[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 1.51

1.51    ! schwarze    1: /*     $Id: roff.c,v 1.50 2013/06/27 09:48:22 schwarze Exp $ */
1.1       schwarze    2: /*
1.48      schwarze    3:  * Copyright (c) 2010, 2011, 2012 Kristaps Dzonsons <kristaps@bsd.lv>
1.50      schwarze    4:  * Copyright (c) 2010, 2011, 2012, 2013 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:  */
                     18: #include <assert.h>
1.3       schwarze   19: #include <ctype.h>
1.51    ! schwarze   20: #include <stdio.h>
1.1       schwarze   21: #include <stdlib.h>
                     22: #include <string.h>
                     23:
                     24: #include "mandoc.h"
1.27      schwarze   25: #include "libroff.h"
1.8       schwarze   26: #include "libmandoc.h"
1.1       schwarze   27:
1.37      schwarze   28: /* Maximum number of nested if-else conditionals. */
1.2       schwarze   29: #define        RSTACK_MAX      128
                     30:
1.43      schwarze   31: /* Maximum number of string expansions per line, to break infinite loops. */
                     32: #define        EXPAND_LIMIT    1000
                     33:
1.1       schwarze   34: enum   rofft {
1.20      schwarze   35:        ROFF_ad,
1.2       schwarze   36:        ROFF_am,
                     37:        ROFF_ami,
                     38:        ROFF_am1,
1.48      schwarze   39:        ROFF_cc,
1.1       schwarze   40:        ROFF_de,
                     41:        ROFF_dei,
1.2       schwarze   42:        ROFF_de1,
                     43:        ROFF_ds,
                     44:        ROFF_el,
1.20      schwarze   45:        ROFF_hy,
1.2       schwarze   46:        ROFF_ie,
                     47:        ROFF_if,
1.1       schwarze   48:        ROFF_ig,
1.30      schwarze   49:        ROFF_it,
1.20      schwarze   50:        ROFF_ne,
                     51:        ROFF_nh,
1.14      schwarze   52:        ROFF_nr,
1.31      schwarze   53:        ROFF_ns,
                     54:        ROFF_ps,
1.2       schwarze   55:        ROFF_rm,
1.14      schwarze   56:        ROFF_so,
1.31      schwarze   57:        ROFF_ta,
1.2       schwarze   58:        ROFF_tr,
1.47      schwarze   59:        ROFF_Dd,
                     60:        ROFF_TH,
1.27      schwarze   61:        ROFF_TS,
                     62:        ROFF_TE,
                     63:        ROFF_T_,
1.32      schwarze   64:        ROFF_EQ,
                     65:        ROFF_EN,
1.2       schwarze   66:        ROFF_cblock,
1.37      schwarze   67:        ROFF_ccond,
1.16      schwarze   68:        ROFF_USERDEF,
1.1       schwarze   69:        ROFF_MAX
                     70: };
                     71:
1.2       schwarze   72: enum   roffrule {
                     73:        ROFFRULE_ALLOW,
                     74:        ROFFRULE_DENY
                     75: };
                     76:
1.41      schwarze   77: /*
                     78:  * A single register entity.  If "set" is zero, the value of the
                     79:  * register should be the default one, which is per-register.
                     80:  * Registers are assumed to be unsigned ints for now.
                     81:  */
                     82: struct reg {
1.42      schwarze   83:        int              set; /* whether set or not */
                     84:        unsigned int     u; /* unsigned integer */
1.41      schwarze   85: };
                     86:
1.42      schwarze   87: /*
                     88:  * An incredibly-simple string buffer.
                     89:  */
1.8       schwarze   90: struct roffstr {
1.42      schwarze   91:        char            *p; /* nil-terminated buffer */
                     92:        size_t           sz; /* saved strlen(p) */
                     93: };
                     94:
                     95: /*
                     96:  * A key-value roffstr pair as part of a singly-linked list.
                     97:  */
                     98: struct roffkv {
                     99:        struct roffstr   key;
                    100:        struct roffstr   val;
                    101:        struct roffkv   *next; /* next in list */
1.8       schwarze  102: };
                    103:
1.1       schwarze  104: struct roff {
1.47      schwarze  105:        enum mparset     parsetype; /* requested parse type */
1.35      schwarze  106:        struct mparse   *parse; /* parse point */
1.1       schwarze  107:        struct roffnode *last; /* leaf of stack */
1.2       schwarze  108:        enum roffrule    rstack[RSTACK_MAX]; /* stack of !`ie' rules */
1.48      schwarze  109:        char             control; /* control character */
1.2       schwarze  110:        int              rstackpos; /* position in rstack */
1.41      schwarze  111:        struct reg       regs[REG__MAX];
1.42      schwarze  112:        struct roffkv   *strtab; /* user-defined strings & macros */
                    113:        struct roffkv   *xmbtab; /* multi-byte trans table (`tr') */
                    114:        struct roffstr  *xtab; /* single-byte trans table (`tr') */
1.16      schwarze  115:        const char      *current_string; /* value of last called user macro */
1.27      schwarze  116:        struct tbl_node *first_tbl; /* first table parsed */
                    117:        struct tbl_node *last_tbl; /* last table parsed */
                    118:        struct tbl_node *tbl; /* current table being parsed */
1.32      schwarze  119:        struct eqn_node *last_eqn; /* last equation parsed */
                    120:        struct eqn_node *first_eqn; /* first equation parsed */
                    121:        struct eqn_node *eqn; /* current equation being parsed */
1.1       schwarze  122: };
                    123:
                    124: struct roffnode {
                    125:        enum rofft       tok; /* type of node */
                    126:        struct roffnode *parent; /* up one in stack */
                    127:        int              line; /* parse line */
                    128:        int              col; /* parse col */
1.16      schwarze  129:        char            *name; /* node name, e.g. macro name */
1.2       schwarze  130:        char            *end; /* end-rules: custom token */
                    131:        int              endspan; /* end-rules: next-line or infty */
                    132:        enum roffrule    rule; /* current evaluation rule */
1.1       schwarze  133: };
                    134:
                    135: #define        ROFF_ARGS        struct roff *r, /* parse ctx */ \
                    136:                         enum rofft tok, /* tok of macro */ \
                    137:                         char **bufp, /* input buffer */ \
                    138:                         size_t *szp, /* size of input buffer */ \
                    139:                         int ln, /* parse line */ \
1.2       schwarze  140:                         int ppos, /* original pos in buffer */ \
                    141:                         int pos, /* current pos in buffer */ \
                    142:                         int *offs /* reset offset of buffer data */
1.1       schwarze  143:
                    144: typedef        enum rofferr (*roffproc)(ROFF_ARGS);
                    145:
                    146: struct roffmac {
                    147:        const char      *name; /* macro name */
1.2       schwarze  148:        roffproc         proc; /* process new macro */
                    149:        roffproc         text; /* process as child text of macro */
                    150:        roffproc         sub; /* process as child of macro */
                    151:        int              flags;
                    152: #define        ROFFMAC_STRUCT  (1 << 0) /* always interpret */
1.3       schwarze  153:        struct roffmac  *next;
1.1       schwarze  154: };
                    155:
1.37      schwarze  156: struct predef {
                    157:        const char      *name; /* predefined input name */
                    158:        const char      *str; /* replacement symbol */
                    159: };
                    160:
                    161: #define        PREDEF(__name, __str) \
                    162:        { (__name), (__str) },
                    163:
1.42      schwarze  164: static enum rofft       roffhash_find(const char *, size_t);
                    165: static void             roffhash_init(void);
                    166: static void             roffnode_cleanscope(struct roff *);
                    167: static void             roffnode_pop(struct roff *);
                    168: static void             roffnode_push(struct roff *, enum rofft,
                    169:                                const char *, int, int);
1.2       schwarze  170: static enum rofferr     roff_block(ROFF_ARGS);
                    171: static enum rofferr     roff_block_text(ROFF_ARGS);
                    172: static enum rofferr     roff_block_sub(ROFF_ARGS);
                    173: static enum rofferr     roff_cblock(ROFF_ARGS);
1.48      schwarze  174: static enum rofferr     roff_cc(ROFF_ARGS);
1.2       schwarze  175: static enum rofferr     roff_ccond(ROFF_ARGS);
                    176: static enum rofferr     roff_cond(ROFF_ARGS);
                    177: static enum rofferr     roff_cond_text(ROFF_ARGS);
                    178: static enum rofferr     roff_cond_sub(ROFF_ARGS);
1.7       schwarze  179: static enum rofferr     roff_ds(ROFF_ARGS);
1.8       schwarze  180: static enum roffrule    roff_evalcond(const char *, int *);
1.42      schwarze  181: static void             roff_free1(struct roff *);
                    182: static void             roff_freestr(struct roffkv *);
1.28      schwarze  183: static char            *roff_getname(struct roff *, char **, int, int);
1.8       schwarze  184: static const char      *roff_getstrn(const struct roff *,
                    185:                                const char *, size_t);
1.51    ! schwarze  186: static enum rofferr     roff_it(ROFF_ARGS);
1.21      schwarze  187: static enum rofferr     roff_line_ignore(ROFF_ARGS);
1.6       schwarze  188: static enum rofferr     roff_nr(ROFF_ARGS);
1.41      schwarze  189: static void             roff_openeqn(struct roff *, const char *,
                    190:                                int, int, const char *);
1.42      schwarze  191: static enum rofft       roff_parse(struct roff *, const char *, int *);
1.51    ! schwarze  192: static enum rofferr     roff_parsetext(char **, size_t *, int, int *);
1.45      schwarze  193: static enum rofferr     roff_res(struct roff *,
1.37      schwarze  194:                                char **, size_t *, int, int);
