------------------------- */
2606:
1.122 schwarze 2607: static enum rofferr
2608: roff_rm(ROFF_ARGS)
2609: {
2610: const char *name;
2611: char *cp;
1.212 schwarze 2612: size_t namesz;
1.122 schwarze 2613:
1.238 schwarze 2614: cp = buf->buf + pos;
2615: while (*cp != '\0') {
1.212 schwarze 2616: name = cp;
1.238 schwarze 2617: namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.212 schwarze 2618: roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.238 schwarze 2619: if (name[namesz] == '\\')
1.212 schwarze 2620: break;
1.122 schwarze 2621: }
1.277 schwarze 2622: return ROFF_IGN;
1.178 schwarze 2623: }
2624:
2625: static enum rofferr
2626: roff_it(ROFF_ARGS)
2627: {
2628: int iv;
2629:
2630: /* Parse the number of lines. */
1.261 schwarze 2631:
2632: if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.222 schwarze 2633: mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.238 schwarze 2634: ln, ppos, buf->buf + 1);
1.277 schwarze 2635: return ROFF_IGN;
1.178 schwarze 2636: }
2637:
1.262 schwarze 2638: while (isspace((unsigned char)buf->buf[pos]))
2639: pos++;
2640:
2641: /*
2642: * Arm the input line trap.
2643: * Special-casing "an-trap" is an ugly workaround to cope
2644: * with DocBook stupidly fiddling with man(7) internals.
2645: */
1.261 schwarze 2646:
1.178 schwarze 2647: roffit_lines = iv;
1.262 schwarze 2648: roffit_macro = mandoc_strdup(iv != 1 ||
2649: strcmp(buf->buf + pos, "an-trap") ?
2650: buf->buf + pos : "br");
1.277 schwarze 2651: return ROFF_IGN;
1.175 schwarze 2652: }
2653:
2654: static enum rofferr
2655: roff_Dd(ROFF_ARGS)
2656: {
2657: const char *const *cp;
2658:
1.227 schwarze 2659: if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.175 schwarze 2660: for (cp = __mdoc_reserved; *cp; cp++)
2661: roff_setstr(r, *cp, NULL, 0);
2662:
1.227 schwarze 2663: if (r->format == 0)
2664: r->format = MPARSE_MDOC;
2665:
1.277 schwarze 2666: return ROFF_CONT;
1.175 schwarze 2667: }
2668:
2669: static enum rofferr
2670: roff_TH(ROFF_ARGS)
2671: {
2672: const char *const *cp;
2673:
1.227 schwarze 2674: if ((r->options & MPARSE_QUICK) == 0)
1.175 schwarze 2675: for (cp = __man_reserved; *cp; cp++)
2676: roff_setstr(r, *cp, NULL, 0);
2677:
1.227 schwarze 2678: if (r->format == 0)
2679: r->format = MPARSE_MAN;
2680:
1.277 schwarze 2681: return ROFF_CONT;
1.109 kristaps 2682: }
2683:
2684: static enum rofferr
2685: roff_TE(ROFF_ARGS)
2686: {
2687:
2688: if (NULL == r->tbl)
1.221 schwarze 2689: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2690: ln, ppos, "TE");
1.258 schwarze 2691: else if ( ! tbl_end(&r->tbl)) {
2692: free(buf->buf);
2693: buf->buf = mandoc_strdup(".sp");
2694: buf->sz = 4;
1.277 schwarze 2695: return ROFF_REPARSE;
1.258 schwarze 2696: }
1.277 schwarze 2697: return ROFF_IGN;
1.112 kristaps 2698: }
2699:
2700: static enum rofferr
2701: roff_T_(ROFF_ARGS)
2702: {
2703:
2704: if (NULL == r->tbl)
1.221 schwarze 2705: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
2706: ln, ppos, "T&");
1.112 kristaps 2707: else
1.296 schwarze 2708: tbl_restart(ln, ppos, r->tbl);
1.112 kristaps 2709:
1.277 schwarze 2710: return ROFF_IGN;
1.109 kristaps 2711: }
2712:
1.230 schwarze 2713: /*
2714: * Handle in-line equation delimiters.
2715: */
2716: static enum rofferr
1.238 schwarze 2717: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.151 kristaps 2718: {
1.233 schwarze 2719: char *cp1, *cp2;
2720: const char *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.151 kristaps 2721:
1.230 schwarze 2722: /*
2723: * Outside equations, look for an opening delimiter.
