[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 1.137

1.137   ! schwarze    1: /*     $OpenBSD: roff.c,v 1.136 2015/04/04 13:52:59 schwarze Exp $ */
1.1       schwarze    2: /*
1.137   ! schwarze    3:  * Copyright (c) 2009-2012, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
1.119     schwarze    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"
1.137   ! schwarze   29: #include "roff.h"
1.98      daniel     30: #include "libmandoc.h"
1.27      schwarze   31: #include "libroff.h"
1.1       schwarze   32:
1.37      schwarze   33: /* Maximum number of nested if-else conditionals. */
1.2       schwarze   34: #define        RSTACK_MAX      128
                     35:
1.43      schwarze   36: /* Maximum number of string expansions per line, to break infinite loops. */
                     37: #define        EXPAND_LIMIT    1000
                     38:
1.1       schwarze   39: enum   rofft {
1.123     schwarze   40:        ROFF_ab,
1.20      schwarze   41:        ROFF_ad,
1.123     schwarze   42:        ROFF_af,
                     43:        ROFF_aln,
                     44:        ROFF_als,
1.2       schwarze   45:        ROFF_am,
1.123     schwarze   46:        ROFF_am1,
1.2       schwarze   47:        ROFF_ami,
1.123     schwarze   48:        ROFF_ami1,
1.66      schwarze   49:        ROFF_as,
1.123     schwarze   50:        ROFF_as1,
                     51:        ROFF_asciify,
                     52:        ROFF_backtrace,
                     53:        ROFF_bd,
                     54:        ROFF_bleedat,
                     55:        ROFF_blm,
                     56:        ROFF_box,
                     57:        ROFF_boxa,
                     58:        ROFF_bp,
                     59:        ROFF_BP,
                     60:        /* MAN_br, MDOC_br */
                     61:        ROFF_break,
                     62:        ROFF_breakchar,
                     63:        ROFF_brnl,
                     64:        ROFF_brp,
                     65:        ROFF_brpnl,
                     66:        ROFF_c2,
1.48      schwarze   67:        ROFF_cc,
1.67      schwarze   68:        ROFF_ce,
1.123     schwarze   69:        ROFF_cf,
                     70:        ROFF_cflags,
                     71:        ROFF_ch,
                     72:        ROFF_char,
                     73:        ROFF_chop,
                     74:        ROFF_class,
                     75:        ROFF_close,
                     76:        ROFF_CL,
                     77:        ROFF_color,
                     78:        ROFF_composite,
                     79:        ROFF_continue,
                     80:        ROFF_cp,
                     81:        ROFF_cropat,
                     82:        ROFF_cs,
                     83:        ROFF_cu,
                     84:        ROFF_da,
                     85:        ROFF_dch,
                     86:        ROFF_Dd,
1.1       schwarze   87:        ROFF_de,
1.123     schwarze   88:        ROFF_de1,
                     89:        ROFF_defcolor,
1.1       schwarze   90:        ROFF_dei,
1.123     schwarze   91:        ROFF_dei1,
                     92:        ROFF_device,
                     93:        ROFF_devicem,
                     94:        ROFF_di,
                     95:        ROFF_do,
1.2       schwarze   96:        ROFF_ds,
1.123     schwarze   97:        ROFF_ds1,
                     98:        ROFF_dwh,
                     99:        ROFF_dt,
                    100:        ROFF_ec,
                    101:        ROFF_ecr,
                    102:        ROFF_ecs,
1.2       schwarze  103:        ROFF_el,
1.123     schwarze  104:        ROFF_em,
                    105:        ROFF_EN,
                    106:        ROFF_eo,
                    107:        ROFF_EP,
                    108:        ROFF_EQ,
                    109:        ROFF_errprint,
                    110:        ROFF_ev,
                    111:        ROFF_evc,
                    112:        ROFF_ex,
                    113:        ROFF_fallback,
1.58      schwarze  114:        ROFF_fam,
1.123     schwarze  115:        ROFF_fc,
                    116:        ROFF_fchar,
                    117:        ROFF_fcolor,
                    118:        ROFF_fdeferlig,
                    119:        ROFF_feature,
                    120:        /* MAN_fi; ignored in mdoc(7) */
                    121:        ROFF_fkern,
                    122:        ROFF_fl,
                    123:        ROFF_flig,
                    124:        ROFF_fp,
                    125:        ROFF_fps,
                    126:        ROFF_fschar,
                    127:        ROFF_fspacewidth,
                    128:        ROFF_fspecial,
                    129:        /* MAN_ft; ignored in mdoc(7) */
                    130:        ROFF_ftr,
                    131:        ROFF_fzoom,
                    132:        ROFF_gcolor,
                    133:        ROFF_hc,
                    134:        ROFF_hcode,
                    135:        ROFF_hidechar,
                    136:        ROFF_hla,
                    137:        ROFF_hlm,
                    138:        ROFF_hpf,
                    139:        ROFF_hpfa,
                    140:        ROFF_hpfcode,
1.59      schwarze  141:        ROFF_hw,
1.20      schwarze  142:        ROFF_hy,
1.123     schwarze  143:        ROFF_hylang,
                    144:        ROFF_hylen,
                    145:        ROFF_hym,
                    146:        ROFF_hypp,
                    147:        ROFF_hys,
1.2       schwarze  148:        ROFF_ie,
                    149:        ROFF_if,
1.1       schwarze  150:        ROFF_ig,
1.123     schwarze  151:        /* MAN_in; ignored in mdoc(7) */
                    152:        ROFF_index,
1.30      schwarze  153:        ROFF_it,
1.123     schwarze  154:        ROFF_itc,
                    155:        ROFF_IX,
                    156:        ROFF_kern,
                    157:        ROFF_kernafter,
                    158:        ROFF_kernbefore,
                    159:        ROFF_kernpair,
                    160:        ROFF_lc,
                    161:        ROFF_lc_ctype,
                    162:        ROFF_lds,
                    163:        ROFF_length,
                    164:        ROFF_letadj,
                    165:        ROFF_lf,
                    166:        ROFF_lg,
                    167:        ROFF_lhang,
                    168:        ROFF_linetabs,
                    169:        /* MAN_ll, MDOC_ll */
                    170:        ROFF_lnr,
                    171:        ROFF_lnrf,
                    172:        ROFF_lpfx,
                    173:        ROFF_ls,
                    174:        ROFF_lsm,
                    175:        ROFF_lt,
                    176:        ROFF_mc,
                    177:        ROFF_mediasize,
                    178:        ROFF_minss,
                    179:        ROFF_mk,
                    180:        ROFF_mso,
1.128     schwarze  181:        ROFF_na,
1.20      schwarze  182:        ROFF_ne,
1.123     schwarze  183:        /* MAN_nf; ignored in mdoc(7) */
1.20      schwarze  184:        ROFF_nh,
1.123     schwarze  185:        ROFF_nhychar,
                    186:        ROFF_nm,
                    187:        ROFF_nn,
                    188:        ROFF_nop,
1.14      schwarze  189:        ROFF_nr,
1.123     schwarze  190:        ROFF_nrf,
                    191:        ROFF_nroff,
1.31      schwarze  192:        ROFF_ns,
1.123     schwarze  193:        ROFF_nx,
                    194:        ROFF_open,
                    195:        ROFF_opena,
                    196:        ROFF_os,
                    197:        ROFF_output,
                    198:        ROFF_padj,
                    199:        ROFF_papersize,
                    200:        ROFF_pc,
                    201:        ROFF_pev,
                    202:        ROFF_pi,
                    203:        ROFF_PI,
1.101     schwarze  204:        ROFF_pl,
1.123     schwarze  205:        ROFF_pm,
                    206:        ROFF_pn,
                    207:        ROFF_pnr,
                    208:        ROFF_po,
1.31      schwarze  209:        ROFF_ps,
1.123     schwarze  210:        ROFF_psbb,
                    211:        ROFF_pshape,
                    212:        ROFF_pso,
                    213:        ROFF_ptr,
                    214:        ROFF_pvs,
                    215:        ROFF_rchar,
                    216:        ROFF_rd,
                    217:        ROFF_recursionlimit,
                    218:        ROFF_return,
                    219:        ROFF_rfschar,
                    220:        ROFF_rhang,
                    221:        ROFF_rj,
1.2       schwarze  222:        ROFF_rm,
1.123     schwarze  223:        ROFF_rn,
                    224:        ROFF_rnn,
1.76      schwarze  225:        ROFF_rr,
1.123     schwarze  226:        ROFF_rs,
                    227:        ROFF_rt,
                    228:        ROFF_schar,
                    229:        ROFF_sentchar,
                    230:        ROFF_shc,
                    231:        ROFF_shift,
                    232:        ROFF_sizes,
1.14      schwarze  233:        ROFF_so,
1.123     schwarze  234:        /* MAN_sp, MDOC_sp */
                    235:        ROFF_spacewidth,
                    236:        ROFF_special,
                    237:        ROFF_spreadwarn,
                    238:        ROFF_ss,
                    239:        ROFF_sty,
                    240:        ROFF_substring,
                    241:        ROFF_sv,
                    242:        ROFF_sy,
                    243:        ROFF_T_,
1.31      schwarze  244:        ROFF_ta,
1.123     schwarze  245:        ROFF_tc,
                    246:        ROFF_TE,
                    247:        ROFF_TH,
                    248:        ROFF_ti,
                    249:        ROFF_tkf,
                    250:        ROFF_tl,
                    251:        ROFF_tm,
                    252:        ROFF_tm1,
                    253:        ROFF_tmc,
1.2       schwarze  254:        ROFF_tr,
1.123     schwarze  255:        ROFF_track,
                    256:        ROFF_transchar,
                    257:        ROFF_trf,
                    258:        ROFF_trimat,
                    259:        ROFF_trin,
                    260:        ROFF_trnt,
                    261:        ROFF_troff,
1.27      schwarze  262:        ROFF_TS,
1.123     schwarze  263:        ROFF_uf,
                    264:        ROFF_ul,
                    265:        ROFF_unformat,
                    266:        ROFF_unwatch,
                    267:        ROFF_unwatchn,
                    268:        ROFF_vpt,
                    269:        ROFF_vs,
                    270:        ROFF_warn,
                    271:        ROFF_warnscale,
                    272:        ROFF_watch,
                    273:        ROFF_watchlength,
                    274:        ROFF_watchn,
                    275:        ROFF_wh,
                    276:        ROFF_while,
                    277:        ROFF_write,
                    278:        ROFF_writec,
                    279:        ROFF_writem,
                    280:        ROFF_xflag,
1.2       schwarze  281:        ROFF_cblock,
1.16      schwarze  282:        ROFF_USERDEF,
1.1       schwarze  283:        ROFF_MAX
                    284: };
                    285:
1.41      schwarze  286: /*
1.42      schwarze  287:  * An incredibly-simple string buffer.
                    288:  */
1.8       schwarze  289: struct roffstr {
1.42      schwarze  290:        char            *p; /* nil-terminated buffer */
                    291:        size_t           sz; /* saved strlen(p) */
                    292: };
                    293:
                    294: /*
                    295:  * A key-value roffstr pair as part of a singly-linked list.
                    296:  */
                    297: struct roffkv {
                    298:        struct roffstr   key;
                    299:        struct roffstr   val;
                    300:        struct roffkv   *next; /* next in list */
1.8       schwarze  301: };
                    302:
1.52      schwarze  303: /*
                    304:  * A single number register as part of a singly-linked list.
                    305:  */
                    306: struct roffreg {
                    307:        struct roffstr   key;
1.53      schwarze  308:        int              val;
1.52      schwarze  309:        struct roffreg  *next;
                    310: };
                    311:
1.1       schwarze  312: struct roff {
1.35      schwarze  313:        struct mparse   *parse; /* parse point */
1.109     schwarze  314:        const struct mchars *mchars; /* character table */
1.1       schwarze  315:        struct roffnode *last; /* leaf of stack */
1.96      schwarze  316:        int             *rstack; /* stack of inverted `ie' values */
1.52      schwarze  317:        struct roffreg  *regtab; /* number registers */
1.42      schwarze  318:        struct roffkv   *strtab; /* user-defined strings & macros */
                    319:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                    320:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.16      schwarze  321:        const char      *current_string; /* value of last called user macro */
1.27      schwarze  322:        struct tbl_node *first_tbl; /* first table parsed */
                    323:        struct tbl_node *last_tbl; /* last table parsed */
                    324:        struct tbl_node *tbl; /* current table being parsed */
1.32      schwarze  325:        struct eqn_node *last_eqn; /* last equation parsed */
                    326:        struct eqn_node *first_eqn; /* first equation parsed */
                    327:        struct eqn_node *eqn; /* current equation being parsed */
1.102     schwarze  328:        int              eqn_inline; /* current equation is inline */
1.96      schwarze  329:        int              options; /* parse options */
                    330:        int              rstacksz; /* current size limit of rstack */
                    331:        int              rstackpos; /* position in rstack */
1.99      schwarze  332:        int              format; /* current file in mdoc or man format */
1.96      schwarze  333:        char             control; /* control character */
1.1       schwarze  334: };
                    335:
                    336: struct roffnode {
                    337:        enum rofft       tok; /* type of node */
                    338:        struct roffnode *parent; /* up one in stack */
                    339:        int              line; /* parse line */
                    340:        int              col; /* parse col */
1.16      schwarze  341:        char            *name; /* node name, e.g. macro name */
1.2       schwarze  342:        char            *end; /* end-rules: custom token */
                    343:        int              endspan; /* end-rules: next-line or infty */
1.71      schwarze  344:        int              rule; /* current evaluation rule */
1.1       schwarze  345: };
                    346:
                    347: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
                    348:                         enum rofft tok, /* tok of macro */ \
1.110     schwarze  349:                         struct buf *buf, /* input buffer */ \
1.1       schwarze  350:                         int ln, /* parse line */ \
1.2       schwarze  351:                         int ppos, /* original pos in buffer */ \
                    352:                         int pos, /* current pos in buffer */ \
                    353:                         int *offs /* reset offset of buffer data */
1.1       schwarze  354:
                    355: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    356:
                    357: struct roffmac {
                    358:        const char      *name; /* macro name */
1.2       schwarze  359:        roffproc         proc; /* process new macro */
                    360:        roffproc         text; /* process as child text of macro */
                    361:        roffproc         sub; /* process as child of macro */
                    362:        int              flags;
                    363: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.3       schwarze  364:        struct roffmac  *next;
1.1       schwarze  365: };
                    366:
1.37      schwarze  367: struct predef {
                    368:        const char      *name; /* predefined input name */
                    369:        const char      *str; /* replacement symbol */
                    370: };
                    371:
                    372: #define        PREDEF(__name, __str) \
                    373:        { (__name), (__str) },
                    374:
1.42      schwarze  375: static enum rofft       roffhash_find(const char *, size_t);
                    376: static void             roffhash_init(void);
                    377: static void             roffnode_cleanscope(struct roff *);
                    378: static void             roffnode_pop(struct roff *);
                    379: static void             roffnode_push(struct roff *, enum rofft,
                    380:                                const char *, int, int);
1.2       schwarze  381: static enum rofferr     roff_block(ROFF_ARGS);
                    382: static enum rofferr     roff_block_text(ROFF_ARGS);
                    383: static enum rofferr     roff_block_sub(ROFF_ARGS);
1.123     schwarze  384: static enum rofferr     roff_brp(ROFF_ARGS);
1.2       schwarze  385: static enum rofferr     roff_cblock(ROFF_ARGS);
1.48      schwarze  386: static enum rofferr     roff_cc(ROFF_ARGS);
1.68      schwarze  387: static void             roff_ccond(struct roff *, int, int);
1.2       schwarze  388: static enum rofferr     roff_cond(ROFF_ARGS);
                    389: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    390: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  391: static enum rofferr     roff_ds(ROFF_ARGS);
1.110     schwarze  392: static enum rofferr     roff_eqndelim(struct roff *, struct buf *, int);
1.106     schwarze  393: static int              roff_evalcond(struct roff *r, int,
                    394:                                const char *, int *);
                    395: static int              roff_evalnum(struct roff *, int,
                    396:                                const char *, int *, int *, int);
                    397: static int              roff_evalpar(struct roff *, int,
1.133     schwarze  398:                                const char *, int *, int *, int);
1.71      schwarze  399: static int              roff_evalstrcond(const char *, int *);
1.42      schwarze  400: static void             roff_free1(struct roff *);
1.52      schwarze  401: static void             roff_freereg(struct roffreg *);
1.42      schwarze  402: static void             roff_freestr(struct roffkv *);
1.85      schwarze  403: static size_t           roff_getname(struct roff *, char **, int, int);
1.133     schwarze  404: static int              roff_getnum(const char *, int *, int *, int);
1.56      schwarze  405: static int              roff_getop(const char *, int *, char *);
1.53      schwarze  406: static int              roff_getregn(const struct roff *,
                    407:                                const char *, size_t);
1.65      schwarze  408: static int              roff_getregro(const char *name);
1.80      schwarze  409: static const char      *roff_getstrn(const struct roff *,
1.8       schwarze  410:                                const char *, size_t);
1.123     schwarze  411: static enum rofferr     roff_insec(ROFF_ARGS);
1.51      schwarze  412: static enum rofferr     roff_it(ROFF_ARGS);
1.21      schwarze  413: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.137   ! schwarze  414: static void             roff_man_alloc1(struct roff_man *);