1.29      schwarze  195: static enum rofferr     roff_rm(ROFF_ARGS);
1.8       schwarze  196: static void             roff_setstr(struct roff *,
1.16      schwarze  197:                                const char *, const char *, int);
1.42      schwarze  198: static void             roff_setstrn(struct roffkv **, const char *,
                    199:                                size_t, const char *, size_t, int);
1.14      schwarze  200: static enum rofferr     roff_so(ROFF_ARGS);
1.42      schwarze  201: static enum rofferr     roff_tr(ROFF_ARGS);
1.47      schwarze  202: static enum rofferr     roff_Dd(ROFF_ARGS);
                    203: static enum rofferr     roff_TH(ROFF_ARGS);
1.27      schwarze  204: static enum rofferr     roff_TE(ROFF_ARGS);
                    205: static enum rofferr     roff_TS(ROFF_ARGS);
1.32      schwarze  206: static enum rofferr     roff_EQ(ROFF_ARGS);
                    207: static enum rofferr     roff_EN(ROFF_ARGS);
1.27      schwarze  208: static enum rofferr     roff_T_(ROFF_ARGS);
1.16      schwarze  209: static enum rofferr     roff_userdef(ROFF_ARGS);
1.1       schwarze  210:
1.42      schwarze  211: /* See roffhash_find() */
1.3       schwarze  212:
                    213: #define        ASCII_HI         126
                    214: #define        ASCII_LO         33
                    215: #define        HASHWIDTH       (ASCII_HI - ASCII_LO + 1)
                    216:
                    217: static struct roffmac  *hash[HASHWIDTH];
                    218:
                    219: static struct roffmac   roffs[ROFF_MAX] = {
1.21      schwarze  220:        { "ad", roff_line_ignore, NULL, NULL, 0, NULL },
1.3       schwarze  221:        { "am", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    222:        { "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    223:        { "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.48      schwarze  224:        { "cc", roff_cc, NULL, NULL, 0, NULL },
1.3       schwarze  225:        { "de", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    226:        { "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL },
                    227:        { "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.7       schwarze  228:        { "ds", roff_ds, NULL, NULL, 0, NULL },
1.3       schwarze  229:        { "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
1.21      schwarze  230:        { "hy", roff_line_ignore, NULL, NULL, 0, NULL },
1.3       schwarze  231:        { "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    232:        { "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL },
                    233:        { "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL },
1.51    ! schwarze  234:        { "it", roff_it, NULL, NULL, 0, NULL },
1.21      schwarze  235:        { "ne", roff_line_ignore, NULL, NULL, 0, NULL },
                    236:        { "nh", roff_line_ignore, NULL, NULL, 0, NULL },
1.14      schwarze  237:        { "nr", roff_nr, NULL, NULL, 0, NULL },
1.31      schwarze  238:        { "ns", roff_line_ignore, NULL, NULL, 0, NULL },
                    239:        { "ps", roff_line_ignore, NULL, NULL, 0, NULL },
1.29      schwarze  240:        { "rm", roff_rm, NULL, NULL, 0, NULL },
1.14      schwarze  241:        { "so", roff_so, NULL, NULL, 0, NULL },
1.31      schwarze  242:        { "ta", roff_line_ignore, NULL, NULL, 0, NULL },
1.42      schwarze  243:        { "tr", roff_tr, NULL, NULL, 0, NULL },
1.47      schwarze  244:        { "Dd", roff_Dd, NULL, NULL, 0, NULL },
                    245:        { "TH", roff_TH, NULL, NULL, 0, NULL },
1.27      schwarze  246:        { "TS", roff_TS, NULL, NULL, 0, NULL },
                    247:        { "TE", roff_TE, NULL, NULL, 0, NULL },
                    248:        { "T&", roff_T_, NULL, NULL, 0, NULL },
1.32      schwarze  249:        { "EQ", roff_EQ, NULL, NULL, 0, NULL },
                    250:        { "EN", roff_EN, NULL, NULL, 0, NULL },
1.3       schwarze  251:        { ".", roff_cblock, NULL, NULL, 0, NULL },
                    252:        { "\\}", roff_ccond, NULL, NULL, 0, NULL },
1.16      schwarze  253:        { NULL, roff_userdef, NULL, NULL, 0, NULL },
1.1       schwarze  254: };
                    255:
1.47      schwarze  256: const  char *const __mdoc_reserved[] = {
                    257:        "Ac", "Ad", "An", "Ao", "Ap", "Aq", "Ar", "At",
                    258:        "Bc", "Bd", "Bf", "Bk", "Bl", "Bo", "Bq",
                    259:        "Brc", "Bro", "Brq", "Bsx", "Bt", "Bx",
                    260:        "Cd", "Cm", "Db", "Dc", "Dd", "Dl", "Do", "Dq",
                    261:        "Ds", "Dt", "Dv", "Dx", "D1",
                    262:        "Ec", "Ed", "Ef", "Ek", "El", "Em", "em",
                    263:        "En", "Eo", "Eq", "Er", "Es", "Ev", "Ex",
                    264:        "Fa", "Fc", "Fd", "Fl", "Fn", "Fo", "Fr", "Ft", "Fx",
                    265:        "Hf", "Ic", "In", "It", "Lb", "Li", "Lk", "Lp", "LP",
                    266:        "Me", "Ms", "Mt", "Nd", "Nm", "No", "Ns", "Nx",
                    267:        "Oc", "Oo", "Op", "Os", "Ot", "Ox",
                    268:        "Pa", "Pc", "Pf", "Po", "Pp", "PP", "pp", "Pq",
                    269:        "Qc", "Ql", "Qo", "Qq", "Or", "Rd", "Re", "Rs", "Rv",
                    270:        "Sc", "Sf", "Sh", "SH", "Sm", "So", "Sq",
                    271:        "Ss", "St", "Sx", "Sy",
                    272:        "Ta", "Tn", "Ud", "Ux", "Va", "Vt", "Xc", "Xo", "Xr",
                    273:        "%A", "%B", "%D", "%I", "%J", "%N", "%O",
                    274:        "%P", "%Q", "%R", "%T", "%U", "%V",
                    275:        NULL
                    276: };
                    277:
                    278: const  char *const __man_reserved[] = {
                    279:        "AT", "B", "BI", "BR", "BT", "DE", "DS", "DT",
                    280:        "EE", "EN", "EQ", "EX", "HF", "HP", "I", "IB", "IP", "IR",
                    281:        "LP", "ME", "MT", "OP", "P", "PD", "PP", "PT",
                    282:        "R", "RB", "RE", "RI", "RS", "SB", "SH", "SM", "SS", "SY",
                    283:        "TE", "TH", "TP", "TQ", "TS", "T&", "UC", "UE", "UR", "YS",
                    284:        NULL
                    285: };
                    286:
1.37      schwarze  287: /* Array of injected predefined strings. */
                    288: #define        PREDEFS_MAX      38
                    289: static const struct predef predefs[PREDEFS_MAX] = {
                    290: #include "predefs.in"
                    291: };
                    292:
1.42      schwarze  293: /* See roffhash_find() */
1.3       schwarze  294: #define        ROFF_HASH(p)    (p[0] - ASCII_LO)
                    295:
1.51    ! schwarze  296: static int      roffit_lines;  /* number of lines to delay */
        !           297: static char    *roffit_macro;  /* nil-terminated macro line */
        !           298:
1.3       schwarze  299: static void
1.42      schwarze  300: roffhash_init(void)
1.3       schwarze  301: {
                    302:        struct roffmac   *n;
                    303:        int               buc, i;
                    304:
1.16      schwarze  305:        for (i = 0; i < (int)ROFF_USERDEF; i++) {
1.3       schwarze  306:                assert(roffs[i].name[0] >= ASCII_LO);
                    307:                assert(roffs[i].name[0] <= ASCII_HI);
                    308:
                    309:                buc = ROFF_HASH(roffs[i].name);
                    310:
                    311:                if (NULL != (n = hash[buc])) {
                    312:                        for ( ; n->next; n = n->next)
                    313:                                /* Do nothing. */ ;
                    314:                        n->next = &roffs[i];
                    315:                } else
                    316:                        hash[buc] = &roffs[i];
                    317:        }
                    318: }
                    319:
1.1       schwarze  320: /*
                    321:  * Look up a roff token by its name.  Returns ROFF_MAX if no macro by
                    322:  * the nil-terminated string name could be found.
                    323:  */
                    324: static enum rofft
1.42      schwarze  325: roffhash_find(const char *p, size_t s)
1.1       schwarze  326: {
1.3       schwarze  327:        int              buc;
                    328:        struct roffmac  *n;
1.1       schwarze  329:
1.3       schwarze  330:        /*
                    331:         * libroff has an extremely simple hashtable, for the time
                    332:         * being, which simply keys on the first character, which must
                    333:         * be printable, then walks a chain.  It works well enough until
                    334:         * optimised.
                    335:         */
                    336:
                    337:        if (p[0] < ASCII_LO || p[0] > ASCII_HI)
                    338:                return(ROFF_MAX);
                    339:
                    340:        buc = ROFF_HASH(p);
                    341:
                    342:        if (NULL == (n = hash[buc]))
                    343:                return(ROFF_MAX);
                    344:        for ( ; n; n = n->next)
1.16      schwarze  345:                if (0 == strncmp(n->name, p, s) && '\0' == n->name[(int)s])
1.3       schwarze  346:                        return((enum rofft)(n - roffs));
1.1       schwarze  347:
                    348:        return(ROFF_MAX);
                    349: }
                    350:
                    351:
                    352: /*
                    353:  * Pop the current node off of the stack of roff instructions currently
                    354:  * pending.
                    355:  */
                    356: static void
                    357: roffnode_pop(struct roff *r)
                    358: {
                    359:        struct roffnode *p;
                    360:
1.2       schwarze  361:        assert(r->last);
                    362:        p = r->last;
                    363:
                    364:        r->last = r->last->parent;
1.16      schwarze  365:        free(p->name);
                    366:        free(p->end);
1.1       schwarze  367:        free(p);
                    368: }
                    369:
                    370:
                    371: /*
                    372:  * Push a roff node onto the instruction stack.  This must later be
                    373:  * removed with roffnode_pop().
