[BACK]Return to roff.c CVS log [TXT][DIR] Up to [local] / src / usr.bin / mandoc

Annotation of src/usr.bin/mandoc/roff.c, Revision 1.124

1.124   ! schwarze    1: /*     $OpenBSD: roff.c,v 1.123 2015/01/20 21:12:46 schwarze Exp $ */
1.1       schwarze    2: /*
1.119     schwarze    3:  * Copyright (c) 2010, 2011, 2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
                      4:  * Copyright (c) 2010-2015 Ingo Schwarze <schwarze@openbsd.org>
1.1       schwarze    5:  *
                      6:  * Permission to use, copy, modify, and distribute this software for any
                      7:  * purpose with or without fee is hereby granted, provided that the above
                      8:  * copyright notice and this permission notice appear in all copies.
                      9:  *
1.16      schwarze   10:  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
1.1       schwarze   11:  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
1.16      schwarze   12:  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
1.1       schwarze   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.
                     17:  */
1.99      schwarze   18: #include <sys/types.h>
                     19:
1.1       schwarze   20: #include <assert.h>
1.3       schwarze   21: #include <ctype.h>
1.117     schwarze   22: #include <limits.h>
1.51      schwarze   23: #include <stdio.h>
1.1       schwarze   24: #include <stdlib.h>
                     25: #include <string.h>
                     26:
                     27: #include "mandoc.h"
1.74      schwarze   28: #include "mandoc_aux.h"
1.98      daniel     29: #include "libmandoc.h"
1.27      schwarze   30: #include "libroff.h"
1.1       schwarze   31:
1.37      schwarze   32: /* Maximum number of nested if-else conditionals. */
1.2       schwarze   33: #define        RSTACK_MAX      128
                     34:
1.43      schwarze   35: /* Maximum number of string expansions per line, to break infinite loops. */
                     36: #define        EXPAND_LIMIT    1000
                     37:
1.1       schwarze   38: enum   rofft {
1.123     schwarze   39:        ROFF_ab,
1.20      schwarze   40:        ROFF_ad,
1.123     schwarze   41:        ROFF_af,
                     42:        ROFF_aln,
                     43:        ROFF_als,
1.2       schwarze   44:        ROFF_am,
1.123     schwarze   45:        ROFF_am1,
1.2       schwarze   46:        ROFF_ami,
1.123     schwarze   47:        ROFF_ami1,
1.66      schwarze   48:        ROFF_as,
1.123     schwarze   49:        ROFF_as1,
                     50:        ROFF_asciify,
                     51:        ROFF_backtrace,
                     52:        ROFF_bd,
                     53:        ROFF_bleedat,
                     54:        ROFF_blm,
                     55:        ROFF_box,
                     56:        ROFF_boxa,
                     57:        ROFF_bp,
                     58:        ROFF_BP,
                     59:        /* MAN_br, MDOC_br */
                     60:        ROFF_break,
                     61:        ROFF_breakchar,
                     62:        ROFF_brnl,
                     63:        ROFF_brp,
                     64:        ROFF_brpnl,
                     65:        ROFF_c2,
1.48      schwarze   66:        ROFF_cc,
1.67      schwarze   67:        ROFF_ce,
1.123     schwarze   68:        ROFF_cf,
                     69:        ROFF_cflags,
                     70:        ROFF_ch,
                     71:        ROFF_char,
                     72:        ROFF_chop,
                     73:        ROFF_class,
                     74:        ROFF_close,
                     75:        ROFF_CL,
                     76:        ROFF_color,
                     77:        ROFF_composite,
                     78:        ROFF_continue,
                     79:        ROFF_cp,
                     80:        ROFF_cropat,
                     81:        ROFF_cs,
                     82:        ROFF_cu,
                     83:        ROFF_da,
                     84:        ROFF_dch,
                     85:        ROFF_Dd,
1.1       schwarze   86:        ROFF_de,
1.123     schwarze   87:        ROFF_de1,
                     88:        ROFF_defcolor,
1.1       schwarze   89:        ROFF_dei,
1.123     schwarze   90:        ROFF_dei1,
                     91:        ROFF_device,
                     92:        ROFF_devicem,
                     93:        ROFF_di,
                     94:        ROFF_do,
1.2       schwarze   95:        ROFF_ds,
1.123     schwarze   96:        ROFF_ds1,
                     97:        ROFF_dwh,
                     98:        ROFF_dt,
                     99:        ROFF_ec,
                    100:        ROFF_ecr,
                    101:        ROFF_ecs,
1.2       schwarze  102:        ROFF_el,
1.123     schwarze  103:        ROFF_em,
                    104:        ROFF_EN,
                    105:        ROFF_eo,
                    106:        ROFF_EP,
                    107:        ROFF_EQ,
                    108:        ROFF_errprint,
                    109:        ROFF_ev,
                    110:        ROFF_evc,
                    111:        ROFF_ex,
                    112:        ROFF_fallback,
1.58      schwarze  113:        ROFF_fam,
1.123     schwarze  114:        ROFF_fc,
                    115:        ROFF_fchar,
                    116:        ROFF_fcolor,
                    117:        ROFF_fdeferlig,
                    118:        ROFF_feature,
                    119:        /* MAN_fi; ignored in mdoc(7) */
                    120:        ROFF_fkern,
                    121:        ROFF_fl,
                    122:        ROFF_flig,
                    123:        ROFF_fp,
                    124:        ROFF_fps,
                    125:        ROFF_fschar,
                    126:        ROFF_fspacewidth,
                    127:        ROFF_fspecial,
                    128:        /* MAN_ft; ignored in mdoc(7) */
                    129:        ROFF_ftr,
                    130:        ROFF_fzoom,
                    131:        ROFF_gcolor,
                    132:        ROFF_hc,
                    133:        ROFF_hcode,
                    134:        ROFF_hidechar,
                    135:        ROFF_hla,
                    136:        ROFF_hlm,
                    137:        ROFF_hpf,
                    138:        ROFF_hpfa,
                    139:        ROFF_hpfcode,
1.59      schwarze  140:        ROFF_hw,
1.20      schwarze  141:        ROFF_hy,
1.123     schwarze  142:        ROFF_hylang,
                    143:        ROFF_hylen,
                    144:        ROFF_hym,
                    145:        ROFF_hypp,
                    146:        ROFF_hys,
1.2       schwarze  147:        ROFF_ie,
                    148:        ROFF_if,
1.1       schwarze  149:        ROFF_ig,
1.123     schwarze  150:        /* MAN_in; ignored in mdoc(7) */
                    151:        ROFF_index,
1.30      schwarze  152:        ROFF_it,
1.123     schwarze  153:        ROFF_itc,
                    154:        ROFF_IX,
                    155:        ROFF_kern,
                    156:        ROFF_kernafter,
                    157:        ROFF_kernbefore,
                    158:        ROFF_kernpair,
                    159:        ROFF_lc,
                    160:        ROFF_lc_ctype,
                    161:        ROFF_lds,
                    162:        ROFF_length,
                    163:        ROFF_letadj,
                    164:        ROFF_lf,
                    165:        ROFF_lg,
                    166:        ROFF_lhang,
                    167:        ROFF_linetabs,
                    168:        /* MAN_ll, MDOC_ll */
                    169:        ROFF_lnr,
                    170:        ROFF_lnrf,
                    171:        ROFF_lpfx,
                    172:        ROFF_ls,
                    173:        ROFF_lsm,
                    174:        ROFF_lt,
                    175:        ROFF_mc,
                    176:        ROFF_mediasize,
                    177:        ROFF_minss,
                    178:        ROFF_mk,
                    179:        ROFF_mso,
                    180:        /* MAN_na; ignored in mdoc(7) */
1.20      schwarze  181:        ROFF_ne,
1.123     schwarze  182:        /* MAN_nf; ignored in mdoc(7) */
1.20      schwarze  183:        ROFF_nh,
1.123     schwarze  184:        ROFF_nhychar,
                    185:        ROFF_nm,
                    186:        ROFF_nn,
                    187:        ROFF_nop,
1.14      schwarze  188:        ROFF_nr,
1.123     schwarze  189:        ROFF_nrf,
                    190:        ROFF_nroff,
1.31      schwarze  191:        ROFF_ns,
1.123     schwarze  192:        ROFF_nx,
                    193:        ROFF_open,
                    194:        ROFF_opena,
                    195:        ROFF_os,
                    196:        ROFF_output,
                    197:        ROFF_padj,
                    198:        ROFF_papersize,
                    199:        ROFF_pc,
                    200:        ROFF_pev,
                    201:        ROFF_pi,
                    202:        ROFF_PI,
1.101     schwarze  203:        ROFF_pl,
1.123     schwarze  204:        ROFF_pm,
                    205:        ROFF_pn,
                    206:        ROFF_pnr,
                    207:        ROFF_po,
1.31      schwarze  208:        ROFF_ps,
1.123     schwarze  209:        ROFF_psbb,
                    210:        ROFF_pshape,
                    211:        ROFF_pso,
                    212:        ROFF_ptr,
                    213:        ROFF_pvs,
                    214:        ROFF_rchar,
                    215:        ROFF_rd,
                    216:        ROFF_recursionlimit,
                    217:        ROFF_return,
                    218:        ROFF_rfschar,
                    219:        ROFF_rhang,
                    220:        ROFF_rj,
1.2       schwarze  221:        ROFF_rm,
1.123     schwarze  222:        ROFF_rn,
                    223:        ROFF_rnn,
1.76      schwarze  224:        ROFF_rr,
1.123     schwarze  225:        ROFF_rs,
                    226:        ROFF_rt,
                    227:        ROFF_schar,
                    228:        ROFF_sentchar,
                    229:        ROFF_shc,
                    230:        ROFF_shift,
                    231:        ROFF_sizes,
1.14      schwarze  232:        ROFF_so,
1.123     schwarze  233:        /* MAN_sp, MDOC_sp */
                    234:        ROFF_spacewidth,
                    235:        ROFF_special,
                    236:        ROFF_spreadwarn,
                    237:        ROFF_ss,
                    238:        ROFF_sty,
                    239:        ROFF_substring,
                    240:        ROFF_sv,
                    241:        ROFF_sy,
                    242:        ROFF_T_,
1.31      schwarze  243:        ROFF_ta,
1.123     schwarze  244:        ROFF_tc,
                    245:        ROFF_TE,
                    246:        ROFF_TH,
                    247:        ROFF_ti,
                    248:        ROFF_tkf,
                    249:        ROFF_tl,
                    250:        ROFF_tm,
                    251:        ROFF_tm1,
                    252:        ROFF_tmc,
1.2       schwarze  253:        ROFF_tr,
1.123     schwarze  254:        ROFF_track,
                    255:        ROFF_transchar,
                    256:        ROFF_trf,
                    257:        ROFF_trimat,
                    258:        ROFF_trin,
                    259:        ROFF_trnt,
                    260:        ROFF_troff,
1.27      schwarze  261:        ROFF_TS,
1.123     schwarze  262:        ROFF_uf,
                    263:        ROFF_ul,
                    264:        ROFF_unformat,
                    265:        ROFF_unwatch,
                    266:        ROFF_unwatchn,
                    267:        ROFF_vpt,
                    268:        ROFF_vs,
                    269:        ROFF_warn,
                    270:        ROFF_warnscale,
                    271:        ROFF_watch,
                    272:        ROFF_watchlength,
                    273:        ROFF_watchn,
                    274:        ROFF_wh,
                    275:        ROFF_while,
                    276:        ROFF_write,
                    277:        ROFF_writec,
                    278:        ROFF_writem,
                    279:        ROFF_xflag,
1.2       schwarze  280:        ROFF_cblock,
1.16      schwarze  281:        ROFF_USERDEF,
1.1       schwarze  282:        ROFF_MAX
                    283: };
                    284:
1.41      schwarze  285: /*
1.42      schwarze  286:  * An incredibly-simple string buffer.
                    287:  */
1.8       schwarze  288: struct roffstr {
1.42      schwarze  289:        char            *p; /* nil-terminated buffer */
                    290:        size_t           sz; /* saved strlen(p) */
                    291: };
                    292:
                    293: /*
                    294:  * A key-value roffstr pair as part of a singly-linked list.
                    295:  */
                    296: struct roffkv {
                    297:        struct roffstr   key;
                    298:        struct roffstr   val;
                    299:        struct roffkv   *next; /* next in list */
1.8       schwarze  300: };
                    301:
1.52      schwarze  302: /*
                    303:  * A single number register as part of a singly-linked list.