2724: * If we are inside an equation, we already know it is
2725: * in-line, or this function wouldn't have been called;
2726: * so look for a closing delimiter.
2727: */
2728:
1.238 schwarze 2729: cp1 = buf->buf + pos;
1.230 schwarze 2730: cp2 = strchr(cp1, r->eqn == NULL ?
2731: r->last_eqn->odelim : r->last_eqn->cdelim);
2732: if (cp2 == NULL)
1.277 schwarze 2733: return ROFF_CONT;
1.230 schwarze 2734:
1.233 schwarze 2735: *cp2++ = '\0';
2736: bef_pr = bef_nl = aft_nl = aft_pr = "";
2737:
2738: /* Handle preceding text, protecting whitespace. */
2739:
1.238 schwarze 2740: if (*buf->buf != '\0') {
1.233 schwarze 2741: if (r->eqn == NULL)
2742: bef_pr = "\\&";
2743: bef_nl = "\n";
2744: }
2745:
2746: /*
2747: * Prepare replacing the delimiter with an equation macro
2748: * and drop leading white space from the equation.
2749: */
1.230 schwarze 2750:
1.233 schwarze 2751: if (r->eqn == NULL) {
2752: while (*cp2 == ' ')
2753: cp2++;
2754: mac = ".EQ";
2755: } else
2756: mac = ".EN";
2757:
2758: /* Handle following text, protecting whitespace. */
2759:
2760: if (*cp2 != '\0') {
2761: aft_nl = "\n";
2762: if (r->eqn != NULL)
2763: aft_pr = "\\&";
2764: }
2765:
2766: /* Do the actual replacement. */
2767:
1.238 schwarze 2768: buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.233 schwarze 2769: bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.238 schwarze 2770: free(buf->buf);
2771: buf->buf = cp1;
1.230 schwarze 2772:
2773: /* Toggle the in-line state of the eqn subsystem. */
2774:
2775: r->eqn_inline = r->eqn == NULL;
1.277 schwarze 2776: return ROFF_REPARSE;
1.151 kristaps 2777: }
2778:
1.235 schwarze 2779: static enum rofferr
2780: roff_EQ(ROFF_ARGS)
1.125 kristaps 2781: {
1.151 kristaps 2782: struct eqn_node *e;
1.125 kristaps 2783:
1.238 schwarze 2784: assert(r->eqn == NULL);
1.235 schwarze 2785: e = eqn_alloc(ppos, ln, r->parse);
1.125 kristaps 2786:
1.230 schwarze 2787: if (r->last_eqn) {
1.125 kristaps 2788: r->last_eqn->next = e;
1.230 schwarze 2789: e->delim = r->last_eqn->delim;
2790: e->odelim = r->last_eqn->odelim;
2791: e->cdelim = r->last_eqn->cdelim;
2792: } else
1.125 kristaps 2793: r->first_eqn = r->last_eqn = e;
2794:
2795: r->eqn = r->last_eqn = e;
1.151 kristaps 2796:
1.238 schwarze 2797: if (buf->buf[pos] != '\0')
1.235 schwarze 2798: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.238 schwarze 2799: ".EQ %s", buf->buf + pos);
1.151 kristaps 2800:
1.277 schwarze 2801: return ROFF_IGN;
1.125 kristaps 2802: }
2803:
2804: static enum rofferr
2805: roff_EN(ROFF_ARGS)
2806: {
2807:
1.221 schwarze 2808: mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.277 schwarze 2809: return ROFF_IGN;
1.125 kristaps 2810: }
2811:
2812: static enum rofferr
1.109 kristaps 2813: roff_TS(ROFF_ARGS)
2814: {
1.176 schwarze 2815: struct tbl_node *tbl;
1.89 kristaps 2816:
1.115 kristaps 2817: if (r->tbl) {