                    374:  */
1.11      schwarze  375: static void
1.16      schwarze  376: roffnode_push(struct roff *r, enum rofft tok, const char *name,
                    377:                int line, int col)
1.1       schwarze  378: {
                    379:        struct roffnode *p;
                    380:
1.11      schwarze  381:        p = mandoc_calloc(1, sizeof(struct roffnode));
1.1       schwarze  382:        p->tok = tok;
1.16      schwarze  383:        if (name)
                    384:                p->name = mandoc_strdup(name);
1.1       schwarze  385:        p->parent = r->last;
                    386:        p->line = line;
                    387:        p->col = col;
1.2       schwarze  388:        p->rule = p->parent ? p->parent->rule : ROFFRULE_DENY;
1.1       schwarze  389:
                    390:        r->last = p;
                    391: }
                    392:
                    393:
                    394: static void
                    395: roff_free1(struct roff *r)
                    396: {
1.49      schwarze  397:        struct tbl_node *tbl;
1.32      schwarze  398:        struct eqn_node *e;
1.42      schwarze  399:        int              i;
1.27      schwarze  400:
1.49      schwarze  401:        while (NULL != (tbl = r->first_tbl)) {
                    402:                r->first_tbl = tbl->next;
                    403:                tbl_free(tbl);
1.27      schwarze  404:        }
                    405:
                    406:        r->first_tbl = r->last_tbl = r->tbl = NULL;
1.1       schwarze  407:
1.32      schwarze  408:        while (NULL != (e = r->first_eqn)) {
                    409:                r->first_eqn = e->next;
                    410:                eqn_free(e);
                    411:        }
                    412:
                    413:        r->first_eqn = r->last_eqn = r->eqn = NULL;
                    414:
1.1       schwarze  415:        while (r->last)
                    416:                roffnode_pop(r);
1.27      schwarze  417:
1.42      schwarze  418:        roff_freestr(r->strtab);
                    419:        roff_freestr(r->xmbtab);
                    420:
                    421:        r->strtab = r->xmbtab = NULL;
                    422:
                    423:        if (r->xtab)
                    424:                for (i = 0; i < 128; i++)
                    425:                        free(r->xtab[i].p);
                    426:
                    427:        free(r->xtab);
                    428:        r->xtab = NULL;
1.1       schwarze  429: }
                    430:
                    431: void
                    432: roff_reset(struct roff *r)
                    433: {
1.38      schwarze  434:        int              i;
1.1       schwarze  435:
                    436:        roff_free1(r);
1.38      schwarze  437:
1.48      schwarze  438:        r->control = 0;
1.41      schwarze  439:        memset(&r->regs, 0, sizeof(struct reg) * REG__MAX);
                    440:
1.38      schwarze  441:        for (i = 0; i < PREDEFS_MAX; i++)
                    442:                roff_setstr(r, predefs[i].name, predefs[i].str, 0);
1.1       schwarze  443: }
                    444:
                    445:
                    446: void
                    447: roff_free(struct roff *r)
                    448: {
                    449:
                    450:        roff_free1(r);
                    451:        free(r);
                    452: }
                    453:
                    454:
                    455: struct roff *
1.47      schwarze  456: roff_alloc(enum mparset type, struct mparse *parse)
1.1       schwarze  457: {
                    458:        struct roff     *r;
1.37      schwarze  459:        int              i;
1.1       schwarze  460:
1.11      schwarze  461:        r = mandoc_calloc(1, sizeof(struct roff));
1.47      schwarze  462:        r->parsetype = type;
1.35      schwarze  463:        r->parse = parse;
1.2       schwarze  464:        r->rstackpos = -1;
1.3       schwarze  465:
1.42      schwarze  466:        roffhash_init();
1.37      schwarze  467:
                    468:        for (i = 0; i < PREDEFS_MAX; i++)
                    469:                roff_setstr(r, predefs[i].name, predefs[i].str, 0);
                    470:
1.1       schwarze  471:        return(r);
                    472: }
                    473:
1.8       schwarze  474: /*
                    475:  * Pre-filter each and every line for reserved words (one beginning with
                    476:  * `\*', e.g., `\*(ab').  These must be handled before the actual line
                    477:  * is processed.
1.42      schwarze  478:  * This also checks the syntax of regular escapes.
1.8       schwarze  479:  */
1.45      schwarze  480: static enum rofferr
1.37      schwarze  481: roff_res(struct roff *r, char **bufp, size_t *szp, int ln, int pos)
1.8       schwarze  482: {
1.42      schwarze  483:        enum mandoc_esc  esc;
1.23      schwarze  484:        const char      *stesc; /* start of an escape sequence ('\\') */
                    485:        const char      *stnam; /* start of the name, after "[(*" */
                    486:        const char      *cp;    /* end of the name, e.g. before ']' */
                    487:        const char      *res;   /* the string to be substituted */
1.43      schwarze  488:        int              i, maxl, expand_count;
1.8       schwarze  489:        size_t           nsz;
                    490:        char            *n;
                    491:
1.43      schwarze  492:        expand_count = 0;
                    493:
1.42      schwarze  494: again:
1.24      schwarze  495:        cp = *bufp + pos;
                    496:        while (NULL != (cp = strchr(cp, '\\'))) {
                    497:                stesc = cp++;
1.23      schwarze  498:
                    499:                /*
                    500:                 * The second character must be an asterisk.
                    501:                 * If it isn't, skip it anyway:  It is escaped,
                    502:                 * so it can't start another escape sequence.
                    503:                 */
                    504:
1.24      schwarze  505:                if ('\0' == *cp)
1.45      schwarze  506:                        return(ROFF_CONT);
1.42      schwarze  507:
                    508:                if ('*' != *cp) {
                    509:                        res = cp;
                    510:                        esc = mandoc_escape(&cp, NULL, NULL);
                    511:                        if (ESCAPE_ERROR != esc)
                    512:                                continue;
                    513:                        cp = res;
                    514:                        mandoc_msg
                    515:                                (MANDOCERR_BADESCAPE, r->parse,
                    516:                                 ln, (int)(stesc - *bufp), NULL);
1.45      schwarze  517:                        return(ROFF_CONT);
1.42      schwarze  518:                }
                    519:
                    520:                cp++;
1.23      schwarze  521:
                    522:                /*
                    523:                 * The third character decides the length
                    524:                 * of the name of the string.
                    525:                 * Save a pointer to the name.
                    526:                 */
                    527:
1.24      schwarze  528:                switch (*cp) {
                    529:                case ('\0'):
1.45      schwarze  530:                        return(ROFF_CONT);
1.8       schwarze  531:                case ('('):
                    532:                        cp++;
                    533:                        maxl = 2;
                    534:                        break;
                    535:                case ('['):
                    536:                        cp++;
                    537:                        maxl = 0;
                    538:                        break;
                    539:                default:
                    540:                        maxl = 1;
                    541:                        break;
                    542:                }
1.23      schwarze  543:                stnam = cp;
1.8       schwarze  544:
1.23      schwarze  545:                /* Advance to the end of the name. */
1.8       schwarze  546:
                    547:                for (i = 0; 0 == maxl || i < maxl; i++, cp++) {
1.42      schwarze  548:                        if ('\0' == *cp) {
                    549:                                mandoc_msg
                    550:                                        (MANDOCERR_BADESCAPE,
                    551:                                         r->parse, ln,
                    552:                                         (int)(stesc - *bufp), NULL);
1.45      schwarze  553:                                return(ROFF_CONT);
1.42      schwarze  554:                        }
1.8       schwarze  555:                        if (0 == maxl && ']' == *cp)
                    556:                                break;
                    557:                }
                    558:
1.23      schwarze  559:                /*
                    560:                 * Retrieve the replacement string; if it is
                    561:                 * undefined, resume searching for escapes.
                    562:                 */
                    563:
                    564:                res = roff_getstrn(r, stnam, (size_t)i);
1.8       schwarze  565:
                    566:                if (NULL == res) {
1.42      schwarze  567:                        mandoc_msg
                    568:                                (MANDOCERR_BADESCAPE, r->parse,
                    569:                                 ln, (int)(stesc - *bufp), NULL);
1.37      schwarze  570:                        res = "";
1.8       schwarze  571:                }
                    572:
1.23      schwarze  573:                /* Replace the escape sequence by the string. */
                    574:
1.42      schwarze  575:                pos = stesc - *bufp;
                    576:
1.8       schwarze  577:                nsz = *szp + strlen(res) + 1;
                    578:                n = mandoc_malloc(nsz);
                    579:
1.23      schwarze  580:                strlcpy(n, *bufp, (size_t)(stesc - *bufp + 1));
1.8       schwarze  581:                strlcat(n, res, nsz);
                    582:                strlcat(n, cp + (maxl ? 0 : 1), nsz);
                    583:
                    584:                free(*bufp);
                    585:
                    586:                *bufp = n;
                    587:                *szp = nsz;
1.43      schwarze  588:
                    589:                if (EXPAND_LIMIT >= ++expand_count)
                    590:                        goto again;
                    591:
                    592:                /* Just leave the string unexpanded. */
                    593:                mandoc_msg(MANDOCERR_ROFFLOOP, r->parse, ln, pos, NULL);
1.45      schwarze  594:                return(ROFF_IGN);
1.42      schwarze  595:        }
1.45      schwarze  596:        return(ROFF_CONT);
1.42      schwarze  597: }
                    598:
                    599: /*
1.51    ! schwarze  600:  * Process text streams:
        !           601:  * Convert all breakable hyphens into ASCII_HYPH.
        !           602:  * Decrement and spring input line trap.
1.42      schwarze  603:  */
                    604: static enum rofferr
1.51    ! schwarze  605: roff_parsetext(char **bufp, size_t *szp, int pos, int *offs)
1.42      schwarze  606: {
                    607:        size_t           sz;
                    608:        const char      *start;
1.51    ! schwarze  609:        char            *p;
        !           610:        int              isz;
1.42      schwarze  611:        enum mandoc_esc  esc;
                    612:
1.51    ! schwarze  613:        start = p = *bufp + pos;
1.42      schwarze  614:
                    615:        while ('\0' != *p) {
                    616:                sz = strcspn(p, "-\\");
                    617:                p += sz;
                    618:
                    619:                if ('\0' == *p)
                    620:                        break;
                    621:
                    622:                if ('\\' == *p) {
                    623:                        /* Skip over escapes. */
                    624:                        p++;
                    625:                        esc = mandoc_escape
                    626:                                ((const char **)&p, NULL, NULL);
                    627:                        if (ESCAPE_ERROR == esc)
                    628:                                break;
                    629:                        continue;
                    630:                } else if (p == start) {
                    631:                        p++;
                    632:                        continue;
                    633:                }
                    634:
1.44      schwarze  635:                if (isalpha((unsigned char)p[-1]) &&
                    636:                    isalpha((unsigned char)p[1]))
1.42      schwarze  637:                        *p = ASCII_HYPH;
                    638:                p++;
1.8       schwarze  639:        }
                    640:
1.51    ! schwarze  641:        /* Spring the input line trap. */
        !           642:        if (1 == roffit_lines) {
        !           643:                isz = asprintf(&p, "%s\n.%s", *bufp, roffit_macro);
        !           644:                if (-1 == isz) {
        !           645:                        perror(NULL);
        !           646:                        exit((int)MANDOCLEVEL_SYSERR);
        !           647:                }
        !           648:                free(*bufp);
        !           649:                *bufp = p;
        !           650:                *szp = isz + 1;
        !           651:                *offs = 0;
        !           652:                free(roffit_macro);
        !           653:                roffit_lines = 0;
        !           654:                return(ROFF_REPARSE);
        !           655:        } else if (1 < roffit_lines)
        !           656:                --roffit_lines;
1.42      schwarze  657:        return(ROFF_CONT);
1.8       schwarze  658: }
                    659:
1.1       schwarze  660: enum rofferr
1.6       schwarze  661: roff_parseln(struct roff *r, int ln, char **bufp,
                    662:                size_t *szp, int pos, int *offs)
1.1       schwarze  663: {
                    664:        enum rofft       t;
1.27      schwarze  665:        enum rofferr     e;
1.35      schwarze  666:        int              ppos, ctl;
1.1       schwarze  667:
1.2       schwarze  668:        /*
1.8       schwarze  669:         * Run the reserved-word filter only if we have some reserved
                    670:         * words to fill in.