        !           415: static void             roff_man_free1(struct roff_man *);
1.6       schwarze  416: static enum rofferr     roff_nr(ROFF_ARGS);
1.87      schwarze  417: static enum rofft       roff_parse(struct roff *, char *, int *,
                    418:                                int, int);
1.110     schwarze  419: static enum rofferr     roff_parsetext(struct buf *, int, int *);
                    420: static enum rofferr     roff_res(struct roff *, struct buf *, int, int);
1.29      schwarze  421: static enum rofferr     roff_rm(ROFF_ARGS);
1.76      schwarze  422: static enum rofferr     roff_rr(ROFF_ARGS);
1.8       schwarze  423: static void             roff_setstr(struct roff *,
1.16      schwarze  424:                                const char *, const char *, int);
1.80      schwarze  425: static void             roff_setstrn(struct roffkv **, const char *,
1.42      schwarze  426:                                size_t, const char *, size_t, int);
1.14      schwarze  427: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  428: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  429: static enum rofferr     roff_Dd(ROFF_ARGS);
                    430: static enum rofferr     roff_TH(ROFF_ARGS);
1.27      schwarze  431: static enum rofferr     roff_TE(ROFF_ARGS);
                    432: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  433: static enum rofferr     roff_EQ(ROFF_ARGS);
                    434: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  435: static enum rofferr     roff_T_(ROFF_ARGS);
1.123     schwarze  436: static enum rofferr     roff_unsupp(ROFF_ARGS);
1.16      schwarze  437: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  438:
1.42      schwarze  439: /* See roffhash_find() */
1.3       schwarze  440:
                    441: #define        ASCII_HI         126
                    442: #define        ASCII_LO         33
                    443: #define        HASHWIDTH       (ASCII_HI - ASCII_LO + 1)
                    444:
1.133     schwarze  445: #define        ROFFNUM_SCALE   (1 << 0)  /* Honour scaling in roff_getnum(). */
                    446: #define        ROFFNUM_WHITE   (1 << 1)  /* Skip whitespace in roff_evalnum(). */
                    447:
1.3       schwarze  448: static struct roffmac  *hash[HASHWIDTH];
                    449:
                    450: static struct roffmac   roffs[ROFF_MAX] = {
1.123     schwarze  451:        { "ab", roff_unsupp, NULL, NULL, 0, NULL },
1.21      schwarze  452:        { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  453:        { "af", roff_line_ignore, NULL, NULL, 0, NULL },
                    454:        { "aln", roff_unsupp, NULL, NULL, 0, NULL },
                    455:        { "als", roff_unsupp, NULL, NULL, 0, NULL },
1.3       schwarze  456:        { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  457:        { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.3       schwarze  458:        { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  459:        { "ami1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.66      schwarze  460:        { "as", roff_ds, NULL, NULL, 0, NULL },
1.123     schwarze  461:        { "as1", roff_ds, NULL, NULL, 0, NULL },
                    462:        { "asciify", roff_unsupp, NULL, NULL, 0, NULL },
                    463:        { "backtrace", roff_line_ignore, NULL, NULL, 0, NULL },
                    464:        { "bd", roff_line_ignore, NULL, NULL, 0, NULL },
                    465:        { "bleedat", roff_line_ignore, NULL, NULL, 0, NULL },
                    466:        { "blm", roff_unsupp, NULL, NULL, 0, NULL },
                    467:        { "box", roff_unsupp, NULL, NULL, 0, NULL },
                    468:        { "boxa", roff_unsupp, NULL, NULL, 0, NULL },
                    469:        { "bp", roff_line_ignore, NULL, NULL, 0, NULL },
                    470:        { "BP", roff_unsupp, NULL, NULL, 0, NULL },
                    471:        { "break", roff_unsupp, NULL, NULL, 0, NULL },
                    472:        { "breakchar", roff_line_ignore, NULL, NULL, 0, NULL },
                    473:        { "brnl", roff_line_ignore, NULL, NULL, 0, NULL },
                    474:        { "brp", roff_brp, NULL, NULL, 0, NULL },
                    475:        { "brpnl", roff_line_ignore, NULL, NULL, 0, NULL },
                    476:        { "c2", roff_unsupp, NULL, NULL, 0, NULL },
1.48      schwarze  477:        { "cc", roff_cc, NULL, NULL, 0, NULL },
1.67      schwarze  478:        { "ce", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  479:        { "cf", roff_insec, NULL, NULL, 0, NULL },
                    480:        { "cflags", roff_line_ignore, NULL, NULL, 0, NULL },
                    481:        { "ch", roff_line_ignore, NULL, NULL, 0, NULL },
                    482:        { "char", roff_unsupp, NULL, NULL, 0, NULL },
                    483:        { "chop", roff_unsupp, NULL, NULL, 0, NULL },
                    484:        { "class", roff_line_ignore, NULL, NULL, 0, NULL },
                    485:        { "close", roff_insec, NULL, NULL, 0, NULL },
                    486:        { "CL", roff_unsupp, NULL, NULL, 0, NULL },
                    487:        { "color", roff_line_ignore, NULL, NULL, 0, NULL },
                    488:        { "composite", roff_unsupp, NULL, NULL, 0, NULL },
                    489:        { "continue", roff_unsupp, NULL, NULL, 0, NULL },
                    490:        { "cp", roff_line_ignore, NULL, NULL, 0, NULL },
                    491:        { "cropat", roff_line_ignore, NULL, NULL, 0, NULL },
                    492:        { "cs", roff_line_ignore, NULL, NULL, 0, NULL },
                    493:        { "cu", roff_line_ignore, NULL, NULL, 0, NULL },
                    494:        { "da", roff_unsupp, NULL, NULL, 0, NULL },
                    495:        { "dch", roff_unsupp, NULL, NULL, 0, NULL },
                    496:        { "Dd", roff_Dd, NULL, NULL, 0, NULL },
1.3       schwarze  497:        { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  498:        { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    499:        { "defcolor", roff_line_ignore, NULL, NULL, 0, NULL },
1.3       schwarze  500:        { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  501:        { "dei1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    502:        { "device", roff_unsupp, NULL, NULL, 0, NULL },
                    503:        { "devicem", roff_unsupp, NULL, NULL, 0, NULL },
                    504:        { "di", roff_unsupp, NULL, NULL, 0, NULL },
                    505:        { "do", roff_unsupp, NULL, NULL, 0, NULL },
1.7       schwarze  506:        { "ds", roff_ds, NULL, NULL, 0, NULL },
1.123     schwarze  507:        { "ds1", roff_ds, NULL, NULL, 0, NULL },
                    508:        { "dwh", roff_unsupp, NULL, NULL, 0, NULL },
                    509:        { "dt", roff_unsupp, NULL, NULL, 0, NULL },
                    510:        { "ec", roff_unsupp, NULL, NULL, 0, NULL },
                    511:        { "ecr", roff_unsupp, NULL, NULL, 0, NULL },
                    512:        { "ecs", roff_unsupp, NULL, NULL, 0, NULL },
1.3       schwarze  513:        { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.123     schwarze  514:        { "em", roff_unsupp, NULL, NULL, 0, NULL },
                    515:        { "EN", roff_EN, NULL, NULL, 0, NULL },
                    516:        { "eo", roff_unsupp, NULL, NULL, 0, NULL },
                    517:        { "EP", roff_unsupp, NULL, NULL, 0, NULL },
                    518:        { "EQ", roff_EQ, NULL, NULL, 0, NULL },
                    519:        { "errprint", roff_line_ignore, NULL, NULL, 0, NULL },
                    520:        { "ev", roff_unsupp, NULL, NULL, 0, NULL },
                    521:        { "evc", roff_unsupp, NULL, NULL, 0, NULL },
                    522:        { "ex", roff_unsupp, NULL, NULL, 0, NULL },
                    523:        { "fallback", roff_line_ignore, NULL, NULL, 0, NULL },
1.58      schwarze  524:        { "fam", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  525:        { "fc", roff_unsupp, NULL, NULL, 0, NULL },
                    526:        { "fchar", roff_unsupp, NULL, NULL, 0, NULL },
                    527:        { "fcolor", roff_line_ignore, NULL, NULL, 0, NULL },
                    528:        { "fdeferlig", roff_line_ignore, NULL, NULL, 0, NULL },
                    529:        { "feature", roff_line_ignore, NULL, NULL, 0, NULL },
                    530:        { "fkern", roff_line_ignore, NULL, NULL, 0, NULL },
                    531:        { "fl", roff_line_ignore, NULL, NULL, 0, NULL },
                    532:        { "flig", roff_line_ignore, NULL, NULL, 0, NULL },
                    533:        { "fp", roff_line_ignore, NULL, NULL, 0, NULL },
                    534:        { "fps", roff_line_ignore, NULL, NULL, 0, NULL },
                    535:        { "fschar", roff_unsupp, NULL, NULL, 0, NULL },
                    536:        { "fspacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
                    537:        { "fspecial", roff_line_ignore, NULL, NULL, 0, NULL },
                    538:        { "ftr", roff_line_ignore, NULL, NULL, 0, NULL },
                    539:        { "fzoom", roff_line_ignore, NULL, NULL, 0, NULL },
                    540:        { "gcolor", roff_line_ignore, NULL, NULL, 0, NULL },
                    541:        { "hc", roff_line_ignore, NULL, NULL, 0, NULL },
                    542:        { "hcode", roff_line_ignore, NULL, NULL, 0, NULL },
                    543:        { "hidechar", roff_line_ignore, NULL, NULL, 0, NULL },
                    544:        { "hla", roff_line_ignore, NULL, NULL, 0, NULL },
                    545:        { "hlm", roff_line_ignore, NULL, NULL, 0, NULL },
                    546:        { "hpf", roff_line_ignore, NULL, NULL, 0, NULL },
                    547:        { "hpfa", roff_line_ignore, NULL, NULL, 0, NULL },
                    548:        { "hpfcode", roff_line_ignore, NULL, NULL, 0, NULL },
1.59      schwarze  549:        { "hw", roff_line_ignore, NULL, NULL, 0, NULL },
1.21      schwarze  550:        { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  551:        { "hylang", roff_line_ignore, NULL, NULL, 0, NULL },
                    552:        { "hylen", roff_line_ignore, NULL, NULL, 0, NULL },
                    553:        { "hym", roff_line_ignore, NULL, NULL, 0, NULL },
                    554:        { "hypp", roff_line_ignore, NULL, NULL, 0, NULL },
                    555:        { "hys", roff_line_ignore, NULL, NULL, 0, NULL },
1.3       schwarze  556:        { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    557:        { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    558:        { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.123     schwarze  559:        { "index", roff_unsupp, NULL, NULL, 0, NULL },
1.51      schwarze  560:        { "it", roff_it, NULL, NULL, 0, NULL },
1.123     schwarze  561:        { "itc", roff_unsupp, NULL, NULL, 0, NULL },
                    562:        { "IX", roff_line_ignore, NULL, NULL, 0, NULL },
                    563:        { "kern", roff_line_ignore, NULL, NULL, 0, NULL },
                    564:        { "kernafter", roff_line_ignore, NULL, NULL, 0, NULL },
                    565:        { "kernbefore", roff_line_ignore, NULL, NULL, 0, NULL },
                    566:        { "kernpair", roff_line_ignore, NULL, NULL, 0, NULL },
                    567:        { "lc", roff_unsupp, NULL, NULL, 0, NULL },
                    568:        { "lc_ctype", roff_unsupp, NULL, NULL, 0, NULL },
                    569:        { "lds", roff_unsupp, NULL, NULL, 0, NULL },
                    570:        { "length", roff_unsupp, NULL, NULL, 0, NULL },
                    571:        { "letadj", roff_line_ignore, NULL, NULL, 0, NULL },
                    572:        { "lf", roff_insec, NULL, NULL, 0, NULL },
                    573:        { "lg", roff_line_ignore, NULL, NULL, 0, NULL },
                    574:        { "lhang", roff_line_ignore, NULL, NULL, 0, NULL },
                    575:        { "linetabs", roff_unsupp, NULL, NULL, 0, NULL },
                    576:        { "lnr", roff_unsupp, NULL, NULL, 0, NULL },
                    577:        { "lnrf", roff_unsupp, NULL, NULL, 0, NULL },
                    578:        { "lpfx", roff_unsupp, NULL, NULL, 0, NULL },
                    579:        { "ls", roff_line_ignore, NULL, NULL, 0, NULL },
                    580:        { "lsm", roff_unsupp, NULL, NULL, 0, NULL },
                    581:        { "lt", roff_line_ignore, NULL, NULL, 0, NULL },
                    582:        { "mc", roff_line_ignore, NULL, NULL, 0, NULL },
                    583:        { "mediasize", roff_line_ignore, NULL, NULL, 0, NULL },
                    584:        { "minss", roff_line_ignore, NULL, NULL, 0, NULL },
                    585:        { "mk", roff_line_ignore, NULL, NULL, 0, NULL },
                    586:        { "mso", roff_insec, NULL, NULL, 0, NULL },
1.128     schwarze  587:        { "na", roff_line_ignore, NULL, NULL, 0, NULL },
1.21      schwarze  588:        { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
                    589:        { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  590:        { "nhychar", roff_line_ignore, NULL, NULL, 0, NULL },
                    591:        { "nm", roff_unsupp, NULL, NULL, 0, NULL },
                    592:        { "nn", roff_unsupp, NULL, NULL, 0, NULL },
                    593:        { "nop", roff_unsupp, NULL, NULL, 0, NULL },
1.14      schwarze  594:        { "nr", roff_nr, NULL, NULL, 0, NULL },
1.123     schwarze  595:        { "nrf", roff_unsupp, NULL, NULL, 0, NULL },
                    596:        { "nroff", roff_line_ignore, NULL, NULL, 0, NULL },
1.31      schwarze  597:        { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  598:        { "nx", roff_insec, NULL, NULL, 0, NULL },
                    599:        { "open", roff_insec, NULL, NULL, 0, NULL },
                    600:        { "opena", roff_insec, NULL, NULL, 0, NULL },
                    601:        { "os", roff_line_ignore, NULL, NULL, 0, NULL },
                    602:        { "output", roff_unsupp, NULL, NULL, 0, NULL },
                    603:        { "padj", roff_line_ignore, NULL, NULL, 0, NULL },
                    604:        { "papersize", roff_line_ignore, NULL, NULL, 0, NULL },
                    605:        { "pc", roff_line_ignore, NULL, NULL, 0, NULL },
                    606:        { "pev", roff_line_ignore, NULL, NULL, 0, NULL },
                    607:        { "pi", roff_insec, NULL, NULL, 0, NULL },
                    608:        { "PI", roff_unsupp, NULL, NULL, 0, NULL },
1.101     schwarze  609:        { "pl", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  610:        { "pm", roff_line_ignore, NULL, NULL, 0, NULL },
                    611:        { "pn", roff_line_ignore, NULL, NULL, 0, NULL },
                    612:        { "pnr", roff_line_ignore, NULL, NULL, 0, NULL },
                    613:        { "po", roff_line_ignore, NULL, NULL, 0, NULL },
1.31      schwarze  614:        { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.123     schwarze  615:        { "psbb", roff_unsupp, NULL, NULL, 0, NULL },
1.127     schwarze  616:        { "pshape", roff_unsupp, NULL, NULL, 0, NULL },
1.123     schwarze  617:        { "pso", roff_insec, NULL, NULL, 0, NULL },
                    618:        { "ptr", roff_line_ignore, NULL, NULL, 0, NULL },
                    619:        { "pvs", roff_line_ignore, NULL, NULL, 0, NULL },
                    620:        { "rchar", roff_unsupp, NULL, NULL, 0, NULL },
                    621:        { "rd", roff_line_ignore, NULL, NULL, 0, NULL },
                    622:        { "recursionlimit", roff_line_ignore, NULL, NULL, 0, NULL },
                    623:        { "return", roff_unsupp, NULL, NULL, 0, NULL },
                    624:        { "rfschar", roff_unsupp, NULL, NULL, 0, NULL },
                    625:        { "rhang", roff_line_ignore, NULL, NULL, 0, NULL },
                    626:        { "rj", roff_line_ignore, NULL, NULL, 0, NULL },
1.29      schwarze  627:        { "rm", roff_rm, NULL, NULL, 0, NULL },
1.123     schwarze  628:        { "rn", roff_unsupp, NULL, NULL, 0, NULL },
                    629:        { "rnn", roff_unsupp, NULL, NULL, 0, NULL },
1.76      schwarze  630:        { "rr", roff_rr, NULL, NULL, 0, NULL },
1.123     schwarze  631:        { "rs", roff_line_ignore, NULL, NULL, 0, NULL },
                    632:        { "rt", roff_line_ignore, NULL, NULL, 0, NULL },
                    633:        { "schar", roff_unsupp, NULL, NULL, 0, NULL },
                    634:        { "sentchar", roff_line_ignore, NULL, NULL, 0, NULL },
                    635:        { "shc", roff_line_ignore, NULL, NULL, 0, NULL },
                    636:        { "shift", roff_unsupp, NULL, NULL, 0, NULL },
                    637:        { "sizes", roff_line_ignore, NULL, NULL, 0, NULL },
1.14      schwarze  638:        { "so", roff_so, NULL, NULL, 0, NULL },
1.123     schwarze  639:        { "spacewidth", roff_line_ignore, NULL, NULL, 0, NULL },
                    640:        { "special", roff_line_ignore, NULL, NULL, 0, NULL },
                    641:        { "spreadwarn", roff_line_ignore, NULL, NULL, 0, NULL },
                    642:        { "ss", roff_line_ignore, NULL, NULL, 0, NULL },
                    643:        { "sty", roff_line_ignore, NULL, NULL, 0, NULL },
                    644:        { "substring", roff_unsupp, NULL, NULL, 0, NULL },
                    645:        { "sv", roff_line_ignore, NULL, NULL, 0, NULL },
                    646:        { "sy", roff_insec, NULL, NULL, 0, NULL },
                    647:        { "T&", roff_T_, NULL, NULL, 0, NULL },
1.127     schwarze  648:        { "ta", roff_unsupp, NULL, NULL, 0, NULL },
1.123     schwarze  649:        { "tc", roff_unsupp, NULL, NULL, 0, NULL },
                    650:        { "TE", roff_TE, NULL, NULL, 0, NULL },
                    651:        { "TH", roff_TH, NULL, NULL, 0, NULL },
1.127     schwarze  652:        { "ti", roff_unsupp, NULL, NULL, 0, NULL },
1.123     schwarze  653:        { "tkf", roff_line_ignore, NULL, NULL, 0, NULL },
                    654:        { "tl", roff_unsupp, NULL, NULL, 0, NULL },
                    655:        { "tm", roff_line_ignore, NULL, NULL, 0, NULL },
                    656:        { "tm1", roff_line_ignore, NULL, NULL, 0, NULL },
                    657:        { "tmc", roff_line_ignore, NULL, NULL, 0, NULL },
1.42      schwarze  658:        { "tr", roff_tr, NULL, NULL, 0, NULL },
1.123     schwarze  659:        { "track", roff_line_ignore, NULL, NULL, 0, NULL },
                    660:        { "transchar", roff_line_ignore, NULL, NULL, 0, NULL },
                    661:        { "trf", roff_insec, NULL, NULL, 0, NULL },
                    662:        { "trimat", roff_line_ignore, NULL, NULL, 0, NULL },
                    663:        { "trin", roff_unsupp, NULL, NULL, 0, NULL },
                    664:        { "trnt", roff_unsupp, NULL, NULL, 0, NULL },
                    665:        { "troff", roff_line_ignore, NULL, NULL, 0, NULL },
1.27      schwarze  666:        { "TS", roff_TS, NULL, NULL, 0, NULL },
1.123     schwarze  667:        { "uf", roff_line_ignore, NULL, NULL, 0, NULL },
                    668:        { "ul", roff_line_ignore, NULL, NULL, 0, NULL },
                    669:        { "unformat", roff_unsupp, NULL, NULL, 0, NULL },
                    670:        { "unwatch", roff_line_ignore, NULL, NULL, 0, NULL },
                    671:        { "unwatchn", roff_line_ignore, NULL, NULL, 0, NULL },
                    672:        { "vpt", roff_line_ignore, NULL, NULL, 0, NULL },
                    673:        { "vs", roff_line_ignore, NULL, NULL, 0, NULL },
                    674:        { "warn", roff_line_ignore, NULL, NULL, 0, NULL },