1.221 schwarze 2818: mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
2819: ln, ppos, "TS breaks TS");
1.151 kristaps 2820: tbl_end(&r->tbl);
1.115 kristaps 2821: }
1.83 schwarze 2822:
1.176 schwarze 2823: tbl = tbl_alloc(ppos, ln, r->parse);
1.113 kristaps 2824:
2825: if (r->last_tbl)
1.176 schwarze 2826: r->last_tbl->next = tbl;
1.113 kristaps 2827: else
1.176 schwarze 2828: r->first_tbl = r->last_tbl = tbl;
1.113 kristaps 2829:
1.176 schwarze 2830: r->tbl = r->last_tbl = tbl;
1.297 schwarze 2831: return ROFF_IGN;
2832: }
2833:
2834: static enum rofferr
2835: roff_onearg(ROFF_ARGS)
2836: {
2837: struct roff_node *n;
2838: char *cp;
2839:
1.302 schwarze 2840: if (r->man->flags & (MAN_BLINE | MAN_ELINE) &&
2841: (tok == ROFF_sp || tok == ROFF_ti))
2842: man_breakscope(r->man, tok);
2843:
1.297 schwarze 2844: roff_elem_alloc(r->man, ln, ppos, tok);
2845: n = r->man->last;
2846:
2847: cp = buf->buf + pos;
2848: if (*cp != '\0') {
2849: while (*cp != '\0' && *cp != ' ')
2850: cp++;
2851: while (*cp == ' ')
2852: *cp++ = '\0';
2853: if (*cp != '\0')
2854: mandoc_vmsg(MANDOCERR_ARG_EXCESS,
2855: r->parse, ln, cp - buf->buf,
2856: "%s ... %s", roff_name[tok], cp);
2857: roff_word_alloc(r->man, ln, pos, buf->buf + pos);
1.300 schwarze 2858: }
2859:
2860: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
2861: r->man->last = n;
2862: r->man->next = ROFF_NEXT_SIBLING;
2863: return ROFF_IGN;
2864: }
2865:
2866: static enum rofferr
2867: roff_manyarg(ROFF_ARGS)
2868: {
2869: struct roff_node *n;
2870: char *sp, *ep;
2871:
2872: roff_elem_alloc(r->man, ln, ppos, tok);
2873: n = r->man->last;
2874:
2875: for (sp = ep = buf->buf + pos; *sp != '\0'; sp = ep) {
2876: while (*ep != '\0' && *ep != ' ')
2877: ep++;
2878: while (*ep == ' ')
2879: *ep++ = '\0';
2880: roff_word_alloc(r->man, ln, sp - buf->buf, sp);
1.297 schwarze 2881: }
2882:
2883: n->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
2884: r->man->last = n;
2885: r->man->next = ROFF_NEXT_SIBLING;
1.277 schwarze 2886: return ROFF_IGN;
1.251 schwarze 2887: }
2888:
2889: static enum rofferr
1.296 schwarze 2890: roff_br(ROFF_ARGS)
1.251 schwarze 2891: {
1.302 schwarze 2892: if (r->man->flags & (MAN_BLINE | MAN_ELINE))
2893: man_breakscope(r->man, ROFF_br);
1.296 schwarze 2894: roff_elem_alloc(r->man, ln, ppos, ROFF_br);
2895: if (buf->buf[pos] != '\0')
2896: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
2897: "%s %s", roff_name[tok], buf->buf + pos);
2898: r->man->last->flags |= NODE_LINE | NODE_VALID | NODE_ENDED;
2899: r->man->next = ROFF_NEXT_SIBLING;
2900: return ROFF_IGN;
1.92 schwarze 2901: }
2902:
1.105 kristaps 2903: static enum rofferr
1.174 kristaps 2904: roff_cc(ROFF_ARGS)
2905: {
2906: const char *p;
2907:
1.238 schwarze 2908: p = buf->buf + pos;
1.174 kristaps 2909:
1.238 schwarze 2910: if (*p == '\0' || (r->control = *p++) == '.')