                    671:         */
                    672:
1.45      schwarze  673:        e = roff_res(r, bufp, szp, ln, pos);
                    674:        if (ROFF_IGN == e)
                    675:                return(e);
                    676:        assert(ROFF_CONT == e);
1.8       schwarze  677:
1.35      schwarze  678:        ppos = pos;
1.48      schwarze  679:        ctl = roff_getcontrol(r, *bufp, &pos);
1.35      schwarze  680:
1.8       schwarze  681:        /*
1.2       schwarze  682:         * First, if a scope is open and we're not a macro, pass the
                    683:         * text through the macro's filter.  If a scope isn't open and
                    684:         * we're not a macro, just let it through.
1.32      schwarze  685:         * Finally, if there's an equation scope open, divert it into it
                    686:         * no matter our state.
1.2       schwarze  687:         */
                    688:
1.35      schwarze  689:        if (r->last && ! ctl) {
1.2       schwarze  690:                t = r->last->tok;
                    691:                assert(roffs[t].text);
1.27      schwarze  692:                e = (*roffs[t].text)
                    693:                        (r, t, bufp, szp, ln, pos, pos, offs);
                    694:                assert(ROFF_IGN == e || ROFF_CONT == e);
1.32      schwarze  695:                if (ROFF_CONT != e)
                    696:                        return(e);
                    697:                if (r->eqn)
1.41      schwarze  698:                        return(eqn_read(&r->eqn, ln, *bufp, pos, offs));
1.32      schwarze  699:                if (r->tbl)
1.35      schwarze  700:                        return(tbl_read(r->tbl, ln, *bufp, pos));
1.51    ! schwarze  701:                return(roff_parsetext(bufp, szp, pos, offs));
1.35      schwarze  702:        } else if ( ! ctl) {
1.32      schwarze  703:                if (r->eqn)
1.41      schwarze  704:                        return(eqn_read(&r->eqn, ln, *bufp, pos, offs));
1.27      schwarze  705:                if (r->tbl)
1.35      schwarze  706:                        return(tbl_read(r->tbl, ln, *bufp, pos));
1.51    ! schwarze  707:                return(roff_parsetext(bufp, szp, pos, offs));
1.32      schwarze  708:        } else if (r->eqn)
1.41      schwarze  709:                return(eqn_read(&r->eqn, ln, *bufp, ppos, offs));
1.2       schwarze  710:
                    711:        /*
                    712:         * If a scope is open, go to the child handler for that macro,
                    713:         * as it may want to preprocess before doing anything with it.
1.32      schwarze  714:         * Don't do so if an equation is open.
1.2       schwarze  715:         */
                    716:
                    717:        if (r->last) {
1.1       schwarze  718:                t = r->last->tok;
                    719:                assert(roffs[t].sub);
1.2       schwarze  720:                return((*roffs[t].sub)
1.8       schwarze  721:                                (r, t, bufp, szp,
1.35      schwarze  722:                                 ln, ppos, pos, offs));
1.2       schwarze  723:        }
                    724:
                    725:        /*
                    726:         * Lastly, as we've no scope open, try to look up and execute
                    727:         * the new macro.  If no macro is found, simply return and let
                    728:         * the compilers handle it.
                    729:         */
                    730:
1.16      schwarze  731:        if (ROFF_MAX == (t = roff_parse(r, *bufp, &pos)))
1.1       schwarze  732:                return(ROFF_CONT);
                    733:
1.2       schwarze  734:        assert(roffs[t].proc);
                    735:        return((*roffs[t].proc)
1.8       schwarze  736:                        (r, t, bufp, szp,
                    737:                         ln, ppos, pos, offs));
1.2       schwarze  738: }
                    739:
1.1       schwarze  740:
1.27      schwarze  741: void
1.2       schwarze  742: roff_endparse(struct roff *r)
                    743: {
1.1       schwarze  744:
1.27      schwarze  745:        if (r->last)
1.35      schwarze  746:                mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse,
1.27      schwarze  747:                                r->last->line, r->last->col, NULL);
                    748:
1.32      schwarze  749:        if (r->eqn) {
1.35      schwarze  750:                mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse,
1.41      schwarze  751:                                r->eqn->eqn.ln, r->eqn->eqn.pos, NULL);
                    752:                eqn_end(&r->eqn);
1.32      schwarze  753:        }
                    754:
1.27      schwarze  755:        if (r->tbl) {
1.35      schwarze  756:                mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse,
1.27      schwarze  757:                                r->tbl->line, r->tbl->pos, NULL);
1.41      schwarze  758:                tbl_end(&r->tbl);
1.27      schwarze  759:        }
1.1       schwarze  760: }
                    761:
                    762: /*
                    763:  * Parse a roff node's type from the input buffer.  This must be in the
                    764:  * form of ".foo xxx" in the usual way.
                    765:  */
                    766: static enum rofft
1.16      schwarze  767: roff_parse(struct roff *r, const char *buf, int *pos)
1.1       schwarze  768: {
1.16      schwarze  769:        const char      *mac;
                    770:        size_t           maclen;
1.1       schwarze  771:        enum rofft       t;
                    772:
1.39      schwarze  773:        if ('\0' == buf[*pos] || '"' == buf[*pos] ||
                    774:                        '\t' == buf[*pos] || ' ' == buf[*pos])
1.1       schwarze  775:                return(ROFF_MAX);
                    776:
1.39      schwarze  777:        /*
                    778:         * We stop the macro parse at an escape, tab, space, or nil.
                    779:         * However, `\}' is also a valid macro, so make sure we don't
                    780:         * clobber it by seeing the `\' as the end of token.
                    781:         */
                    782:
1.16      schwarze  783:        mac = buf + *pos;
1.39      schwarze  784:        maclen = strcspn(mac + 1, " \\\t\0") + 1;
1.1       schwarze  785:
1.16      schwarze  786:        t = (r->current_string = roff_getstrn(r, mac, maclen))
1.42      schwarze  787:            ? ROFF_USERDEF : roffhash_find(mac, maclen);
1.1       schwarze  788:
1.34      schwarze  789:        *pos += (int)maclen;
1.35      schwarze  790:
1.1       schwarze  791:        while (buf[*pos] && ' ' == buf[*pos])
                    792:                (*pos)++;
                    793:
                    794:        return(t);
                    795: }
                    796:
                    797: /* ARGSUSED */
                    798: static enum rofferr
1.2       schwarze  799: roff_cblock(ROFF_ARGS)
1.1       schwarze  800: {
                    801:
1.2       schwarze  802:        /*
                    803:         * A block-close `..' should only be invoked as a child of an
                    804:         * ignore macro, otherwise raise a warning and just ignore it.
                    805:         */
                    806:
                    807:        if (NULL == r->last) {
1.35      schwarze  808:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.2       schwarze  809:                return(ROFF_IGN);
                    810:        }
1.1       schwarze  811:
1.2       schwarze  812:        switch (r->last->tok) {
                    813:        case (ROFF_am):
                    814:                /* FALLTHROUGH */
                    815:        case (ROFF_ami):
                    816:                /* FALLTHROUGH */
                    817:        case (ROFF_am1):
                    818:                /* FALLTHROUGH */
                    819:        case (ROFF_de):
1.23      schwarze  820:                /* ROFF_de1 is remapped to ROFF_de in roff_block(). */
1.2       schwarze  821:                /* FALLTHROUGH */
                    822:        case (ROFF_dei):
                    823:                /* FALLTHROUGH */
                    824:        case (ROFF_ig):
                    825:                break;
                    826:        default:
1.35      schwarze  827:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.1       schwarze  828:                return(ROFF_IGN);
1.2       schwarze  829:        }
                    830:
                    831:        if ((*bufp)[pos])
1.35      schwarze  832:                mandoc_msg(MANDOCERR_ARGSLOST, r->parse, ln, pos, NULL);
1.2       schwarze  833:
                    834:        roffnode_pop(r);
                    835:        roffnode_cleanscope(r);
                    836:        return(ROFF_IGN);
                    837:
                    838: }
1.1       schwarze  839:
                    840:
1.2       schwarze  841: static void
                    842: roffnode_cleanscope(struct roff *r)
                    843: {
1.1       schwarze  844:
1.2       schwarze  845:        while (r->last) {
1.46      schwarze  846:                if (--r->last->endspan != 0)
1.2       schwarze  847:                        break;
                    848:                roffnode_pop(r);
                    849:        }
                    850: }
1.1       schwarze  851:
                    852:
1.2       schwarze  853: /* ARGSUSED */
                    854: static enum rofferr
                    855: roff_ccond(ROFF_ARGS)
                    856: {
1.1       schwarze  857:
1.2       schwarze  858:        if (NULL == r->last) {
1.35      schwarze  859:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.1       schwarze  860:                return(ROFF_IGN);
1.2       schwarze  861:        }
1.1       schwarze  862:
1.2       schwarze  863:        switch (r->last->tok) {
                    864:        case (ROFF_el):
                    865:                /* FALLTHROUGH */
                    866:        case (ROFF_ie):
                    867:                /* FALLTHROUGH */
                    868:        case (ROFF_if):
                    869:                break;
                    870:        default:
1.35      schwarze  871:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.2       schwarze  872:                return(ROFF_IGN);
                    873:        }
1.1       schwarze  874:
1.2       schwarze  875:        if (r->last->endspan > -1) {
1.35      schwarze  876:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.1       schwarze  877:                return(ROFF_IGN);
1.2       schwarze  878:        }
                    879:
                    880:        if ((*bufp)[pos])
1.35      schwarze  881:                mandoc_msg(MANDOCERR_ARGSLOST, r->parse, ln, pos, NULL);
1.1       schwarze  882:
1.2       schwarze  883:        roffnode_pop(r);
                    884:        roffnode_cleanscope(r);
1.1       schwarze  885:        return(ROFF_IGN);
                    886: }
                    887:
                    888:
                    889: /* ARGSUSED */
                    890: static enum rofferr
1.2       schwarze  891: roff_block(ROFF_ARGS)
1.1       schwarze  892: {
1.2       schwarze  893:        int             sv;
                    894:        size_t          sz;
1.16      schwarze  895:        char            *name;
                    896:
                    897:        name = NULL;
1.2       schwarze  898:
1.16      schwarze  899:        if (ROFF_ig != tok) {
                    900:                if ('\0' == (*bufp)[pos]) {
1.35      schwarze  901:                        mandoc_msg(MANDOCERR_NOARGS, r->parse, ln, ppos, NULL);
1.16      schwarze  902:                        return(ROFF_IGN);
                    903:                }
1.22      schwarze  904:
                    905:                /*
                    906:                 * Re-write `de1', since we don't really care about
                    907:                 * groff's strange compatibility mode, into `de'.