                    304:  */
                    305: struct roffreg {
                    306:        struct roffstr   key;
1.53      schwarze  307:        int              val;
1.52      schwarze  308:        struct roffreg  *next;
                    309: };
                    310:
1.1       schwarze  311: struct roff {
1.35      schwarze  312:        struct mparse   *parse; /* parse point */
1.109     schwarze  313:        const struct mchars *mchars; /* character table */
1.1       schwarze  314:        struct roffnode *last; /* leaf of stack */
1.96      schwarze  315:        int             *rstack; /* stack of inverted `ie' values */
1.52      schwarze  316:        struct roffreg  *regtab; /* number registers */
1.42      schwarze  317:        struct roffkv   *strtab; /* user-defined strings & macros */
                    318:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                    319:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.16      schwarze  320:        const char      *current_string; /* value of last called user macro */
1.27      schwarze  321:        struct tbl_node *first_tbl; /* first table parsed */
                    322:        struct tbl_node *last_tbl; /* last table parsed */
                    323:        struct tbl_node *tbl; /* current table being parsed */
1.32      schwarze  324:        struct eqn_node *last_eqn; /* last equation parsed */
                    325:        struct eqn_node *first_eqn; /* first equation parsed */
                    326:        struct eqn_node *eqn; /* current equation being parsed */
1.102     schwarze  327:        int              eqn_inline; /* current equation is inline */
1.96      schwarze  328:        int              options; /* parse options */
                    329:        int              rstacksz; /* current size limit of rstack */
                    330:        int              rstackpos; /* position in rstack */
1.99      schwarze  331:        int              format; /* current file in mdoc or man format */
1.96      schwarze  332:        char             control; /* control character */
1.1       schwarze  333: };
                    334:
                    335: struct roffnode {
                    336:        enum rofft       tok; /* type of node */
                    337:        struct roffnode *parent; /* up one in stack */
                    338:        int              line; /* parse line */
                    339:        int              col; /* parse col */
1.16      schwarze  340:        char            *name; /* node name, e.g. macro name */
1.2       schwarze  341:        char            *end; /* end-rules: custom token */
                    342:        int              endspan; /* end-rules: next-line or infty */
1.71      schwarze  343:        int              rule; /* current evaluation rule */
1.1       schwarze  344: };
                    345:
                    346: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
                    347:                         enum rofft tok, /* tok of macro */ \
1.110     schwarze  348:                         struct buf *buf, /* input buffer */ \
1.1       schwarze  349:                         int ln, /* parse line */ \
1.2       schwarze  350:                         int ppos, /* original pos in buffer */ \
                    351:                         int pos, /* current pos in buffer */ \
                    352:                         int *offs /* reset offset of buffer data */
1.1       schwarze  353:
                    354: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    355:
                    356: struct roffmac {
                    357:        const char      *name; /* macro name */
1.2       schwarze  358:        roffproc         proc; /* process new macro */
                    359:        roffproc         text; /* process as child text of macro */
                    360:        roffproc         sub; /* process as child of macro */
                    361:        int              flags;
                    362: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.3       schwarze  363:        struct roffmac  *next;
1.1       schwarze  364: };
                    365:
1.37      schwarze  366: struct predef {
                    367:        const char      *name; /* predefined input name */
                    368:        const char      *str; /* replacement symbol */
                    369: };
                    370:
                    371: #define        PREDEF(__name, __str) \
                    372:        { (__name), (__str) },
                    373:
1.42      schwarze  374: static enum rofft       roffhash_find(const char *, size_t);
                    375: static void             roffhash_init(void);
                    376: static void             roffnode_cleanscope(struct roff *);
                    377: static void             roffnode_pop(struct roff *);
                    378: static void             roffnode_push(struct roff *, enum rofft,
                    379:                                const char *, int, int);
1.2       schwarze  380: static enum rofferr     roff_block(ROFF_ARGS);
                    381: static enum rofferr     roff_block_text(ROFF_ARGS);
                    382: static enum rofferr     roff_block_sub(ROFF_ARGS);
1.123     schwarze  383: static enum rofferr     roff_brp(ROFF_ARGS);
1.2       schwarze  384: static enum rofferr     roff_cblock(ROFF_ARGS);
1.48      schwarze  385: static enum rofferr     roff_cc(ROFF_ARGS);
1.68      schwarze  386: static void             roff_ccond(struct roff *, int, int);
1.2       schwarze  387: static enum rofferr     roff_cond(ROFF_ARGS);
                    388: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    389: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  390: static enum rofferr     roff_ds(ROFF_ARGS);
1.110     schwarze  391: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.106     schwarze  392: static int              roff_evalcond(struct roff *r, int,
                    393:                                const char *, int *);
                    394: static int              roff_evalnum(struct roff *, int,
                    395:                                const char *, int *, int *, int);
                    396: static int              roff_evalpar(struct roff *, int,
                    397:                                const char *, int *, int *);
1.71      schwarze  398: static int              roff_evalstrcond(const char *, int *);
1.42      schwarze  399: static void             roff_free1(struct roff *);
1.52      schwarze  400: static void             roff_freereg(struct roffreg *);
1.42      schwarze  401: static void             roff_freestr(struct roffkv *);
1.85      schwarze  402: static size_t           roff_getname(struct roff *, char **, int, int);
1.56      schwarze  403: static int              roff_getnum(const char *, int *, int *);
                    404: static int              roff_getop(const char *, int *, char *);
1.53      schwarze  405: static int              roff_getregn(const struct roff *,
                    406:                                const char *, size_t);
1.65      schwarze  407: static int              roff_getregro(const char *name);
1.80      schwarze  408: static const char      *roff_getstrn(const struct roff *,
1.8       schwarze  409:                                const char *, size_t);
1.123     schwarze  410: static enum rofferr     roff_insec(ROFF_ARGS);
1.51      schwarze  411: static enum rofferr     roff_it(ROFF_ARGS);
1.21      schwarze  412: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.6       schwarze  413: static enum rofferr     roff_nr(ROFF_ARGS);
1.87      schwarze  414: static enum rofft       roff_parse(struct roff *, char *, int *,
                    415:                                int, int);
1.110     schwarze  416: static enum rofferr     roff_parsetext(struct buf *, int, int *);
                    417: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.29      schwarze  418: static enum rofferr     roff_rm(ROFF_ARGS);
1.76      schwarze  419: static enum rofferr     roff_rr(ROFF_ARGS);
1.8       schwarze  420: static void             roff_setstr(struct roff *,
1.16      schwarze  421:                                const char *, const char *, int);
1.80      schwarze  422: static void             roff_setstrn(struct roffkv **, const char *,
1.42      schwarze  423:                                size_t, const char *, size_t, int);
1.14      schwarze  424: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  425: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  426: static enum rofferr     roff_Dd(ROFF_ARGS);
                    427: static enum rofferr     roff_TH(ROFF_ARGS);
1.27      schwarze  428: static enum rofferr     roff_TE(ROFF_ARGS);
                    429: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  430: static enum rofferr     roff_EQ(ROFF_ARGS);
                    431: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  432: static enum rofferr     roff_T_(ROFF_ARGS);
1.123     schwarze  433: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.16      schwarze  434: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  435:
1.42      schwarze  436: /* See roffhash_find() */
1.3       schwarze  437:
                    438: #define        ASCII_HI         126
                    439: #define        ASCII_LO         33
                    440: #define        HASHWIDTH       (ASCII_HI - ASCII_LO + 1)
                    441:
                    442: static struct roffmac  *hash[HASHWIDTH];
                    443:
                    444: static struct roffmac   roffs[ROFF_MAX] = {
1.123     schwarze  445:        { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.21      schwarze  446:        { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  447:        { "af", roff_line_ignore, NULL, NULL, 0, NULL },
                    448:        { "aln", roff_unsupp, NULL, NULL, 0, NULL },
                    449:        { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.3       schwarze  450:        { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  451:        { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.3       schwarze  452:        { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  453:        { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.66      schwarze  454:        { "as", roff_ds, NULL, NULL, 0, NULL },
1.123     schwarze  455:        { "as1", roff_ds, NULL, NULL, 0, NULL },
                    456:        { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
                    457:        { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
                    458:        { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
                    459:        { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
                    460:        { "blm", roff_unsupp, NULL, NULL, 0, NULL },
                    461:        { "box", roff_unsupp, NULL, NULL, 0, NULL },
                    462:        { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
                    463:        { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
                    464:        { "BP", roff_unsupp, NULL, NULL, 0, NULL },
                    465:        { "break", roff_unsupp, NULL, NULL, 0, NULL },
                    466:        { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
                    467:        { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
                    468:        { "brp", roff_brp, NULL, NULL, 0, NULL },
                    469:        { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
                    470:        { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.48      schwarze  471:        { "cc", roff_cc, NULL, NULL, 0, NULL },
1.67      schwarze  472:        { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  473:        { "cf", roff_insec, NULL, NULL, 0, NULL },
                    474:        { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
                    475:        { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
                    476:        { "char", roff_unsupp, NULL, NULL, 0, NULL },
                    477:        { "chop", roff_unsupp, NULL, NULL, 0, NULL },
                    478:        { "class", roff_line_ignore, NULL, NULL, 0, NULL },
                    479:        { "close", roff_insec, NULL, NULL, 0, NULL },
                    480:        { "CL", roff_unsupp, NULL, NULL, 0, NULL },
                    481:        { "color", roff_line_ignore, NULL, NULL, 0, NULL },
                    482:        { "composite", roff_unsupp, NULL, NULL, 0, NULL },
                    483:        { "continue", roff_unsupp, NULL, NULL, 0, NULL },
                    484:        { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
                    485:        { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
                    486:        { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
                    487:        { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
                    488:        { "da", roff_unsupp, NULL, NULL, 0, NULL },
                    489:        { "dch", roff_unsupp, NULL, NULL, 0, NULL },
                    490:        { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.3       schwarze  491:        { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  492:        { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    493:        { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.3       schwarze  494:        { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  495:        { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    496:        { "device", roff_unsupp, NULL, NULL, 0, NULL },
                    497:        { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
                    498:        { "di", roff_unsupp, NULL, NULL, 0, NULL },
                    499:        { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.7       schwarze  500:        { "ds", roff_ds, NULL, NULL, 0, NULL },
1.123     schwarze  501:        { "ds1", roff_ds, NULL, NULL, 0, NULL },
                    502:        { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
                    503:        { "dt", roff_unsupp, NULL, NULL, 0, NULL },
                    504:        { "ec", roff_unsupp, NULL, NULL, 0, NULL },
                    505:        { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
                    506:        { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.3       schwarze  507:        { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.123     schwarze  508:        { "em", roff_unsupp, NULL, NULL, 0, NULL },
                    509:        { "EN", roff_EN, NULL, NULL, 0, NULL },
                    510:        { "eo", roff_unsupp, NULL, NULL, 0, NULL },
                    511:        { "EP", roff_unsupp, NULL, NULL, 0, NULL },
                    512:        { "EQ", roff_EQ, NULL, NULL, 0, NULL },
                    513:        { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
                    514:        { "ev", roff_unsupp, NULL, NULL, 0, NULL },
                    515:        { "evc", roff_unsupp, NULL, NULL, 0, NULL },
                    516:        { "ex", roff_unsupp, NULL, NULL, 0, NULL },
                    517:        { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.58      schwarze  518:        { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  519:        { "fc", roff_unsupp, NULL, NULL, 0, NULL },
                    520:        { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
                    521:        { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
                    522:        { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
                    523:        { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
                    524:        { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
                    525:        { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
                    526:        { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
                    527:        { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
                    528:        { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
                    529:        { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
                    530:        { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
                    531:        { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
                    532:        { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
                    533:        { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
                    534:        { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
                    535:        { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
                    536:        { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
                    537:        { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
                    538:        { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
                    539:        { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
                    540:        { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
                    541:        { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
                    542:        { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.59      schwarze  543:        { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.21      schwarze  544:        { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  545:        { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
                    546:        { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
                    547:        { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
                    548:        { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
                    549:        { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.3       schwarze  550:        { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    551:        { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    552:        { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  553:        { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.51      schwarze  554:        { "it", roff_it, NULL, NULL, 0, NULL },
1.123     schwarze  555:        { "itc", roff_unsupp, NULL, NULL, 0, NULL },
                    556:        { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
                    557:        { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
                    558:        { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
                    559:        { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
                    560:        { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
                    561:        { "lc", roff_unsupp, NULL, NULL, 0, NULL },
                    562:        { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
                    563:        { "lds", roff_unsupp, NULL, NULL, 0, NULL },
                    564:        { "length", roff_unsupp, NULL, NULL, 0, NULL },
                    565:        { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
                    566:        { "lf", roff_insec, NULL, NULL, 0, NULL },
                    567:        { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
                    568:        { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
                    569:        { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
                    570:        { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
                    571:        { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
                    572:        { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
                    573:        { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
                    574:        { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
                    575:        { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
                    576:        { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
                    577:        { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
                    578:        { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
                    579:        { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
                    580:        { "mso", roff_insec, NULL, NULL, 0, NULL },
1.21      schwarze  581:        { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
                    582:        { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  583:        { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
                    584:        { "nm", roff_unsupp, NULL, NULL, 0, NULL },
                    585:        { "nn", roff_unsupp, NULL, NULL, 0, NULL },
                    586:        { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.14      schwarze  587:        { "nr", roff_nr, NULL, NULL, 0, NULL },
1.123     schwarze  588:        { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
                    589:        { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.31      schwarze  590:        { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  591:        { "nx", roff_insec, NULL, NULL, 0, NULL },
                    592:        { "open", roff_insec, NULL, NULL, 0, NULL },
                    593:        { "opena", roff_insec, NULL, NULL, 0, NULL },
                    594:        { "os", roff_line_ignore, NULL, NULL, 0, NULL },
                    595:        { "output", roff_unsupp, NULL, NULL, 0, NULL },
                    596:        { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
                    597:        { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
                    598:        { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
                    599:        { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
                    600:        { "pi", roff_insec, NULL, NULL, 0, NULL },
                    601:        { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.101     schwarze  602:        { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  603:        { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
                    604:        { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
                    605:        { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
                    606:        { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.31      schwarze  607:        { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  608:        { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
                    609:        { "pshape", roff_line_ignore, NULL, NULL, 0, NULL },
                    610:        { "pso", roff_insec, NULL, NULL, 0, NULL },
                    611:        { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
                    612:        { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
                    613:        { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
                    614:        { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
                    615:        { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
                    616:        { "return", roff_unsupp, NULL, NULL, 0, NULL },
                    617:        { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
                    618:        { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
                    619:        { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.29      schwarze  620:        { "rm", roff_rm, NULL, NULL, 0, NULL },
1.123     schwarze  621:        { "rn", roff_unsupp, NULL, NULL, 0, NULL },
                    622:        { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.76      schwarze  623:        { "rr", roff_rr, NULL, NULL, 0, NULL },
1.123     schwarze  624:        { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
                    625:        { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
                    626:        { "schar", roff_unsupp, NULL, NULL, 0, NULL },
                    627:        { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
                    628:        { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
                    629:        { "shift", roff_unsupp, NULL, NULL, 0, NULL },
                    630:        { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.14      schwarze  631:        { "so", roff_so, NULL, NULL, 0, NULL },
1.123     schwarze  632:        { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
                    633:        { "special", roff_line_ignore, NULL, NULL, 0, NULL },
                    634:        { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
                    635:        { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
                    636:        { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
                    637:        { "substring", roff_unsupp, NULL, NULL, 0, NULL },
                    638:        { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
                    639:        { "sy", roff_insec, NULL, NULL, 0, NULL },
                    640:        { "T&", roff_T_, NULL, NULL, 0, NULL },
1.31      schwarze  641:        { "ta", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  642:        { "tc", roff_unsupp, NULL, NULL, 0, NULL },
                    643:        { "TE", roff_TE, NULL, NULL, 0, NULL },
                    644:        { "TH", roff_TH, NULL, NULL, 0, NULL },
                    645:        { "ti", roff_line_ignore, NULL, NULL, 0, NULL },
                    646:        { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
                    647:        { "tl", roff_unsupp, NULL, NULL, 0, NULL },
                    648:        { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
                    649:        { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
                    650:        { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.42      schwarze  651:        { "tr", roff_tr, NULL, NULL, 0, NULL },
1.123     schwarze  652:        { "track", roff_line_ignore, NULL, NULL, 0, NULL },
                    653:        { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
                    654:        { "trf", roff_insec, NULL, NULL, 0, NULL },
                    655:        { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
                    656:        { "trin", roff_unsupp, NULL, NULL, 0, NULL },
                    657:        { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
                    658:        { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.27      schwarze  659:        { "TS", roff_TS, NULL, NULL, 0, NULL },
1.123     schwarze  660:        { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
                    661:        { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
                    662:        { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
                    663:        { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
                    664:        { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
                    665:        { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
                    666:        { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
                    667:        { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
                    668:        { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
                    669:        { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
                    670:        { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
                    671:        { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
                    672:        { "wh", roff_unsupp, NULL, NULL, 0, NULL },
                    673:        { "while", roff_unsupp, NULL, NULL, 0, NULL },
                    674:        { "write", roff_insec, NULL, NULL, 0, NULL },
                    675:        { "writec", roff_insec, NULL, NULL, 0, NULL },
                    676:        { "writem", roff_insec, NULL, NULL, 0, NULL },
                    677:        { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.3       schwarze  678:        { ".", roff_cblock, NULL, NULL, 0, NULL },
1.16      schwarze  679:        { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.1       schwarze  680: };
                    681:
1.73      schwarze  682: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47      schwarze  683: const  char *const __mdoc_reserved[] = {
                    684:        "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
                    685:        "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
                    686:        "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
                    687:        "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73      schwarze  688:        "Dt", "Dv", "Dx", "D1",
                    689:        "Ec", "Ed", "Ef", "Ek", "El", "Em",
                    690:        "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47      schwarze  691:        "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73      schwarze  692:        "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
                    693:        "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47      schwarze  694:        "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73      schwarze  695:        "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
                    696:        "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
                    697:        "Sc", "Sh", "Sm", "So", "Sq",
1.47      schwarze  698:        "Ss", "St", "Sx", "Sy",
                    699:        "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73      schwarze  700:        "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47      schwarze  701:        "%P", "%Q", "%R", "%T", "%U", "%V",
                    702:        NULL
                    703: };
                    704:
1.73      schwarze  705: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47      schwarze  706: const  char *const __man_reserved[] = {
1.73      schwarze  707:        "AT", "B", "BI", "BR", "DT",
                    708:        "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
                    709:        "LP", "OP", "P", "PD", "PP",
                    710:        "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
                    711:        "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47      schwarze  712:        NULL
                    713: };
                    714:
1.37      schwarze  715: /* Array of injected predefined strings. */
                    716: #define        PREDEFS_MAX      38
                    717: static const struct predef predefs[PREDEFS_MAX] = {
                    718: #include "predefs.in"
                    719: };
                    720:
1.42      schwarze  721: /* See roffhash_find() */
1.3       schwarze  722: #define        ROFF_HASH(p)    (p[0] - ASCII_LO)
                    723:
1.51      schwarze  724: static int      roffit_lines;  /* number of lines to delay */
                    725: static char    *roffit_macro;  /* nil-terminated macro line */
                    726:
1.80      schwarze  727:
1.3       schwarze  728: static void
1.42      schwarze  729: roffhash_init(void)
1.3       schwarze  730: {
                    731:        struct roffmac   *n;
                    732:        int               buc, i;
                    733:
1.16      schwarze  734:        for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.3       schwarze  735:                assert(roffs[i].name[0] >= ASCII_LO);
                    736:                assert(roffs[i].name[0] <= ASCII_HI);
                    737:
                    738:                buc = ROFF_HASH(roffs[i].name);
                    739:
                    740:                if (NULL != (n = hash[buc])) {
                    741:                        for ( ; n->next; n = n->next)
                    742:                                /* Do nothing. */ ;
                    743:                        n->next = &roffs[i];
                    744:                } else
                    745:                        hash[buc] = &roffs[i];
                    746:        }
                    747: }
                    748:
1.1       schwarze  749: /*
                    750:  * Look up a roff token by its name.  Returns ROFF_MAX if no macro by
                    751:  * the nil-terminated string name could be found.