                    675:        { "warnscale", roff_line_ignore, NULL, NULL, 0, NULL },
                    676:        { "watch", roff_line_ignore, NULL, NULL, 0, NULL },
                    677:        { "watchlength", roff_line_ignore, NULL, NULL, 0, NULL },
                    678:        { "watchn", roff_line_ignore, NULL, NULL, 0, NULL },
                    679:        { "wh", roff_unsupp, NULL, NULL, 0, NULL },
                    680:        { "while", roff_unsupp, NULL, NULL, 0, NULL },
                    681:        { "write", roff_insec, NULL, NULL, 0, NULL },
                    682:        { "writec", roff_insec, NULL, NULL, 0, NULL },
                    683:        { "writem", roff_insec, NULL, NULL, 0, NULL },
                    684:        { "xflag", roff_line_ignore, NULL, NULL, 0, NULL },
1.3       schwarze  685:        { ".", roff_cblock, NULL, NULL, 0, NULL },
1.16      schwarze  686:        { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.1       schwarze  687: };
                    688:
1.73      schwarze  689: /* not currently implemented: Ds em Eq LP Me PP pp Or Rd Sf SH */
1.47      schwarze  690: const  char *const __mdoc_reserved[] = {
                    691:        "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
                    692:        "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
                    693:        "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
                    694:        "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
1.73      schwarze  695:        "Dt", "Dv", "Dx", "D1",
                    696:        "Ec", "Ed", "Ef", "Ek", "El", "Em",
                    697:        "En", "Eo", "Er", "Es", "Ev", "Ex",
1.47      schwarze  698:        "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
1.73      schwarze  699:        "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp",
                    700:        "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
1.47      schwarze  701:        "Oc", "Oo", "Op", "Os", "Ot", "Ox",
1.73      schwarze  702:        "Pa", "Pc", "Pf", "Po", "Pp", "Pq",
                    703:        "Qc", "Ql", "Qo", "Qq", "Re", "Rs", "Rv",
                    704:        "Sc", "Sh", "Sm", "So", "Sq",
1.47      schwarze  705:        "Ss", "St", "Sx", "Sy",
                    706:        "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
1.73      schwarze  707:        "%A", "%B", "%C", "%D", "%I", "%J", "%N", "%O",
1.47      schwarze  708:        "%P", "%Q", "%R", "%T", "%U", "%V",
                    709:        NULL
                    710: };
                    711:
1.73      schwarze  712: /* not currently implemented: BT DE DS ME MT PT SY TQ YS */
1.47      schwarze  713: const  char *const __man_reserved[] = {
1.73      schwarze  714:        "AT", "B", "BI", "BR", "DT",
                    715:        "EE", "EN", "EQ", "EX", "HP", "I", "IB", "IP", "IR",
                    716:        "LP", "OP", "P", "PD", "PP",
                    717:        "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS",
                    718:        "TE", "TH", "TP", "TS", "T&", "UC", "UE", "UR",
1.47      schwarze  719:        NULL
                    720: };
                    721:
1.37      schwarze  722: /* Array of injected predefined strings. */
                    723: #define        PREDEFS_MAX      38
                    724: static const struct predef predefs[PREDEFS_MAX] = {
                    725: #include "predefs.in"
                    726: };
                    727:
1.42      schwarze  728: /* See roffhash_find() */
1.3       schwarze  729: #define        ROFF_HASH(p)    (p[0] - ASCII_LO)
                    730:
1.51      schwarze  731: static int      roffit_lines;  /* number of lines to delay */
                    732: static char    *roffit_macro;  /* nil-terminated macro line */
                    733:
1.80      schwarze  734:
1.3       schwarze  735: static void
1.42      schwarze  736: roffhash_init(void)
1.3       schwarze  737: {
                    738:        struct roffmac   *n;
                    739:        int               buc, i;
                    740:
1.16      schwarze  741:        for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.3       schwarze  742:                assert(roffs[i].name[0] >= ASCII_LO);
                    743:                assert(roffs[i].name[0] <= ASCII_HI);
                    744:
                    745:                buc = ROFF_HASH(roffs[i].name);
                    746:
                    747:                if (NULL != (n = hash[buc])) {
                    748:                        for ( ; n->next; n = n->next)
                    749:                                /* Do nothing. */ ;
                    750:                        n->next = &roffs[i];
                    751:                } else
                    752:                        hash[buc] = &roffs[i];
                    753:        }
                    754: }
                    755:
1.1       schwarze  756: /*
                    757:  * Look up a roff token by its name.  Returns ROFF_MAX if no macro by
                    758:  * the nil-terminated string name could be found.
                    759:  */
                    760: static enum rofft
1.42      schwarze  761: roffhash_find(const char *p, size_t s)
1.1       schwarze  762: {
1.3       schwarze  763:        int              buc;
                    764:        struct roffmac  *n;
1.1       schwarze  765:
1.3       schwarze  766:        /*
                    767:         * libroff has an extremely simple hashtable, for the time
                    768:         * being, which simply keys on the first character, which must
                    769:         * be printable, then walks a chain.  It works well enough until
                    770:         * optimised.
                    771:         */
                    772:
                    773:        if (p[0] < ASCII_LO || p[0] > ASCII_HI)
                    774:                return(ROFF_MAX);
                    775:
                    776:        buc = ROFF_HASH(p);
                    777:
                    778:        if (NULL == (n = hash[buc]))
                    779:                return(ROFF_MAX);
                    780:        for ( ; n; n = n->next)
1.16      schwarze  781:                if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.3       schwarze  782:                        return((enum rofft)(n - roffs));
1.1       schwarze  783:
                    784:        return(ROFF_MAX);
                    785: }
                    786:
                    787: /*
                    788:  * Pop the current node off of the stack of roff instructions currently
                    789:  * pending.
                    790:  */
                    791: static void
                    792: roffnode_pop(struct roff *r)
                    793: {
                    794:        struct roffnode *p;
                    795:
1.2       schwarze  796:        assert(r->last);
1.80      schwarze  797:        p = r->last;
1.2       schwarze  798:
                    799:        r->last = r->last->parent;
1.16      schwarze  800:        free(p->name);
                    801:        free(p->end);
1.1       schwarze  802:        free(p);
                    803: }
                    804:
                    805: /*
                    806:  * Push a roff node onto the instruction stack.  This must later be
                    807:  * removed with roffnode_pop().
                    808:  */
1.11      schwarze  809: static void
1.16      schwarze  810: roffnode_push(struct roff *r, enum rofft tok, const char *name,
                    811:                int line, int col)
1.1       schwarze  812: {
                    813:        struct roffnode *p;
                    814:
1.11      schwarze  815:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  816:        p->tok = tok;
1.16      schwarze  817:        if (name)
                    818:                p->name = mandoc_strdup(name);
1.1       schwarze  819:        p->parent = r->last;
                    820:        p->line = line;
                    821:        p->col = col;
1.71      schwarze  822:        p->rule = p->parent ? p->parent->rule : 0;
1.1       schwarze  823:
                    824:        r->last = p;
                    825: }
                    826:
                    827: static void
                    828: roff_free1(struct roff *r)
                    829: {
1.49      schwarze  830:        struct tbl_node *tbl;
1.32      schwarze  831:        struct eqn_node *e;
1.42      schwarze  832:        int              i;
1.27      schwarze  833:
1.49      schwarze  834:        while (NULL != (tbl = r->first_tbl)) {
                    835:                r->first_tbl = tbl->next;
                    836:                tbl_free(tbl);
1.27      schwarze  837:        }
                    838:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  839:
1.32      schwarze  840:        while (NULL != (e = r->first_eqn)) {
                    841:                r->first_eqn = e->next;
                    842:                eqn_free(e);
                    843:        }
                    844:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    845:
1.1       schwarze  846:        while (r->last)
                    847:                roffnode_pop(r);
1.27      schwarze  848:
1.96      schwarze  849:        free (r->rstack);
                    850:        r->rstack = NULL;
                    851:        r->rstacksz = 0;
                    852:        r->rstackpos = -1;
                    853:
                    854:        roff_freereg(r->regtab);
                    855:        r->regtab = NULL;
                    856:
1.42      schwarze  857:        roff_freestr(r->strtab);
                    858:        roff_freestr(r->xmbtab);
                    859:        r->strtab = r->xmbtab = NULL;
                    860:
                    861:        if (r->xtab)
                    862:                for (i = 0; i < 128; i++)
                    863:                        free(r->xtab[i].p);
                    864:        free(r->xtab);
                    865:        r->xtab = NULL;
1.1       schwarze  866: }
                    867:
                    868: void
                    869: roff_reset(struct roff *r)
                    870: {
                    871:
                    872:        roff_free1(r);
1.99      schwarze  873:        r->format = r->options & (MPARSE_MDOC | MPARSE_MAN);
1.48      schwarze  874:        r->control = 0;
1.1       schwarze  875: }
                    876:
                    877: void
                    878: roff_free(struct roff *r)
                    879: {
                    880:
                    881:        roff_free1(r);
                    882:        free(r);
                    883: }
                    884:
                    885: struct roff *
1.109     schwarze  886: roff_alloc(struct mparse *parse, const struct mchars *mchars, int options)
1.1       schwarze  887: {
                    888:        struct roff     *r;
                    889:
1.11      schwarze  890:        r = mandoc_calloc(1, sizeof(struct roff));
1.35      schwarze  891:        r->parse = parse;
1.109     schwarze  892:        r->mchars = mchars;
1.72      schwarze  893:        r->options = options;
1.99      schwarze  894:        r->format = options & (MPARSE_MDOC | MPARSE_MAN);
1.2       schwarze  895:        r->rstackpos = -1;
1.80      schwarze  896:
1.42      schwarze  897:        roffhash_init();
1.37      schwarze  898:
1.1       schwarze  899:        return(r);
1.137   ! schwarze  900: }
        !           901:
        !           902: static void
        !           903: roff_man_free1(struct roff_man *man)
        !           904: {
        !           905:
        !           906:        if (man->first != NULL) {
        !           907:                if (man->macroset == MACROSET_MDOC)
        !           908:                        mdoc_node_delete(man, man->first);
        !           909:                else
        !           910:                        man_node_delete(man, man->first);
        !           911:        }
        !           912:        free(man->meta.msec);
        !           913:        free(man->meta.vol);
        !           914:        free(man->meta.os);
        !           915:        free(man->meta.arch);
        !           916:        free(man->meta.title);
        !           917:        free(man->meta.name);
        !           918:        free(man->meta.date);
        !           919: }
        !           920:
        !           921: static void
        !           922: roff_man_alloc1(struct roff_man *man)
        !           923: {
        !           924:
        !           925:        memset(&man->meta, 0, sizeof(man->meta));
        !           926:        man->first = mandoc_calloc(1, sizeof(*man->first));
        !           927:        man->first->type = ROFFT_ROOT;
        !           928:        man->last = man->first;
        !           929:        man->last_es = NULL;
        !           930:        man->flags = 0;
        !           931:        man->macroset = MACROSET_NONE;
        !           932:        man->lastsec = man->lastnamed = SEC_NONE;
        !           933:        man->next = ROFF_NEXT_CHILD;
        !           934: }
        !           935:
        !           936: void
        !           937: roff_man_reset(struct roff_man *man)
        !           938: {
        !           939:
        !           940:        roff_man_free1(man);
        !           941:        roff_man_alloc1(man);
        !           942: }
        !           943:
        !           944: void
        !           945: roff_man_free(struct roff_man *man)
        !           946: {
        !           947:
        !           948:        roff_man_free1(man);
        !           949:        free(man);
        !           950: }
        !           951:
        !           952: struct roff_man *
        !           953: roff_man_alloc(struct roff *roff, struct mparse *parse,
        !           954:        const char *defos, int quick)
        !           955: {
        !           956:        struct roff_man *man;
        !           957:
        !           958:        man = mandoc_calloc(1, sizeof(*man));
        !           959:        man->parse = parse;
        !           960:        man->roff = roff;
        !           961:        man->defos = defos;
        !           962:        man->quick = quick;
        !           963:        roff_man_alloc1(man);
        !           964:        return(man);
1.1       schwarze  965: }
                    966:
1.8       schwarze  967: /*
1.79      schwarze  968:  * In the current line, expand escape sequences that tend to get
                    969:  * used in numerical expressions and conditional requests.
                    970:  * Also check the syntax of the remaining escape sequences.
1.8       schwarze  971:  */
1.45      schwarze  972: static enum rofferr
1.110     schwarze  973: roff_res(struct roff *r, struct buf *buf, int ln, int pos)
1.8       schwarze  974: {
1.81      schwarze  975:        char             ubuf[24]; /* buffer to print the number */
1.78      schwarze  976:        const char      *start; /* start of the string to process */
1.82      schwarze  977:        char            *stesc; /* start of an escape sequence ('\\') */
1.23      schwarze  978:        const char      *stnam; /* start of the name, after "[(*" */
                    979:        const char      *cp;    /* end of the name, e.g. before ']' */
                    980:        const char      *res;   /* the string to be substituted */
1.110     schwarze  981:        char            *nbuf;  /* new buffer to copy buf->buf to */
1.53      schwarze  982:        size_t           maxl;  /* expected length of the escape name */
                    983:        size_t           naml;  /* actual length of the escape name */
1.109     schwarze  984:        enum mandoc_esc  esc;   /* type of the escape sequence */
                    985:        int              inaml; /* length returned from mandoc_escape() */
1.53      schwarze  986:        int              expand_count;  /* to avoid infinite loops */
1.79      schwarze  987:        int              npos;  /* position in numeric expression */
1.91      schwarze  988:        int              arg_complete; /* argument not interrupted by eol */
1.79      schwarze  989:        char             term;  /* character terminating the escape */
1.8       schwarze  990:
1.43      schwarze  991:        expand_count = 0;
1.110     schwarze  992:        start = buf->buf + pos;
1.78      schwarze  993:        stesc = strchr(start, '\0') - 1;
                    994:        while (stesc-- > start) {
1.43      schwarze  995:
1.78      schwarze  996:                /* Search backwards for the next backslash. */
                    997:
1.110     schwarze  998:                if (*stesc != '\\')
1.78      schwarze  999:                        continue;
                   1000:
                   1001:                /* If it is escaped, skip it. */
                   1002:
                   1003:                for (cp = stesc - 1; cp >= start; cp--)
1.110     schwarze 1004:                        if (*cp != '\\')
1.78      schwarze 1005:                                break;
                   1006:
1.110     schwarze 1007:                if ((stesc - cp) % 2 == 0) {
1.82      schwarze 1008:                        stesc = (char *)cp;
1.78      schwarze 1009:                        continue;
                   1010:                }
1.23      schwarze 1011:
1.79      schwarze 1012:                /* Decide whether to expand or to check only. */
1.23      schwarze 1013:
1.79      schwarze 1014:                term = '\0';
1.78      schwarze 1015:                cp = stesc + 1;
1.53      schwarze 1016:                switch (*cp) {
1.80      schwarze 1017:                case '*':
1.53      schwarze 1018:                        res = NULL;
                   1019:                        break;
1.80      schwarze 1020:                case 'B':
1.79      schwarze 1021:                        /* FALLTHROUGH */
1.80      schwarze 1022:                case 'w':
1.79      schwarze 1023:                        term = cp[1];
                   1024:                        /* FALLTHROUGH */
1.80      schwarze 1025:                case 'n':
1.53      schwarze 1026:                        res = ubuf;
                   1027:                        break;
                   1028:                default:
1.109     schwarze 1029:                        esc = mandoc_escape(&cp, &stnam, &inaml);
                   1030:                        if (esc == ESCAPE_ERROR ||
                   1031:                            (esc == ESCAPE_SPECIAL &&
                   1032:                             mchars_spec2cp(r->mchars, stnam, inaml) < 0))
1.92      schwarze 1033:                                mandoc_vmsg(MANDOCERR_ESC_BAD,
1.110     schwarze 1034:                                    r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1035:                                    "%.*s", (int)(cp - stesc), stesc);
1.78      schwarze 1036:                        continue;
1.42      schwarze 1037:                }
                   1038:
1.78      schwarze 1039:                if (EXPAND_LIMIT < ++expand_count) {
                   1040:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
1.110     schwarze 1041:                            ln, (int)(stesc - buf->buf), NULL);
1.78      schwarze 1042:                        return(ROFF_IGN);
                   1043:                }
1.23      schwarze 1044:
                   1045:                /*
                   1046:                 * The third character decides the length
1.53      schwarze 1047:                 * of the name of the string or register.
1.23      schwarze 1048:                 * Save a pointer to the name.