                    908:                 */
                    909:
1.18      schwarze  910:                if (ROFF_de1 == tok)
                    911:                        tok = ROFF_de;
1.16      schwarze  912:                if (ROFF_de == tok)
                    913:                        name = *bufp + pos;
1.21      schwarze  914:                else
1.35      schwarze  915:                        mandoc_msg(MANDOCERR_REQUEST, r->parse, ln, ppos,
1.21      schwarze  916:                            roffs[tok].name);
1.22      schwarze  917:
1.33      schwarze  918:                while ((*bufp)[pos] && ! isspace((unsigned char)(*bufp)[pos]))
1.2       schwarze  919:                        pos++;
1.22      schwarze  920:
1.33      schwarze  921:                while (isspace((unsigned char)(*bufp)[pos]))
1.16      schwarze  922:                        (*bufp)[pos++] = '\0';
1.2       schwarze  923:        }
                    924:
1.16      schwarze  925:        roffnode_push(r, tok, name, ln, ppos);
                    926:
                    927:        /*
                    928:         * At the beginning of a `de' macro, clear the existing string
                    929:         * with the same name, if there is one.  New content will be
                    930:         * added from roff_block_text() in multiline mode.
                    931:         */
1.22      schwarze  932:
1.16      schwarze  933:        if (ROFF_de == tok)
1.19      schwarze  934:                roff_setstr(r, name, "", 0);
1.2       schwarze  935:
                    936:        if ('\0' == (*bufp)[pos])
                    937:                return(ROFF_IGN);
1.1       schwarze  938:
1.22      schwarze  939:        /* If present, process the custom end-of-line marker. */
                    940:
1.2       schwarze  941:        sv = pos;
1.33      schwarze  942:        while ((*bufp)[pos] && ! isspace((unsigned char)(*bufp)[pos]))
1.2       schwarze  943:                pos++;
                    944:
                    945:        /*
                    946:         * Note: groff does NOT like escape characters in the input.
                    947:         * Instead of detecting this, we're just going to let it fly and
                    948:         * to hell with it.
                    949:         */
                    950:
                    951:        assert(pos > sv);
                    952:        sz = (size_t)(pos - sv);
                    953:
                    954:        if (1 == sz && '.' == (*bufp)[sv])
                    955:                return(ROFF_IGN);
                    956:
1.11      schwarze  957:        r->last->end = mandoc_malloc(sz + 1);
1.2       schwarze  958:
                    959:        memcpy(r->last->end, *bufp + sv, sz);
                    960:        r->last->end[(int)sz] = '\0';
                    961:
                    962:        if ((*bufp)[pos])
1.35      schwarze  963:                mandoc_msg(MANDOCERR_ARGSLOST, r->parse, ln, pos, NULL);
1.1       schwarze  964:
                    965:        return(ROFF_IGN);
                    966: }
                    967:
                    968:
                    969: /* ARGSUSED */
                    970: static enum rofferr
1.2       schwarze  971: roff_block_sub(ROFF_ARGS)
1.1       schwarze  972: {
1.2       schwarze  973:        enum rofft      t;
                    974:        int             i, j;
                    975:
                    976:        /*
                    977:         * First check whether a custom macro exists at this level.  If
                    978:         * it does, then check against it.  This is some of groff's
                    979:         * stranger behaviours.  If we encountered a custom end-scope
                    980:         * tag and that tag also happens to be a "real" macro, then we
                    981:         * need to try interpreting it again as a real macro.  If it's
                    982:         * not, then return ignore.  Else continue.
                    983:         */
                    984:
                    985:        if (r->last->end) {
1.35      schwarze  986:                for (i = pos, j = 0; r->last->end[j]; j++, i++)
1.2       schwarze  987:                        if ((*bufp)[i] != r->last->end[j])
                    988:                                break;
1.1       schwarze  989:
1.2       schwarze  990:                if ('\0' == r->last->end[j] &&
                    991:                                ('\0' == (*bufp)[i] ||
                    992:                                 ' ' == (*bufp)[i] ||
                    993:                                 '\t' == (*bufp)[i])) {
                    994:                        roffnode_pop(r);
                    995:                        roffnode_cleanscope(r);
1.1       schwarze  996:
1.35      schwarze  997:                        while (' ' == (*bufp)[i] || '\t' == (*bufp)[i])
                    998:                                i++;
                    999:
                   1000:                        pos = i;
1.16      schwarze 1001:                        if (ROFF_MAX != roff_parse(r, *bufp, &pos))
1.2       schwarze 1002:                                return(ROFF_RERUN);
                   1003:                        return(ROFF_IGN);
                   1004:                }
1.1       schwarze 1005:        }
                   1006:
1.2       schwarze 1007:        /*
                   1008:         * If we have no custom end-query or lookup failed, then try
                   1009:         * pulling it out of the hashtable.
                   1010:         */
1.1       schwarze 1011:
1.36      schwarze 1012:        t = roff_parse(r, *bufp, &pos);
1.1       schwarze 1013:
1.16      schwarze 1014:        /*
                   1015:         * Macros other than block-end are only significant
                   1016:         * in `de' blocks; elsewhere, simply throw them away.
                   1017:         */
                   1018:        if (ROFF_cblock != t) {
                   1019:                if (ROFF_de == tok)
                   1020:                        roff_setstr(r, r->last->name, *bufp + ppos, 1);
1.1       schwarze 1021:                return(ROFF_IGN);
1.16      schwarze 1022:        }
1.1       schwarze 1023:
1.2       schwarze 1024:        assert(roffs[t].proc);
1.6       schwarze 1025:        return((*roffs[t].proc)(r, t, bufp, szp,
                   1026:                                ln, ppos, pos, offs));
1.2       schwarze 1027: }
                   1028:
                   1029:
                   1030: /* ARGSUSED */
                   1031: static enum rofferr
                   1032: roff_block_text(ROFF_ARGS)
                   1033: {
                   1034:
1.16      schwarze 1035:        if (ROFF_de == tok)
                   1036:                roff_setstr(r, r->last->name, *bufp + pos, 1);
                   1037:
1.2       schwarze 1038:        return(ROFF_IGN);
                   1039: }
                   1040:
                   1041:
                   1042: /* ARGSUSED */
                   1043: static enum rofferr
                   1044: roff_cond_sub(ROFF_ARGS)
                   1045: {
                   1046:        enum rofft       t;
                   1047:        enum roffrule    rr;
1.37      schwarze 1048:        char            *ep;
1.2       schwarze 1049:
                   1050:        rr = r->last->rule;
1.37      schwarze 1051:        roffnode_cleanscope(r);
1.50      schwarze 1052:        t = roff_parse(r, *bufp, &pos);
1.2       schwarze 1053:
1.37      schwarze 1054:        /*
1.50      schwarze 1055:         * Fully handle known macros when they are structurally
                   1056:         * required or when the conditional evaluated to true.
1.5       schwarze 1057:         */
                   1058:
1.50      schwarze 1059:        if ((ROFF_MAX != t) &&
                   1060:            (ROFF_ccond == t || ROFFRULE_ALLOW == rr ||
                   1061:             ROFFMAC_STRUCT & roffs[t].flags)) {
                   1062:                assert(roffs[t].proc);
                   1063:                return((*roffs[t].proc)(r, t, bufp, szp,
                   1064:                                        ln, ppos, pos, offs));
                   1065:        }
1.39      schwarze 1066:
1.50      schwarze 1067:        /* Always check for the closing delimiter `\}'. */
1.39      schwarze 1068:
1.50      schwarze 1069:        ep = &(*bufp)[pos];
                   1070:        while (NULL != (ep = strchr(ep, '\\'))) {
                   1071:                if ('}' != *(++ep))
                   1072:                        continue;
1.2       schwarze 1073:
1.50      schwarze 1074:                /*
                   1075:                 * If we're at the end of line, then just chop
                   1076:                 * off the \} and resize the buffer.
                   1077:                 * If we aren't, then convert it to spaces.
                   1078:                 */
1.37      schwarze 1079:
1.50      schwarze 1080:                if ('\0' == *(ep + 1)) {
                   1081:                        *--ep = '\0';
                   1082:                        *szp -= 2;
                   1083:                } else
                   1084:                        *(ep - 1) = *ep = ' ';
1.2       schwarze 1085:
1.50      schwarze 1086:                roff_ccond(r, ROFF_ccond, bufp, szp,
                   1087:                                ln, pos, pos + 2, offs);
                   1088:                break;
                   1089:        }
                   1090:        return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT);
1.2       schwarze 1091: }
                   1092:
                   1093: /* ARGSUSED */
                   1094: static enum rofferr
                   1095: roff_cond_text(ROFF_ARGS)
                   1096: {
1.37      schwarze 1097:        char            *ep;
1.2       schwarze 1098:        enum roffrule    rr;
                   1099:
                   1100:        rr = r->last->rule;
1.37      schwarze 1101:        roffnode_cleanscope(r);
1.1       schwarze 1102:
1.37      schwarze 1103:        ep = &(*bufp)[pos];
                   1104:        for ( ; NULL != (ep = strchr(ep, '\\')); ep++) {
                   1105:                ep++;
                   1106:                if ('}' != *ep)
                   1107:                        continue;
                   1108:                *ep = '&';
                   1109:                roff_ccond(r, ROFF_ccond, bufp, szp,
                   1110:                                ln, pos, pos + 2, offs);
1.2       schwarze 1111:        }
                   1112:        return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT);
                   1113: }
                   1114:
1.5       schwarze 1115: static enum roffrule
                   1116: roff_evalcond(const char *v, int *pos)
                   1117: {
                   1118:
                   1119:        switch (v[*pos]) {
                   1120:        case ('n'):
                   1121:                (*pos)++;
                   1122:                return(ROFFRULE_ALLOW);
                   1123:        case ('e'):
                   1124:                /* FALLTHROUGH */
                   1125:        case ('o'):
                   1126:                /* FALLTHROUGH */
                   1127:        case ('t'):
                   1128:                (*pos)++;
                   1129:                return(ROFFRULE_DENY);
                   1130:        default:
                   1131:                break;
                   1132:        }
                   1133:
                   1134:        while (v[*pos] && ' ' != v[*pos])
                   1135:                (*pos)++;
                   1136:        return(ROFFRULE_DENY);
                   1137: }
                   1138:
1.2       schwarze 1139: /* ARGSUSED */
                   1140: static enum rofferr
1.21      schwarze 1141: roff_line_ignore(ROFF_ARGS)
1.6       schwarze 1142: {
1.30      schwarze 1143:
1.21      schwarze 1144:        return(ROFF_IGN);
                   1145: }
                   1146:
                   1147: /* ARGSUSED */
                   1148: static enum rofferr
1.2       schwarze 1149: roff_cond(ROFF_ARGS)
                   1150: {
1.46      schwarze 1151:
                   1152:        roffnode_push(r, tok, NULL, ln, ppos);
1.2       schwarze 1153:
1.35      schwarze 1154:        /*
                   1155:         * An `.el' has no conditional body: it will consume the value
                   1156:         * of the current rstack entry set in prior `ie' calls or
                   1157:         * defaults to DENY.