                    752:  */
                    753: static enum rofft
1.42      schwarze  754: roffhash_find(const char *p, size_t s)
1.1       schwarze  755: {
1.3       schwarze  756:        int              buc;
                    757:        struct roffmac  *n;
1.1       schwarze  758:
1.3       schwarze  759:        /*
                    760:         * libroff has an extremely simple hashtable, for the time
                    761:         * being, which simply keys on the first character, which must
                    762:         * be printable, then walks a chain.  It works well enough until
                    763:         * optimised.
                    764:         */
                    765:
                    766:        if (p[0] < ASCII_LO || p[0] > ASCII_HI)
                    767:                return(ROFF_MAX);
                    768:
                    769:        buc = ROFF_HASH(p);
                    770:
                    771:        if (NULL == (n = hash[buc]))
                    772:                return(ROFF_MAX);
                    773:        for ( ; n; n = n->next)
1.16      schwarze  774:                if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.3       schwarze  775:                        return((enum rofft)(n - roffs));
1.1       schwarze  776:
                    777:        return(ROFF_MAX);
                    778: }
                    779:
                    780: /*
                    781:  * Pop the current node off of the stack of roff instructions currently
                    782:  * pending.
                    783:  */
                    784: static void
                    785: roffnode_pop(struct roff *r)
                    786: {
                    787:        struct roffnode *p;
                    788:
1.2       schwarze  789:        assert(r->last);
1.80      schwarze  790:        p = r->last;
1.2       schwarze  791:
                    792:        r->last = r->last->parent;
1.16      schwarze  793:        free(p->name);
                    794:        free(p->end);
1.1       schwarze  795:        free(p);
                    796: }
                    797:
                    798: /*
                    799:  * Push a roff node onto the instruction stack.  This must later be
                    800:  * removed with roffnode_pop().
                    801:  */
1.11      schwarze  802: static void
1.16      schwarze  803: roffnode_push(struct roff *r, enum rofft tok, const char *name,
                    804:                int line, int col)
1.1       schwarze  805: {
                    806:        struct roffnode *p;
                    807:
1.11      schwarze  808:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  809:        p->tok = tok;
1.16      schwarze  810:        if (name)
                    811:                p->name = mandoc_strdup(name);
1.1       schwarze  812:        p->parent = r->last;
                    813:        p->line = line;
                    814:        p->col = col;
1.71      schwarze  815:        p->rule = p->parent ? p->parent->rule : 0;
1.1       schwarze  816:
                    817:        r->last = p;
                    818: }
                    819:
                    820: static void
                    821: roff_free1(struct roff *r)
                    822: {
1.49      schwarze  823:        struct tbl_node *tbl;
1.32      schwarze  824:        struct eqn_node *e;
1.42      schwarze  825:        int              i;
1.27      schwarze  826:
1.49      schwarze  827:        while (NULL != (tbl = r->first_tbl)) {
                    828:                r->first_tbl = tbl->next;
                    829:                tbl_free(tbl);
1.27      schwarze  830:        }
                    831:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  832:
1.32      schwarze  833:        while (NULL != (e = r->first_eqn)) {
                    834:                r->first_eqn = e->next;
                    835:                eqn_free(e);
                    836:        }
                    837:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    838:
1.1       schwarze  839:        while (r->last)
                    840:                roffnode_pop(r);
1.27      schwarze  841:
1.96      schwarze  842:        free (r->rstack);
                    843:        r->rstack = NULL;
                    844:        r->rstacksz = 0;
                    845:        r->rstackpos = -1;
                    846:
                    847:        roff_freereg(r->regtab);
                    848:        r->regtab = NULL;
                    849:
1.42      schwarze  850:        roff_freestr(r->strtab);
                    851:        roff_freestr(r->xmbtab);
                    852:        r->strtab = r->xmbtab = NULL;
                    853:
                    854:        if (r->xtab)
                    855:                for (i = 0; i < 128; i++)
                    856:                        free(r->xtab[i].p);
                    857:        free(r->xtab);
                    858:        r->xtab = NULL;
1.1       schwarze  859: }
                    860:
                    861: void
                    862: roff_reset(struct roff *r)
                    863: {
                    864:
                    865:        roff_free1(r);
1.99      schwarze  866:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.48      schwarze  867:        r->control = 0;
1.1       schwarze  868: }
                    869:
                    870: void
                    871: roff_free(struct roff *r)
                    872: {
                    873:
                    874:        roff_free1(r);
                    875:        free(r);
                    876: }
                    877:
                    878: struct roff *
1.109     schwarze  879: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.1       schwarze  880: {
                    881:        struct roff     *r;
                    882:
1.11      schwarze  883:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  884:        r->parse = parse;
1.109     schwarze  885:        r->mchars = mchars;
1.72      schwarze  886:        r->options = options;
1.99      schwarze  887:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2       schwarze  888:        r->rstackpos = -1;
1.80      schwarze  889:
1.42      schwarze  890:        roffhash_init();
1.37      schwarze  891:
1.1       schwarze  892:        return(r);
                    893: }
                    894:
1.8       schwarze  895: /*
1.79      schwarze  896:  * In the current line, expand escape sequences that tend to get
                    897:  * used in numerical expressions and conditional requests.
                    898:  * Also check the syntax of the remaining escape sequences.
1.8       schwarze  899:  */
1.45      schwarze  900: static enum rofferr
1.110     schwarze  901: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8       schwarze  902: {
1.81      schwarze  903:        char             ubuf[24]; /* buffer to print the number */
1.78      schwarze  904:        const char      *start; /* start of the string to process */
1.82      schwarze  905:        char            *stesc; /* start of an escape sequence ('\\') */
1.23      schwarze  906:        const char      *stnam; /* start of the name, after "[(*" */
                    907:        const char      *cp;    /* end of the name, e.g. before ']' */
                    908:        const char      *res;   /* the string to be substituted */
1.110     schwarze  909:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.53      schwarze  910:        size_t           maxl;  /* expected length of the escape name */
                    911:        size_t           naml;  /* actual length of the escape name */
1.109     schwarze  912:        enum mandoc_esc  esc;   /* type of the escape sequence */
                    913:        int              inaml; /* length returned from mandoc_escape() */
1.53      schwarze  914:        int              expand_count;  /* to avoid infinite loops */
1.79      schwarze  915:        int              npos;  /* position in numeric expression */
1.91      schwarze  916:        int              arg_complete; /* argument not interrupted by eol */
1.79      schwarze  917:        char             term;  /* character terminating the escape */
1.8       schwarze  918:
1.43      schwarze  919:        expand_count = 0;
1.110     schwarze  920:        start = buf->buf + pos;
1.78      schwarze  921:        stesc = strchr(start, '\0') - 1;
                    922:        while (stesc-- > start) {
1.43      schwarze  923:
1.78      schwarze  924:                /* Search backwards for the next backslash. */
                    925:
1.110     schwarze  926:                if (*stesc != '\\')
1.78      schwarze  927:                        continue;
                    928:
                    929:                /* If it is escaped, skip it. */
                    930:
                    931:                for (cp = stesc - 1; cp >= start; cp--)
1.110     schwarze  932:                        if (*cp != '\\')
1.78      schwarze  933:                                break;
                    934:
1.110     schwarze  935:                if ((stesc - cp) % 2 == 0) {
1.82      schwarze  936:                        stesc = (char *)cp;
1.78      schwarze  937:                        continue;
                    938:                }
1.23      schwarze  939:
1.79      schwarze  940:                /* Decide whether to expand or to check only. */
1.23      schwarze  941:
1.79      schwarze  942:                term = '\0';
1.78      schwarze  943:                cp = stesc + 1;
1.53      schwarze  944:                switch (*cp) {
1.80      schwarze  945:                case '*':
1.53      schwarze  946:                        res = NULL;
                    947:                        break;
1.80      schwarze  948:                case 'B':
1.79      schwarze  949:                        /* FALLTHROUGH */
1.80      schwarze  950:                case 'w':
1.79      schwarze  951:                        term = cp[1];
                    952:                        /* FALLTHROUGH */
1.80      schwarze  953:                case 'n':
1.53      schwarze  954:                        res = ubuf;
                    955:                        break;
                    956:                default:
1.109     schwarze  957:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                    958:                        if (esc == ESCAPE_ERROR ||
                    959:                            (esc == ESCAPE_SPECIAL &&
                    960:                             mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.92      schwarze  961:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110     schwarze  962:                                    r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze  963:                                    "%.*s", (int)(cp - stesc), stesc);
1.78      schwarze  964:                        continue;
1.42      schwarze  965:                }
                    966:
1.78      schwarze  967:                if (EXPAND_LIMIT < ++expand_count) {
                    968:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110     schwarze  969:                            ln, (int)(stesc - buf->buf), NULL);
1.78      schwarze  970:                        return(ROFF_IGN);
                    971:                }
1.23      schwarze  972:
                    973:                /*
                    974:                 * The third character decides the length
1.53      schwarze  975:                 * of the name of the string or register.
1.23      schwarze  976:                 * Save a pointer to the name.
                    977:                 */
                    978:
1.110     schwarze  979:                if (term == '\0') {
1.79      schwarze  980:                        switch (*++cp) {
1.80      schwarze  981:                        case '\0':
1.79      schwarze  982:                                maxl = 0;
                    983:                                break;
1.80      schwarze  984:                        case '(':
1.79      schwarze  985:                                cp++;
                    986:                                maxl = 2;
                    987:                                break;
1.80      schwarze  988:                        case '[':
1.79      schwarze  989:                                cp++;
                    990:                                term = ']';
                    991:                                maxl = 0;
                    992:                                break;
                    993:                        default:
                    994:                                maxl = 1;
                    995:                                break;
                    996:                        }
                    997:                } else {
                    998:                        cp += 2;
1.8       schwarze  999:                        maxl = 0;
                   1000:                }
1.23      schwarze 1001:                stnam = cp;
1.8       schwarze 1002:
1.23      schwarze 1003:                /* Advance to the end of the name. */
1.8       schwarze 1004:
1.91      schwarze 1005:                arg_complete = 1;
1.110     schwarze 1006:                for (naml = 0; maxl == 0 || naml < maxl; naml++, cp++) {
                   1007:                        if (*cp == '\0') {
1.92      schwarze 1008:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 1009:                                    ln, (int)(stesc - buf->buf), stesc);
1.91      schwarze 1010:                                arg_complete = 0;
1.79      schwarze 1011:                                break;
1.42      schwarze 1012:                        }
1.110     schwarze 1013:                        if (maxl == 0 && *cp == term) {
1.79      schwarze 1014:                                cp++;
1.8       schwarze 1015:                                break;
1.79      schwarze 1016:                        }
1.8       schwarze 1017:                }
                   1018:
1.23      schwarze 1019:                /*
                   1020:                 * Retrieve the replacement string; if it is
                   1021:                 * undefined, resume searching for escapes.
                   1022:                 */
                   1023:
1.79      schwarze 1024:                switch (stesc[1]) {
1.80      schwarze 1025:                case '*':
1.91      schwarze 1026:                        if (arg_complete)
                   1027:                                res = roff_getstrn(r, stnam, naml);
1.79      schwarze 1028:                        break;
1.80      schwarze 1029:                case 'B':
1.79      schwarze 1030:                        npos = 0;
1.91      schwarze 1031:                        ubuf[0] = arg_complete &&
1.106     schwarze 1032:                            roff_evalnum(r, ln, stnam, &npos, NULL, 0) &&
1.91      schwarze 1033:                            stnam + npos + 1 == cp ? '1' : '0';
1.79      schwarze 1034:                        ubuf[1] = '\0';
                   1035:                        break;
1.80      schwarze 1036:                case 'n':
1.91      schwarze 1037:                        if (arg_complete)
                   1038:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1039:                                    roff_getregn(r, stnam, naml));
                   1040:                        else
                   1041:                                ubuf[0] = '\0';
1.79      schwarze 1042:                        break;
1.80      schwarze 1043:                case 'w':
1.91      schwarze 1044:                        /* use even incomplete args */
1.81      schwarze 1045:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79      schwarze 1046:                            24 * (int)naml);
                   1047:                        break;
                   1048:                }
1.8       schwarze 1049:
1.110     schwarze 1050:                if (res == NULL) {
1.92      schwarze 1051:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1052:                            r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1053:                            "%.*s", (int)naml, stnam);
1.37      schwarze 1054:                        res = "";
1.118     schwarze 1055:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1056:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1057:                            ln, (int)(stesc - buf->buf), NULL);
                   1058:                        return(ROFF_IGN);
1.8       schwarze 1059:                }
                   1060:
1.23      schwarze 1061:                /* Replace the escape sequence by the string. */
                   1062:
1.82      schwarze 1063:                *stesc = '\0';
1.110     schwarze 1064:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1065:                    buf->buf, res, cp) + 1;
1.8       schwarze 1066:
1.78      schwarze 1067:                /* Prepare for the next replacement. */
1.8       schwarze 1068:
1.78      schwarze 1069:                start = nbuf + pos;
1.110     schwarze 1070:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1071:                free(buf->buf);
                   1072:                buf->buf = nbuf;
1.42      schwarze 1073:        }
1.45      schwarze 1074:        return(ROFF_CONT);
1.42      schwarze 1075: }
                   1076:
                   1077: /*
1.51      schwarze 1078:  * Process text streams:
                   1079:  * Convert all breakable hyphens into ASCII_HYPH.