1.303 schwarze 2911: r->control = '\0';
1.174 kristaps 2912:
1.238 schwarze 2913: if (*p != '\0')
1.260 schwarze 2914: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2915: ln, p - buf->buf, "cc ... %s", p);
1.174 kristaps 2916:
1.277 schwarze 2917: return ROFF_IGN;
1.174 kristaps 2918: }
2919:
2920: static enum rofferr
1.303 schwarze 2921: roff_ec(ROFF_ARGS)
2922: {
2923: const char *p;
2924:
2925: p = buf->buf + pos;
2926: if (*p == '\0')
2927: r->escape = '\\';
2928: else {
2929: r->escape = *p;
2930: if (*++p != '\0')
2931: mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
2932: ln, p - buf->buf, "ec ... %s", p);
2933: }
2934: return ROFF_IGN;
2935: }
2936:
2937: static enum rofferr
2938: roff_eo(ROFF_ARGS)
2939: {
2940: r->escape = '\0';
2941: if (buf->buf[pos] != '\0')
2942: mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse,
2943: ln, pos, "eo %s", buf->buf + pos);
2944: return ROFF_IGN;
2945: }
2946:
2947: static enum rofferr
1.164 kristaps 2948: roff_tr(ROFF_ARGS)
2949: {
2950: const char *p, *first, *second;
2951: size_t fsz, ssz;
2952: enum mandoc_esc esc;
2953:
1.238 schwarze 2954: p = buf->buf + pos;
1.164 kristaps 2955:
1.238 schwarze 2956: if (*p == '\0') {
1.260 schwarze 2957: mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.277 schwarze 2958: return ROFF_IGN;
1.164 kristaps 2959: }
2960:
1.238 schwarze 2961: while (*p != '\0') {
1.164 kristaps 2962: fsz = ssz = 1;
2963:
2964: first = p++;
1.238 schwarze 2965: if (*first == '\\') {
1.164 kristaps 2966: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2967: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2968: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2969: ln, (int)(p - buf->buf), first);
1.277 schwarze 2970: return ROFF_IGN;
1.164 kristaps 2971: }
2972: fsz = (size_t)(p - first);
2973: }
2974:
2975: second = p++;
1.238 schwarze 2976: if (*second == '\\') {
1.164 kristaps 2977: esc = mandoc_escape(&p, NULL, NULL);
1.238 schwarze 2978: if (esc == ESCAPE_ERROR) {
1.219 schwarze 2979: mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.238 schwarze 2980: ln, (int)(p - buf->buf), second);
1.277 schwarze 2981: return ROFF_IGN;
1.164 kristaps 2982: }
2983: ssz = (size_t)(p - second);
1.238 schwarze 2984: } else if (*second == '\0') {
1.260 schwarze 2985: mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
2986: ln, first - buf->buf, "tr %s", first);
1.164 kristaps 2987: second = " ";
1.165 kristaps 2988: p--;
1.164 kristaps 2989: }
2990:
1.167 kristaps 2991: if (fsz > 1) {
1.207 schwarze 2992: roff_setstrn(&r->xmbtab, first, fsz,
2993: second, ssz, 0);
1.167 kristaps 2994: continue;
2995: }
2996:
1.238 schwarze 2997: if (r->xtab == NULL)
1.207 schwarze 2998: r->xtab = mandoc_calloc(128,
2999: sizeof(struct roffstr));
1.167 kristaps 3000:
3001: free(r->xtab[(int)*first].p);
3002: r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
3003: r->xtab[(int)*first].sz = ssz;
1.164 kristaps 3004: }
3005:
1.277 schwarze 3006: return ROFF_IGN;
1.164 kristaps 3007: }
3008:
3009: static enum rofferr
1.105 kristaps 3010: roff_so(ROFF_ARGS)
3011: {
1.249 schwarze 3012: char *name, *cp;
1.105 kristaps 3013:
1.238 schwarze 3014: name = buf->buf + pos;
1.224 schwarze 3015: mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.105 kristaps 3016:
3017: /*
3018: * Handle `so'. Be EXTREMELY careful, as we shouldn't be
3019: * opening anything that's not in our cwd or anything beneath
3020: * it. Thus, explicitly disallow traversing up the file-system
3021: * or using absolute paths.
3022: */
3023:
1.238 schwarze 3024: if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.210 schwarze 3025: mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
3026: ".so %s", name);
1.249 schwarze 3027: buf->sz = mandoc_asprintf(&cp,
3028: ".sp\nSee the file %s.\n.sp", name) + 1;
3029: free(buf->buf);
3030: buf->buf = cp;
3031: *offs = 0;
1.277 schwarze 3032: return ROFF_REPARSE;
1.105 kristaps 3033: }
3034:
3035: *offs = pos;
1.277 schwarze 3036: return ROFF_SO;
1.105 kristaps 3037: }
1.92 schwarze 3038:
1.266 schwarze 3039: /* --- user defined strings and macros ------------------------------------ */
3040:
1.106 kristaps 3041: static enum rofferr
3042: roff_userdef(ROFF_ARGS)
1.99 kristaps 3043: {
1.263 schwarze 3044: const char *arg[9], *ap;
1.106 kristaps 3045: char *cp, *n1, *n2;
1.292 schwarze 3046: int expand_count, i, ib, ie;
1.263 schwarze 3047: size_t asz, rsz;
1.106 kristaps 3048:
3049: /*
3050: * Collect pointers to macro argument strings
1.188 schwarze 3051: * and NUL-terminate them.