                   1049:                 */
                   1050:
1.110     schwarze 1051:                if (term == '\0') {
1.79      schwarze 1052:                        switch (*++cp) {
1.80      schwarze 1053:                        case '\0':
1.79      schwarze 1054:                                maxl = 0;
                   1055:                                break;
1.80      schwarze 1056:                        case '(':
1.79      schwarze 1057:                                cp++;
                   1058:                                maxl = 2;
                   1059:                                break;
1.80      schwarze 1060:                        case '[':
1.79      schwarze 1061:                                cp++;
                   1062:                                term = ']';
                   1063:                                maxl = 0;
                   1064:                                break;
                   1065:                        default:
                   1066:                                maxl = 1;
                   1067:                                break;
                   1068:                        }
                   1069:                } else {
                   1070:                        cp += 2;
1.8       schwarze 1071:                        maxl = 0;
                   1072:                }
1.23      schwarze 1073:                stnam = cp;
1.8       schwarze 1074:
1.23      schwarze 1075:                /* Advance to the end of the name. */
1.8       schwarze 1076:
1.125     schwarze 1077:                naml = 0;
1.91      schwarze 1078:                arg_complete = 1;
1.125     schwarze 1079:                while (maxl == 0 || naml < maxl) {
1.110     schwarze 1080:                        if (*cp == '\0') {
1.92      schwarze 1081:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 1082:                                    ln, (int)(stesc - buf->buf), stesc);
1.91      schwarze 1083:                                arg_complete = 0;
1.79      schwarze 1084:                                break;
1.42      schwarze 1085:                        }
1.110     schwarze 1086:                        if (maxl == 0 && *cp == term) {
1.79      schwarze 1087:                                cp++;
1.125     schwarze 1088:                                break;
                   1089:                        }
                   1090:                        if (*cp++ != '\\' || stesc[1] != 'w') {
                   1091:                                naml++;
                   1092:                                continue;
                   1093:                        }
                   1094:                        switch (mandoc_escape(&cp, NULL, NULL)) {
                   1095:                        case ESCAPE_SPECIAL:
                   1096:                                /* FALLTHROUGH */
                   1097:                        case ESCAPE_UNICODE:
                   1098:                                /* FALLTHROUGH */
                   1099:                        case ESCAPE_NUMBERED:
                   1100:                                /* FALLTHROUGH */
                   1101:                        case ESCAPE_OVERSTRIKE:
                   1102:                                naml++;
                   1103:                                break;
                   1104:                        default:
1.8       schwarze 1105:                                break;
1.79      schwarze 1106:                        }
1.8       schwarze 1107:                }
                   1108:
1.23      schwarze 1109:                /*
                   1110:                 * Retrieve the replacement string; if it is
                   1111:                 * undefined, resume searching for escapes.
                   1112:                 */
                   1113:
1.79      schwarze 1114:                switch (stesc[1]) {
1.80      schwarze 1115:                case '*':
1.91      schwarze 1116:                        if (arg_complete)
                   1117:                                res = roff_getstrn(r, stnam, naml);
1.79      schwarze 1118:                        break;
1.80      schwarze 1119:                case 'B':
1.79      schwarze 1120:                        npos = 0;
1.91      schwarze 1121:                        ubuf[0] = arg_complete &&
1.133     schwarze 1122:                            roff_evalnum(r, ln, stnam, &npos,
                   1123:                              NULL, ROFFNUM_SCALE) &&
1.91      schwarze 1124:                            stnam + npos + 1 == cp ? '1' : '0';
1.79      schwarze 1125:                        ubuf[1] = '\0';
                   1126:                        break;
1.80      schwarze 1127:                case 'n':
1.91      schwarze 1128:                        if (arg_complete)
                   1129:                                (void)snprintf(ubuf, sizeof(ubuf), "%d",
                   1130:                                    roff_getregn(r, stnam, naml));
                   1131:                        else
                   1132:                                ubuf[0] = '\0';
1.79      schwarze 1133:                        break;
1.80      schwarze 1134:                case 'w':
1.91      schwarze 1135:                        /* use even incomplete args */
1.81      schwarze 1136:                        (void)snprintf(ubuf, sizeof(ubuf), "%d",
1.79      schwarze 1137:                            24 * (int)naml);
                   1138:                        break;
                   1139:                }
1.8       schwarze 1140:
1.110     schwarze 1141:                if (res == NULL) {
1.92      schwarze 1142:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1143:                            r->parse, ln, (int)(stesc - buf->buf),
1.92      schwarze 1144:                            "%.*s", (int)naml, stnam);
1.37      schwarze 1145:                        res = "";
1.118     schwarze 1146:                } else if (buf->sz + strlen(res) > SHRT_MAX) {
                   1147:                        mandoc_msg(MANDOCERR_ROFFLOOP, r->parse,
                   1148:                            ln, (int)(stesc - buf->buf), NULL);
                   1149:                        return(ROFF_IGN);
1.8       schwarze 1150:                }
                   1151:
1.23      schwarze 1152:                /* Replace the escape sequence by the string. */
                   1153:
1.82      schwarze 1154:                *stesc = '\0';
1.110     schwarze 1155:                buf->sz = mandoc_asprintf(&nbuf, "%s%s%s",
                   1156:                    buf->buf, res, cp) + 1;
1.8       schwarze 1157:
1.78      schwarze 1158:                /* Prepare for the next replacement. */
1.8       schwarze 1159:
1.78      schwarze 1160:                start = nbuf + pos;
1.110     schwarze 1161:                stesc = nbuf + (stesc - buf->buf) + strlen(res);
                   1162:                free(buf->buf);
                   1163:                buf->buf = nbuf;
1.42      schwarze 1164:        }
1.45      schwarze 1165:        return(ROFF_CONT);
1.42      schwarze 1166: }
                   1167:
                   1168: /*
1.51      schwarze 1169:  * Process text streams:
                   1170:  * Convert all breakable hyphens into ASCII_HYPH.
                   1171:  * Decrement and spring input line trap.
1.42      schwarze 1172:  */
                   1173: static enum rofferr
1.110     schwarze 1174: roff_parsetext(struct buf *buf, int pos, int *offs)
1.42      schwarze 1175: {
                   1176:        size_t           sz;
                   1177:        const char      *start;
1.51      schwarze 1178:        char            *p;
                   1179:        int              isz;
1.42      schwarze 1180:        enum mandoc_esc  esc;
                   1181:
1.110     schwarze 1182:        start = p = buf->buf + pos;
1.42      schwarze 1183:
1.110     schwarze 1184:        while (*p != '\0') {
1.42      schwarze 1185:                sz = strcspn(p, "-\\");
                   1186:                p += sz;
                   1187:
1.110     schwarze 1188:                if (*p == '\0')
1.42      schwarze 1189:                        break;
                   1190:
1.110     schwarze 1191:                if (*p == '\\') {
1.42      schwarze 1192:                        /* Skip over escapes. */
                   1193:                        p++;
1.62      schwarze 1194:                        esc = mandoc_escape((const char **)&p, NULL, NULL);
1.110     schwarze 1195:                        if (esc == ESCAPE_ERROR)
1.42      schwarze 1196:                                break;
1.136     schwarze 1197:                        while (*p == '-')
                   1198:                                p++;
1.42      schwarze 1199:                        continue;
                   1200:                } else if (p == start) {
                   1201:                        p++;
                   1202:                        continue;
                   1203:                }
                   1204:
1.44      schwarze 1205:                if (isalpha((unsigned char)p[-1]) &&
                   1206:                    isalpha((unsigned char)p[1]))
1.42      schwarze 1207:                        *p = ASCII_HYPH;
                   1208:                p++;
1.8       schwarze 1209:        }
                   1210:
1.51      schwarze 1211:        /* Spring the input line trap. */
1.110     schwarze 1212:        if (roffit_lines == 1) {
                   1213:                isz = mandoc_asprintf(&p, "%s\n.%s", buf->buf, roffit_macro);
                   1214:                free(buf->buf);
                   1215:                buf->buf = p;
                   1216:                buf->sz = isz + 1;
1.51      schwarze 1217:                *offs = 0;
                   1218:                free(roffit_macro);
                   1219:                roffit_lines = 0;
                   1220:                return(ROFF_REPARSE);
1.110     schwarze 1221:        } else if (roffit_lines > 1)
1.51      schwarze 1222:                --roffit_lines;
1.42      schwarze 1223:        return(ROFF_CONT);
1.8       schwarze 1224: }
                   1225:
1.1       schwarze 1226: enum rofferr
1.110     schwarze 1227: roff_parseln(struct roff *r, int ln, struct buf *buf, int *offs)
1.1       schwarze 1228: {
                   1229:        enum rofft       t;
1.27      schwarze 1230:        enum rofferr     e;
1.110     schwarze 1231:        int              pos;   /* parse point */
1.115     schwarze 1232:        int              spos;  /* saved parse point for messages */
1.110     schwarze 1233:        int              ppos;  /* original offset in buf->buf */
                   1234:        int              ctl;   /* macro line (boolean) */
                   1235:
                   1236:        ppos = pos = *offs;
1.1       schwarze 1237:
1.102     schwarze 1238:        /* Handle in-line equation delimiters. */
                   1239:
1.108     schwarze 1240:        if (r->tbl == NULL &&
                   1241:            r->last_eqn != NULL && r->last_eqn->delim &&
1.102     schwarze 1242:            (r->eqn == NULL || r->eqn_inline)) {
1.110     schwarze 1243:                e = roff_eqndelim(r, buf, pos);
1.102     schwarze 1244:                if (e == ROFF_REPARSE)
                   1245:                        return(e);
                   1246:                assert(e == ROFF_CONT);
                   1247:        }
                   1248:
                   1249:        /* Expand some escape sequences. */
1.8       schwarze 1250:
1.110     schwarze 1251:        e = roff_res(r, buf, ln, pos);
                   1252:        if (e == ROFF_IGN)
1.45      schwarze 1253:                return(e);
1.110     schwarze 1254:        assert(e == ROFF_CONT);
1.8       schwarze 1255:
1.110     schwarze 1256:        ctl = roff_getcontrol(r, buf->buf, &pos);
1.35      schwarze 1257:
1.8       schwarze 1258:        /*
1.2       schwarze 1259:         * First, if a scope is open and we're not a macro, pass the
1.124     schwarze 1260:         * text through the macro's filter.
                   1261:         * Equations process all content themselves.
                   1262:         * Tables process almost all content themselves, but we want
                   1263:         * to warn about macros before passing it there.
1.2       schwarze 1264:         */
                   1265:
1.124     schwarze 1266:        if (r->last != NULL && ! ctl) {
1.2       schwarze 1267:                t = r->last->tok;
                   1268:                assert(roffs[t].text);
1.110     schwarze 1269:                e = (*roffs[t].text)(r, t, buf, ln, pos, pos, offs);
                   1270:                assert(e == ROFF_IGN || e == ROFF_CONT);
                   1271:                if (e != ROFF_CONT)
1.32      schwarze 1272:                        return(e);
1.54      schwarze 1273:        }
1.124     schwarze 1274:        if (r->eqn != NULL)
1.110     schwarze 1275:                return(eqn_read(&r->eqn, ln, buf->buf, ppos, offs));
1.124     schwarze 1276:        if (r->tbl != NULL && ( ! ctl || buf->buf[pos] == '\0'))
1.131     schwarze 1277:                return(tbl_read(r->tbl, ln, buf->buf, ppos));
1.124     schwarze 1278:        if ( ! ctl)
1.110     schwarze 1279:                return(roff_parsetext(buf, pos, offs));
1.100     schwarze 1280:
                   1281:        /* Skip empty request lines. */
                   1282:
1.110     schwarze 1283:        if (buf->buf[pos] == '"') {
1.100     schwarze 1284:                mandoc_msg(MANDOCERR_COMMENT_BAD, r->parse,
                   1285:                    ln, pos, NULL);
                   1286:                return(ROFF_IGN);
1.110     schwarze 1287:        } else if (buf->buf[pos] == '\0')
1.100     schwarze 1288:                return(ROFF_IGN);
1.2       schwarze 1289:
                   1290:        /*
                   1291:         * If a scope is open, go to the child handler for that macro,
                   1292:         * as it may want to preprocess before doing anything with it.
1.32      schwarze 1293:         * Don't do so if an equation is open.
1.2       schwarze 1294:         */
                   1295:
                   1296:        if (r->last) {
1.1       schwarze 1297:                t = r->last->tok;
                   1298:                assert(roffs[t].sub);
1.110     schwarze 1299:                return((*roffs[t].sub)(r, t, buf, ln, ppos, pos, offs));
1.2       schwarze 1300:        }
                   1301:
1.115     schwarze 1302:        /* No scope is open.  This is a new request or macro. */
                   1303:
                   1304:        spos = pos;
                   1305:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
                   1306:
                   1307:        /* Tables ignore most macros. */
                   1308:
                   1309:        if (r->tbl != NULL && (t == ROFF_MAX || t == ROFF_TS)) {
                   1310:                mandoc_msg(MANDOCERR_TBLMACRO, r->parse,
                   1311:                    ln, pos, buf->buf + spos);
1.129     schwarze 1312:                if (t == ROFF_TS)
                   1313:                        return(ROFF_IGN);
                   1314:                while (buf->buf[pos] != '\0' && buf->buf[pos] != ' ')
                   1315:                        pos++;
                   1316:                while (buf->buf[pos] != '\0' && buf->buf[pos] == ' ')
                   1317:                        pos++;
                   1318:                return(tbl_read(r->tbl, ln, buf->buf, pos));
1.115     schwarze 1319:        }
                   1320:
1.2       schwarze 1321:        /*
1.115     schwarze 1322:         * This is neither a roff request nor a user-defined macro.
                   1323:         * Let the standard macro set parsers handle it.
1.2       schwarze 1324:         */
                   1325:
1.115     schwarze 1326:        if (t == ROFF_MAX)
1.1       schwarze 1327:                return(ROFF_CONT);
1.115     schwarze 1328:
                   1329:        /* Execute a roff request or a user defined macro. */
1.1       schwarze 1330:
1.2       schwarze 1331:        assert(roffs[t].proc);
1.110     schwarze 1332:        return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2       schwarze 1333: }
                   1334:
1.27      schwarze 1335: void
1.2       schwarze 1336: roff_endparse(struct roff *r)
                   1337: {
1.1       schwarze 1338:
1.27      schwarze 1339:        if (r->last)
1.94      schwarze 1340:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1341:                    r->last->line, r->last->col,
                   1342:                    roffs[r->last->tok].name);
1.27      schwarze 1343:
1.32      schwarze 1344:        if (r->eqn) {
1.94      schwarze 1345:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1346:                    r->eqn->eqn.ln, r->eqn->eqn.pos, "EQ");
1.41      schwarze 1347:                eqn_end(&r->eqn);
1.32      schwarze 1348:        }
                   1349:
1.27      schwarze 1350:        if (r->tbl) {
1.94      schwarze 1351:                mandoc_msg(MANDOCERR_BLK_NOEND, r->parse,
                   1352:                    r->tbl->line, r->tbl->pos, "TS");
1.41      schwarze 1353:                tbl_end(&r->tbl);
1.27      schwarze 1354:        }
1.1       schwarze 1355: }
                   1356:
                   1357: /*
                   1358:  * Parse a roff node's type from the input buffer.  This must be in the
                   1359:  * form of ".foo xxx" in the usual way.
                   1360:  */
                   1361: static enum rofft
1.87      schwarze 1362: roff_parse(struct roff *r, char *buf, int *pos, int ln, int ppos)
1.1       schwarze 1363: {
1.87      schwarze 1364:        char            *cp;
1.16      schwarze 1365:        const char      *mac;
                   1366:        size_t           maclen;
1.1       schwarze 1367:        enum rofft       t;
                   1368:
1.87      schwarze 1369:        cp = buf + *pos;
                   1370:
                   1371:        if ('\0' == *cp || '"' == *cp || '\t' == *cp || ' ' == *cp)
1.1       schwarze 1372:                return(ROFF_MAX);
                   1373:
1.87      schwarze 1374:        mac = cp;
                   1375:        maclen = roff_getname(r, &cp, ln, ppos);
1.1       schwarze 1376:
1.16      schwarze 1377:        t = (r->current_string = roff_getstrn(r, mac, maclen))
1.42      schwarze 1378:            ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.1       schwarze 1379:
1.87      schwarze 1380:        if (ROFF_MAX != t)
                   1381:                *pos = cp - buf;
1.1       schwarze 1382:
                   1383:        return(t);
                   1384: }
                   1385:
                   1386: static enum rofferr
1.2       schwarze 1387: roff_cblock(ROFF_ARGS)
1.1       schwarze 1388: {
                   1389:
1.2       schwarze 1390:        /*
                   1391:         * A block-close `..' should only be invoked as a child of an
                   1392:         * ignore macro, otherwise raise a warning and just ignore it.