                   1080:  * Decrement and spring input line trap.
1.42      schwarze 1081:  */
                   1082: static enum rofferr
1.110     schwarze 1083: roff_parsetext(struct buf *buf, int pos, int *offs)
1.42      schwarze 1084: {
                   1085:        size_t           sz;
                   1086:        const char      *start;
1.51      schwarze 1087:        char            *p;
                   1088:        int              isz;
1.42      schwarze 1089:        enum mandoc_esc  esc;
                   1090:
1.110     schwarze 1091:        start = p = buf->buf + pos;
1.42      schwarze 1092:
1.110     schwarze 1093:        while (*p != '\0') {
1.42      schwarze 1094:                sz = strcspn(p, "-\\");
                   1095:                p += sz;
                   1096:
1.110     schwarze 1097:                if (*p == '\0')
1.42      schwarze 1098:                        break;
                   1099:
1.110     schwarze 1100:                if (*p == '\\') {
1.42      schwarze 1101:                        /* Skip over escapes. */
                   1102:                        p++;
1.62      schwarze 1103:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110     schwarze 1104:                        if (esc == ESCAPE_ERROR)
1.42      schwarze 1105:                                break;
                   1106:                        continue;
                   1107:                } else if (p == start) {
                   1108:                        p++;
                   1109:                        continue;
                   1110:                }
                   1111:
1.44      schwarze 1112:                if (isalpha((unsigned char)p[-1]) &&
                   1113:                    isalpha((unsigned char)p[1]))
1.42      schwarze 1114:                        *p = ASCII_HYPH;
                   1115:                p++;
1.8       schwarze 1116:        }
                   1117:
1.51      schwarze 1118:        /* Spring the input line trap. */
1.110     schwarze 1119:        if (roffit_lines == 1) {
                   1120:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1121:                free(buf->buf);
                   1122:                buf->buf = p;
                   1123:                buf->sz = isz + 1;
1.51      schwarze 1124:                *offs = 0;
                   1125:                free(roffit_macro);
                   1126:                roffit_lines = 0;
                   1127:                return(ROFF_REPARSE);
1.110     schwarze 1128:        } else if (roffit_lines > 1)
1.51      schwarze 1129:                --roffit_lines;
1.42      schwarze 1130:        return(ROFF_CONT);
1.8       schwarze 1131: }
                   1132:
1.1       schwarze 1133: enum rofferr
1.110     schwarze 1134: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1       schwarze 1135: {
                   1136:        enum rofft       t;
1.27      schwarze 1137:        enum rofferr     e;
1.110     schwarze 1138:        int              pos;   /* parse point */
1.115     schwarze 1139:        int              spos;  /* saved parse point for messages */
1.110     schwarze 1140:        int              ppos;  /* original offset in buf->buf */
                   1141:        int              ctl;   /* macro line (boolean) */
                   1142:
                   1143:        ppos = pos = *offs;
1.1       schwarze 1144:
1.102     schwarze 1145:        /* Handle in-line equation delimiters. */
                   1146:
1.108     schwarze 1147:        if (r->tbl == NULL &&
                   1148:            r->last_eqn != NULL && r->last_eqn->delim &&
1.102     schwarze 1149:            (r->eqn == NULL || r->eqn_inline)) {
1.110     schwarze 1150:                e = roff_eqndelim(r, buf, pos);
1.102     schwarze 1151:                if (e == ROFF_REPARSE)
                   1152:                        return(e);
                   1153:                assert(e == ROFF_CONT);
                   1154:        }
                   1155:
                   1156:        /* Expand some escape sequences. */
1.8       schwarze 1157:
1.110     schwarze 1158:        e = roff_res(r, buf, ln, pos);
                   1159:        if (e == ROFF_IGN)
1.45      schwarze 1160:                return(e);
1.110     schwarze 1161:        assert(e == ROFF_CONT);
1.8       schwarze 1162:
1.110     schwarze 1163:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.35      schwarze 1164:
1.8       schwarze 1165:        /*
1.2       schwarze 1166:         * First, if a scope is open and we're not a macro, pass the
1.124   ! schwarze 1167:         * text through the macro's filter.
        !          1168:         * Equations process all content themselves.
        !          1169:         * Tables process almost all content themselves, but we want
        !          1170:         * to warn about macros before passing it there.
1.2       schwarze 1171:         */
                   1172:
1.124   ! schwarze 1173:        if (r->last != NULL && ! ctl) {
1.2       schwarze 1174:                t = r->last->tok;
                   1175:                assert(roffs[t].text);
1.110     schwarze 1176:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
                   1177:                assert(e == ROFF_IGN || e == ROFF_CONT);
                   1178:                if (e != ROFF_CONT)
1.32      schwarze 1179:                        return(e);
1.54      schwarze 1180:        }
1.124   ! schwarze 1181:        if (r->eqn != NULL)
1.110     schwarze 1182:                return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.124   ! schwarze 1183:        if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
        !          1184:                return(tbl_read(r->tbl, ln, buf->buf, pos));
        !          1185:        if ( ! ctl)
1.110     schwarze 1186:                return(roff_parsetext(buf, pos, offs));
1.100     schwarze 1187:
                   1188:        /* Skip empty request lines. */
                   1189:
1.110     schwarze 1190:        if (buf->buf[pos] == '"') {
1.100     schwarze 1191:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1192:                    ln, pos, NULL);
                   1193:                return(ROFF_IGN);
1.110     schwarze 1194:        } else if (buf->buf[pos] == '\0')
1.100     schwarze 1195:                return(ROFF_IGN);
1.2       schwarze 1196:
                   1197:        /*
                   1198:         * If a scope is open, go to the child handler for that macro,
                   1199:         * as it may want to preprocess before doing anything with it.
1.32      schwarze 1200:         * Don't do so if an equation is open.
1.2       schwarze 1201:         */
                   1202:
                   1203:        if (r->last) {
1.1       schwarze 1204:                t = r->last->tok;
                   1205:                assert(roffs[t].sub);
1.110     schwarze 1206:                return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.2       schwarze 1207:        }
                   1208:
1.115     schwarze 1209:        /* No scope is open.  This is a new request or macro. */
                   1210:
                   1211:        spos = pos;
                   1212:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1213:
                   1214:        /* Tables ignore most macros. */
                   1215:
                   1216:        if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
                   1217:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1218:                    ln, pos, buf->buf + spos);
                   1219:                return(ROFF_IGN);
                   1220:        }
                   1221:
1.2       schwarze 1222:        /*
1.115     schwarze 1223:         * This is neither a roff request nor a user-defined macro.
                   1224:         * Let the standard macro set parsers handle it.
1.2       schwarze 1225:         */
                   1226:
1.115     schwarze 1227:        if (t == ROFF_MAX)
1.1       schwarze 1228:                return(ROFF_CONT);
1.115     schwarze 1229:
                   1230:        /* Execute a roff request or a user defined macro. */
1.1       schwarze 1231:
1.2       schwarze 1232:        assert(roffs[t].proc);
1.110     schwarze 1233:        return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2       schwarze 1234: }
                   1235:
1.27      schwarze 1236: void
1.2       schwarze 1237: roff_endparse(struct roff *r)
                   1238: {
1.1       schwarze 1239:
1.27      schwarze 1240:        if (r->last)
1.94      schwarze 1241:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1242:                    r->last->line, r->last->col,
                   1243:                    roffs[r->last->tok].name);
1.27      schwarze 1244:
1.32      schwarze 1245:        if (r->eqn) {
1.94      schwarze 1246:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1247:                    r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41      schwarze 1248:                eqn_end(&r->eqn);
1.32      schwarze 1249:        }
                   1250:
1.27      schwarze 1251:        if (r->tbl) {
1.94      schwarze 1252:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1253:                    r->tbl->line, r->tbl->pos, "TS");
1.41      schwarze 1254:                tbl_end(&r->tbl);
1.27      schwarze 1255:        }
1.1       schwarze 1256: }
                   1257:
                   1258: /*
                   1259:  * Parse a roff node's type from the input buffer.  This must be in the
                   1260:  * form of ".foo xxx" in the usual way.
                   1261:  */
                   1262: static enum rofft
1.87      schwarze 1263: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1       schwarze 1264: {
1.87      schwarze 1265:        char            *cp;
1.16      schwarze 1266:        const char      *mac;
                   1267:        size_t           maclen;
1.1       schwarze 1268:        enum rofft       t;
                   1269:
1.87      schwarze 1270:        cp = buf + *pos;
                   1271:
                   1272:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.1       schwarze 1273:                return(ROFF_MAX);
                   1274:
1.87      schwarze 1275:        mac = cp;
                   1276:        maclen = roff_getname(r, &cp, ln, ppos);
1.1       schwarze 1277:
1.16      schwarze 1278:        t = (r->current_string = roff_getstrn(r, mac, maclen))
1.42      schwarze 1279:            ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.1       schwarze 1280:
1.87      schwarze 1281:        if (ROFF_MAX != t)
                   1282:                *pos = cp - buf;
1.1       schwarze 1283:
                   1284:        return(t);
                   1285: }
                   1286:
                   1287: static enum rofferr
1.2       schwarze 1288: roff_cblock(ROFF_ARGS)
1.1       schwarze 1289: {
                   1290:
1.2       schwarze 1291:        /*
                   1292:         * A block-close `..' should only be invoked as a child of an
                   1293:         * ignore macro, otherwise raise a warning and just ignore it.
                   1294:         */
                   1295:
1.110     schwarze 1296:        if (r->last == NULL) {
1.94      schwarze 1297:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1298:                    ln, ppos, "..");
1.2       schwarze 1299:                return(ROFF_IGN);
                   1300:        }
1.1       schwarze 1301:
1.2       schwarze 1302:        switch (r->last->tok) {
1.80      schwarze 1303:        case ROFF_am:
1.93      schwarze 1304:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.2       schwarze 1305:                /* FALLTHROUGH */
1.80      schwarze 1306:        case ROFF_ami:
1.2       schwarze 1307:                /* FALLTHROUGH */
1.80      schwarze 1308:        case ROFF_de:
1.23      schwarze 1309:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.2       schwarze 1310:                /* FALLTHROUGH */
1.80      schwarze 1311:        case ROFF_dei:
1.2       schwarze 1312:                /* FALLTHROUGH */
1.80      schwarze 1313:        case ROFF_ig:
1.2       schwarze 1314:                break;
                   1315:        default:
1.94      schwarze 1316:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1317:                    ln, ppos, "..");
1.1       schwarze 1318:                return(ROFF_IGN);
1.2       schwarze 1319:        }
                   1320:
1.110     schwarze 1321:        if (buf->buf[pos] != '\0')
1.90      schwarze 1322:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 1323:                    ".. %s", buf->buf + pos);
1.2       schwarze 1324:
                   1325:        roffnode_pop(r);
                   1326:        roffnode_cleanscope(r);
                   1327:        return(ROFF_IGN);
                   1328:
                   1329: }
1.1       schwarze 1330:
1.2       schwarze 1331: static void
                   1332: roffnode_cleanscope(struct roff *r)
                   1333: {
1.1       schwarze 1334:
1.2       schwarze 1335:        while (r->last) {
1.46      schwarze 1336:                if (--r->last->endspan != 0)
1.2       schwarze 1337:                        break;
                   1338:                roffnode_pop(r);
                   1339:        }
                   1340: }
1.1       schwarze 1341:
1.68      schwarze 1342: static void
                   1343: roff_ccond(struct roff *r, int ln, int ppos)
1.2       schwarze 1344: {
1.1       schwarze 1345:
1.2       schwarze 1346:        if (NULL == r->last) {
1.94      schwarze 1347:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1348:                    ln, ppos, "\\}");
1.68      schwarze 1349:                return;
1.2       schwarze 1350:        }
1.1       schwarze 1351:
1.2       schwarze 1352:        switch (r->last->tok) {
1.80      schwarze 1353:        case ROFF_el:
1.2       schwarze 1354:                /* FALLTHROUGH */
1.80      schwarze 1355:        case ROFF_ie:
1.2       schwarze 1356:                /* FALLTHROUGH */
1.80      schwarze 1357:        case ROFF_if:
1.2       schwarze 1358:                break;
                   1359:        default:
1.94      schwarze 1360:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1361:                    ln, ppos, "\\}");
1.68      schwarze 1362:                return;
1.2       schwarze 1363:        }
1.1       schwarze 1364:
1.2       schwarze 1365:        if (r->last->endspan > -1) {
1.94      schwarze 1366:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1367:                    ln, ppos, "\\}");
1.68      schwarze 1368:                return;
1.2       schwarze 1369:        }
                   1370:
                   1371:        roffnode_pop(r);
                   1372:        roffnode_cleanscope(r);
1.68      schwarze 1373:        return;
1.1       schwarze 1374: }
                   1375:
                   1376: static enum rofferr
1.2       schwarze 1377: roff_block(ROFF_ARGS)
1.1       schwarze 1378: {
1.93      schwarze 1379:        const char      *name;
                   1380:        char            *iname, *cp;
1.86      schwarze 1381:        size_t           namesz;
1.16      schwarze 1382:
1.93      schwarze 1383:        /* Ignore groff compatibility mode for now. */
1.2       schwarze 1384:
1.110     schwarze 1385:        if (tok == ROFF_de1)
1.93      schwarze 1386:                tok = ROFF_de;
1.123     schwarze 1387:        else if (tok == ROFF_dei1)
                   1388:                tok = ROFF_dei;
1.110     schwarze 1389:        else if (tok == ROFF_am1)
1.93      schwarze 1390:                tok = ROFF_am;
1.123     schwarze 1391:        else if (tok == ROFF_ami1)
                   1392:                tok = ROFF_ami;
1.93      schwarze 1393:
                   1394:        /* Parse the macro name argument. */
                   1395:
1.110     schwarze 1396:        cp = buf->buf + pos;
                   1397:        if (tok == ROFF_ig) {
1.93      schwarze 1398:                iname = NULL;
                   1399:                namesz = 0;
                   1400:        } else {
                   1401:                iname = cp;
                   1402:                namesz = roff_getname(r, &cp, ln, ppos);
                   1403:                iname[namesz] = '\0';
                   1404:        }
1.22      schwarze 1405:
1.93      schwarze 1406:        /* Resolve the macro name argument if it is indirect. */
1.22      schwarze 1407:
1.110     schwarze 1408:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1409:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93      schwarze 1410:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1411:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1412:                            "%.*s", (int)namesz, iname);
                   1413:                        namesz = 0;
                   1414:                } else
                   1415:                        namesz = strlen(name);
                   1416:        } else
                   1417:                name = iname;
1.22      schwarze 1418:
1.110     schwarze 1419:        if (namesz == 0 && tok != ROFF_ig) {
1.93      schwarze 1420:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
                   1421:                    ln, ppos, roffs[tok].name);
                   1422:                return(ROFF_IGN);
                   1423:        }
1.2       schwarze 1424:
1.16      schwarze 1425:        roffnode_push(r, tok, name, ln, ppos);
                   1426:
                   1427:        /*
                   1428:         * At the beginning of a `de' macro, clear the existing string
                   1429:         * with the same name, if there is one.  New content will be
1.66      schwarze 1430:         * appended from roff_block_text() in multiline mode.