                   1158:         *
                   1159:         * If we're not an `el', however, then evaluate the conditional.
                   1160:         */
1.1       schwarze 1161:
1.46      schwarze 1162:        r->last->rule = ROFF_el == tok ?
1.35      schwarze 1163:                (r->rstackpos < 0 ?
                   1164:                 ROFFRULE_DENY : r->rstack[r->rstackpos--]) :
                   1165:                roff_evalcond(*bufp, &pos);
1.2       schwarze 1166:
1.35      schwarze 1167:        /*
                   1168:         * An if-else will put the NEGATION of the current evaluated
                   1169:         * conditional into the stack of rules.
                   1170:         */
                   1171:
1.2       schwarze 1172:        if (ROFF_ie == tok) {
1.35      schwarze 1173:                if (r->rstackpos == RSTACK_MAX - 1) {
                   1174:                        mandoc_msg(MANDOCERR_MEM,
                   1175:                                r->parse, ln, ppos, NULL);
                   1176:                        return(ROFF_ERR);
                   1177:                }
                   1178:                r->rstack[++r->rstackpos] =
                   1179:                        ROFFRULE_DENY == r->last->rule ?
                   1180:                        ROFFRULE_ALLOW : ROFFRULE_DENY;
1.2       schwarze 1181:        }
1.5       schwarze 1182:
                   1183:        /* If the parent has false as its rule, then so do we. */
                   1184:
1.2       schwarze 1185:        if (r->last->parent && ROFFRULE_DENY == r->last->parent->rule)
                   1186:                r->last->rule = ROFFRULE_DENY;
1.5       schwarze 1187:
                   1188:        /*
1.46      schwarze 1189:         * Determine scope.
                   1190:         * If there is nothing on the line after the conditional,
                   1191:         * not even whitespace, use next-line scope.
1.5       schwarze 1192:         */
1.2       schwarze 1193:
1.46      schwarze 1194:        if ('\0' == (*bufp)[pos]) {
                   1195:                r->last->endspan = 2;
                   1196:                goto out;
                   1197:        }
                   1198:
                   1199:        while (' ' == (*bufp)[pos])
                   1200:                pos++;
                   1201:
                   1202:        /* An opening brace requests multiline scope. */
1.2       schwarze 1203:
                   1204:        if ('\\' == (*bufp)[pos] && '{' == (*bufp)[pos + 1]) {
                   1205:                r->last->endspan = -1;
                   1206:                pos += 2;
1.46      schwarze 1207:                goto out;
1.2       schwarze 1208:        }
                   1209:
                   1210:        /*
1.46      schwarze 1211:         * Anything else following the conditional causes
                   1212:         * single-line scope.  Warn if the scope contains
                   1213:         * nothing but trailing whitespace.
1.2       schwarze 1214:         */
                   1215:
                   1216:        if ('\0' == (*bufp)[pos])
1.46      schwarze 1217:                mandoc_msg(MANDOCERR_NOARGS, r->parse, ln, ppos, NULL);
1.2       schwarze 1218:
1.46      schwarze 1219:        r->last->endspan = 1;
1.1       schwarze 1220:
1.46      schwarze 1221: out:
1.2       schwarze 1222:        *offs = pos;
                   1223:        return(ROFF_RERUN);
1.1       schwarze 1224: }
                   1225:
                   1226:
1.2       schwarze 1227: /* ARGSUSED */
                   1228: static enum rofferr
1.7       schwarze 1229: roff_ds(ROFF_ARGS)
                   1230: {
1.10      schwarze 1231:        char            *name, *string;
                   1232:
                   1233:        /*
                   1234:         * A symbol is named by the first word following the macro
                   1235:         * invocation up to a space.  Its value is anything after the
                   1236:         * name's trailing whitespace and optional double-quote.  Thus,
                   1237:         *
                   1238:         *  [.ds foo "bar  "     ]
                   1239:         *
                   1240:         * will have `bar  "     ' as its value.
                   1241:         */
1.7       schwarze 1242:
1.28      schwarze 1243:        string = *bufp + pos;
                   1244:        name = roff_getname(r, &string, ln, pos);
1.7       schwarze 1245:        if ('\0' == *name)
                   1246:                return(ROFF_IGN);
                   1247:
1.28      schwarze 1248:        /* Read past initial double-quote. */
                   1249:        if ('"' == *string)
1.7       schwarze 1250:                string++;
                   1251:
1.10      schwarze 1252:        /* The rest is the value. */
1.16      schwarze 1253:        roff_setstr(r, name, string, 0);
1.7       schwarze 1254:        return(ROFF_IGN);
                   1255: }
                   1256:
1.41      schwarze 1257: int
                   1258: roff_regisset(const struct roff *r, enum regs reg)
                   1259: {
                   1260:
                   1261:        return(r->regs[(int)reg].set);
                   1262: }
                   1263:
                   1264: unsigned int
                   1265: roff_regget(const struct roff *r, enum regs reg)
                   1266: {
                   1267:
                   1268:        return(r->regs[(int)reg].u);
                   1269: }
                   1270:
                   1271: void
                   1272: roff_regunset(struct roff *r, enum regs reg)
                   1273: {
                   1274:
                   1275:        r->regs[(int)reg].set = 0;
                   1276: }
1.7       schwarze 1277:
                   1278: /* ARGSUSED */
                   1279: static enum rofferr
1.6       schwarze 1280: roff_nr(ROFF_ARGS)
1.1       schwarze 1281: {
1.28      schwarze 1282:        const char      *key;
                   1283:        char            *val;
1.37      schwarze 1284:        int              iv;
1.6       schwarze 1285:
1.28      schwarze 1286:        val = *bufp + pos;
                   1287:        key = roff_getname(r, &val, ln, pos);
1.6       schwarze 1288:
                   1289:        if (0 == strcmp(key, "nS")) {
1.41      schwarze 1290:                r->regs[(int)REG_nS].set = 1;
                   1291:                if ((iv = mandoc_strntoi(val, strlen(val), 10)) >= 0)
                   1292:                        r->regs[(int)REG_nS].u = (unsigned)iv;
1.37      schwarze 1293:                else
1.41      schwarze 1294:                        r->regs[(int)REG_nS].u = 0u;
1.6       schwarze 1295:        }
1.1       schwarze 1296:
1.29      schwarze 1297:        return(ROFF_IGN);
                   1298: }
                   1299:
                   1300: /* ARGSUSED */
                   1301: static enum rofferr
                   1302: roff_rm(ROFF_ARGS)
                   1303: {
                   1304:        const char       *name;
                   1305:        char             *cp;
                   1306:
                   1307:        cp = *bufp + pos;
                   1308:        while ('\0' != *cp) {
1.34      schwarze 1309:                name = roff_getname(r, &cp, ln, (int)(cp - *bufp));
1.29      schwarze 1310:                if ('\0' != *name)
                   1311:                        roff_setstr(r, name, NULL, 0);
                   1312:        }
1.51    ! schwarze 1313:        return(ROFF_IGN);
        !          1314: }
        !          1315:
        !          1316: /* ARGSUSED */
        !          1317: static enum rofferr
        !          1318: roff_it(ROFF_ARGS)
        !          1319: {
        !          1320:        char            *cp;
        !          1321:        size_t           len;
        !          1322:        int              iv;
        !          1323:
        !          1324:        /* Parse the number of lines. */
        !          1325:        cp = *bufp + pos;
        !          1326:        len = strcspn(cp, " \t");
        !          1327:        cp[len] = '\0';
        !          1328:        if ((iv = mandoc_strntoi(cp, len, 10)) <= 0) {
        !          1329:                mandoc_msg(MANDOCERR_NUMERIC, r->parse,
        !          1330:                                ln, ppos, *bufp + 1);
        !          1331:                return(ROFF_IGN);
        !          1332:        }
        !          1333:        cp += len + 1;
        !          1334:
        !          1335:        /* Arm the input line trap. */
        !          1336:        roffit_lines = iv;
        !          1337:        roffit_macro = mandoc_strdup(cp);
1.2       schwarze 1338:        return(ROFF_IGN);
1.47      schwarze 1339: }
                   1340:
                   1341: /* ARGSUSED */
                   1342: static enum rofferr
                   1343: roff_Dd(ROFF_ARGS)
                   1344: {
                   1345:        const char *const       *cp;
                   1346:
                   1347:        if (MPARSE_MDOC != r->parsetype)
                   1348:                for (cp = __mdoc_reserved; *cp; cp++)
                   1349:                        roff_setstr(r, *cp, NULL, 0);
                   1350:
                   1351:        return(ROFF_CONT);
                   1352: }
                   1353:
                   1354: /* ARGSUSED */
                   1355: static enum rofferr
                   1356: roff_TH(ROFF_ARGS)
                   1357: {
                   1358:        const char *const       *cp;
                   1359:
                   1360:        if (MPARSE_MDOC != r->parsetype)
                   1361:                for (cp = __man_reserved; *cp; cp++)
                   1362:                        roff_setstr(r, *cp, NULL, 0);
                   1363:
                   1364:        return(ROFF_CONT);
1.14      schwarze 1365: }
                   1366:
                   1367: /* ARGSUSED */
                   1368: static enum rofferr
1.27      schwarze 1369: roff_TE(ROFF_ARGS)
                   1370: {
                   1371:
                   1372:        if (NULL == r->tbl)
1.35      schwarze 1373:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.27      schwarze 1374:        else
1.41      schwarze 1375:                tbl_end(&r->tbl);
1.27      schwarze 1376:
                   1377:        return(ROFF_IGN);
                   1378: }
                   1379:
                   1380: /* ARGSUSED */
                   1381: static enum rofferr
                   1382: roff_T_(ROFF_ARGS)
                   1383: {
                   1384:
                   1385:        if (NULL == r->tbl)
1.35      schwarze 1386:                mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.27      schwarze 1387:        else
                   1388:                tbl_restart(ppos, ln, r->tbl);
                   1389:
                   1390:        return(ROFF_IGN);
                   1391: }
                   1392:
1.41      schwarze 1393: #if 0
                   1394: static int
                   1395: roff_closeeqn(struct roff *r)
                   1396: {
                   1397:
                   1398:        return(r->eqn && ROFF_EQN == eqn_end(&r->eqn) ? 1 : 0);
                   1399: }
                   1400: #endif
                   1401:
                   1402: static void
                   1403: roff_openeqn(struct roff *r, const char *name, int line,
                   1404:                int offs, const char *buf)
1.32      schwarze 1405: {
1.41      schwarze 1406:        struct eqn_node *e;
                   1407:        int              poff;
1.32      schwarze 1408:
                   1409:        assert(NULL == r->eqn);
1.41      schwarze 1410:        e = eqn_alloc(name, offs, line, r->parse);
1.32      schwarze 1411:
                   1412:        if (r->last_eqn)
                   1413:                r->last_eqn->next = e;
                   1414:        else
                   1415:                r->first_eqn = r->last_eqn = e;
                   1416:
                   1417:        r->eqn = r->last_eqn = e;
1.41      schwarze 1418:
                   1419:        if (buf) {
                   1420:                poff = 0;
                   1421:                eqn_read(&r->eqn, line, buf, offs, &poff);
                   1422:        }
                   1423: }
                   1424:
                   1425: /* ARGSUSED */
                   1426: static enum rofferr
                   1427: roff_EQ(ROFF_ARGS)
                   1428: {
                   1429:
                   1430:        roff_openeqn(r, *bufp + pos, ln, ppos, NULL);
1.32      schwarze 1431:        return(ROFF_IGN);
                   1432: }
                   1433:
                   1434: /* ARGSUSED */
                   1435: static enum rofferr
                   1436: roff_EN(ROFF_ARGS)
                   1437: {
                   1438:
1.35      schwarze 1439:        mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
1.32      schwarze 1440:        return(ROFF_IGN);
                   1441: }
                   1442:
                   1443: /* ARGSUSED */
                   1444: static enum rofferr
1.27      schwarze 1445: roff_TS(ROFF_ARGS)
                   1446: {
1.49      schwarze 1447:        struct tbl_node *tbl;
1.27      schwarze 1448:
                   1449:        if (r->tbl) {
1.35      schwarze 1450:                mandoc_msg(MANDOCERR_SCOPEBROKEN, r->parse, ln, ppos, NULL);
1.41      schwarze 1451:                tbl_end(&r->tbl);
1.27      schwarze 1452:        }
                   1453:
1.49      schwarze 1454:        tbl = tbl_alloc(ppos, ln, r->parse);
1.27      schwarze 1455:
                   1456:        if (r->last_tbl)
1.49      schwarze 1457:                r->last_tbl->next = tbl;
1.27      schwarze 1458:        else
1.49      schwarze 1459:                r->first_tbl = r->last_tbl = tbl;
1.27      schwarze 1460:
1.49      schwarze 1461:        r->tbl = r->last_tbl = tbl;
1.27      schwarze 1462:        return(ROFF_IGN);
                   1463: }
                   1464:
                   1465: /* ARGSUSED */
                   1466: static enum rofferr
1.48      schwarze 1467: roff_cc(ROFF_ARGS)
                   1468: {
                   1469:        const char      *p;
                   1470:
                   1471:        p = *bufp + pos;
                   1472:
                   1473:        if ('\0' == *p || '.' == (r->control = *p++))
                   1474:                r->control = 0;
                   1475:
                   1476:        if ('\0' != *p)
                   1477:                mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
                   1478:
                   1479:        return(ROFF_IGN);
                   1480: }
                   1481:
                   1482: /* ARGSUSED */
                   1483: static enum rofferr
1.42      schwarze 1484: roff_tr(ROFF_ARGS)
                   1485: {
                   1486:        const char      *p, *first, *second;
                   1487:        size_t           fsz, ssz;
                   1488:        enum mandoc_esc  esc;
                   1489:
                   1490:        p = *bufp + pos;
                   1491:
                   1492:        if ('\0' == *p) {
                   1493:                mandoc_msg(MANDOCERR_ARGCOUNT, r->parse, ln, ppos, NULL);
                   1494:                return(ROFF_IGN);
                   1495:        }
                   1496:
                   1497:        while ('\0' != *p) {
                   1498:                fsz = ssz = 1;
                   1499:
                   1500:                first = p++;
                   1501:                if ('\\' == *first) {
                   1502:                        esc = mandoc_escape(&p, NULL, NULL);
                   1503:                        if (ESCAPE_ERROR == esc) {
                   1504:                                mandoc_msg
                   1505:                                        (MANDOCERR_BADESCAPE, r->parse,
                   1506:                                         ln, (int)(p - *bufp), NULL);
                   1507:                                return(ROFF_IGN);
                   1508:                        }
                   1509:                        fsz = (size_t)(p - first);
                   1510:                }
                   1511:
                   1512:                second = p++;
                   1513:                if ('\\' == *second) {
                   1514:                        esc = mandoc_escape(&p, NULL, NULL);
                   1515:                        if (ESCAPE_ERROR == esc) {
                   1516:                                mandoc_msg
                   1517:                                        (MANDOCERR_BADESCAPE, r->parse,
                   1518:                                         ln, (int)(p - *bufp), NULL);
                   1519:                                return(ROFF_IGN);
                   1520:                        }
                   1521:                        ssz = (size_t)(p - second);
                   1522:                } else if ('\0' == *second) {
                   1523:                        mandoc_msg(MANDOCERR_ARGCOUNT, r->parse,
                   1524:                                        ln, (int)(p - *bufp), NULL);
                   1525:                        second = " ";
                   1526:                        p--;
                   1527:                }
                   1528:
                   1529:                if (fsz > 1) {
                   1530:                        roff_setstrn(&r->xmbtab, first,
                   1531:                                        fsz, second, ssz, 0);
                   1532:                        continue;
                   1533:                }
                   1534:
                   1535:                if (NULL == r->xtab)
                   1536:                        r->xtab = mandoc_calloc
                   1537:                                (128, sizeof(struct roffstr));
                   1538:
                   1539:                free(r->xtab[(int)*first].p);
                   1540:                r->xtab[(int)*first].p = mandoc_strndup(second, ssz);
                   1541:                r->xtab[(int)*first].sz = ssz;
                   1542:        }
                   1543:
                   1544:        return(ROFF_IGN);
                   1545: }
                   1546:
                   1547: /* ARGSUSED */
                   1548: static enum rofferr
1.14      schwarze 1549: roff_so(ROFF_ARGS)
                   1550: {
                   1551:        char *name;
1.15      schwarze 1552:
1.35      schwarze 1553:        mandoc_msg(MANDOCERR_SO, r->parse, ln, ppos, NULL);
1.14      schwarze 1554:
1.22      schwarze 1555:        /*
                   1556:         * Handle `so'.  Be EXTREMELY careful, as we shouldn't be
                   1557:         * opening anything that's not in our cwd or anything beneath
                   1558:         * it.  Thus, explicitly disallow traversing up the file-system
                   1559:         * or using absolute paths.
                   1560:         */
                   1561:
1.14      schwarze 1562:        name = *bufp + pos;
                   1563:        if ('/' == *name || strstr(name, "../") || strstr(name, "/..")) {
1.35      schwarze 1564:                mandoc_msg(MANDOCERR_SOPATH, r->parse, ln, pos, NULL);
1.14      schwarze 1565:                return(ROFF_ERR);
                   1566:        }
                   1567:
                   1568:        *offs = pos;
                   1569:        return(ROFF_SO);
1.7       schwarze 1570: }
                   1571:
1.16      schwarze 1572: /* ARGSUSED */
                   1573: static enum rofferr
                   1574: roff_userdef(ROFF_ARGS)
1.12      schwarze 1575: {
1.16      schwarze 1576:        const char       *arg[9];
                   1577:        char             *cp, *n1, *n2;
1.25      schwarze 1578:        int               i;
1.12      schwarze 1579:
1.16      schwarze 1580:        /*
                   1581:         * Collect pointers to macro argument strings
                   1582:         * and null-terminate them.
                   1583:         */
                   1584:        cp = *bufp + pos;
1.25      schwarze 1585:        for (i = 0; i < 9; i++)
1.26      schwarze 1586:                arg[i] = '\0' == *cp ? "" :
1.35      schwarze 1587:                    mandoc_getarg(r->parse, &cp, ln, &pos);
1.16      schwarze 1588:
                   1589:        /*
                   1590:         * Expand macro arguments.
1.12      schwarze 1591:         */
1.16      schwarze 1592:        *szp = 0;
                   1593:        n1 = cp = mandoc_strdup(r->current_string);
                   1594:        while (NULL != (cp = strstr(cp, "\\$"))) {
                   1595:                i = cp[2] - '1';
                   1596:                if (0 > i || 8 < i) {
                   1597:                        /* Not an argument invocation. */
                   1598:                        cp += 2;
                   1599:                        continue;
                   1600:                }
                   1601:
                   1602:                *szp = strlen(n1) - 3 + strlen(arg[i]) + 1;
                   1603:                n2 = mandoc_malloc(*szp);
                   1604:
                   1605:                strlcpy(n2, n1, (size_t)(cp - n1 + 1));
                   1606:                strlcat(n2, arg[i], *szp);
                   1607:                strlcat(n2, cp + 3, *szp);
                   1608:
                   1609:                cp = n2 + (cp - n1);
                   1610:                free(n1);
                   1611:                n1 = n2;
1.12      schwarze 1612:        }
                   1613:
1.16      schwarze 1614:        /*
                   1615:         * Replace the macro invocation
                   1616:         * by the expanded macro.
                   1617:         */
                   1618:        free(*bufp);
                   1619:        *bufp = n1;
                   1620:        if (0 == *szp)
                   1621:                *szp = strlen(*bufp) + 1;
                   1622:
1.19      schwarze 1623:        return(*szp > 1 && '\n' == (*bufp)[(int)*szp - 2] ?