1.106 kristaps 3052: */
1.263 schwarze 3053:
1.273 schwarze 3054: r->argc = 0;
1.238 schwarze 3055: cp = buf->buf + pos;
1.273 schwarze 3056: for (i = 0; i < 9; i++) {
3057: if (*cp == '\0')
3058: arg[i] = "";
3059: else {
3060: arg[i] = mandoc_getarg(r->parse, &cp, ln, &pos);
3061: r->argc = i + 1;
3062: }
3063: }
1.99 kristaps 3064:
1.106 kristaps 3065: /*
3066: * Expand macro arguments.
1.99 kristaps 3067: */
1.263 schwarze 3068:
3069: buf->sz = strlen(r->current_string) + 1;
1.292 schwarze 3070: n1 = n2 = cp = mandoc_malloc(buf->sz);
1.263 schwarze 3071: memcpy(n1, r->current_string, buf->sz);
1.292 schwarze 3072: expand_count = 0;
1.263 schwarze 3073: while (*cp != '\0') {
3074:
3075: /* Scan ahead for the next argument invocation. */
3076:
3077: if (*cp++ != '\\')
3078: continue;
3079: if (*cp++ != '$')
3080: continue;
1.274 schwarze 3081: if (*cp == '*') { /* \\$* inserts all arguments */
3082: ib = 0;
3083: ie = r->argc - 1;
3084: } else { /* \\$1 .. \\$9 insert one argument */
3085: ib = ie = *cp - '1';
3086: if (ib < 0 || ib > 8)
3087: continue;
3088: }
1.263 schwarze 3089: cp -= 2;
1.292 schwarze 3090:
3091: /*
3092: * Prevent infinite recursion.
3093: */
3094:
3095: if (cp >= n2)
3096: expand_count = 1;
3097: else if (++expand_count > EXPAND_LIMIT) {
3098: mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
3099: ln, (int)(cp - n1), NULL);
1.293 schwarze 3100: free(buf->buf);
3101: buf->buf = n1;
1.292 schwarze 3102: return ROFF_IGN;
3103: }
1.263 schwarze 3104:
3105: /*
3106: * Determine the size of the expanded argument,
3107: * taking escaping of quotes into account.
3108: */
3109:
1.274 schwarze 3110: asz = ie > ib ? ie - ib : 0; /* for blanks */
3111: for (i = ib; i <= ie; i++) {
3112: for (ap = arg[i]; *ap != '\0'; ap++) {
3113: asz++;
3114: if (*ap == '"')
3115: asz += 3;
3116: }
1.263 schwarze 3117: }
3118: if (asz != 3) {
3119:
3120: /*
3121: * Determine the size of the rest of the
3122: * unexpanded macro, including the NUL.
3123: */
3124:
3125: rsz = buf->sz - (cp - n1) - 3;
3126:
3127: /*
3128: * When shrinking, move before
3129: * releasing the storage.
3130: */
3131:
3132: if (asz < 3)
3133: memmove(cp + asz, cp + 3, rsz);
3134:
3135: /*
3136: * Resize the storage for the macro
3137: * and readjust the parse pointer.
3138: */
3139:
3140: buf->sz += asz - 3;
3141: n2 = mandoc_realloc(n1, buf->sz);
3142: cp = n2 + (cp - n1);
3143: n1 = n2;
3144:
3145: /*
3146: * When growing, make room
3147: * for the expanded argument.
3148: */
3149:
3150: if (asz > 3)
3151: memmove(cp + asz, cp + 3, rsz);
3152: }
3153:
3154: /* Copy the expanded argument, escaping quotes. */
3155:
3156: n2 = cp;
1.274 schwarze 3157: for (i = ib; i <= ie; i++) {
3158: for (ap = arg[i]; *ap != '\0'; ap++) {
3159: if (*ap == '"') {
3160: memcpy(n2, "\\(dq", 4);
3161: n2 += 4;
3162: } else
3163: *n2++ = *ap;
3164: }
3165: if (i < ie)
3166: *n2++ = ' ';
1.106 kristaps 3167: }
1.99 kristaps 3168: }
3169:
1.106 kristaps 3170: /*
3171: * Replace the macro invocation
3172: * by the expanded macro.