                   1393:         */
                   1394:
1.110     schwarze 1395:        if (r->last == NULL) {
1.94      schwarze 1396:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1397:                    ln, ppos, "..");
1.2       schwarze 1398:                return(ROFF_IGN);
                   1399:        }
1.1       schwarze 1400:
1.2       schwarze 1401:        switch (r->last->tok) {
1.80      schwarze 1402:        case ROFF_am:
1.93      schwarze 1403:                /* ROFF_am1 is remapped to ROFF_am in roff_block(). */
1.2       schwarze 1404:                /* FALLTHROUGH */
1.80      schwarze 1405:        case ROFF_ami:
1.2       schwarze 1406:                /* FALLTHROUGH */
1.80      schwarze 1407:        case ROFF_de:
1.23      schwarze 1408:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.2       schwarze 1409:                /* FALLTHROUGH */
1.80      schwarze 1410:        case ROFF_dei:
1.2       schwarze 1411:                /* FALLTHROUGH */
1.80      schwarze 1412:        case ROFF_ig:
1.2       schwarze 1413:                break;
                   1414:        default:
1.94      schwarze 1415:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1416:                    ln, ppos, "..");
1.1       schwarze 1417:                return(ROFF_IGN);
1.2       schwarze 1418:        }
                   1419:
1.110     schwarze 1420:        if (buf->buf[pos] != '\0')
1.90      schwarze 1421:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 1422:                    ".. %s", buf->buf + pos);
1.2       schwarze 1423:
                   1424:        roffnode_pop(r);
                   1425:        roffnode_cleanscope(r);
                   1426:        return(ROFF_IGN);
                   1427:
                   1428: }
1.1       schwarze 1429:
1.2       schwarze 1430: static void
                   1431: roffnode_cleanscope(struct roff *r)
                   1432: {
1.1       schwarze 1433:
1.2       schwarze 1434:        while (r->last) {
1.46      schwarze 1435:                if (--r->last->endspan != 0)
1.2       schwarze 1436:                        break;
                   1437:                roffnode_pop(r);
                   1438:        }
                   1439: }
1.1       schwarze 1440:
1.68      schwarze 1441: static void
                   1442: roff_ccond(struct roff *r, int ln, int ppos)
1.2       schwarze 1443: {
1.1       schwarze 1444:
1.2       schwarze 1445:        if (NULL == r->last) {
1.94      schwarze 1446:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1447:                    ln, ppos, "\\}");
1.68      schwarze 1448:                return;
1.2       schwarze 1449:        }
1.1       schwarze 1450:
1.2       schwarze 1451:        switch (r->last->tok) {
1.80      schwarze 1452:        case ROFF_el:
1.2       schwarze 1453:                /* FALLTHROUGH */
1.80      schwarze 1454:        case ROFF_ie:
1.2       schwarze 1455:                /* FALLTHROUGH */
1.80      schwarze 1456:        case ROFF_if:
1.2       schwarze 1457:                break;
                   1458:        default:
1.94      schwarze 1459:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1460:                    ln, ppos, "\\}");
1.68      schwarze 1461:                return;
1.2       schwarze 1462:        }
1.1       schwarze 1463:
1.2       schwarze 1464:        if (r->last->endspan > -1) {
1.94      schwarze 1465:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   1466:                    ln, ppos, "\\}");
1.68      schwarze 1467:                return;
1.2       schwarze 1468:        }
                   1469:
                   1470:        roffnode_pop(r);
                   1471:        roffnode_cleanscope(r);
1.68      schwarze 1472:        return;
1.1       schwarze 1473: }
                   1474:
                   1475: static enum rofferr
1.2       schwarze 1476: roff_block(ROFF_ARGS)
1.1       schwarze 1477: {
1.93      schwarze 1478:        const char      *name;
                   1479:        char            *iname, *cp;
1.86      schwarze 1480:        size_t           namesz;
1.16      schwarze 1481:
1.93      schwarze 1482:        /* Ignore groff compatibility mode for now. */
1.2       schwarze 1483:
1.110     schwarze 1484:        if (tok == ROFF_de1)
1.93      schwarze 1485:                tok = ROFF_de;
1.123     schwarze 1486:        else if (tok == ROFF_dei1)
                   1487:                tok = ROFF_dei;
1.110     schwarze 1488:        else if (tok == ROFF_am1)
1.93      schwarze 1489:                tok = ROFF_am;
1.123     schwarze 1490:        else if (tok == ROFF_ami1)
                   1491:                tok = ROFF_ami;
1.93      schwarze 1492:
                   1493:        /* Parse the macro name argument. */
                   1494:
1.110     schwarze 1495:        cp = buf->buf + pos;
                   1496:        if (tok == ROFF_ig) {
1.93      schwarze 1497:                iname = NULL;
                   1498:                namesz = 0;
                   1499:        } else {
                   1500:                iname = cp;
                   1501:                namesz = roff_getname(r, &cp, ln, ppos);
                   1502:                iname[namesz] = '\0';
                   1503:        }
1.22      schwarze 1504:
1.93      schwarze 1505:        /* Resolve the macro name argument if it is indirect. */
1.22      schwarze 1506:
1.110     schwarze 1507:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1508:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93      schwarze 1509:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1510:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1511:                            "%.*s", (int)namesz, iname);
                   1512:                        namesz = 0;
                   1513:                } else
                   1514:                        namesz = strlen(name);
                   1515:        } else
                   1516:                name = iname;
1.22      schwarze 1517:
1.110     schwarze 1518:        if (namesz == 0 && tok != ROFF_ig) {
1.93      schwarze 1519:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse,
                   1520:                    ln, ppos, roffs[tok].name);
                   1521:                return(ROFF_IGN);
                   1522:        }
1.2       schwarze 1523:
1.16      schwarze 1524:        roffnode_push(r, tok, name, ln, ppos);
                   1525:
                   1526:        /*
                   1527:         * At the beginning of a `de' macro, clear the existing string
                   1528:         * with the same name, if there is one.  New content will be
1.66      schwarze 1529:         * appended from roff_block_text() in multiline mode.
1.16      schwarze 1530:         */
1.22      schwarze 1531:
1.110     schwarze 1532:        if (tok == ROFF_de || tok == ROFF_dei)
1.86      schwarze 1533:                roff_setstrn(&r->strtab, name, namesz, "", 0, 0);
1.2       schwarze 1534:
1.110     schwarze 1535:        if (*cp == '\0')
1.2       schwarze 1536:                return(ROFF_IGN);
1.1       schwarze 1537:
1.93      schwarze 1538:        /* Get the custom end marker. */
1.22      schwarze 1539:
1.93      schwarze 1540:        iname = cp;
1.86      schwarze 1541:        namesz = roff_getname(r, &cp, ln, ppos);
1.93      schwarze 1542:
                   1543:        /* Resolve the end marker if it is indirect. */
                   1544:
1.110     schwarze 1545:        if (namesz && (tok == ROFF_dei || tok == ROFF_ami)) {
                   1546:                if ((name = roff_getstrn(r, iname, namesz)) == NULL) {
1.93      schwarze 1547:                        mandoc_vmsg(MANDOCERR_STR_UNDEF,
1.110     schwarze 1548:                            r->parse, ln, (int)(iname - buf->buf),
1.93      schwarze 1549:                            "%.*s", (int)namesz, iname);
                   1550:                        namesz = 0;
                   1551:                } else
                   1552:                        namesz = strlen(name);
                   1553:        } else
                   1554:                name = iname;
                   1555:
1.86      schwarze 1556:        if (namesz)
                   1557:                r->last->end = mandoc_strndup(name, namesz);
1.2       schwarze 1558:
1.110     schwarze 1559:        if (*cp != '\0')
1.90      schwarze 1560:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   1561:                    ln, pos, ".%s ... %s", roffs[tok].name, cp);
1.1       schwarze 1562:
                   1563:        return(ROFF_IGN);
                   1564: }
                   1565:
                   1566: static enum rofferr
1.2       schwarze 1567: roff_block_sub(ROFF_ARGS)
1.1       schwarze 1568: {
1.2       schwarze 1569:        enum rofft      t;
                   1570:        int             i, j;
                   1571:
                   1572:        /*
                   1573:         * First check whether a custom macro exists at this level.  If
                   1574:         * it does, then check against it.  This is some of groff's
                   1575:         * stranger behaviours.  If we encountered a custom end-scope
                   1576:         * tag and that tag also happens to be a "real" macro, then we
                   1577:         * need to try interpreting it again as a real macro.  If it's
                   1578:         * not, then return ignore.  Else continue.
                   1579:         */
                   1580:
                   1581:        if (r->last->end) {
1.35      schwarze 1582:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.110     schwarze 1583:                        if (buf->buf[i] != r->last->end[j])
1.2       schwarze 1584:                                break;
1.1       schwarze 1585:
1.110     schwarze 1586:                if (r->last->end[j] == '\0' &&
                   1587:                    (buf->buf[i] == '\0' ||
                   1588:                     buf->buf[i] == ' ' ||
                   1589:                     buf->buf[i] == '\t')) {
1.2       schwarze 1590:                        roffnode_pop(r);
                   1591:                        roffnode_cleanscope(r);
1.1       schwarze 1592:
1.110     schwarze 1593:                        while (buf->buf[i] == ' ' || buf->buf[i] == '\t')
1.35      schwarze 1594:                                i++;
                   1595:
                   1596:                        pos = i;
1.110     schwarze 1597:                        if (roff_parse(r, buf->buf, &pos, ln, ppos) !=
                   1598:                            ROFF_MAX)
1.2       schwarze 1599:                                return(ROFF_RERUN);
                   1600:                        return(ROFF_IGN);
                   1601:                }
1.1       schwarze 1602:        }
                   1603:
1.2       schwarze 1604:        /*
                   1605:         * If we have no custom end-query or lookup failed, then try
                   1606:         * pulling it out of the hashtable.
                   1607:         */
1.1       schwarze 1608:
1.110     schwarze 1609:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.1       schwarze 1610:
1.110     schwarze 1611:        if (t != ROFF_cblock) {
                   1612:                if (tok != ROFF_ig)
                   1613:                        roff_setstr(r, r->last->name, buf->buf + ppos, 2);
1.1       schwarze 1614:                return(ROFF_IGN);
1.16      schwarze 1615:        }
1.1       schwarze 1616:
1.2       schwarze 1617:        assert(roffs[t].proc);
1.110     schwarze 1618:        return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.2       schwarze 1619: }
                   1620:
                   1621: static enum rofferr
                   1622: roff_block_text(ROFF_ARGS)
                   1623: {
                   1624:
1.110     schwarze 1625:        if (tok != ROFF_ig)
                   1626:                roff_setstr(r, r->last->name, buf->buf + pos, 2);
1.16      schwarze 1627:
1.2       schwarze 1628:        return(ROFF_IGN);
                   1629: }
                   1630:
                   1631: static enum rofferr
                   1632: roff_cond_sub(ROFF_ARGS)
                   1633: {
                   1634:        enum rofft       t;
1.37      schwarze 1635:        char            *ep;
1.71      schwarze 1636:        int              rr;
1.2       schwarze 1637:
                   1638:        rr = r->last->rule;
1.37      schwarze 1639:        roffnode_cleanscope(r);
1.110     schwarze 1640:        t = roff_parse(r, buf->buf, &pos, ln, ppos);
1.2       schwarze 1641:
1.37      schwarze 1642:        /*
1.50      schwarze 1643:         * Fully handle known macros when they are structurally
                   1644:         * required or when the conditional evaluated to true.
1.5       schwarze 1645:         */
                   1646:
1.110     schwarze 1647:        if ((t != ROFF_MAX) &&
                   1648:            (rr || roffs[t].flags & ROFFMAC_STRUCT)) {
1.50      schwarze 1649:                assert(roffs[t].proc);
1.110     schwarze 1650:                return((*roffs[t].proc)(r, t, buf, ln, ppos, pos, offs));
1.50      schwarze 1651:        }
1.39      schwarze 1652:
1.69      schwarze 1653:        /*
                   1654:         * If `\}' occurs on a macro line without a preceding macro,
                   1655:         * drop the line completely.
                   1656:         */
                   1657:
1.110     schwarze 1658:        ep = buf->buf + pos;
                   1659:        if (ep[0] == '\\' && ep[1] == '}')
1.71      schwarze 1660:                rr = 0;
1.69      schwarze 1661:
1.50      schwarze 1662:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1663:
1.110     schwarze 1664:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1665:                if (*(++ep) == '}') {
1.70      schwarze 1666:                        *ep = '&';
1.110     schwarze 1667:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1668:                }
1.119     schwarze 1669:                if (*ep != '\0')
                   1670:                        ++ep;
1.50      schwarze 1671:        }
1.71      schwarze 1672:        return(rr ? ROFF_CONT : ROFF_IGN);
1.2       schwarze 1673: }
                   1674:
                   1675: static enum rofferr
                   1676: roff_cond_text(ROFF_ARGS)
                   1677: {
1.37      schwarze 1678:        char            *ep;
1.71      schwarze 1679:        int              rr;
1.2       schwarze 1680:
                   1681:        rr = r->last->rule;
1.37      schwarze 1682:        roffnode_cleanscope(r);
1.1       schwarze 1683:
1.110     schwarze 1684:        ep = buf->buf + pos;
                   1685:        while ((ep = strchr(ep, '\\')) != NULL) {
                   1686:                if (*(++ep) == '}') {
1.70      schwarze 1687:                        *ep = '&';
1.110     schwarze 1688:                        roff_ccond(r, ln, ep - buf->buf - 1);
1.70      schwarze 1689:                }
1.119     schwarze 1690:                if (*ep != '\0')
                   1691:                        ++ep;
1.2       schwarze 1692:        }
1.71      schwarze 1693:        return(rr ? ROFF_CONT : ROFF_IGN);
1.2       schwarze 1694: }
                   1695:
1.77      schwarze 1696: /*
                   1697:  * Parse a single signed integer number.  Stop at the first non-digit.
                   1698:  * If there is at least one digit, return success and advance the
                   1699:  * parse point, else return failure and let the parse point unchanged.
                   1700:  * Ignore overflows, treat them just like the C language.
                   1701:  */
1.56      schwarze 1702: static int
1.133     schwarze 1703: roff_getnum(const char *v, int *pos, int *res, int flags)
1.56      schwarze 1704: {
1.133     schwarze 1705:        int      myres, scaled, n, p;
1.79      schwarze 1706:
                   1707:        if (NULL == res)
                   1708:                res = &myres;
1.56      schwarze 1709:
                   1710:        p = *pos;
                   1711:        n = v[p] == '-';
1.133     schwarze 1712:        if (n || v[p] == '+')
1.56      schwarze 1713:                p++;
                   1714:
1.133     schwarze 1715:        if (flags & ROFFNUM_WHITE)
                   1716:                while (isspace((unsigned char)v[p]))
                   1717:                        p++;
                   1718:
1.56      schwarze 1719:        for (*res = 0; isdigit((unsigned char)v[p]); p++)
1.77      schwarze 1720:                *res = 10 * *res + v[p] - '0';
1.56      schwarze 1721:        if (p == *pos + n)
                   1722:                return 0;
                   1723:
                   1724:        if (n)
                   1725:                *res = -*res;
                   1726:
1.126     schwarze 1727:        /* Each number may be followed by one optional scaling unit. */
                   1728:
                   1729:        switch (v[p]) {
                   1730:        case 'f':
1.133     schwarze 1731:                scaled = *res * 65536;
1.126     schwarze 1732:                break;
                   1733:        case 'i':
1.133     schwarze 1734:                scaled = *res * 240;
1.126     schwarze 1735:                break;
                   1736:        case 'c':
1.133     schwarze 1737:                scaled = *res * 240 / 2.54;
1.126     schwarze 1738:                break;
                   1739:        case 'v':
                   1740:                /* FALLTROUGH */
                   1741:        case 'P':
1.133     schwarze 1742:                scaled = *res * 40;
1.126     schwarze 1743:                break;
                   1744:        case 'm':
                   1745:                /* FALLTROUGH */
                   1746:        case 'n':
1.133     schwarze 1747:                scaled = *res * 24;
1.126     schwarze 1748:                break;
                   1749:        case 'p':
1.133     schwarze 1750:                scaled = *res * 10 / 3;
1.126     schwarze 1751:                break;
                   1752:        case 'u':
1.133     schwarze 1753:                scaled = *res;
1.126     schwarze 1754:                break;
                   1755:        case 'M':
1.133     schwarze 1756:                scaled = *res * 6 / 25;
1.126     schwarze 1757:                break;
                   1758:        default:
1.133     schwarze 1759:                scaled = *res;
1.126     schwarze 1760:                p--;
                   1761:                break;
                   1762:        }
1.133     schwarze 1763:        if (flags & ROFFNUM_SCALE)
                   1764:                *res = scaled;
1.126     schwarze 1765:
                   1766:        *pos = p + 1;
                   1767:        return(1);
1.56      schwarze 1768: }
                   1769:
1.71      schwarze 1770: /*
                   1771:  * Evaluate a string comparison condition.
                   1772:  * The first character is the delimiter.
                   1773:  * Succeed if the string up to its second occurrence
                   1774:  * matches the string up to its third occurence.
                   1775:  * Advance the cursor after the third occurrence
                   1776:  * or lacking that, to the end of the line.
                   1777:  */
                   1778: static int
                   1779: roff_evalstrcond(const char *v, int *pos)
                   1780: {
                   1781:        const char      *s1, *s2, *s3;
                   1782:        int              match;
                   1783:
                   1784:        match = 0;
                   1785:        s1 = v + *pos;          /* initial delimiter */
                   1786:        s2 = s1 + 1;            /* for scanning the first string */
                   1787:        s3 = strchr(s2, *s1);   /* for scanning the second string */
                   1788:
                   1789:        if (NULL == s3)         /* found no middle delimiter */
                   1790:                goto out;
                   1791:
                   1792:        while ('\0' != *++s3) {
                   1793:                if (*s2 != *s3) {  /* mismatch */
                   1794:                        s3 = strchr(s3, *s1);
                   1795:                        break;
                   1796:                }
                   1797:                if (*s3 == *s1) {  /* found the final delimiter */
                   1798:                        match = 1;
                   1799:                        break;
                   1800:                }
                   1801:                s2++;
                   1802:        }
                   1803:
                   1804: out:
                   1805:        if (NULL == s3)
                   1806:                s3 = strchr(s2, '\0');
1.114     schwarze 1807:        else if (*s3 != '\0')
1.71      schwarze 1808:                s3++;
                   1809:        *pos = s3 - v;
                   1810:        return(match);
                   1811: }
                   1812:
1.77      schwarze 1813: /*
                   1814:  * Evaluate an optionally negated single character, numerical,
                   1815:  * or string condition.