1.16      schwarze 1431:         */
1.22      schwarze 1432:
1.110     schwarze 1433:        if (tok == ROFF_de || tok == ROFF_dei)
1.86      schwarze 1434:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.2       schwarze 1435:
1.110     schwarze 1436:        if (*cp == '\0')
1.2       schwarze 1437:                return(ROFF_IGN);
1.1       schwarze 1438:
1.93      schwarze 1439:        /* Get the custom end marker. */
1.22      schwarze 1440:
1.93      schwarze 1441:        iname = cp;
1.86      schwarze 1442:        namesz = roff_getname(r, &cp, ln, ppos);
1.93      schwarze 1443:
                   1444:        /* Resolve the end marker if it is indirect. */
                   1445:
1.110     schwarze 1446:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1447:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93      schwarze 1448:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1449:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1450:                            "%.*s", (int)namesz, iname);
                   1451:                        namesz = 0;
                   1452:                } else
                   1453:                        namesz = strlen(name);
                   1454:        } else
                   1455:                name = iname;
                   1456:
1.86      schwarze 1457:        if (namesz)
                   1458:                r->last->end = mandoc_strndup(name, namesz);
1.2       schwarze 1459:
1.110     schwarze 1460:        if (*cp != '\0')
1.90      schwarze 1461:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   1462:                    ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.1       schwarze 1463:
                   1464:        return(ROFF_IGN);
                   1465: }
                   1466:
                   1467: static enum rofferr
1.2       schwarze 1468: roff_block_sub(ROFF_ARGS)
1.1       schwarze 1469: {
1.2       schwarze 1470:        enum rofft      t;
                   1471:        int             i, j;
                   1472:
                   1473:        /*
                   1474:         * First check whether a custom macro exists at this level.  If
                   1475:         * it does, then check against it.  This is some of groff's
                   1476:         * stranger behaviours.  If we encountered a custom end-scope
                   1477:         * tag and that tag also happens to be a "real" macro, then we
                   1478:         * need to try interpreting it again as a real macro.  If it's
                   1479:         * not, then return ignore.  Else continue.
                   1480:         */
                   1481:
                   1482:        if (r->last->end) {
1.35      schwarze 1483:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110     schwarze 1484:                        if (buf->buf[i] != r->last->end[j])
1.2       schwarze 1485:                                break;
1.1       schwarze 1486:
1.110     schwarze 1487:                if (r->last->end[j] == '\0' &&
                   1488:                    (buf->buf[i] == '\0' ||
                   1489:                     buf->buf[i] == ' ' ||
                   1490:                     buf->buf[i] == '\t')) {
1.2       schwarze 1491:                        roffnode_pop(r);
                   1492:                        roffnode_cleanscope(r);
1.1       schwarze 1493:
1.110     schwarze 1494:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35      schwarze 1495:                                i++;
                   1496:
                   1497:                        pos = i;
1.110     schwarze 1498:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
                   1499:                            ROFF_MAX)
1.2       schwarze 1500:                                return(ROFF_RERUN);
                   1501:                        return(ROFF_IGN);
                   1502:                }
1.1       schwarze 1503:        }
                   1504:
1.2       schwarze 1505:        /*
                   1506:         * If we have no custom end-query or lookup failed, then try
                   1507:         * pulling it out of the hashtable.
                   1508:         */
1.1       schwarze 1509:
1.110     schwarze 1510:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1       schwarze 1511:
1.110     schwarze 1512:        if (t != ROFF_cblock) {
                   1513:                if (tok != ROFF_ig)
                   1514:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.1       schwarze 1515:                return(ROFF_IGN);
1.16      schwarze 1516:        }
1.1       schwarze 1517:
1.2       schwarze 1518:        assert(roffs[t].proc);
1.110     schwarze 1519:        return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2       schwarze 1520: }
                   1521:
                   1522: static enum rofferr
                   1523: roff_block_text(ROFF_ARGS)
                   1524: {
                   1525:
1.110     schwarze 1526:        if (tok != ROFF_ig)
                   1527:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16      schwarze 1528:
1.2       schwarze 1529:        return(ROFF_IGN);
                   1530: }
                   1531:
                   1532: static enum rofferr
                   1533: roff_cond_sub(ROFF_ARGS)
                   1534: {
                   1535:        enum rofft       t;
1.37      schwarze 1536:        char            *ep;
1.71      schwarze 1537:        int              rr;
1.2       schwarze 1538:
                   1539:        rr = r->last->rule;
1.37      schwarze 1540:        roffnode_cleanscope(r);
1.110     schwarze 1541:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2       schwarze 1542:
1.37      schwarze 1543:        /*
1.50      schwarze 1544:         * Fully handle known macros when they are structurally
                   1545:         * required or when the conditional evaluated to true.
1.5       schwarze 1546:         */
                   1547:
1.110     schwarze 1548:        if ((t != ROFF_MAX) &&
                   1549:            (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.50      schwarze 1550:                assert(roffs[t].proc);
1.110     schwarze 1551:                return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.50      schwarze 1552:        }
1.39      schwarze 1553:
1.69      schwarze 1554:        /*
                   1555:         * If `\}' occurs on a macro line without a preceding macro,
                   1556:         * drop the line completely.
                   1557:         */
                   1558:
1.110     schwarze 1559:        ep = buf->buf + pos;
                   1560:        if (ep[0] == '\\' && ep[1] == '}')
1.71      schwarze 1561:                rr = 0;
1.69      schwarze 1562:
1.50      schwarze 1563:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1564:
1.110     schwarze 1565:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1566:                if (*(++ep) == '}') {
1.70      schwarze 1567:                        *ep = '&';
1.110     schwarze 1568:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1569:                }
1.119     schwarze 1570:                if (*ep != '\0')
                   1571:                        ++ep;
1.50      schwarze 1572:        }
1.71      schwarze 1573:        return(rr ? ROFF_CONT : ROFF_IGN);
1.2       schwarze 1574: }
                   1575:
                   1576: static enum rofferr
                   1577: roff_cond_text(ROFF_ARGS)
                   1578: {
1.37      schwarze 1579:        char            *ep;
1.71      schwarze 1580:        int              rr;
1.2       schwarze 1581:
                   1582:        rr = r->last->rule;
1.37      schwarze 1583:        roffnode_cleanscope(r);
1.1       schwarze 1584:
1.110     schwarze 1585:        ep = buf->buf + pos;
                   1586:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1587:                if (*(++ep) == '}') {
1.70      schwarze 1588:                        *ep = '&';
1.110     schwarze 1589:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1590:                }
1.119     schwarze 1591:                if (*ep != '\0')
                   1592:                        ++ep;
1.2       schwarze 1593:        }
1.71      schwarze 1594:        return(rr ? ROFF_CONT : ROFF_IGN);
1.2       schwarze 1595: }
                   1596:
1.77      schwarze 1597: /*
                   1598:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1599:  * If there is at least one digit, return success and advance the
                   1600:  * parse point, else return failure and let the parse point unchanged.
                   1601:  * Ignore overflows, treat them just like the C language.
                   1602:  */
1.56      schwarze 1603: static int
                   1604: roff_getnum(const char *v, int *pos, int *res)
                   1605: {
1.79      schwarze 1606:        int      myres, n, p;
                   1607:
                   1608:        if (NULL == res)
                   1609:                res = &myres;
1.56      schwarze 1610:
                   1611:        p = *pos;
                   1612:        n = v[p] == '-';
                   1613:        if (n)
                   1614:                p++;
                   1615:
                   1616:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77      schwarze 1617:                *res = 10 * *res + v[p] - '0';
1.56      schwarze 1618:        if (p == *pos + n)
                   1619:                return 0;
                   1620:
                   1621:        if (n)
                   1622:                *res = -*res;
                   1623:
                   1624:        *pos = p;
                   1625:        return 1;
                   1626: }
                   1627:
1.71      schwarze 1628: /*
                   1629:  * Evaluate a string comparison condition.
                   1630:  * The first character is the delimiter.
                   1631:  * Succeed if the string up to its second occurrence
                   1632:  * matches the string up to its third occurence.
                   1633:  * Advance the cursor after the third occurrence
                   1634:  * or lacking that, to the end of the line.
                   1635:  */
                   1636: static int
                   1637: roff_evalstrcond(const char *v, int *pos)
                   1638: {
                   1639:        const char      *s1, *s2, *s3;
                   1640:        int              match;
                   1641:
                   1642:        match = 0;
                   1643:        s1 = v + *pos;          /* initial delimiter */
                   1644:        s2 = s1 + 1;            /* for scanning the first string */
                   1645:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   1646:
                   1647:        if (NULL == s3)         /* found no middle delimiter */
                   1648:                goto out;
                   1649:
                   1650:        while ('\0' != *++s3) {
                   1651:                if (*s2 != *s3) {  /* mismatch */
                   1652:                        s3 = strchr(s3, *s1);
                   1653:                        break;
                   1654:                }
                   1655:                if (*s3 == *s1) {  /* found the final delimiter */
                   1656:                        match = 1;
                   1657:                        break;
                   1658:                }
                   1659:                s2++;
                   1660:        }
                   1661:
                   1662: out:
                   1663:        if (NULL == s3)
                   1664:                s3 = strchr(s2, '\0');
1.114     schwarze 1665:        else if (*s3 != '\0')
1.71      schwarze 1666:                s3++;
                   1667:        *pos = s3 - v;
                   1668:        return(match);
                   1669: }
                   1670:
1.77      schwarze 1671: /*
                   1672:  * Evaluate an optionally negated single character, numerical,
                   1673:  * or string condition.
                   1674:  */
1.71      schwarze 1675: static int
1.106     schwarze 1676: roff_evalcond(struct roff *r, int ln, const char *v, int *pos)
1.5       schwarze 1677: {
1.113     schwarze 1678:        int      number, savepos, wanttrue;
1.5       schwarze 1679:
1.71      schwarze 1680:        if ('!' == v[*pos]) {
                   1681:                wanttrue = 0;
                   1682:                (*pos)++;
                   1683:        } else
                   1684:                wanttrue = 1;
                   1685:
1.5       schwarze 1686:        switch (v[*pos]) {
1.112     schwarze 1687:        case '\0':
                   1688:                return(0);
1.80      schwarze 1689:        case 'n':
1.71      schwarze 1690:                /* FALLTHROUGH */
1.80      schwarze 1691:        case 'o':
1.5       schwarze 1692:                (*pos)++;
1.71      schwarze 1693:                return(wanttrue);
1.80      schwarze 1694:        case 'c':
1.71      schwarze 1695:                /* FALLTHROUGH */
1.80      schwarze 1696:        case 'd':
1.71      schwarze 1697:                /* FALLTHROUGH */
1.80      schwarze 1698:        case 'e':
1.5       schwarze 1699:                /* FALLTHROUGH */
1.80      schwarze 1700:        case 'r':
1.5       schwarze 1701:                /* FALLTHROUGH */
1.80      schwarze 1702:        case 't':
1.111     schwarze 1703:                /* FALLTHROUGH */
                   1704:        case 'v':
1.5       schwarze 1705:                (*pos)++;
1.71      schwarze 1706:                return(!wanttrue);
1.5       schwarze 1707:        default:
                   1708:                break;
                   1709:        }
                   1710:
1.113     schwarze 1711:        savepos = *pos;
1.106     schwarze 1712:        if (roff_evalnum(r, ln, v, pos, &number, 0))
1.77      schwarze 1713:                return((number > 0) == wanttrue);
1.113     schwarze 1714:        else if (*pos == savepos)
                   1715:                return(roff_evalstrcond(v, pos) == wanttrue);
1.77      schwarze 1716:        else
1.113     schwarze 1717:                return (0);
1.5       schwarze 1718: }
                   1719:
1.2       schwarze 1720: static enum rofferr
1.21      schwarze 1721: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 1722: {
1.30      schwarze 1723:
1.21      schwarze 1724:        return(ROFF_IGN);
                   1725: }
                   1726:
                   1727: static enum rofferr
1.123     schwarze 1728: roff_insec(ROFF_ARGS)
                   1729: {
                   1730:
                   1731:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
                   1732:            ln, ppos, roffs[tok].name);
                   1733:        return(ROFF_IGN);
                   1734: }
                   1735:
                   1736: static enum rofferr
                   1737: roff_unsupp(ROFF_ARGS)
                   1738: {
                   1739:
                   1740:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
                   1741:            ln, ppos, roffs[tok].name);
                   1742:        return(ROFF_IGN);
                   1743: }
                   1744:
                   1745: static enum rofferr
1.2       schwarze 1746: roff_cond(ROFF_ARGS)
                   1747: {
1.46      schwarze 1748:
                   1749:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 1750:
1.80      schwarze 1751:        /*
1.35      schwarze 1752:         * An `.el' has no conditional body: it will consume the value
                   1753:         * of the current rstack entry set in prior `ie' calls or
1.80      schwarze 1754:         * defaults to DENY.
1.35      schwarze 1755:         *
                   1756:         * If we're not an `el', however, then evaluate the conditional.
                   1757:         */
1.1       schwarze 1758:
1.110     schwarze 1759:        r->last->rule = tok == ROFF_el ?
1.80      schwarze 1760:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110     schwarze 1761:            roff_evalcond(r, ln, buf->buf, &pos);
1.2       schwarze 1762:
1.35      schwarze 1763:        /*
                   1764:         * An if-else will put the NEGATION of the current evaluated
                   1765:         * conditional into the stack of rules.
                   1766:         */
                   1767:
1.110     schwarze 1768:        if (tok == ROFF_ie) {
1.96      schwarze 1769:                if (r->rstackpos + 1 == r->rstacksz) {
                   1770:                        r->rstacksz += 16;
                   1771:                        r->rstack = mandoc_reallocarray(r->rstack,
                   1772:                            r->rstacksz, sizeof(int));
1.35      schwarze 1773:                }
1.71      schwarze 1774:                r->rstack[++r->rstackpos] = !r->last->rule;
1.2       schwarze 1775:        }
1.5       schwarze 1776:
                   1777:        /* If the parent has false as its rule, then so do we. */
                   1778:
1.71      schwarze 1779:        if (r->last->parent && !r->last->parent->rule)
                   1780:                r->last->rule = 0;
1.5       schwarze 1781:
                   1782:        /*
1.46      schwarze 1783:         * Determine scope.
                   1784:         * If there is nothing on the line after the conditional,
                   1785:         * not even whitespace, use next-line scope.
1.5       schwarze 1786:         */
1.2       schwarze 1787:
1.110     schwarze 1788:        if (buf->buf[pos] == '\0') {
1.46      schwarze 1789:                r->last->endspan = 2;
                   1790:                goto out;
                   1791:        }
                   1792:
1.110     schwarze 1793:        while (buf->buf[pos] == ' ')
1.46      schwarze 1794:                pos++;
                   1795:
                   1796:        /* An opening brace requests multiline scope. */
1.2       schwarze 1797:
1.110     schwarze 1798:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2       schwarze 1799:                r->last->endspan = -1;
                   1800:                pos += 2;
1.46      schwarze 1801:                goto out;
1.80      schwarze 1802:        }
1.2       schwarze 1803:
                   1804:        /*
1.46      schwarze 1805:         * Anything else following the conditional causes
                   1806:         * single-line scope.  Warn if the scope contains
                   1807:         * nothing but trailing whitespace.