3173: */
1.263 schwarze 3174:
1.238 schwarze 3175: free(buf->buf);
3176: buf->buf = n1;
1.248 schwarze 3177: *offs = 0;
1.106 kristaps 3178:
1.277 schwarze 3179: return buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
3180: ROFF_REPARSE : ROFF_APPEND;
1.99 kristaps 3181: }
1.121 schwarze 3182:
1.212 schwarze 3183: static size_t
1.121 schwarze 3184: roff_getname(struct roff *r, char **cpp, int ln, int pos)
3185: {
3186: char *name, *cp;
1.212 schwarze 3187: size_t namesz;
1.121 schwarze 3188:
3189: name = *cpp;
3190: if ('\0' == *name)
1.277 schwarze 3191: return 0;
1.121 schwarze 3192:
1.212 schwarze 3193: /* Read until end of name and terminate it with NUL. */
3194: for (cp = name; 1; cp++) {
3195: if ('\0' == *cp || ' ' == *cp) {
3196: namesz = cp - name;
3197: break;
3198: }
1.121 schwarze 3199: if ('\\' != *cp)
3200: continue;
1.215 schwarze 3201: namesz = cp - name;
3202: if ('{' == cp[1] || '}' == cp[1])
3203: break;
1.121 schwarze 3204: cp++;
3205: if ('\\' == *cp)
3206: continue;
1.224 schwarze 3207: mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
3208: "%.*s", (int)(cp - name + 1), name);
1.212 schwarze 3209: mandoc_escape((const char **)&cp, NULL, NULL);
3210: break;
1.121 schwarze 3211: }
3212:
3213: /* Read past spaces. */
3214: while (' ' == *cp)
3215: cp++;
3216:
3217: *cpp = cp;
1.277 schwarze 3218: return namesz;
1.121 schwarze 3219: }
3220:
1.106 kristaps 3221: /*
3222: * Store *string into the user-defined string called *name.
3223: * To clear an existing entry, call with (*r, *name, NULL, 0).
1.193 schwarze 3224: * append == 0: replace mode
3225: * append == 1: single-line append mode
3226: * append == 2: multiline append mode, append '\n' after each call
1.106 kristaps 3227: */
1.94 kristaps 3228: static void
1.106 kristaps 3229: roff_setstr(struct roff *r, const char *name, const char *string,
1.193 schwarze 3230: int append)
1.92 schwarze 3231: {
1.164 kristaps 3232:
3233: roff_setstrn(&r->strtab, name, strlen(name), string,
1.207 schwarze 3234: string ? strlen(string) : 0, append);
1.164 kristaps 3235: }
3236:
3237: static void
1.166 kristaps 3238: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.193 schwarze 3239: const char *string, size_t stringsz, int append)
1.164 kristaps 3240: {
1.166 kristaps 3241: struct roffkv *n;
1.164 kristaps 3242: char *c;
3243: int i;
3244: size_t oldch, newch;
1.92 schwarze 3245:
1.106 kristaps 3246: /* Search for an existing string with the same name. */
1.164 kristaps 3247: n = *r;
3248:
1.211 schwarze 3249: while (n && (namesz != n->key.sz ||
3250: strncmp(n->key.p, name, namesz)))
1.92 schwarze 3251: n = n->next;
1.94 kristaps 3252:
3253: if (NULL == n) {
1.106 kristaps 3254: /* Create a new string table entry. */
1.166 kristaps 3255: n = mandoc_malloc(sizeof(struct roffkv));
3256: n->key.p = mandoc_strndup(name, namesz);
3257: n->key.sz = namesz;
3258: n->val.p = NULL;
3259: n->val.sz = 0;
1.164 kristaps 3260: n->next = *r;
3261: *r = n;
1.193 schwarze 3262: } else if (0 == append) {
1.166 kristaps 3263: free(n->val.p);
3264: n->val.p = NULL;
3265: n->val.sz = 0;
1.106 kristaps 3266: }
3267:
3268: if (NULL == string)
3269: return;
3270:
3271: /*
3272: * One additional byte for the '\n' in multiline mode,
3273: * and one for the terminating '\0'.