                   1816:  */
1.71      schwarze 1817: static int
1.106     schwarze 1818: roff_evalcond(struct roff *r, int ln, const char *v, int *pos)
1.5       schwarze 1819: {
1.113     schwarze 1820:        int      number, savepos, wanttrue;
1.5       schwarze 1821:
1.71      schwarze 1822:        if ('!' == v[*pos]) {
                   1823:                wanttrue = 0;
                   1824:                (*pos)++;
                   1825:        } else
                   1826:                wanttrue = 1;
                   1827:
1.5       schwarze 1828:        switch (v[*pos]) {
1.112     schwarze 1829:        case '\0':
                   1830:                return(0);
1.80      schwarze 1831:        case 'n':
1.71      schwarze 1832:                /* FALLTHROUGH */
1.80      schwarze 1833:        case 'o':
1.5       schwarze 1834:                (*pos)++;
1.71      schwarze 1835:                return(wanttrue);
1.80      schwarze 1836:        case 'c':
1.71      schwarze 1837:                /* FALLTHROUGH */
1.80      schwarze 1838:        case 'd':
1.71      schwarze 1839:                /* FALLTHROUGH */
1.80      schwarze 1840:        case 'e':
1.5       schwarze 1841:                /* FALLTHROUGH */
1.80      schwarze 1842:        case 'r':
1.5       schwarze 1843:                /* FALLTHROUGH */
1.80      schwarze 1844:        case 't':
1.111     schwarze 1845:                /* FALLTHROUGH */
                   1846:        case 'v':
1.5       schwarze 1847:                (*pos)++;
1.71      schwarze 1848:                return(!wanttrue);
1.5       schwarze 1849:        default:
                   1850:                break;
                   1851:        }
                   1852:
1.113     schwarze 1853:        savepos = *pos;
1.133     schwarze 1854:        if (roff_evalnum(r, ln, v, pos, &number, ROFFNUM_SCALE))
1.77      schwarze 1855:                return((number > 0) == wanttrue);
1.113     schwarze 1856:        else if (*pos == savepos)
                   1857:                return(roff_evalstrcond(v, pos) == wanttrue);
1.77      schwarze 1858:        else
1.113     schwarze 1859:                return (0);
1.5       schwarze 1860: }
                   1861:
1.2       schwarze 1862: static enum rofferr
1.21      schwarze 1863: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 1864: {
1.30      schwarze 1865:
1.21      schwarze 1866:        return(ROFF_IGN);
                   1867: }
                   1868:
                   1869: static enum rofferr
1.123     schwarze 1870: roff_insec(ROFF_ARGS)
                   1871: {
                   1872:
                   1873:        mandoc_msg(MANDOCERR_REQ_INSEC, r->parse,
                   1874:            ln, ppos, roffs[tok].name);
                   1875:        return(ROFF_IGN);
                   1876: }
                   1877:
                   1878: static enum rofferr
                   1879: roff_unsupp(ROFF_ARGS)
                   1880: {
                   1881:
                   1882:        mandoc_msg(MANDOCERR_REQ_UNSUPP, r->parse,
                   1883:            ln, ppos, roffs[tok].name);
                   1884:        return(ROFF_IGN);
                   1885: }
                   1886:
                   1887: static enum rofferr
1.2       schwarze 1888: roff_cond(ROFF_ARGS)
                   1889: {
1.46      schwarze 1890:
                   1891:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 1892:
1.80      schwarze 1893:        /*
1.35      schwarze 1894:         * An `.el' has no conditional body: it will consume the value
                   1895:         * of the current rstack entry set in prior `ie' calls or
1.80      schwarze 1896:         * defaults to DENY.
1.35      schwarze 1897:         *
                   1898:         * If we're not an `el', however, then evaluate the conditional.
                   1899:         */
1.1       schwarze 1900:
1.110     schwarze 1901:        r->last->rule = tok == ROFF_el ?
1.80      schwarze 1902:            (r->rstackpos < 0 ? 0 : r->rstack[r->rstackpos--]) :
1.110     schwarze 1903:            roff_evalcond(r, ln, buf->buf, &pos);
1.2       schwarze 1904:
1.35      schwarze 1905:        /*
                   1906:         * An if-else will put the NEGATION of the current evaluated
                   1907:         * conditional into the stack of rules.
                   1908:         */
                   1909:
1.110     schwarze 1910:        if (tok == ROFF_ie) {
1.96      schwarze 1911:                if (r->rstackpos + 1 == r->rstacksz) {
                   1912:                        r->rstacksz += 16;
                   1913:                        r->rstack = mandoc_reallocarray(r->rstack,
                   1914:                            r->rstacksz, sizeof(int));
1.35      schwarze 1915:                }
1.71      schwarze 1916:                r->rstack[++r->rstackpos] = !r->last->rule;
1.2       schwarze 1917:        }
1.5       schwarze 1918:
                   1919:        /* If the parent has false as its rule, then so do we. */
                   1920:
1.71      schwarze 1921:        if (r->last->parent && !r->last->parent->rule)
                   1922:                r->last->rule = 0;
1.5       schwarze 1923:
                   1924:        /*
1.46      schwarze 1925:         * Determine scope.
                   1926:         * If there is nothing on the line after the conditional,
                   1927:         * not even whitespace, use next-line scope.
1.5       schwarze 1928:         */
1.2       schwarze 1929:
1.110     schwarze 1930:        if (buf->buf[pos] == '\0') {
1.46      schwarze 1931:                r->last->endspan = 2;
                   1932:                goto out;
                   1933:        }
                   1934:
1.110     schwarze 1935:        while (buf->buf[pos] == ' ')
1.46      schwarze 1936:                pos++;
                   1937:
                   1938:        /* An opening brace requests multiline scope. */
1.2       schwarze 1939:
1.110     schwarze 1940:        if (buf->buf[pos] == '\\' && buf->buf[pos + 1] == '{') {
1.2       schwarze 1941:                r->last->endspan = -1;
                   1942:                pos += 2;
1.46      schwarze 1943:                goto out;
1.80      schwarze 1944:        }
1.2       schwarze 1945:
                   1946:        /*
1.46      schwarze 1947:         * Anything else following the conditional causes
                   1948:         * single-line scope.  Warn if the scope contains
                   1949:         * nothing but trailing whitespace.
1.2       schwarze 1950:         */
                   1951:
1.110     schwarze 1952:        if (buf->buf[pos] == '\0')
1.89      schwarze 1953:                mandoc_msg(MANDOCERR_COND_EMPTY, r->parse,
                   1954:                    ln, ppos, roffs[tok].name);
1.2       schwarze 1955:
1.46      schwarze 1956:        r->last->endspan = 1;
1.1       schwarze 1957:
1.46      schwarze 1958: out:
1.2       schwarze 1959:        *offs = pos;
                   1960:        return(ROFF_RERUN);
1.1       schwarze 1961: }
                   1962:
1.2       schwarze 1963: static enum rofferr
1.7       schwarze 1964: roff_ds(ROFF_ARGS)
                   1965: {
1.85      schwarze 1966:        char            *string;
                   1967:        const char      *name;
                   1968:        size_t           namesz;
1.10      schwarze 1969:
1.123     schwarze 1970:        /* Ignore groff compatibility mode for now. */
                   1971:
                   1972:        if (tok == ROFF_ds1)
                   1973:                tok = ROFF_ds;
                   1974:        else if (tok == ROFF_as1)
                   1975:                tok = ROFF_as;
                   1976:
1.10      schwarze 1977:        /*
1.85      schwarze 1978:         * The first word is the name of the string.
                   1979:         * If it is empty or terminated by an escape sequence,
                   1980:         * abort the `ds' request without defining anything.
1.10      schwarze 1981:         */
1.7       schwarze 1982:
1.110     schwarze 1983:        name = string = buf->buf + pos;
                   1984:        if (*name == '\0')
1.7       schwarze 1985:                return(ROFF_IGN);
                   1986:
1.85      schwarze 1987:        namesz = roff_getname(r, &string, ln, pos);
1.110     schwarze 1988:        if (name[namesz] == '\\')
1.85      schwarze 1989:                return(ROFF_IGN);
                   1990:
                   1991:        /* Read past the initial double-quote, if any. */
1.110     schwarze 1992:        if (*string == '"')
1.7       schwarze 1993:                string++;
                   1994:
1.10      schwarze 1995:        /* The rest is the value. */
1.85      schwarze 1996:        roff_setstrn(&r->strtab, name, namesz, string, strlen(string),
                   1997:            ROFF_as == tok);
1.7       schwarze 1998:        return(ROFF_IGN);
                   1999: }
                   2000:
1.77      schwarze 2001: /*
                   2002:  * Parse a single operator, one or two characters long.
                   2003:  * If the operator is recognized, return success and advance the
                   2004:  * parse point, else return failure and let the parse point unchanged.
                   2005:  */
                   2006: static int
                   2007: roff_getop(const char *v, int *pos, char *res)
                   2008: {
                   2009:
                   2010:        *res = v[*pos];
                   2011:
                   2012:        switch (*res) {
1.80      schwarze 2013:        case '+':
1.77      schwarze 2014:                /* FALLTHROUGH */
1.80      schwarze 2015:        case '-':
1.77      schwarze 2016:                /* FALLTHROUGH */
1.80      schwarze 2017:        case '*':
1.77      schwarze 2018:                /* FALLTHROUGH */
1.80      schwarze 2019:        case '/':
1.77      schwarze 2020:                /* FALLTHROUGH */
1.80      schwarze 2021:        case '%':
1.77      schwarze 2022:                /* FALLTHROUGH */
1.80      schwarze 2023:        case '&':
1.77      schwarze 2024:                /* FALLTHROUGH */
1.80      schwarze 2025:        case ':':
1.77      schwarze 2026:                break;
                   2027:        case '<':
                   2028:                switch (v[*pos + 1]) {
1.80      schwarze 2029:                case '=':
1.77      schwarze 2030:                        *res = 'l';
                   2031:                        (*pos)++;
                   2032:                        break;
1.80      schwarze 2033:                case '>':
1.77      schwarze 2034:                        *res = '!';
                   2035:                        (*pos)++;
                   2036:                        break;
1.80      schwarze 2037:                case '?':
1.77      schwarze 2038:                        *res = 'i';
                   2039:                        (*pos)++;
                   2040:                        break;
                   2041:                default:
                   2042:                        break;
                   2043:                }
                   2044:                break;
                   2045:        case '>':
                   2046:                switch (v[*pos + 1]) {
1.80      schwarze 2047:                case '=':
1.77      schwarze 2048:                        *res = 'g';
                   2049:                        (*pos)++;
                   2050:                        break;
1.80      schwarze 2051:                case '?':
1.77      schwarze 2052:                        *res = 'a';
                   2053:                        (*pos)++;
                   2054:                        break;
                   2055:                default:
                   2056:                        break;
                   2057:                }
                   2058:                break;
                   2059:        case '=':
                   2060:                if ('=' == v[*pos + 1])
                   2061:                        (*pos)++;
                   2062:                break;
                   2063:        default:
                   2064:                return(0);
                   2065:        }
                   2066:        (*pos)++;
                   2067:
                   2068:        return(*res);
                   2069: }
                   2070:
                   2071: /*
                   2072:  * Evaluate either a parenthesized numeric expression
                   2073:  * or a single signed integer number.
                   2074:  */
                   2075: static int
1.106     schwarze 2076: roff_evalpar(struct roff *r, int ln,
1.133     schwarze 2077:        const char *v, int *pos, int *res, int flags)
1.77      schwarze 2078: {
                   2079:
                   2080:        if ('(' != v[*pos])
1.133     schwarze 2081:                return(roff_getnum(v, pos, res, flags));
1.77      schwarze 2082:
                   2083:        (*pos)++;
1.133     schwarze 2084:        if ( ! roff_evalnum(r, ln, v, pos, res, flags | ROFFNUM_WHITE))
1.77      schwarze 2085:                return(0);
                   2086:
1.79      schwarze 2087:        /*
                   2088:         * Omission of the closing parenthesis
                   2089:         * is an error in validation mode,
                   2090:         * but ignored in evaluation mode.
                   2091:         */
                   2092:
1.77      schwarze 2093:        if (')' == v[*pos])
                   2094:                (*pos)++;
1.79      schwarze 2095:        else if (NULL == res)
                   2096:                return(0);
1.77      schwarze 2097:
                   2098:        return(1);
                   2099: }
                   2100:
                   2101: /*
                   2102:  * Evaluate a complete numeric expression.
                   2103:  * Proceed left to right, there is no concept of precedence.
                   2104:  */
                   2105: static int
1.106     schwarze 2106: roff_evalnum(struct roff *r, int ln, const char *v,
1.133     schwarze 2107:        int *pos, int *res, int flags)
1.77      schwarze 2108: {
                   2109:        int              mypos, operand2;
                   2110:        char             operator;
                   2111:
                   2112:        if (NULL == pos) {
                   2113:                mypos = 0;
                   2114:                pos = &mypos;
                   2115:        }
                   2116:
1.133     schwarze 2117:        if (flags & ROFFNUM_WHITE)
1.77      schwarze 2118:                while (isspace((unsigned char)v[*pos]))
                   2119:                        (*pos)++;
                   2120:
1.133     schwarze 2121:        if ( ! roff_evalpar(r, ln, v, pos, res, flags))
1.77      schwarze 2122:                return(0);
                   2123:
                   2124:        while (1) {
1.133     schwarze 2125:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2126:                        while (isspace((unsigned char)v[*pos]))
                   2127:                                (*pos)++;
                   2128:
                   2129:                if ( ! roff_getop(v, pos, &operator))
                   2130:                        break;
                   2131:
1.133     schwarze 2132:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2133:                        while (isspace((unsigned char)v[*pos]))
                   2134:                                (*pos)++;
                   2135:
1.133     schwarze 2136:                if ( ! roff_evalpar(r, ln, v, pos, &operand2, flags))
1.77      schwarze 2137:                        return(0);
                   2138:
1.133     schwarze 2139:                if (flags & ROFFNUM_WHITE)
1.77      schwarze 2140:                        while (isspace((unsigned char)v[*pos]))
                   2141:                                (*pos)++;
1.79      schwarze 2142:
                   2143:                if (NULL == res)
                   2144:                        continue;
1.77      schwarze 2145:
                   2146:                switch (operator) {
1.80      schwarze 2147:                case '+':
1.77      schwarze 2148:                        *res += operand2;
                   2149:                        break;
1.80      schwarze 2150:                case '-':
1.77      schwarze 2151:                        *res -= operand2;
                   2152:                        break;
1.80      schwarze 2153:                case '*':
1.77      schwarze 2154:                        *res *= operand2;
                   2155:                        break;
1.80      schwarze 2156:                case '/':
1.116     schwarze 2157:                        if (operand2 == 0) {
1.106     schwarze 2158:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2159:                                        r->parse, ln, *pos, v);
                   2160:                                *res = 0;
                   2161:                                break;
                   2162:                        }
1.77      schwarze 2163:                        *res /= operand2;
                   2164:                        break;
1.80      schwarze 2165:                case '%':
1.116     schwarze 2166:                        if (operand2 == 0) {
                   2167:                                mandoc_msg(MANDOCERR_DIVZERO,
                   2168:                                        r->parse, ln, *pos, v);
                   2169:                                *res = 0;
                   2170:                                break;
                   2171:                        }
1.77      schwarze 2172:                        *res %= operand2;
                   2173:                        break;
1.80      schwarze 2174:                case '<':
1.77      schwarze 2175:                        *res = *res < operand2;
                   2176:                        break;
1.80      schwarze 2177:                case '>':
1.77      schwarze 2178:                        *res = *res > operand2;
                   2179:                        break;
1.80      schwarze 2180:                case 'l':
1.77      schwarze 2181:                        *res = *res <= operand2;
                   2182:                        break;
1.80      schwarze 2183:                case 'g':
1.77      schwarze 2184:                        *res = *res >= operand2;
                   2185:                        break;
1.80      schwarze 2186:                case '=':
1.77      schwarze 2187:                        *res = *res == operand2;
                   2188:                        break;
1.80      schwarze 2189:                case '!':
1.77      schwarze 2190:                        *res = *res != operand2;
                   2191:                        break;
1.80      schwarze 2192:                case '&':
1.77      schwarze 2193:                        *res = *res && operand2;
                   2194:                        break;
1.80      schwarze 2195:                case ':':
1.77      schwarze 2196:                        *res = *res || operand2;
                   2197:                        break;
1.80      schwarze 2198:                case 'i':
1.77      schwarze 2199:                        if (operand2 < *res)
                   2200:                                *res = operand2;
                   2201:                        break;
1.80      schwarze 2202:                case 'a':
1.77      schwarze 2203:                        if (operand2 > *res)
                   2204:                                *res = operand2;
                   2205:                        break;
                   2206:                default:
                   2207:                        abort();
                   2208:                }
                   2209:        }
                   2210:        return(1);
                   2211: }
                   2212:
1.52      schwarze 2213: void
1.60      schwarze 2214: roff_setreg(struct roff *r, const char *name, int val, char sign)
1.41      schwarze 2215: {
1.52      schwarze 2216:        struct roffreg  *reg;
                   2217:
                   2218:        /* Search for an existing register with the same name. */
                   2219:        reg = r->regtab;
                   2220:
                   2221:        while (reg && strcmp(name, reg->key.p))
                   2222:                reg = reg->next;
1.41      schwarze 2223:
1.52      schwarze 2224:        if (NULL == reg) {
                   2225:                /* Create a new register. */
                   2226:                reg = mandoc_malloc(sizeof(struct roffreg));
                   2227:                reg->key.p = mandoc_strdup(name);
                   2228:                reg->key.sz = strlen(name);
1.60      schwarze 2229:                reg->val = 0;
1.52      schwarze 2230:                reg->next = r->regtab;
                   2231:                r->regtab = reg;
                   2232:        }
                   2233:
1.60      schwarze 2234:        if ('+' == sign)
                   2235:                reg->val += val;
                   2236:        else if ('-' == sign)
                   2237:                reg->val -= val;
                   2238:        else
                   2239:                reg->val = val;
1.41      schwarze 2240: }
                   2241:
1.65      schwarze 2242: /*
                   2243:  * Handle some predefined read-only number registers.
                   2244:  * For now, return -1 if the requested register is not predefined;
                   2245:  * in case a predefined read-only register having the value -1
                   2246:  * were to turn up, another special value would have to be chosen.