1.2       schwarze 1808:         */
                   1809:
1.110     schwarze 1810:        if (buf->buf[pos] == '\0')
1.89      schwarze 1811:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
                   1812:                    ln, ppos, roffs[tok].name);
1.2       schwarze 1813:
1.46      schwarze 1814:        r->last->endspan = 1;
1.1       schwarze 1815:
1.46      schwarze 1816: out:
1.2       schwarze 1817:        *offs = pos;
                   1818:        return(ROFF_RERUN);
1.1       schwarze 1819: }
                   1820:
1.2       schwarze 1821: static enum rofferr
1.7       schwarze 1822: roff_ds(ROFF_ARGS)
                   1823: {
1.85      schwarze 1824:        char            *string;
                   1825:        const char      *name;
                   1826:        size_t           namesz;
1.10      schwarze 1827:
1.123     schwarze 1828:        /* Ignore groff compatibility mode for now. */
                   1829:
                   1830:        if (tok == ROFF_ds1)
                   1831:                tok = ROFF_ds;
                   1832:        else if (tok == ROFF_as1)
                   1833:                tok = ROFF_as;
                   1834:
1.10      schwarze 1835:        /*
1.85      schwarze 1836:         * The first word is the name of the string.
                   1837:         * If it is empty or terminated by an escape sequence,
                   1838:         * abort the `ds' request without defining anything.
1.10      schwarze 1839:         */
1.7       schwarze 1840:
1.110     schwarze 1841:        name = string = buf->buf + pos;
                   1842:        if (*name == '\0')
1.7       schwarze 1843:                return(ROFF_IGN);
                   1844:
1.85      schwarze 1845:        namesz = roff_getname(r, &string, ln, pos);
1.110     schwarze 1846:        if (name[namesz] == '\\')
1.85      schwarze 1847:                return(ROFF_IGN);
                   1848:
                   1849:        /* Read past the initial double-quote, if any. */
1.110     schwarze 1850:        if (*string == '"')
1.7       schwarze 1851:                string++;
                   1852:
1.10      schwarze 1853:        /* The rest is the value. */
1.85      schwarze 1854:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   1855:            ROFF_as == tok);
1.7       schwarze 1856:        return(ROFF_IGN);
                   1857: }
                   1858:
1.77      schwarze 1859: /*
                   1860:  * Parse a single operator, one or two characters long.
                   1861:  * If the operator is recognized, return success and advance the
                   1862:  * parse point, else return failure and let the parse point unchanged.
                   1863:  */
                   1864: static int
                   1865: roff_getop(const char *v, int *pos, char *res)
                   1866: {
                   1867:
                   1868:        *res = v[*pos];
                   1869:
                   1870:        switch (*res) {
1.80      schwarze 1871:        case '+':
1.77      schwarze 1872:                /* FALLTHROUGH */
1.80      schwarze 1873:        case '-':
1.77      schwarze 1874:                /* FALLTHROUGH */
1.80      schwarze 1875:        case '*':
1.77      schwarze 1876:                /* FALLTHROUGH */
1.80      schwarze 1877:        case '/':
1.77      schwarze 1878:                /* FALLTHROUGH */
1.80      schwarze 1879:        case '%':
1.77      schwarze 1880:                /* FALLTHROUGH */
1.80      schwarze 1881:        case '&':
1.77      schwarze 1882:                /* FALLTHROUGH */
1.80      schwarze 1883:        case ':':
1.77      schwarze 1884:                break;
                   1885:        case '<':
                   1886:                switch (v[*pos + 1]) {
1.80      schwarze 1887:                case '=':
1.77      schwarze 1888:                        *res = 'l';
                   1889:                        (*pos)++;
                   1890:                        break;
1.80      schwarze 1891:                case '>':
1.77      schwarze 1892:                        *res = '!';
                   1893:                        (*pos)++;
                   1894:                        break;
1.80      schwarze 1895:                case '?':
1.77      schwarze 1896:                        *res = 'i';
                   1897:                        (*pos)++;
                   1898:                        break;
                   1899:                default:
                   1900:                        break;
                   1901:                }
                   1902:                break;
                   1903:        case '>':
                   1904:                switch (v[*pos + 1]) {
1.80      schwarze 1905:                case '=':
1.77      schwarze 1906:                        *res = 'g';
                   1907:                        (*pos)++;
                   1908:                        break;
1.80      schwarze 1909:                case '?':
1.77      schwarze 1910:                        *res = 'a';
                   1911:                        (*pos)++;
                   1912:                        break;
                   1913:                default:
                   1914:                        break;
                   1915:                }
                   1916:                break;
                   1917:        case '=':
                   1918:                if ('=' == v[*pos + 1])
                   1919:                        (*pos)++;
                   1920:                break;
                   1921:        default:
                   1922:                return(0);
                   1923:        }
                   1924:        (*pos)++;
                   1925:
                   1926:        return(*res);
                   1927: }
                   1928:
                   1929: /*
                   1930:  * Evaluate either a parenthesized numeric expression
                   1931:  * or a single signed integer number.
                   1932:  */
                   1933: static int
1.106     schwarze 1934: roff_evalpar(struct roff *r, int ln,
                   1935:        const char *v, int *pos, int *res)
1.77      schwarze 1936: {
                   1937:
                   1938:        if ('(' != v[*pos])
                   1939:                return(roff_getnum(v, pos, res));
                   1940:
                   1941:        (*pos)++;
1.106     schwarze 1942:        if ( ! roff_evalnum(r, ln, v, pos, res, 1))
1.77      schwarze 1943:                return(0);
                   1944:
1.79      schwarze 1945:        /*
                   1946:         * Omission of the closing parenthesis
                   1947:         * is an error in validation mode,
                   1948:         * but ignored in evaluation mode.
                   1949:         */
                   1950:
1.77      schwarze 1951:        if (')' == v[*pos])
                   1952:                (*pos)++;
1.79      schwarze 1953:        else if (NULL == res)
                   1954:                return(0);
1.77      schwarze 1955:
                   1956:        return(1);
                   1957: }
                   1958:
                   1959: /*
                   1960:  * Evaluate a complete numeric expression.
                   1961:  * Proceed left to right, there is no concept of precedence.
                   1962:  */
                   1963: static int
1.106     schwarze 1964: roff_evalnum(struct roff *r, int ln, const char *v,
                   1965:        int *pos, int *res, int skipwhite)
1.77      schwarze 1966: {
                   1967:        int              mypos, operand2;
                   1968:        char             operator;
                   1969:
                   1970:        if (NULL == pos) {
                   1971:                mypos = 0;
                   1972:                pos = &mypos;
                   1973:        }
                   1974:
                   1975:        if (skipwhite)
                   1976:                while (isspace((unsigned char)v[*pos]))
                   1977:                        (*pos)++;
                   1978:
1.106     schwarze 1979:        if ( ! roff_evalpar(r, ln, v, pos, res))
1.77      schwarze 1980:                return(0);
                   1981:
                   1982:        while (1) {
                   1983:                if (skipwhite)
                   1984:                        while (isspace((unsigned char)v[*pos]))
                   1985:                                (*pos)++;
                   1986:
                   1987:                if ( ! roff_getop(v, pos, &operator))
                   1988:                        break;
                   1989:
                   1990:                if (skipwhite)
                   1991:                        while (isspace((unsigned char)v[*pos]))
                   1992:                                (*pos)++;
                   1993:
1.106     schwarze 1994:                if ( ! roff_evalpar(r, ln, v, pos, &operand2))
1.77      schwarze 1995:                        return(0);
                   1996:
                   1997:                if (skipwhite)
                   1998:                        while (isspace((unsigned char)v[*pos]))
                   1999:                                (*pos)++;
1.79      schwarze 2000:
                   2001:                if (NULL == res)
                   2002:                        continue;
1.77      schwarze 2003:
                   2004:                switch (operator) {
1.80      schwarze 2005:                case '+':
1.77      schwarze 2006:                        *res += operand2;
                   2007:                        break;
1.80      schwarze 2008:                case '-':
1.77      schwarze 2009:                        *res -= operand2;
                   2010:                        break;
1.80      schwarze 2011:                case '*':
1.77      schwarze 2012:                        *res *= operand2;
                   2013:                        break;
1.80      schwarze 2014:                case '/':
1.116     schwarze 2015:                        if (operand2 == 0) {
1.106     schwarze 2016:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2017:                                        r->parse, ln, *pos, v);
                   2018:                                *res = 0;
                   2019:                                break;
                   2020:                        }
1.77      schwarze 2021:                        *res /= operand2;
                   2022:                        break;
1.80      schwarze 2023:                case '%':
1.116     schwarze 2024:                        if (operand2 == 0) {
                   2025:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2026:                                        r->parse, ln, *pos, v);
                   2027:                                *res = 0;
                   2028:                                break;
                   2029:                        }
1.77      schwarze 2030:                        *res %= operand2;
                   2031:                        break;
1.80      schwarze 2032:                case '<':
1.77      schwarze 2033:                        *res = *res < operand2;
                   2034:                        break;
1.80      schwarze 2035:                case '>':
1.77      schwarze 2036:                        *res = *res > operand2;
                   2037:                        break;
1.80      schwarze 2038:                case 'l':
1.77      schwarze 2039:                        *res = *res <= operand2;
                   2040:                        break;
1.80      schwarze 2041:                case 'g':
1.77      schwarze 2042:                        *res = *res >= operand2;
                   2043:                        break;
1.80      schwarze 2044:                case '=':
1.77      schwarze 2045:                        *res = *res == operand2;
                   2046:                        break;
1.80      schwarze 2047:                case '!':
1.77      schwarze 2048:                        *res = *res != operand2;
                   2049:                        break;
1.80      schwarze 2050:                case '&':
1.77      schwarze 2051:                        *res = *res && operand2;
                   2052:                        break;
1.80      schwarze 2053:                case ':':
1.77      schwarze 2054:                        *res = *res || operand2;
                   2055:                        break;
1.80      schwarze 2056:                case 'i':
1.77      schwarze 2057:                        if (operand2 < *res)
                   2058:                                *res = operand2;
                   2059:                        break;
1.80      schwarze 2060:                case 'a':
1.77      schwarze 2061:                        if (operand2 > *res)
                   2062:                                *res = operand2;
                   2063:                        break;
                   2064:                default:
                   2065:                        abort();
                   2066:                }
                   2067:        }
                   2068:        return(1);
                   2069: }
                   2070:
1.52      schwarze 2071: void
1.60      schwarze 2072: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41      schwarze 2073: {
1.52      schwarze 2074:        struct roffreg  *reg;
                   2075:
                   2076:        /* Search for an existing register with the same name. */
                   2077:        reg = r->regtab;
                   2078:
                   2079:        while (reg && strcmp(name, reg->key.p))
                   2080:                reg = reg->next;
1.41      schwarze 2081:
1.52      schwarze 2082:        if (NULL == reg) {
                   2083:                /* Create a new register. */
                   2084:                reg = mandoc_malloc(sizeof(struct roffreg));
                   2085:                reg->key.p = mandoc_strdup(name);
                   2086:                reg->key.sz = strlen(name);
1.60      schwarze 2087:                reg->val = 0;
1.52      schwarze 2088:                reg->next = r->regtab;
                   2089:                r->regtab = reg;
                   2090:        }
                   2091:
1.60      schwarze 2092:        if ('+' == sign)
                   2093:                reg->val += val;
                   2094:        else if ('-' == sign)
                   2095:                reg->val -= val;
                   2096:        else
                   2097:                reg->val = val;
1.41      schwarze 2098: }
                   2099:
1.65      schwarze 2100: /*
                   2101:  * Handle some predefined read-only number registers.
                   2102:  * For now, return -1 if the requested register is not predefined;
                   2103:  * in case a predefined read-only register having the value -1
                   2104:  * were to turn up, another special value would have to be chosen.