3274: */
1.193 schwarze 3275: newch = stringsz + (1 < append ? 2u : 1u);
1.164 kristaps 3276:
1.166 kristaps 3277: if (NULL == n->val.p) {
3278: n->val.p = mandoc_malloc(newch);
3279: *n->val.p = '\0';
1.106 kristaps 3280: oldch = 0;
3281: } else {
1.166 kristaps 3282: oldch = n->val.sz;
3283: n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.106 kristaps 3284: }
3285:
3286: /* Skip existing content in the destination buffer. */
1.166 kristaps 3287: c = n->val.p + (int)oldch;
1.106 kristaps 3288:
3289: /* Append new content to the destination buffer. */
1.164 kristaps 3290: i = 0;
3291: while (i < (int)stringsz) {
1.106 kristaps 3292: /*
3293: * Rudimentary roff copy mode:
3294: * Handle escaped backslashes.
3295: */
1.164 kristaps 3296: if ('\\' == string[i] && '\\' == string[i + 1])
3297: i++;
3298: *c++ = string[i++];
1.106 kristaps 3299: }
1.94 kristaps 3300:
1.106 kristaps 3301: /* Append terminating bytes. */
1.193 schwarze 3302: if (1 < append)
1.106 kristaps 3303: *c++ = '\n';
1.163 kristaps 3304:
1.106 kristaps 3305: *c = '\0';
1.166 kristaps 3306: n->val.sz = (int)(c - n->val.p);
1.92 schwarze 3307: }
3308:
1.94 kristaps 3309: static const char *
3310: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.92 schwarze 3311: {
1.166 kristaps 3312: const struct roffkv *n;
1.191 schwarze 3313: int i;
1.92 schwarze 3314:
1.164 kristaps 3315: for (n = r->strtab; n; n = n->next)
1.207 schwarze 3316: if (0 == strncmp(name, n->key.p, len) &&
3317: '\0' == n->key.p[(int)len])
1.277 schwarze 3318: return n->val.p;
1.191 schwarze 3319:
3320: for (i = 0; i < PREDEFS_MAX; i++)
3321: if (0 == strncmp(name, predefs[i].name, len) &&
3322: '\0' == predefs[i].name[(int)len])
1.277 schwarze 3323: return predefs[i].str;
1.94 kristaps 3324:
1.277 schwarze 3325: return NULL;
1.92 schwarze 3326: }
3327:
1.94 kristaps 3328: static void
1.167 kristaps 3329: roff_freestr(struct roffkv *r)
1.92 schwarze 3330: {
1.166 kristaps 3331: struct roffkv *n, *nn;
1.92 schwarze 3332:
1.167 kristaps 3333: for (n = r; n; n = nn) {
1.166 kristaps 3334: free(n->key.p);
3335: free(n->val.p);
1.92 schwarze 3336: nn = n->next;
3337: free(n);
3338: }
1.114 kristaps 3339: }
1.266 schwarze 3340:
3341: /* --- accessors and utility functions ------------------------------------ */
1.114 kristaps 3342:
3343: const struct tbl_span *
3344: roff_span(const struct roff *r)
3345: {
1.207 schwarze 3346:
1.277 schwarze 3347: return r->tbl ? tbl_span(r->tbl) : NULL;
1.125 kristaps 3348: }
3349:
3350: const struct eqn *
3351: roff_eqn(const struct roff *r)
3352: {
1.207 schwarze 3353:
1.277 schwarze 3354: return r->last_eqn ? &r->last_eqn->eqn : NULL;
1.164 kristaps 3355: }
3356:
3357: /*
3358: * Duplicate an input string, making the appropriate character
3359: * conversations (as stipulated by `tr') along the way.
3360: * Returns a heap-allocated string with all the replacements made.
3361: */
3362: char *
3363: roff_strdup(const struct roff *r, const char *p)
3364: {
1.166 kristaps 3365: const struct roffkv *cp;
1.164 kristaps 3366: char *res;
3367: const char *pp;
3368: size_t ssz, sz;
3369: enum mandoc_esc esc;
3370:
1.167 kristaps 3371: if (NULL == r->xmbtab && NULL == r->xtab)
1.277 schwarze 3372: return mandoc_strdup(p);
1.164 kristaps 3373: else if ('\0' == *p)
1.277 schwarze 3374: return mandoc_strdup("");
1.164 kristaps 3375:
3376: /*
3377: * Step through each character looking for term matches
3378: * (remember that a `tr' can be invoked with an escape, which is
3379: * a glyph but the escape is multi-character).