                   2247:  */
                   2248: static int
                   2249: roff_getregro(const char *name)
                   2250: {
                   2251:
                   2252:        switch (*name) {
1.80      schwarze 2253:        case 'A':  /* ASCII approximation mode is always off. */
1.65      schwarze 2254:                return(0);
1.80      schwarze 2255:        case 'g':  /* Groff compatibility mode is always on. */
1.65      schwarze 2256:                return(1);
1.80      schwarze 2257:        case 'H':  /* Fixed horizontal resolution. */
1.65      schwarze 2258:                return (24);
1.80      schwarze 2259:        case 'j':  /* Always adjust left margin only. */
1.65      schwarze 2260:                return(0);
1.80      schwarze 2261:        case 'T':  /* Some output device is always defined. */
1.65      schwarze 2262:                return(1);
1.80      schwarze 2263:        case 'V':  /* Fixed vertical resolution. */
1.65      schwarze 2264:                return (40);
                   2265:        default:
                   2266:                return (-1);
                   2267:        }
                   2268: }
                   2269:
1.53      schwarze 2270: int
1.52      schwarze 2271: roff_getreg(const struct roff *r, const char *name)
1.41      schwarze 2272: {
1.52      schwarze 2273:        struct roffreg  *reg;
1.65      schwarze 2274:        int              val;
                   2275:
                   2276:        if ('.' == name[0] && '\0' != name[1] && '\0' == name[2]) {
                   2277:                val = roff_getregro(name + 1);
                   2278:                if (-1 != val)
                   2279:                        return (val);
                   2280:        }
1.52      schwarze 2281:
                   2282:        for (reg = r->regtab; reg; reg = reg->next)
                   2283:                if (0 == strcmp(name, reg->key.p))
1.53      schwarze 2284:                        return(reg->val);
                   2285:
                   2286:        return(0);
                   2287: }
                   2288:
                   2289: static int
                   2290: roff_getregn(const struct roff *r, const char *name, size_t len)
                   2291: {
                   2292:        struct roffreg  *reg;
1.65      schwarze 2293:        int              val;
                   2294:
                   2295:        if ('.' == name[0] && 2 == len) {
                   2296:                val = roff_getregro(name + 1);
                   2297:                if (-1 != val)
                   2298:                        return (val);
                   2299:        }
1.53      schwarze 2300:
                   2301:        for (reg = r->regtab; reg; reg = reg->next)
                   2302:                if (len == reg->key.sz &&
                   2303:                    0 == strncmp(name, reg->key.p, len))
                   2304:                        return(reg->val);
1.41      schwarze 2305:
1.52      schwarze 2306:        return(0);
1.41      schwarze 2307: }
                   2308:
1.52      schwarze 2309: static void
                   2310: roff_freereg(struct roffreg *reg)
1.41      schwarze 2311: {
1.52      schwarze 2312:        struct roffreg  *old_reg;
1.41      schwarze 2313:
1.52      schwarze 2314:        while (NULL != reg) {
                   2315:                free(reg->key.p);
                   2316:                old_reg = reg;
                   2317:                reg = reg->next;
                   2318:                free(old_reg);
                   2319:        }
1.41      schwarze 2320: }
1.7       schwarze 2321:
                   2322: static enum rofferr
1.6       schwarze 2323: roff_nr(ROFF_ARGS)
1.1       schwarze 2324: {
1.85      schwarze 2325:        char            *key, *val;
                   2326:        size_t           keysz;
1.37      schwarze 2327:        int              iv;
1.60      schwarze 2328:        char             sign;
1.6       schwarze 2329:
1.110     schwarze 2330:        key = val = buf->buf + pos;
                   2331:        if (*key == '\0')
1.85      schwarze 2332:                return(ROFF_IGN);
                   2333:
                   2334:        keysz = roff_getname(r, &val, ln, pos);
1.110     schwarze 2335:        if (key[keysz] == '\\')
1.85      schwarze 2336:                return(ROFF_IGN);
                   2337:        key[keysz] = '\0';
1.6       schwarze 2338:
1.60      schwarze 2339:        sign = *val;
1.110     schwarze 2340:        if (sign == '+' || sign == '-')
1.60      schwarze 2341:                val++;
                   2342:
1.133     schwarze 2343:        if (roff_evalnum(r, ln, val, NULL, &iv, ROFFNUM_SCALE))
1.77      schwarze 2344:                roff_setreg(r, key, iv, sign);
1.1       schwarze 2345:
1.76      schwarze 2346:        return(ROFF_IGN);
                   2347: }
                   2348:
                   2349: static enum rofferr
                   2350: roff_rr(ROFF_ARGS)
                   2351: {
                   2352:        struct roffreg  *reg, **prev;
1.85      schwarze 2353:        char            *name, *cp;
                   2354:        size_t           namesz;
1.76      schwarze 2355:
1.110     schwarze 2356:        name = cp = buf->buf + pos;
                   2357:        if (*name == '\0')
1.85      schwarze 2358:                return(ROFF_IGN);
                   2359:        namesz = roff_getname(r, &cp, ln, pos);
                   2360:        name[namesz] = '\0';
1.76      schwarze 2361:
                   2362:        prev = &r->regtab;
                   2363:        while (1) {
                   2364:                reg = *prev;
1.110     schwarze 2365:                if (reg == NULL || !strcmp(name, reg->key.p))
1.76      schwarze 2366:                        break;
                   2367:                prev = &reg->next;
                   2368:        }
1.110     schwarze 2369:        if (reg != NULL) {
1.76      schwarze 2370:                *prev = reg->next;
                   2371:                free(reg->key.p);
                   2372:                free(reg);
                   2373:        }
1.29      schwarze 2374:        return(ROFF_IGN);
                   2375: }
                   2376:
                   2377: static enum rofferr
                   2378: roff_rm(ROFF_ARGS)
                   2379: {
                   2380:        const char       *name;
                   2381:        char             *cp;
1.85      schwarze 2382:        size_t            namesz;
1.29      schwarze 2383:
1.110     schwarze 2384:        cp = buf->buf + pos;
                   2385:        while (*cp != '\0') {
1.85      schwarze 2386:                name = cp;
1.110     schwarze 2387:                namesz = roff_getname(r, &cp, ln, (int)(cp - buf->buf));
1.85      schwarze 2388:                roff_setstrn(&r->strtab, name, namesz, NULL, 0, 0);
1.110     schwarze 2389:                if (name[namesz] == '\\')
1.85      schwarze 2390:                        break;
1.29      schwarze 2391:        }
1.51      schwarze 2392:        return(ROFF_IGN);
                   2393: }
                   2394:
                   2395: static enum rofferr
                   2396: roff_it(ROFF_ARGS)
                   2397: {
                   2398:        int              iv;
                   2399:
                   2400:        /* Parse the number of lines. */
1.133     schwarze 2401:
                   2402:        if ( ! roff_evalnum(r, ln, buf->buf, &pos, &iv, 0)) {
1.95      schwarze 2403:                mandoc_msg(MANDOCERR_IT_NONUM, r->parse,
1.110     schwarze 2404:                    ln, ppos, buf->buf + 1);
1.51      schwarze 2405:                return(ROFF_IGN);
                   2406:        }
                   2407:
1.134     schwarze 2408:        while (isspace((unsigned char)buf->buf[pos]))
                   2409:                pos++;
                   2410:
                   2411:        /*
                   2412:         * Arm the input line trap.
                   2413:         * Special-casing "an-trap" is an ugly workaround to cope
                   2414:         * with DocBook stupidly fiddling with man(7) internals.
                   2415:         */
1.133     schwarze 2416:
1.51      schwarze 2417:        roffit_lines = iv;
1.134     schwarze 2418:        roffit_macro = mandoc_strdup(iv != 1 ||
                   2419:            strcmp(buf->buf + pos, "an-trap") ?
                   2420:            buf->buf + pos : "br");
1.2       schwarze 2421:        return(ROFF_IGN);
1.47      schwarze 2422: }
                   2423:
                   2424: static enum rofferr
                   2425: roff_Dd(ROFF_ARGS)
                   2426: {
                   2427:        const char *const       *cp;
                   2428:
1.99      schwarze 2429:        if ((r->options & (MPARSE_MDOC | MPARSE_QUICK)) == 0)
1.47      schwarze 2430:                for (cp = __mdoc_reserved; *cp; cp++)
                   2431:                        roff_setstr(r, *cp, NULL, 0);
                   2432:
1.99      schwarze 2433:        if (r->format == 0)
                   2434:                r->format = MPARSE_MDOC;
                   2435:
1.47      schwarze 2436:        return(ROFF_CONT);
                   2437: }
                   2438:
                   2439: static enum rofferr
                   2440: roff_TH(ROFF_ARGS)
                   2441: {
                   2442:        const char *const       *cp;
                   2443:
1.99      schwarze 2444:        if ((r->options & MPARSE_QUICK) == 0)
1.47      schwarze 2445:                for (cp = __man_reserved; *cp; cp++)
                   2446:                        roff_setstr(r, *cp, NULL, 0);
                   2447:
1.99      schwarze 2448:        if (r->format == 0)
                   2449:                r->format = MPARSE_MAN;
                   2450:
1.47      schwarze 2451:        return(ROFF_CONT);
1.14      schwarze 2452: }
                   2453:
                   2454: static enum rofferr
1.27      schwarze 2455: roff_TE(ROFF_ARGS)
                   2456: {
                   2457:
                   2458:        if (NULL == r->tbl)
1.94      schwarze 2459:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2460:                    ln, ppos, "TE");
1.130     schwarze 2461:        else if ( ! tbl_end(&r->tbl)) {
                   2462:                free(buf->buf);
                   2463:                buf->buf = mandoc_strdup(".sp");
                   2464:                buf->sz = 4;
                   2465:                return(ROFF_REPARSE);
                   2466:        }
1.27      schwarze 2467:        return(ROFF_IGN);
                   2468: }
                   2469:
                   2470: static enum rofferr
                   2471: roff_T_(ROFF_ARGS)
                   2472: {
                   2473:
                   2474:        if (NULL == r->tbl)
1.94      schwarze 2475:                mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse,
                   2476:                    ln, ppos, "T&");
1.27      schwarze 2477:        else
                   2478:                tbl_restart(ppos, ln, r->tbl);
                   2479:
                   2480:        return(ROFF_IGN);
                   2481: }
                   2482:
1.102     schwarze 2483: /*
                   2484:  * Handle in-line equation delimiters.
                   2485:  */
                   2486: static enum rofferr
1.110     schwarze 2487: roff_eqndelim(struct roff *r, struct buf *buf, int pos)
1.41      schwarze 2488: {
1.105     schwarze 2489:        char            *cp1, *cp2;
                   2490:        const char      *bef_pr, *bef_nl, *mac, *aft_nl, *aft_pr;
1.41      schwarze 2491:
1.102     schwarze 2492:        /*
                   2493:         * Outside equations, look for an opening delimiter.
                   2494:         * If we are inside an equation, we already know it is
                   2495:         * in-line, or this function wouldn't have been called;
                   2496:         * so look for a closing delimiter.
                   2497:         */
                   2498:
1.110     schwarze 2499:        cp1 = buf->buf + pos;
1.102     schwarze 2500:        cp2 = strchr(cp1, r->eqn == NULL ?
                   2501:            r->last_eqn->odelim : r->last_eqn->cdelim);
                   2502:        if (cp2 == NULL)
                   2503:                return(ROFF_CONT);
                   2504:
1.105     schwarze 2505:        *cp2++ = '\0';
                   2506:        bef_pr = bef_nl = aft_nl = aft_pr = "";
                   2507:
                   2508:        /* Handle preceding text, protecting whitespace. */
                   2509:
1.110     schwarze 2510:        if (*buf->buf != '\0') {
1.105     schwarze 2511:                if (r->eqn == NULL)
                   2512:                        bef_pr = "\\&";
                   2513:                bef_nl = "\n";
                   2514:        }
                   2515:
                   2516:        /*
                   2517:         * Prepare replacing the delimiter with an equation macro
                   2518:         * and drop leading white space from the equation.
                   2519:         */
1.102     schwarze 2520:
1.105     schwarze 2521:        if (r->eqn == NULL) {
                   2522:                while (*cp2 == ' ')
                   2523:                        cp2++;
                   2524:                mac = ".EQ";
                   2525:        } else
                   2526:                mac = ".EN";
                   2527:
                   2528:        /* Handle following text, protecting whitespace. */
                   2529:
                   2530:        if (*cp2 != '\0') {
                   2531:                aft_nl = "\n";
                   2532:                if (r->eqn != NULL)
                   2533:                        aft_pr = "\\&";
                   2534:        }
                   2535:
                   2536:        /* Do the actual replacement. */
                   2537:
1.110     schwarze 2538:        buf->sz = mandoc_asprintf(&cp1, "%s%s%s%s%s%s%s", buf->buf,
1.105     schwarze 2539:            bef_pr, bef_nl, mac, aft_nl, aft_pr, cp2) + 1;
1.110     schwarze 2540:        free(buf->buf);
                   2541:        buf->buf = cp1;
1.102     schwarze 2542:
                   2543:        /* Toggle the in-line state of the eqn subsystem. */
                   2544:
                   2545:        r->eqn_inline = r->eqn == NULL;
                   2546:        return(ROFF_REPARSE);
1.41      schwarze 2547: }
                   2548:
1.107     schwarze 2549: static enum rofferr
                   2550: roff_EQ(ROFF_ARGS)
1.32      schwarze 2551: {
1.41      schwarze 2552:        struct eqn_node *e;
1.32      schwarze 2553:
1.110     schwarze 2554:        assert(r->eqn == NULL);
1.107     schwarze 2555:        e = eqn_alloc(ppos, ln, r->parse);
1.32      schwarze 2556:
1.102     schwarze 2557:        if (r->last_eqn) {
1.32      schwarze 2558:                r->last_eqn->next = e;
1.102     schwarze 2559:                e->delim = r->last_eqn->delim;
                   2560:                e->odelim = r->last_eqn->odelim;
                   2561:                e->cdelim = r->last_eqn->cdelim;
                   2562:        } else
1.32      schwarze 2563:                r->first_eqn = r->last_eqn = e;
                   2564:
                   2565:        r->eqn = r->last_eqn = e;
1.41      schwarze 2566:
1.110     schwarze 2567:        if (buf->buf[pos] != '\0')
1.107     schwarze 2568:                mandoc_vmsg(MANDOCERR_ARG_SKIP, r->parse, ln, pos,
1.110     schwarze 2569:                    ".EQ %s", buf->buf + pos);
1.41      schwarze 2570:
1.32      schwarze 2571:        return(ROFF_IGN);
                   2572: }
                   2573:
                   2574: static enum rofferr
                   2575: roff_EN(ROFF_ARGS)
                   2576: {
                   2577:
1.94      schwarze 2578:        mandoc_msg(MANDOCERR_BLK_NOTOPEN, r->parse, ln, ppos, "EN");
1.32      schwarze 2579:        return(ROFF_IGN);
                   2580: }
                   2581:
                   2582: static enum rofferr
1.27      schwarze 2583: roff_TS(ROFF_ARGS)
                   2584: {
1.49      schwarze 2585:        struct tbl_node *tbl;
1.27      schwarze 2586:
                   2587:        if (r->tbl) {
1.94      schwarze 2588:                mandoc_msg(MANDOCERR_BLK_BROKEN, r->parse,
                   2589:                    ln, ppos, "TS breaks TS");
1.41      schwarze 2590:                tbl_end(&r->tbl);
1.27      schwarze 2591:        }
                   2592:
1.49      schwarze 2593:        tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 2594:
                   2595:        if (r->last_tbl)
1.49      schwarze 2596:                r->last_tbl->next = tbl;
1.27      schwarze 2597:        else
1.49      schwarze 2598:                r->first_tbl = r->last_tbl = tbl;
1.27      schwarze 2599:
1.49      schwarze 2600:        r->tbl = r->last_tbl = tbl;
1.27      schwarze 2601:        return(ROFF_IGN);
1.123     schwarze 2602: }
                   2603:
                   2604: static enum rofferr
                   2605: roff_brp(ROFF_ARGS)
                   2606: {
                   2607:
                   2608:        buf->buf[pos - 1] = '\0';
                   2609:        return(ROFF_CONT);
1.27      schwarze 2610: }
                   2611:
                   2612: static enum rofferr
1.48      schwarze 2613: roff_cc(ROFF_ARGS)
                   2614: {
                   2615:        const char      *p;
                   2616:
1.110     schwarze 2617:        p = buf->buf + pos;
1.48      schwarze 2618:
1.110     schwarze 2619:        if (*p == '\0' || (r->control = *p++) == '.')
1.48      schwarze 2620:                r->control = 0;
                   2621:
1.110     schwarze 2622:        if (*p != '\0')
1.132     schwarze 2623:                mandoc_vmsg(MANDOCERR_ARG_EXCESS, r->parse,
                   2624:                    ln, p - buf->buf, "cc ... %s", p);
1.48      schwarze 2625:
                   2626:        return(ROFF_IGN);
                   2627: }
                   2628:
                   2629: static enum rofferr
1.42      schwarze 2630: roff_tr(ROFF_ARGS)
                   2631: {
                   2632:        const char      *p, *first, *second;
                   2633:        size_t           fsz, ssz;
                   2634:        enum mandoc_esc  esc;
                   2635:
1.110     schwarze 2636:        p = buf->buf + pos;
1.42      schwarze 2637:
1.110     schwarze 2638:        if (*p == '\0') {
1.132     schwarze 2639:                mandoc_msg(MANDOCERR_REQ_EMPTY, r->parse, ln, ppos, "tr");
1.42      schwarze 2640:                return(ROFF_IGN);
                   2641:        }
                   2642:
1.110     schwarze 2643:        while (*p != '\0') {
1.42      schwarze 2644:                fsz = ssz = 1;
                   2645:
                   2646:                first = p++;
1.110     schwarze 2647:                if (*first == '\\') {
1.42      schwarze 2648:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 2649:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 2650:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 2651:                                    ln, (int)(p - buf->buf), first);
1.42      schwarze 2652:                                return(ROFF_IGN);
                   2653:                        }
                   2654:                        fsz = (size_t)(p - first);
                   2655:                }
                   2656:
                   2657:                second = p++;
1.110     schwarze 2658:                if (*second == '\\') {
1.42      schwarze 2659:                        esc = mandoc_escape(&p, NULL, NULL);
1.110     schwarze 2660:                        if (esc == ESCAPE_ERROR) {
1.92      schwarze 2661:                                mandoc_msg(MANDOCERR_ESC_BAD, r->parse,
1.110     schwarze 2662:                                    ln, (int)(p - buf->buf), second);
1.42      schwarze 2663:                                return(ROFF_IGN);
                   2664:                        }
                   2665:                        ssz = (size_t)(p - second);
1.110     schwarze 2666:                } else if (*second == '\0') {
1.132     schwarze 2667:                        mandoc_vmsg(MANDOCERR_TR_ODD, r->parse,
                   2668:                            ln, first - buf->buf, "tr %s", first);
1.42      schwarze 2669:                        second = " ";
                   2670:                        p--;
                   2671:                }
                   2672:
                   2673:                if (fsz > 1) {
1.80      schwarze 2674:                        roff_setstrn(&r->xmbtab, first, fsz,
                   2675:                            second, ssz, 0);
1.42      schwarze 2676:                        continue;
                   2677:                }
                   2678:
1.110     schwarze 2679:                if (r->xtab == NULL)
1.80      schwarze 2680:                        r->xtab = mandoc_calloc(128,
                   2681:                            sizeof(struct roffstr));
1.42      schwarze 2682:
                   2683:                free(r->xtab[(int)*first].p);
                   2684:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   2685:                r->xtab[(int)*first].sz = ssz;
                   2686:        }
                   2687:
                   2688:        return(ROFF_IGN);
                   2689: }
                   2690:
                   2691: static enum rofferr
1.14      schwarze 2692: roff_so(ROFF_ARGS)
                   2693: {
1.121     schwarze 2694:        char *name, *cp;
1.15      schwarze 2695:
1.110     schwarze 2696:        name = buf->buf + pos;
1.97      schwarze 2697:        mandoc_vmsg(MANDOCERR_SO, r->parse, ln, ppos, "so %s", name);
1.14      schwarze 2698:
1.22      schwarze 2699:        /*
                   2700:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   2701:         * opening anything that's not in our cwd or anything beneath
                   2702:         * it.  Thus, explicitly disallow traversing up the file-system
                   2703:         * or using absolute paths.