                   2105:  */
                   2106: static int
                   2107: roff_getregro(const char *name)
                   2108: {
                   2109:
                   2110:        switch (*name) {
1.80      schwarze 2111:        case 'A':  /* ASCII approximation mode is always off. */
1.65      schwarze 2112:                return(0);
1.80      schwarze 2113:        case 'g':  /* Groff compatibility mode is always on. */
1.65      schwarze 2114:                return(1);
1.80      schwarze 2115:        case 'H':  /* Fixed horizontal resolution. */
1.65      schwarze 2116:                return (24);
1.80      schwarze 2117:        case 'j':  /* Always adjust left margin only. */
1.65      schwarze 2118:                return(0);
1.80      schwarze 2119:        case 'T':  /* Some output device is always defined. */
1.65      schwarze 2120:                return(1);
1.80      schwarze 2121:        case 'V':  /* Fixed vertical resolution. */
1.65      schwarze 2122:                return (40);
                   2123:        default:
                   2124:                return (-1);
                   2125:        }
                   2126: }
                   2127:
1.53      schwarze 2128: int
1.52      schwarze 2129: roff_getreg(const struct roff *r, const char *name)
1.41      schwarze 2130: {
1.52      schwarze 2131:        struct roffreg  *reg;
1.65      schwarze 2132:        int              val;
                   2133:
                   2134:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
                   2135:                val = roff_getregro(name + 1);
                   2136:                if (-1 != val)
                   2137:                        return (val);
                   2138:        }
1.52      schwarze 2139:
                   2140:        for (reg = r->regtab; reg; reg = reg->next)
                   2141:                if (0 == strcmp(name, reg->key.p))
1.53      schwarze 2142:                        return(reg->val);
                   2143:
                   2144:        return(0);
                   2145: }
                   2146:
                   2147: static int
                   2148: roff_getregn(const struct roff *r, const char *name, size_t len)
                   2149: {
                   2150:        struct roffreg  *reg;
1.65      schwarze 2151:        int              val;
                   2152:
                   2153:        if ('.' == name[0] && 2 == len) {
                   2154:                val = roff_getregro(name + 1);
                   2155:                if (-1 != val)
                   2156:                        return (val);
                   2157:        }
1.53      schwarze 2158:
                   2159:        for (reg = r->regtab; reg; reg = reg->next)
                   2160:                if (len == reg->key.sz &&
                   2161:                    0 == strncmp(name, reg->key.p, len))
                   2162:                        return(reg->val);
1.41      schwarze 2163:
1.52      schwarze 2164:        return(0);
1.41      schwarze 2165: }
                   2166:
1.52      schwarze 2167: static void
                   2168: roff_freereg(struct roffreg *reg)
1.41      schwarze 2169: {
1.52      schwarze 2170:        struct roffreg  *old_reg;
1.41      schwarze 2171:
1.52      schwarze 2172:        while (NULL != reg) {
                   2173:                free(reg->key.p);
                   2174:                old_reg = reg;
                   2175:                reg = reg->next;
                   2176:                free(old_reg);
                   2177:        }
1.41      schwarze 2178: }
1.7       schwarze 2179:
                   2180: static enum rofferr
1.6       schwarze 2181: roff_nr(ROFF_ARGS)
1.1       schwarze 2182: {
1.85      schwarze 2183:        char            *key, *val;
                   2184:        size_t           keysz;
1.37      schwarze 2185:        int              iv;
1.60      schwarze 2186:        char             sign;
1.6       schwarze 2187:
1.110     schwarze 2188:        key = val = buf->buf + pos;
                   2189:        if (*key == '\0')
1.85      schwarze 2190:                return(ROFF_IGN);
                   2191:
                   2192:        keysz = roff_getname(r, &val, ln, pos);
1.110     schwarze 2193:        if (key[keysz] == '\\')
1.85      schwarze 2194:                return(ROFF_IGN);
                   2195:        key[keysz] = '\0';
1.6       schwarze 2196:
1.60      schwarze 2197:        sign = *val;
1.110     schwarze 2198:        if (sign == '+' || sign == '-')
1.60      schwarze 2199:                val++;
                   2200:
1.106     schwarze 2201:        if (roff_evalnum(r, ln, val, NULL, &iv, 0))
1.77      schwarze 2202:                roff_setreg(r, key, iv, sign);
1.1       schwarze 2203:
1.76      schwarze 2204:        return(ROFF_IGN);
                   2205: }
                   2206:
                   2207: static enum rofferr
                   2208: roff_rr(ROFF_ARGS)
                   2209: {
                   2210:        struct roffreg  *reg, **prev;
1.85      schwarze 2211:        char            *name, *cp;
                   2212:        size_t           namesz;
1.76      schwarze 2213:
1.110     schwarze 2214:        name = cp = buf->buf + pos;
                   2215:        if (*name == '\0')
1.85      schwarze 2216:                return(ROFF_IGN);
                   2217:        namesz = roff_getname(r, &cp, ln, pos);
                   2218:        name[namesz] = '\0';
1.76      schwarze 2219:
                   2220:        prev = &r->regtab;
                   2221:        while (1) {
                   2222:                reg = *prev;
1.110     schwarze 2223:                if (reg == NULL || !strcmp(name, reg->key.p))
1.76      schwarze 2224:                        break;
                   2225:                prev = &reg->next;
                   2226:        }
1.110     schwarze 2227:        if (reg != NULL) {
1.76      schwarze 2228:                *prev = reg->next;
                   2229:                free(reg->key.p);
                   2230:                free(reg);
                   2231:        }
1.29      schwarze 2232:        return(ROFF_IGN);
                   2233: }
                   2234:
                   2235: static enum rofferr
                   2236: roff_rm(ROFF_ARGS)
                   2237: {
                   2238:        const char       *name;
                   2239:        char             *cp;
1.85      schwarze 2240:        size_t            namesz;
1.29      schwarze 2241:
1.110     schwarze 2242:        cp = buf->buf + pos;
                   2243:        while (*cp != '\0') {
1.85      schwarze 2244:                name = cp;
1.110     schwarze 2245:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85      schwarze 2246:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.110     schwarze 2247:                if (name[namesz] == '\\')
1.85      schwarze 2248:                        break;
1.29      schwarze 2249:        }
1.51      schwarze 2250:        return(ROFF_IGN);
                   2251: }
                   2252:
                   2253: static enum rofferr
                   2254: roff_it(ROFF_ARGS)
                   2255: {
                   2256:        char            *cp;
                   2257:        size_t           len;
                   2258:        int              iv;
                   2259:
                   2260:        /* Parse the number of lines. */
1.110     schwarze 2261:        cp = buf->buf + pos;
1.51      schwarze 2262:        len = strcspn(cp, " \t");
                   2263:        cp[len] = '\0';
                   2264:        if ((iv = mandoc_strntoi(cp, len, 10)) <= 0) {
1.95      schwarze 2265:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110     schwarze 2266:                    ln, ppos, buf->buf + 1);
1.51      schwarze 2267:                return(ROFF_IGN);
                   2268:        }
                   2269:        cp += len + 1;
                   2270:
                   2271:        /* Arm the input line trap. */
                   2272:        roffit_lines = iv;
                   2273:        roffit_macro = mandoc_strdup(cp);
1.2       schwarze 2274:        return(ROFF_IGN);
1.47      schwarze 2275: }
                   2276:
                   2277: static enum rofferr
                   2278: roff_Dd(ROFF_ARGS)
                   2279: {
                   2280:        const char *const       *cp;
                   2281:
1.99      schwarze 2282:        if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47      schwarze 2283:                for (cp = __mdoc_reserved; *cp; cp++)
                   2284:                        roff_setstr(r, *cp, NULL, 0);
                   2285:
1.99      schwarze 2286:        if (r->format == 0)
                   2287:                r->format = MPARSE_MDOC;
                   2288:
1.47      schwarze 2289:        return(ROFF_CONT);
                   2290: }
                   2291:
                   2292: static enum rofferr
                   2293: roff_TH(ROFF_ARGS)
                   2294: {
                   2295:        const char *const       *cp;
                   2296:
1.99      schwarze 2297:        if ((r->options & MPARSE_QUICK) == 0)
1.47      schwarze 2298:                for (cp = __man_reserved; *cp; cp++)
                   2299:                        roff_setstr(r, *cp, NULL, 0);
                   2300:
1.99      schwarze 2301:        if (r->format == 0)
                   2302:                r->format = MPARSE_MAN;
                   2303:
1.47      schwarze 2304:        return(ROFF_CONT);
1.14      schwarze 2305: }
                   2306:
                   2307: static enum rofferr
1.27      schwarze 2308: roff_TE(ROFF_ARGS)
                   2309: {
                   2310:
                   2311:        if (NULL == r->tbl)
1.94      schwarze 2312:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2313:                    ln, ppos, "TE");
1.27      schwarze 2314:        else
1.41      schwarze 2315:                tbl_end(&r->tbl);
1.27      schwarze 2316:
                   2317:        return(ROFF_IGN);
                   2318: }
                   2319:
                   2320: static enum rofferr
                   2321: roff_T_(ROFF_ARGS)
                   2322: {
                   2323:
                   2324:        if (NULL == r->tbl)
1.94      schwarze 2325:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2326:                    ln, ppos, "T&");
1.27      schwarze 2327:        else
                   2328:                tbl_restart(ppos, ln, r->tbl);
                   2329:
                   2330:        return(ROFF_IGN);
                   2331: }
                   2332:
1.102     schwarze 2333: /*
                   2334:  * Handle in-line equation delimiters.
                   2335:  */
                   2336: static enum rofferr
1.110     schwarze 2337: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41      schwarze 2338: {
1.105     schwarze 2339:        char            *cp1, *cp2;
                   2340:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41      schwarze 2341:
1.102     schwarze 2342:        /*
                   2343:         * Outside equations, look for an opening delimiter.
                   2344:         * If we are inside an equation, we already know it is
                   2345:         * in-line, or this function wouldn't have been called;
                   2346:         * so look for a closing delimiter.
                   2347:         */
                   2348:
1.110     schwarze 2349:        cp1 = buf->buf + pos;
1.102     schwarze 2350:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2351:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2352:        if (cp2 == NULL)
                   2353:                return(ROFF_CONT);
                   2354:
1.105     schwarze 2355:        *cp2++ = '\0';
                   2356:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2357:
                   2358:        /* Handle preceding text, protecting whitespace. */
                   2359:
1.110     schwarze 2360:        if (*buf->buf != '\0') {
1.105     schwarze 2361:                if (r->eqn == NULL)
                   2362:                        bef_pr = "\\&";
                   2363:                bef_nl = "\n";
                   2364:        }
                   2365:
                   2366:        /*
                   2367:         * Prepare replacing the delimiter with an equation macro
                   2368:         * and drop leading white space from the equation.
                   2369:         */
1.102     schwarze 2370:
1.105     schwarze 2371:        if (r->eqn == NULL) {
                   2372:                while (*cp2 == ' ')
                   2373:                        cp2++;
                   2374:                mac = ".EQ";
                   2375:        } else
                   2376:                mac = ".EN";
                   2377:
                   2378:        /* Handle following text, protecting whitespace. */
                   2379:
                   2380:        if (*cp2 != '\0') {
                   2381:                aft_nl = "\n";
                   2382:                if (r->eqn != NULL)
                   2383:                        aft_pr = "\\&";
                   2384:        }
                   2385:
                   2386:        /* Do the actual replacement. */
                   2387:
1.110     schwarze 2388:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105     schwarze 2389:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110     schwarze 2390:        free(buf->buf);
                   2391:        buf->buf = cp1;
1.102     schwarze 2392:
                   2393:        /* Toggle the in-line state of the eqn subsystem. */
                   2394:
                   2395:        r->eqn_inline = r->eqn == NULL;
                   2396:        return(ROFF_REPARSE);
1.41      schwarze 2397: }
                   2398:
1.107     schwarze 2399: static enum rofferr
                   2400: roff_EQ(ROFF_ARGS)
1.32      schwarze 2401: {
1.41      schwarze 2402:        struct eqn_node *e;
1.32      schwarze 2403:
1.110     schwarze 2404:        assert(r->eqn == NULL);
1.107     schwarze 2405:        e = eqn_alloc(ppos, ln, r->parse);
1.32      schwarze 2406:
1.102     schwarze 2407:        if (r->last_eqn) {
1.32      schwarze 2408:                r->last_eqn->next = e;
1.102     schwarze 2409:                e->delim = r->last_eqn->delim;
                   2410:                e->odelim = r->last_eqn->odelim;
                   2411:                e->cdelim = r->last_eqn->cdelim;
                   2412:        } else
1.32      schwarze 2413:                r->first_eqn = r->last_eqn = e;
                   2414:
                   2415:        r->eqn = r->last_eqn = e;
1.41      schwarze 2416:
1.110     schwarze 2417:        if (buf->buf[pos] != '\0')
1.107     schwarze 2418:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 2419:                    ".EQ %s", buf->buf + pos);
1.41      schwarze 2420:
1.32      schwarze 2421:        return(ROFF_IGN);
                   2422: }
                   2423:
                   2424: static enum rofferr
                   2425: roff_EN(ROFF_ARGS)
                   2426: {
                   2427:
1.94      schwarze 2428:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.32      schwarze 2429:        return(ROFF_IGN);
                   2430: }
                   2431:
                   2432: static enum rofferr
1.27      schwarze 2433: roff_TS(ROFF_ARGS)
                   2434: {
1.49      schwarze 2435:        struct tbl_node *tbl;
1.27      schwarze 2436:
                   2437:        if (r->tbl) {
1.94      schwarze 2438:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2439:                    ln, ppos, "TS breaks TS");
1.41      schwarze 2440:                tbl_end(&r->tbl);
1.27      schwarze 2441:        }
                   2442:
1.49      schwarze 2443:        tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 2444:
                   2445:        if (r->last_tbl)
1.49      schwarze 2446:                r->last_tbl->next = tbl;
1.27      schwarze 2447:        else
1.49      schwarze 2448:                r->first_tbl = r->last_tbl = tbl;
1.27      schwarze 2449:
1.49      schwarze 2450:        r->tbl = r->last_tbl = tbl;
1.27      schwarze 2451:        return(ROFF_IGN);
1.123     schwarze 2452: }
                   2453:
                   2454: static enum rofferr
                   2455: roff_brp(ROFF_ARGS)
                   2456: {
                   2457:
                   2458:        buf->buf[pos - 1] = '\0';
                   2459:        return(ROFF_CONT);
1.27      schwarze 2460: }
                   2461:
                   2462: static enum rofferr
1.48      schwarze 2463: roff_cc(ROFF_ARGS)
                   2464: {
                   2465:        const char      *p;
                   2466:
1.110     schwarze 2467:        p = buf->buf + pos;
1.48      schwarze 2468:
1.110     schwarze 2469:        if (*p == '\0' || (r->control = *p++) == '.')
1.48      schwarze 2470:                r->control = 0;
                   2471:
1.110     schwarze 2472:        if (*p != '\0')
1.48      schwarze 2473:                mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
                   2474:
                   2475:        return(ROFF_IGN);
                   2476: }
                   2477:
                   2478: static enum rofferr
1.42      schwarze 2479: roff_tr(ROFF_ARGS)
                   2480: {
                   2481:        const char      *p, *first, *second;
                   2482:        size_t           fsz, ssz;
                   2483:        enum mandoc_esc  esc;
                   2484:
1.110     schwarze 2485:        p = buf->buf + pos;
1.42      schwarze 2486:
1.110     schwarze 2487:        if (*p == '\0') {
1.42      schwarze 2488:                mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
                   2489:                return(ROFF_IGN);
                   2490:        }
                   2491:
1.110     schwarze 2492:        while (*p != '\0') {
1.42      schwarze 2493:                fsz = ssz = 1;
                   2494:
                   2495:                first = p++;
1.110     schwarze 2496:                if (*first == '\\') {
1.42      schwarze 2497:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 2498:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 2499:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 2500:                                    ln, (int)(p - buf->buf), first);
1.42      schwarze 2501:                                return(ROFF_IGN);
                   2502:                        }
                   2503:                        fsz = (size_t)(p - first);
                   2504:                }
                   2505:
                   2506:                second = p++;
1.110     schwarze 2507:                if (*second == '\\') {
1.42      schwarze 2508:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 2509:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 2510:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 2511:                                    ln, (int)(p - buf->buf), second);
1.42      schwarze 2512:                                return(ROFF_IGN);
                   2513:                        }
                   2514:                        ssz = (size_t)(p - second);
1.110     schwarze 2515:                } else if (*second == '\0') {
1.80      schwarze 2516:                        mandoc_msg(MANDOCERR_ARGCOUNT, r->parse,
1.110     schwarze 2517:                            ln, (int)(p - buf->buf), NULL);
1.42      schwarze 2518:                        second = " ";
                   2519:                        p--;
                   2520:                }
                   2521:
                   2522:                if (fsz > 1) {
1.80      schwarze 2523:                        roff_setstrn(&r->xmbtab, first, fsz,
                   2524:                            second, ssz, 0);
1.42      schwarze 2525:                        continue;
                   2526:                }
                   2527:
1.110     schwarze 2528:                if (r->xtab == NULL)
1.80      schwarze 2529:                        r->xtab = mandoc_calloc(128,
                   2530:                            sizeof(struct roffstr));
1.42      schwarze 2531:
                   2532:                free(r->xtab[(int)*first].p);
                   2533:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   2534:                r->xtab[(int)*first].sz = ssz;
                   2535:        }
                   2536:
                   2537:        return(ROFF_IGN);
                   2538: }
                   2539:
                   2540: static enum rofferr
1.14      schwarze 2541: roff_so(ROFF_ARGS)
                   2542: {
1.121     schwarze 2543:        char *name, *cp;
1.15      schwarze 2544:
1.110     schwarze 2545:        name = buf->buf + pos;
1.97      schwarze 2546:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14      schwarze 2547:
1.22      schwarze 2548:        /*
                   2549:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   2550:         * opening anything that's not in our cwd or anything beneath
                   2551:         * it.  Thus, explicitly disallow traversing up the file-system
                   2552:         * or using absolute paths.