3380: * We only do this if the character hash has been initialised
3381: * and the string is >0 length.
3382: */
3383:
3384: res = NULL;
3385: ssz = 0;
3386:
3387: while ('\0' != *p) {
1.289 schwarze 3388: assert((unsigned int)*p < 128);
3389: if ('\\' != *p && r->xtab && r->xtab[(unsigned int)*p].p) {
1.167 kristaps 3390: sz = r->xtab[(int)*p].sz;
3391: res = mandoc_realloc(res, ssz + sz + 1);
3392: memcpy(res + ssz, r->xtab[(int)*p].p, sz);
3393: ssz += sz;
3394: p++;
3395: continue;
3396: } else if ('\\' != *p) {
3397: res = mandoc_realloc(res, ssz + 2);
3398: res[ssz++] = *p++;
3399: continue;
3400: }
3401:
1.164 kristaps 3402: /* Search for term matches. */
1.167 kristaps 3403: for (cp = r->xmbtab; cp; cp = cp->next)
1.166 kristaps 3404: if (0 == strncmp(p, cp->key.p, cp->key.sz))
1.164 kristaps 3405: break;
3406:
3407: if (NULL != cp) {
3408: /*
3409: * A match has been found.
3410: * Append the match to the array and move
3411: * forward by its keysize.
3412: */
1.207 schwarze 3413: res = mandoc_realloc(res,
3414: ssz + cp->val.sz + 1);
1.166 kristaps 3415: memcpy(res + ssz, cp->val.p, cp->val.sz);
3416: ssz += cp->val.sz;
3417: p += (int)cp->key.sz;
1.164 kristaps 3418: continue;
3419: }
3420:
1.167 kristaps 3421: /*
3422: * Handle escapes carefully: we need to copy
3423: * over just the escape itself, or else we might
3424: * do replacements within the escape itself.
3425: * Make sure to pass along the bogus string.
3426: */
3427: pp = p++;
3428: esc = mandoc_escape(&p, NULL, NULL);
3429: if (ESCAPE_ERROR == esc) {
3430: sz = strlen(pp);
1.164 kristaps 3431: res = mandoc_realloc(res, ssz + sz + 1);
3432: memcpy(res + ssz, pp, sz);
1.167 kristaps 3433: break;
1.164 kristaps 3434: }
1.207 schwarze 3435: /*
3436: * We bail out on bad escapes.
1.167 kristaps 3437: * No need to warn: we already did so when
3438: * roff_res() was called.
3439: */
3440: sz = (int)(p - pp);
3441: res = mandoc_realloc(res, ssz + sz + 1);
3442: memcpy(res + ssz, pp, sz);
3443: ssz += sz;
1.164 kristaps 3444: }
3445:
3446: res[(int)ssz] = '\0';
1.277 schwarze 3447: return res;
1.227 schwarze 3448: }
3449:
3450: int
3451: roff_getformat(const struct roff *r)
3452: {
3453:
1.277 schwarze 3454: return r->format;
1.174 kristaps 3455: }
3456:
3457: /*
1.207 schwarze 3458: * Find out whether a line is a macro line or not.
1.174 kristaps 3459: * If it is, adjust the current position and return one; if it isn't,
3460: * return zero and don't change the current position.
3461: * If the control character has been set with `.cc', then let that grain
3462: * precedence.
3463: * This is slighly contrary to groff, where using the non-breaking
3464: * control character when `cc' has been invoked will cause the
3465: * non-breaking macro contents to be printed verbatim.
3466: */
3467: int
3468: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
3469: {
3470: int pos;
3471:
3472: pos = *ppos;
3473:
1.303 schwarze 3474: if (r->control != '\0' && cp[pos] == r->control)
1.174 kristaps 3475: pos++;
1.303 schwarze 3476: else if (r->control != '\0')
1.277 schwarze 3477: return 0;
1.174 kristaps 3478: else if ('\\' == cp[pos] && '.' == cp[pos + 1])
3479: pos += 2;
3480: else if ('.' == cp[pos] || '\'' == cp[pos])
3481: pos++;
3482: else
1.277 schwarze 3483: return 0;
1.174 kristaps 3484:
3485: while (' ' == cp[pos] || '\t' == cp[pos])
3486: pos++;
3487:
3488: *ppos = pos;
1.277 schwarze 3489: return 1;
1.74 kristaps 3490: }
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