                   2704:         */
                   2705:
1.110     schwarze 2706:        if (*name == '/' || strstr(name, "../") || strstr(name, "/..")) {
1.83      schwarze 2707:                mandoc_vmsg(MANDOCERR_SO_PATH, r->parse, ln, ppos,
                   2708:                    ".so %s", name);
1.121     schwarze 2709:                buf->sz = mandoc_asprintf(&cp,
                   2710:                    ".sp\nSee the file %s.\n.sp", name) + 1;
                   2711:                free(buf->buf);
                   2712:                buf->buf = cp;
                   2713:                *offs = 0;
                   2714:                return(ROFF_REPARSE);
1.14      schwarze 2715:        }
                   2716:
                   2717:        *offs = pos;
                   2718:        return(ROFF_SO);
1.7       schwarze 2719: }
                   2720:
1.16      schwarze 2721: static enum rofferr
                   2722: roff_userdef(ROFF_ARGS)
1.12      schwarze 2723: {
1.135     schwarze 2724:        const char       *arg[9], *ap;
1.16      schwarze 2725:        char             *cp, *n1, *n2;
1.25      schwarze 2726:        int               i;
1.135     schwarze 2727:        size_t            asz, rsz;
1.12      schwarze 2728:
1.16      schwarze 2729:        /*
                   2730:         * Collect pointers to macro argument strings
1.61      schwarze 2731:         * and NUL-terminate them.
1.16      schwarze 2732:         */
1.135     schwarze 2733:
1.110     schwarze 2734:        cp = buf->buf + pos;
1.25      schwarze 2735:        for (i = 0; i < 9; i++)
1.110     schwarze 2736:                arg[i] = *cp == '\0' ? "" :
1.35      schwarze 2737:                    mandoc_getarg(r->parse, &cp, ln, &pos);
1.16      schwarze 2738:
                   2739:        /*
                   2740:         * Expand macro arguments.
1.12      schwarze 2741:         */
1.135     schwarze 2742:
                   2743:        buf->sz = strlen(r->current_string) + 1;
                   2744:        n1 = cp = mandoc_malloc(buf->sz);
                   2745:        memcpy(n1, r->current_string, buf->sz);
                   2746:        while (*cp != '\0') {
                   2747:
                   2748:                /* Scan ahead for the next argument invocation. */
                   2749:
                   2750:                if (*cp++ != '\\')
                   2751:                        continue;
                   2752:                if (*cp++ != '$')
                   2753:                        continue;
                   2754:                i = *cp - '1';
                   2755:                if (0 > i || 8 < i)
1.16      schwarze 2756:                        continue;
1.135     schwarze 2757:                cp -= 2;
                   2758:
                   2759:                /*
                   2760:                 * Determine the size of the expanded argument,
                   2761:                 * taking escaping of quotes into account.
                   2762:                 */
                   2763:
                   2764:                asz = 0;
                   2765:                for (ap = arg[i]; *ap != '\0'; ap++) {
                   2766:                        asz++;
                   2767:                        if (*ap == '"')
                   2768:                                asz += 3;
                   2769:                }
                   2770:                if (asz != 3) {
                   2771:
                   2772:                        /*
                   2773:                         * Determine the size of the rest of the
                   2774:                         * unexpanded macro, including the NUL.
                   2775:                         */
                   2776:
                   2777:                        rsz = buf->sz - (cp - n1) - 3;
                   2778:
                   2779:                        /*
                   2780:                         * When shrinking, move before
                   2781:                         * releasing the storage.
                   2782:                         */
                   2783:
                   2784:                        if (asz < 3)
                   2785:                                memmove(cp + asz, cp + 3, rsz);
                   2786:
                   2787:                        /*
                   2788:                         * Resize the storage for the macro
                   2789:                         * and readjust the parse pointer.
                   2790:                         */
                   2791:
                   2792:                        buf->sz += asz - 3;
                   2793:                        n2 = mandoc_realloc(n1, buf->sz);
                   2794:                        cp = n2 + (cp - n1);
                   2795:                        n1 = n2;
                   2796:
                   2797:                        /*
                   2798:                         * When growing, make room
                   2799:                         * for the expanded argument.
                   2800:                         */
                   2801:
                   2802:                        if (asz > 3)
                   2803:                                memmove(cp + asz, cp + 3, rsz);
                   2804:                }
                   2805:
                   2806:                /* Copy the expanded argument, escaping quotes. */
                   2807:
                   2808:                n2 = cp;
                   2809:                for (ap = arg[i]; *ap != '\0'; ap++) {
                   2810:                        if (*ap == '"') {
                   2811:                                memcpy(n2, "\\(dq", 4);
                   2812:                                n2 += 4;
                   2813:                        } else
                   2814:                                *n2++ = *ap;
1.16      schwarze 2815:                }
1.12      schwarze 2816:        }
                   2817:
1.16      schwarze 2818:        /*
                   2819:         * Replace the macro invocation
                   2820:         * by the expanded macro.
                   2821:         */
1.135     schwarze 2822:
1.110     schwarze 2823:        free(buf->buf);
                   2824:        buf->buf = n1;
1.120     schwarze 2825:        *offs = 0;
1.16      schwarze 2826:
1.110     schwarze 2827:        return(buf->sz > 1 && buf->buf[buf->sz - 2] == '\n' ?
1.16      schwarze 2828:           ROFF_REPARSE : ROFF_APPEND);
1.12      schwarze 2829: }
1.28      schwarze 2830:
1.85      schwarze 2831: static size_t
1.28      schwarze 2832: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   2833: {
                   2834:        char     *name, *cp;
1.85      schwarze 2835:        size_t    namesz;
1.28      schwarze 2836:
                   2837:        name = *cpp;
                   2838:        if ('\0' == *name)
1.85      schwarze 2839:                return(0);
1.28      schwarze 2840:
1.85      schwarze 2841:        /* Read until end of name and terminate it with NUL. */
                   2842:        for (cp = name; 1; cp++) {
                   2843:                if ('\0' == *cp || ' ' == *cp) {
                   2844:                        namesz = cp - name;
                   2845:                        break;
                   2846:                }
1.28      schwarze 2847:                if ('\\' != *cp)
                   2848:                        continue;
1.88      schwarze 2849:                namesz = cp - name;
                   2850:                if ('{' == cp[1] || '}' == cp[1])
                   2851:                        break;
1.28      schwarze 2852:                cp++;
                   2853:                if ('\\' == *cp)
                   2854:                        continue;
1.97      schwarze 2855:                mandoc_vmsg(MANDOCERR_NAMESC, r->parse, ln, pos,
                   2856:                    "%.*s", (int)(cp - name + 1), name);
1.85      schwarze 2857:                mandoc_escape((const char **)&cp, NULL, NULL);
                   2858:                break;
1.28      schwarze 2859:        }
                   2860:
                   2861:        /* Read past spaces. */
                   2862:        while (' ' == *cp)
                   2863:                cp++;
                   2864:
                   2865:        *cpp = cp;
1.85      schwarze 2866:        return(namesz);
1.28      schwarze 2867: }
                   2868:
1.16      schwarze 2869: /*
                   2870:  * Store *string into the user-defined string called *name.
                   2871:  * To clear an existing entry, call with (*r, *name, NULL, 0).
1.66      schwarze 2872:  * append == 0: replace mode
                   2873:  * append == 1: single-line append mode
                   2874:  * append == 2: multiline append mode, append '\n' after each call
1.16      schwarze 2875:  */
1.8       schwarze 2876: static void
1.16      schwarze 2877: roff_setstr(struct roff *r, const char *name, const char *string,
1.66      schwarze 2878:        int append)
1.7       schwarze 2879: {
1.42      schwarze 2880:
                   2881:        roff_setstrn(&r->strtab, name, strlen(name), string,
1.80      schwarze 2882:            string ? strlen(string) : 0, append);
1.42      schwarze 2883: }
                   2884:
                   2885: static void
                   2886: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
1.66      schwarze 2887:                const char *string, size_t stringsz, int append)
1.42      schwarze 2888: {
                   2889:        struct roffkv   *n;
                   2890:        char            *c;
                   2891:        int              i;
                   2892:        size_t           oldch, newch;
1.7       schwarze 2893:
1.16      schwarze 2894:        /* Search for an existing string with the same name. */
1.42      schwarze 2895:        n = *r;
                   2896:
1.84      schwarze 2897:        while (n && (namesz != n->key.sz ||
                   2898:                        strncmp(n->key.p, name, namesz)))
1.7       schwarze 2899:                n = n->next;
1.8       schwarze 2900:
                   2901:        if (NULL == n) {
1.16      schwarze 2902:                /* Create a new string table entry. */
1.42      schwarze 2903:                n = mandoc_malloc(sizeof(struct roffkv));
                   2904:                n->key.p = mandoc_strndup(name, namesz);
                   2905:                n->key.sz = namesz;
                   2906:                n->val.p = NULL;
                   2907:                n->val.sz = 0;
                   2908:                n->next = *r;
                   2909:                *r = n;
1.66      schwarze 2910:        } else if (0 == append) {
1.42      schwarze 2911:                free(n->val.p);
                   2912:                n->val.p = NULL;
                   2913:                n->val.sz = 0;
1.16      schwarze 2914:        }
                   2915:
                   2916:        if (NULL == string)
                   2917:                return;
                   2918:
                   2919:        /*
                   2920:         * One additional byte for the '\n' in multiline mode,
                   2921:         * and one for the terminating '\0'.
                   2922:         */
1.66      schwarze 2923:        newch = stringsz + (1 < append ? 2u : 1u);
1.42      schwarze 2924:
                   2925:        if (NULL == n->val.p) {
                   2926:                n->val.p = mandoc_malloc(newch);
                   2927:                *n->val.p = '\0';
1.16      schwarze 2928:                oldch = 0;
                   2929:        } else {
1.42      schwarze 2930:                oldch = n->val.sz;
                   2931:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 2932:        }
                   2933:
                   2934:        /* Skip existing content in the destination buffer. */
1.42      schwarze 2935:        c = n->val.p + (int)oldch;
1.16      schwarze 2936:
                   2937:        /* Append new content to the destination buffer. */
1.42      schwarze 2938:        i = 0;
                   2939:        while (i < (int)stringsz) {
1.16      schwarze 2940:                /*
                   2941:                 * Rudimentary roff copy mode:
                   2942:                 * Handle escaped backslashes.
                   2943:                 */
1.42      schwarze 2944:                if ('\\' == string[i] && '\\' == string[i + 1])
                   2945:                        i++;
                   2946:                *c++ = string[i++];
1.16      schwarze 2947:        }
1.8       schwarze 2948:
1.16      schwarze 2949:        /* Append terminating bytes. */
1.66      schwarze 2950:        if (1 < append)
1.16      schwarze 2951:                *c++ = '\n';
1.42      schwarze 2952:
1.16      schwarze 2953:        *c = '\0';
1.42      schwarze 2954:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 2955: }
                   2956:
1.8       schwarze 2957: static const char *
                   2958: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7       schwarze 2959: {
1.42      schwarze 2960:        const struct roffkv *n;
1.64      schwarze 2961:        int i;
1.7       schwarze 2962:
1.42      schwarze 2963:        for (n = r->strtab; n; n = n->next)
1.80      schwarze 2964:                if (0 == strncmp(name, n->key.p, len) &&
                   2965:                    '\0' == n->key.p[(int)len])
1.42      schwarze 2966:                        return(n->val.p);
1.64      schwarze 2967:
                   2968:        for (i = 0; i < PREDEFS_MAX; i++)
                   2969:                if (0 == strncmp(name, predefs[i].name, len) &&
                   2970:                                '\0' == predefs[i].name[(int)len])
                   2971:                        return(predefs[i].str);
1.8       schwarze 2972:
1.42      schwarze 2973:        return(NULL);
1.7       schwarze 2974: }
                   2975:
1.8       schwarze 2976: static void
1.42      schwarze 2977: roff_freestr(struct roffkv *r)
1.7       schwarze 2978: {
1.42      schwarze 2979:        struct roffkv    *n, *nn;
1.7       schwarze 2980:
1.42      schwarze 2981:        for (n = r; n; n = nn) {
                   2982:                free(n->key.p);
                   2983:                free(n->val.p);
1.7       schwarze 2984:                nn = n->next;
                   2985:                free(n);
                   2986:        }
1.27      schwarze 2987: }
                   2988:
                   2989: const struct tbl_span *
                   2990: roff_span(const struct roff *r)
                   2991: {
1.80      schwarze 2992:
1.27      schwarze 2993:        return(r->tbl ? tbl_span(r->tbl) : NULL);
1.32      schwarze 2994: }
                   2995:
                   2996: const struct eqn *
                   2997: roff_eqn(const struct roff *r)
                   2998: {
1.80      schwarze 2999:
1.32      schwarze 3000:        return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.42      schwarze 3001: }
                   3002:
                   3003: /*
                   3004:  * Duplicate an input string, making the appropriate character
                   3005:  * conversations (as stipulated by `tr') along the way.
                   3006:  * Returns a heap-allocated string with all the replacements made.
                   3007:  */
                   3008: char *
                   3009: roff_strdup(const struct roff *r, const char *p)
                   3010: {
                   3011:        const struct roffkv *cp;
                   3012:        char            *res;
                   3013:        const char      *pp;
                   3014:        size_t           ssz, sz;
                   3015:        enum mandoc_esc  esc;
                   3016:
                   3017:        if (NULL == r->xmbtab && NULL == r->xtab)
                   3018:                return(mandoc_strdup(p));
                   3019:        else if ('\0' == *p)
                   3020:                return(mandoc_strdup(""));
                   3021:
                   3022:        /*
                   3023:         * Step through each character looking for term matches
                   3024:         * (remember that a `tr' can be invoked with an escape, which is
                   3025:         * a glyph but the escape is multi-character).
                   3026:         * We only do this if the character hash has been initialised
                   3027:         * and the string is >0 length.
                   3028:         */
                   3029:
                   3030:        res = NULL;
                   3031:        ssz = 0;
                   3032:
                   3033:        while ('\0' != *p) {
                   3034:                if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
                   3035:                        sz = r->xtab[(int)*p].sz;
                   3036:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3037:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   3038:                        ssz += sz;
                   3039:                        p++;
                   3040:                        continue;
                   3041:                } else if ('\\' != *p) {
                   3042:                        res = mandoc_realloc(res, ssz + 2);
                   3043:                        res[ssz++] = *p++;
                   3044:                        continue;
                   3045:                }
                   3046:
                   3047:                /* Search for term matches. */
                   3048:                for (cp = r->xmbtab; cp; cp = cp->next)
                   3049:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   3050:                                break;
                   3051:
                   3052:                if (NULL != cp) {
                   3053:                        /*
                   3054:                         * A match has been found.
                   3055:                         * Append the match to the array and move
                   3056:                         * forward by its keysize.
                   3057:                         */
1.80      schwarze 3058:                        res = mandoc_realloc(res,
                   3059:                            ssz + cp->val.sz + 1);
1.42      schwarze 3060:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   3061:                        ssz += cp->val.sz;
                   3062:                        p += (int)cp->key.sz;
                   3063:                        continue;
                   3064:                }
                   3065:
                   3066:                /*
                   3067:                 * Handle escapes carefully: we need to copy
                   3068:                 * over just the escape itself, or else we might
                   3069:                 * do replacements within the escape itself.
                   3070:                 * Make sure to pass along the bogus string.
                   3071:                 */
                   3072:                pp = p++;
                   3073:                esc = mandoc_escape(&p, NULL, NULL);
                   3074:                if (ESCAPE_ERROR == esc) {
                   3075:                        sz = strlen(pp);
                   3076:                        res = mandoc_realloc(res, ssz + sz + 1);
                   3077:                        memcpy(res + ssz, pp, sz);
                   3078:                        break;
                   3079:                }
1.80      schwarze 3080:                /*
                   3081:                 * We bail out on bad escapes.
1.42      schwarze 3082:                 * No need to warn: we already did so when
                   3083:                 * roff_res() was called.
                   3084:                 */
                   3085:                sz = (int)(p - pp);
                   3086:                res = mandoc_realloc(res, ssz + sz + 1);
                   3087:                memcpy(res + ssz, pp, sz);
                   3088:                ssz += sz;
                   3089:        }
                   3090:
                   3091:        res[(int)ssz] = '\0';
                   3092:        return(res);
1.99      schwarze 3093: }
                   3094:
                   3095: int
                   3096: roff_getformat(const struct roff *r)
                   3097: {
                   3098:
                   3099:        return(r->format);
1.48      schwarze 3100: }
                   3101:
                   3102: /*
1.80      schwarze 3103:  * Find out whether a line is a macro line or not.
1.48      schwarze 3104:  * If it is, adjust the current position and return one; if it isn't,
                   3105:  * return zero and don't change the current position.
                   3106:  * If the control character has been set with `.cc', then let that grain
                   3107:  * precedence.
                   3108:  * This is slighly contrary to groff, where using the non-breaking
                   3109:  * control character when `cc' has been invoked will cause the
                   3110:  * non-breaking macro contents to be printed verbatim.
                   3111:  */
                   3112: int
                   3113: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   3114: {
                   3115:        int             pos;
                   3116:
                   3117:        pos = *ppos;
                   3118:
                   3119:        if (0 != r->control && cp[pos] == r->control)
                   3120:                pos++;
                   3121:        else if (0 != r->control)
                   3122:                return(0);
                   3123:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   3124:                pos += 2;
                   3125:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   3126:                pos++;
                   3127:        else
                   3128:                return(0);
                   3129:
                   3130:        while (' ' == cp[pos] || '\t' == cp[pos])
                   3131:                pos++;
                   3132:
                   3133:        *ppos = pos;
                   3134:        return(1);
1.1       schwarze 3135: }