                   2553:         */
                   2554:
1.110     schwarze 2555:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83      schwarze 2556:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   2557:                    ".so %s", name);
1.121     schwarze 2558:                buf->sz = mandoc_asprintf(&cp,
                   2559:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   2560:                free(buf->buf);
                   2561:                buf->buf = cp;
                   2562:                *offs = 0;
                   2563:                return(ROFF_REPARSE);
1.14      schwarze 2564:        }
                   2565:
                   2566:        *offs = pos;
                   2567:        return(ROFF_SO);
1.7       schwarze 2568: }
                   2569:
1.16      schwarze 2570: static enum rofferr
                   2571: roff_userdef(ROFF_ARGS)
1.12      schwarze 2572: {
1.16      schwarze 2573:        const char       *arg[9];
                   2574:        char             *cp, *n1, *n2;
1.25      schwarze 2575:        int               i;
1.12      schwarze 2576:
1.16      schwarze 2577:        /*
                   2578:         * Collect pointers to macro argument strings
1.61      schwarze 2579:         * and NUL-terminate them.
1.16      schwarze 2580:         */
1.110     schwarze 2581:        cp = buf->buf + pos;
1.25      schwarze 2582:        for (i = 0; i < 9; i++)
1.110     schwarze 2583:                arg[i] = *cp == '\0' ? "" :
1.35      schwarze 2584:                    mandoc_getarg(r->parse, &cp, ln, &pos);
1.16      schwarze 2585:
                   2586:        /*
                   2587:         * Expand macro arguments.
1.12      schwarze 2588:         */
1.110     schwarze 2589:        buf->sz = 0;
1.16      schwarze 2590:        n1 = cp = mandoc_strdup(r->current_string);
1.110     schwarze 2591:        while ((cp = strstr(cp, "\\$")) != NULL) {
1.16      schwarze 2592:                i = cp[2] - '1';
                   2593:                if (0 > i || 8 < i) {
                   2594:                        /* Not an argument invocation. */
                   2595:                        cp += 2;
                   2596:                        continue;
                   2597:                }
1.82      schwarze 2598:                *cp = '\0';
1.110     schwarze 2599:                buf->sz = mandoc_asprintf(&n2, "%s%s%s",
1.82      schwarze 2600:                    n1, arg[i], cp + 3) + 1;
1.16      schwarze 2601:                cp = n2 + (cp - n1);
                   2602:                free(n1);
                   2603:                n1 = n2;
1.12      schwarze 2604:        }
                   2605:
1.16      schwarze 2606:        /*
                   2607:         * Replace the macro invocation
                   2608:         * by the expanded macro.
                   2609:         */
1.110     schwarze 2610:        free(buf->buf);
                   2611:        buf->buf = n1;
                   2612:        if (buf->sz == 0)
                   2613:                buf->sz = strlen(buf->buf) + 1;
1.120     schwarze 2614:        *offs = 0;
1.16      schwarze 2615:
1.110     schwarze 2616:        return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.16      schwarze 2617:           ROFF_REPARSE : ROFF_APPEND);
1.12      schwarze 2618: }
1.28      schwarze 2619:
1.85      schwarze 2620: static size_t
1.28      schwarze 2621: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   2622: {
                   2623:        char     *name, *cp;
1.85      schwarze 2624:        size_t    namesz;
1.28      schwarze 2625:
                   2626:        name = *cpp;
                   2627:        if ('\0' == *name)
1.85      schwarze 2628:                return(0);
1.28      schwarze 2629:
1.85      schwarze 2630:        /* Read until end of name and terminate it with NUL. */
                   2631:        for (cp = name; 1; cp++) {
                   2632:                if ('\0' == *cp || ' ' == *cp) {
                   2633:                        namesz = cp - name;
                   2634:                        break;
                   2635:                }
1.28      schwarze 2636:                if ('\\' != *cp)
                   2637:                        continue;
1.88      schwarze 2638:                namesz = cp - name;
                   2639:                if ('{' == cp[1] || '}' == cp[1])
                   2640:                        break;
1.28      schwarze 2641:                cp++;
                   2642:                if ('\\' == *cp)
                   2643:                        continue;
1.97      schwarze 2644:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   2645:                    "%.*s", (int)(cp - name + 1), name);
1.85      schwarze 2646:                mandoc_escape((const char **)&cp, NULL, NULL);
                   2647:                break;
1.28      schwarze 2648:        }
                   2649:
                   2650:        /* Read past spaces. */
                   2651:        while (' ' == *cp)
                   2652:                cp++;
                   2653:
                   2654:        *cpp = cp;
1.85      schwarze 2655:        return(namesz);
1.28      schwarze 2656: }
                   2657:
1.16      schwarze 2658: /*
                   2659:  * Store *string into the user-defined string called *name.
                   2660:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66      schwarze 2661:  * append == 0: replace mode
                   2662:  * append == 1: single-line append mode
                   2663:  * append == 2: multiline append mode, append '\n' after each call
1.16      schwarze 2664:  */
1.8       schwarze 2665: static void
1.16      schwarze 2666: roff_setstr(struct roff *r, const char *name, const char *string,
1.66      schwarze 2667:        int append)
1.7       schwarze 2668: {
1.42      schwarze 2669:
                   2670:        roff_setstrn(&r->strtab, name, strlen(name), string,
1.80      schwarze 2671:            string ? strlen(string) : 0, append);
1.42      schwarze 2672: }
                   2673:
                   2674: static void
                   2675: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66      schwarze 2676:                const char *string, size_t stringsz, int append)
1.42      schwarze 2677: {
                   2678:        struct roffkv   *n;
                   2679:        char            *c;
                   2680:        int              i;
                   2681:        size_t           oldch, newch;
1.7       schwarze 2682:
1.16      schwarze 2683:        /* Search for an existing string with the same name. */
1.42      schwarze 2684:        n = *r;
                   2685:
1.84      schwarze 2686:        while (n && (namesz != n->key.sz ||
                   2687:                        strncmp(n->key.p, name, namesz)))
1.7       schwarze 2688:                n = n->next;
1.8       schwarze 2689:
                   2690:        if (NULL == n) {
1.16      schwarze 2691:                /* Create a new string table entry. */
1.42      schwarze 2692:                n = mandoc_malloc(sizeof(struct roffkv));
                   2693:                n->key.p = mandoc_strndup(name, namesz);
                   2694:                n->key.sz = namesz;
                   2695:                n->val.p = NULL;
                   2696:                n->val.sz = 0;
                   2697:                n->next = *r;
                   2698:                *r = n;
1.66      schwarze 2699:        } else if (0 == append) {
1.42      schwarze 2700:                free(n->val.p);
                   2701:                n->val.p = NULL;
                   2702:                n->val.sz = 0;
1.16      schwarze 2703:        }
                   2704:
                   2705:        if (NULL == string)
                   2706:                return;
                   2707:
                   2708:        /*
                   2709:         * One additional byte for the '\n' in multiline mode,
                   2710:         * and one for the terminating '\0'.
                   2711:         */
1.66      schwarze 2712:        newch = stringsz + (1 < append ? 2u : 1u);
1.42      schwarze 2713:
                   2714:        if (NULL == n->val.p) {
                   2715:                n->val.p = mandoc_malloc(newch);
                   2716:                *n->val.p = '\0';
1.16      schwarze 2717:                oldch = 0;
                   2718:        } else {
1.42      schwarze 2719:                oldch = n->val.sz;
                   2720:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 2721:        }
                   2722:
                   2723:        /* Skip existing content in the destination buffer. */
1.42      schwarze 2724:        c = n->val.p + (int)oldch;
1.16      schwarze 2725:
                   2726:        /* Append new content to the destination buffer. */
1.42      schwarze 2727:        i = 0;
                   2728:        while (i < (int)stringsz) {
1.16      schwarze 2729:                /*
                   2730:                 * Rudimentary roff copy mode:
                   2731:                 * Handle escaped backslashes.
                   2732:                 */
1.42      schwarze 2733:                if ('\\' == string[i] && '\\' == string[i + 1])
                   2734:                        i++;
                   2735:                *c++ = string[i++];
1.16      schwarze 2736:        }
1.8       schwarze 2737:
1.16      schwarze 2738:        /* Append terminating bytes. */
1.66      schwarze 2739:        if (1 < append)
1.16      schwarze 2740:                *c++ = '\n';
1.42      schwarze 2741:
1.16      schwarze 2742:        *c = '\0';
1.42      schwarze 2743:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 2744: }
                   2745:
1.8       schwarze 2746: static const char *
                   2747: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7       schwarze 2748: {
1.42      schwarze 2749:        const struct roffkv *n;
1.64      schwarze 2750:        int i;
1.7       schwarze 2751:
1.42      schwarze 2752:        for (n = r->strtab; n; n = n->next)
1.80      schwarze 2753:                if (0 == strncmp(name, n->key.p, len) &&
                   2754:                    '\0' == n->key.p[(int)len])
1.42      schwarze 2755:                        return(n->val.p);
1.64      schwarze 2756:
                   2757:        for (i = 0; i < PREDEFS_MAX; i++)
                   2758:                if (0 == strncmp(name, predefs[i].name, len) &&
                   2759:                                '\0' == predefs[i].name[(int)len])
                   2760:                        return(predefs[i].str);
1.8       schwarze 2761:
1.42      schwarze 2762:        return(NULL);
1.7       schwarze 2763: }
                   2764:
1.8       schwarze 2765: static void
1.42      schwarze 2766: roff_freestr(struct roffkv *r)
1.7       schwarze 2767: {
1.42      schwarze 2768:        struct roffkv    *n, *nn;
1.7       schwarze 2769:
1.42      schwarze 2770:        for (n = r; n; n = nn) {
                   2771:                free(n->key.p);
                   2772:                free(n->val.p);
1.7       schwarze 2773:                nn = n->next;
                   2774:                free(n);
                   2775:        }
1.27      schwarze 2776: }
                   2777:
                   2778: const struct tbl_span *
                   2779: roff_span(const struct roff *r)
                   2780: {
1.80      schwarze 2781:
1.27      schwarze 2782:        return(r->tbl ? tbl_span(r->tbl) : NULL);
1.32      schwarze 2783: }
                   2784:
                   2785: const struct eqn *
                   2786: roff_eqn(const struct roff *r)
                   2787: {
1.80      schwarze 2788:
1.32      schwarze 2789:        return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.42      schwarze 2790: }
                   2791:
                   2792: /*
                   2793:  * Duplicate an input string, making the appropriate character
                   2794:  * conversations (as stipulated by `tr') along the way.
                   2795:  * Returns a heap-allocated string with all the replacements made.
                   2796:  */
                   2797: char *
                   2798: roff_strdup(const struct roff *r, const char *p)
                   2799: {
                   2800:        const struct roffkv *cp;
                   2801:        char            *res;
                   2802:        const char      *pp;
                   2803:        size_t           ssz, sz;
                   2804:        enum mandoc_esc  esc;
                   2805:
                   2806:        if (NULL == r->xmbtab && NULL == r->xtab)
                   2807:                return(mandoc_strdup(p));
                   2808:        else if ('\0' == *p)
                   2809:                return(mandoc_strdup(""));
                   2810:
                   2811:        /*
                   2812:         * Step through each character looking for term matches
                   2813:         * (remember that a `tr' can be invoked with an escape, which is
                   2814:         * a glyph but the escape is multi-character).
                   2815:         * We only do this if the character hash has been initialised
                   2816:         * and the string is >0 length.
                   2817:         */
                   2818:
                   2819:        res = NULL;
                   2820:        ssz = 0;
                   2821:
                   2822:        while ('\0' != *p) {
                   2823:                if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
                   2824:                        sz = r->xtab[(int)*p].sz;
                   2825:                        res = mandoc_realloc(res, ssz + sz + 1);
                   2826:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   2827:                        ssz += sz;
                   2828:                        p++;
                   2829:                        continue;
                   2830:                } else if ('\\' != *p) {
                   2831:                        res = mandoc_realloc(res, ssz + 2);
                   2832:                        res[ssz++] = *p++;
                   2833:                        continue;
                   2834:                }
                   2835:
                   2836:                /* Search for term matches. */
                   2837:                for (cp = r->xmbtab; cp; cp = cp->next)
                   2838:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   2839:                                break;
                   2840:
                   2841:                if (NULL != cp) {
                   2842:                        /*
                   2843:                         * A match has been found.
                   2844:                         * Append the match to the array and move
                   2845:                         * forward by its keysize.
                   2846:                         */
1.80      schwarze 2847:                        res = mandoc_realloc(res,
                   2848:                            ssz + cp->val.sz + 1);
1.42      schwarze 2849:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   2850:                        ssz += cp->val.sz;
                   2851:                        p += (int)cp->key.sz;
                   2852:                        continue;
                   2853:                }
                   2854:
                   2855:                /*
                   2856:                 * Handle escapes carefully: we need to copy
                   2857:                 * over just the escape itself, or else we might
                   2858:                 * do replacements within the escape itself.
                   2859:                 * Make sure to pass along the bogus string.
                   2860:                 */
                   2861:                pp = p++;
                   2862:                esc = mandoc_escape(&p, NULL, NULL);
                   2863:                if (ESCAPE_ERROR == esc) {
                   2864:                        sz = strlen(pp);
                   2865:                        res = mandoc_realloc(res, ssz + sz + 1);
                   2866:                        memcpy(res + ssz, pp, sz);
                   2867:                        break;
                   2868:                }
1.80      schwarze 2869:                /*
                   2870:                 * We bail out on bad escapes.
1.42      schwarze 2871:                 * No need to warn: we already did so when
                   2872:                 * roff_res() was called.
                   2873:                 */
                   2874:                sz = (int)(p - pp);
                   2875:                res = mandoc_realloc(res, ssz + sz + 1);
                   2876:                memcpy(res + ssz, pp, sz);
                   2877:                ssz += sz;
                   2878:        }
                   2879:
                   2880:        res[(int)ssz] = '\0';
                   2881:        return(res);
1.99      schwarze 2882: }
                   2883:
                   2884: int
                   2885: roff_getformat(const struct roff *r)
                   2886: {
                   2887:
                   2888:        return(r->format);
1.48      schwarze 2889: }
                   2890:
                   2891: /*
1.80      schwarze 2892:  * Find out whether a line is a macro line or not.
1.48      schwarze 2893:  * If it is, adjust the current position and return one; if it isn't,
                   2894:  * return zero and don't change the current position.
                   2895:  * If the control character has been set with `.cc', then let that grain
                   2896:  * precedence.
                   2897:  * This is slighly contrary to groff, where using the non-breaking
                   2898:  * control character when `cc' has been invoked will cause the
                   2899:  * non-breaking macro contents to be printed verbatim.
                   2900:  */
                   2901: int
                   2902: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   2903: {
                   2904:        int             pos;
                   2905:
                   2906:        pos = *ppos;
                   2907:
                   2908:        if (0 != r->control && cp[pos] == r->control)
                   2909:                pos++;
                   2910:        else if (0 != r->control)
                   2911:                return(0);
                   2912:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   2913:                pos += 2;
                   2914:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   2915:                pos++;
                   2916:        else
                   2917:                return(0);
                   2918:
                   2919:        while (' ' == cp[pos] || '\t' == cp[pos])
                   2920:                pos++;
                   2921:
                   2922:        *ppos = pos;
                   2923:        return(1);
1.1       schwarze 2924: }