1.16      schwarze 1624:           ROFF_REPARSE : ROFF_APPEND);
1.12      schwarze 1625: }
1.28      schwarze 1626:
                   1627: static char *
                   1628: roff_getname(struct roff *r, char **cpp, int ln, int pos)
                   1629: {
                   1630:        char     *name, *cp;
                   1631:
                   1632:        name = *cpp;
                   1633:        if ('\0' == *name)
                   1634:                return(name);
                   1635:
                   1636:        /* Read until end of name. */
                   1637:        for (cp = name; '\0' != *cp && ' ' != *cp; cp++) {
                   1638:                if ('\\' != *cp)
                   1639:                        continue;
                   1640:                cp++;
                   1641:                if ('\\' == *cp)
                   1642:                        continue;
1.35      schwarze 1643:                mandoc_msg(MANDOCERR_NAMESC, r->parse, ln, pos, NULL);
1.28      schwarze 1644:                *cp = '\0';
                   1645:                name = cp;
                   1646:        }
                   1647:
                   1648:        /* Nil-terminate name. */
                   1649:        if ('\0' != *cp)
                   1650:                *(cp++) = '\0';
                   1651:
                   1652:        /* Read past spaces. */
                   1653:        while (' ' == *cp)
                   1654:                cp++;
                   1655:
                   1656:        *cpp = cp;
                   1657:        return(name);
                   1658: }
                   1659:
1.16      schwarze 1660: /*
                   1661:  * Store *string into the user-defined string called *name.
                   1662:  * In multiline mode, append to an existing entry and append '\n';
                   1663:  * else replace the existing entry, if there is one.
                   1664:  * To clear an existing entry, call with (*r, *name, NULL, 0).
                   1665:  */
1.8       schwarze 1666: static void
1.16      schwarze 1667: roff_setstr(struct roff *r, const char *name, const char *string,
                   1668:        int multiline)
1.7       schwarze 1669: {
1.42      schwarze 1670:
                   1671:        roff_setstrn(&r->strtab, name, strlen(name), string,
                   1672:                        string ? strlen(string) : 0, multiline);
                   1673: }
                   1674:
                   1675: static void
                   1676: roff_setstrn(struct roffkv **r, const char *name, size_t namesz,
                   1677:                const char *string, size_t stringsz, int multiline)
                   1678: {
                   1679:        struct roffkv   *n;
                   1680:        char            *c;
                   1681:        int              i;
                   1682:        size_t           oldch, newch;
1.7       schwarze 1683:
1.16      schwarze 1684:        /* Search for an existing string with the same name. */
1.42      schwarze 1685:        n = *r;
                   1686:
                   1687:        while (n && strcmp(name, n->key.p))
1.7       schwarze 1688:                n = n->next;
1.8       schwarze 1689:
                   1690:        if (NULL == n) {
1.16      schwarze 1691:                /* Create a new string table entry. */
1.42      schwarze 1692:                n = mandoc_malloc(sizeof(struct roffkv));
                   1693:                n->key.p = mandoc_strndup(name, namesz);
                   1694:                n->key.sz = namesz;
                   1695:                n->val.p = NULL;
                   1696:                n->val.sz = 0;
                   1697:                n->next = *r;
                   1698:                *r = n;
1.16      schwarze 1699:        } else if (0 == multiline) {
                   1700:                /* In multiline mode, append; else replace. */
1.42      schwarze 1701:                free(n->val.p);
                   1702:                n->val.p = NULL;
                   1703:                n->val.sz = 0;
1.16      schwarze 1704:        }
                   1705:
                   1706:        if (NULL == string)
                   1707:                return;
                   1708:
                   1709:        /*
                   1710:         * One additional byte for the '\n' in multiline mode,
                   1711:         * and one for the terminating '\0'.
                   1712:         */
1.42      schwarze 1713:        newch = stringsz + (multiline ? 2u : 1u);
                   1714:
                   1715:        if (NULL == n->val.p) {
                   1716:                n->val.p = mandoc_malloc(newch);
                   1717:                *n->val.p = '\0';
1.16      schwarze 1718:                oldch = 0;
                   1719:        } else {
1.42      schwarze 1720:                oldch = n->val.sz;
                   1721:                n->val.p = mandoc_realloc(n->val.p, oldch + newch);
1.16      schwarze 1722:        }
                   1723:
                   1724:        /* Skip existing content in the destination buffer. */
1.42      schwarze 1725:        c = n->val.p + (int)oldch;
1.16      schwarze 1726:
                   1727:        /* Append new content to the destination buffer. */
1.42      schwarze 1728:        i = 0;
                   1729:        while (i < (int)stringsz) {
1.16      schwarze 1730:                /*
                   1731:                 * Rudimentary roff copy mode:
                   1732:                 * Handle escaped backslashes.
                   1733:                 */
1.42      schwarze 1734:                if ('\\' == string[i] && '\\' == string[i + 1])
                   1735:                        i++;
                   1736:                *c++ = string[i++];
1.16      schwarze 1737:        }
1.8       schwarze 1738:
1.16      schwarze 1739:        /* Append terminating bytes. */
                   1740:        if (multiline)
                   1741:                *c++ = '\n';
1.42      schwarze 1742:
1.16      schwarze 1743:        *c = '\0';
1.42      schwarze 1744:        n->val.sz = (int)(c - n->val.p);
1.7       schwarze 1745: }
                   1746:
1.8       schwarze 1747: static const char *
                   1748: roff_getstrn(const struct roff *r, const char *name, size_t len)
1.7       schwarze 1749: {
1.42      schwarze 1750:        const struct roffkv *n;
1.7       schwarze 1751:
1.42      schwarze 1752:        for (n = r->strtab; n; n = n->next)
                   1753:                if (0 == strncmp(name, n->key.p, len) &&
                   1754:                                '\0' == n->key.p[(int)len])
                   1755:                        return(n->val.p);
1.8       schwarze 1756:
1.42      schwarze 1757:        return(NULL);
1.7       schwarze 1758: }
                   1759:
1.8       schwarze 1760: static void
1.42      schwarze 1761: roff_freestr(struct roffkv *r)
1.7       schwarze 1762: {
1.42      schwarze 1763:        struct roffkv    *n, *nn;
1.7       schwarze 1764:
1.42      schwarze 1765:        for (n = r; n; n = nn) {
                   1766:                free(n->key.p);
                   1767:                free(n->val.p);
1.7       schwarze 1768:                nn = n->next;
                   1769:                free(n);
                   1770:        }
1.27      schwarze 1771: }
                   1772:
                   1773: const struct tbl_span *
                   1774: roff_span(const struct roff *r)
                   1775: {
                   1776:
                   1777:        return(r->tbl ? tbl_span(r->tbl) : NULL);
1.32      schwarze 1778: }
                   1779:
                   1780: const struct eqn *
                   1781: roff_eqn(const struct roff *r)
                   1782: {
                   1783:
                   1784:        return(r->last_eqn ? &r->last_eqn->eqn : NULL);
1.42      schwarze 1785: }
                   1786:
                   1787: /*
                   1788:  * Duplicate an input string, making the appropriate character
                   1789:  * conversations (as stipulated by `tr') along the way.
                   1790:  * Returns a heap-allocated string with all the replacements made.
                   1791:  */
                   1792: char *
                   1793: roff_strdup(const struct roff *r, const char *p)
                   1794: {
                   1795:        const struct roffkv *cp;
                   1796:        char            *res;
                   1797:        const char      *pp;
                   1798:        size_t           ssz, sz;
                   1799:        enum mandoc_esc  esc;
                   1800:
                   1801:        if (NULL == r->xmbtab && NULL == r->xtab)
                   1802:                return(mandoc_strdup(p));
                   1803:        else if ('\0' == *p)
                   1804:                return(mandoc_strdup(""));
                   1805:
                   1806:        /*
                   1807:         * Step through each character looking for term matches
                   1808:         * (remember that a `tr' can be invoked with an escape, which is
                   1809:         * a glyph but the escape is multi-character).
                   1810:         * We only do this if the character hash has been initialised
                   1811:         * and the string is >0 length.
                   1812:         */
                   1813:
                   1814:        res = NULL;
                   1815:        ssz = 0;
                   1816:
                   1817:        while ('\0' != *p) {
                   1818:                if ('\\' != *p && r->xtab && r->xtab[(int)*p].p) {
                   1819:                        sz = r->xtab[(int)*p].sz;
                   1820:                        res = mandoc_realloc(res, ssz + sz + 1);
                   1821:                        memcpy(res + ssz, r->xtab[(int)*p].p, sz);
                   1822:                        ssz += sz;
                   1823:                        p++;
                   1824:                        continue;
                   1825:                } else if ('\\' != *p) {
                   1826:                        res = mandoc_realloc(res, ssz + 2);
                   1827:                        res[ssz++] = *p++;
                   1828:                        continue;
                   1829:                }
                   1830:
                   1831:                /* Search for term matches. */
                   1832:                for (cp = r->xmbtab; cp; cp = cp->next)
                   1833:                        if (0 == strncmp(p, cp->key.p, cp->key.sz))
                   1834:                                break;
                   1835:
                   1836:                if (NULL != cp) {
                   1837:                        /*
                   1838:                         * A match has been found.
                   1839:                         * Append the match to the array and move
                   1840:                         * forward by its keysize.
                   1841:                         */
                   1842:                        res = mandoc_realloc
                   1843:                                (res, ssz + cp->val.sz + 1);
                   1844:                        memcpy(res + ssz, cp->val.p, cp->val.sz);
                   1845:                        ssz += cp->val.sz;
                   1846:                        p += (int)cp->key.sz;
                   1847:                        continue;
                   1848:                }
                   1849:
                   1850:                /*
                   1851:                 * Handle escapes carefully: we need to copy
                   1852:                 * over just the escape itself, or else we might
                   1853:                 * do replacements within the escape itself.
                   1854:                 * Make sure to pass along the bogus string.
                   1855:                 */
                   1856:                pp = p++;
                   1857:                esc = mandoc_escape(&p, NULL, NULL);
                   1858:                if (ESCAPE_ERROR == esc) {
                   1859:                        sz = strlen(pp);
                   1860:                        res = mandoc_realloc(res, ssz + sz + 1);
                   1861:                        memcpy(res + ssz, pp, sz);
                   1862:                        break;
                   1863:                }
                   1864:                /*
                   1865:                 * We bail out on bad escapes.
                   1866:                 * No need to warn: we already did so when
                   1867:                 * roff_res() was called.
                   1868:                 */
                   1869:                sz = (int)(p - pp);
                   1870:                res = mandoc_realloc(res, ssz + sz + 1);
                   1871:                memcpy(res + ssz, pp, sz);
                   1872:                ssz += sz;
                   1873:        }
                   1874:
                   1875:        res[(int)ssz] = '\0';
                   1876:        return(res);
1.48      schwarze 1877: }
                   1878:
                   1879: /*
                   1880:  * Find out whether a line is a macro line or not.
                   1881:  * If it is, adjust the current position and return one; if it isn't,
                   1882:  * return zero and don't change the current position.
                   1883:  * If the control character has been set with `.cc', then let that grain
                   1884:  * precedence.
                   1885:  * This is slighly contrary to groff, where using the non-breaking
                   1886:  * control character when `cc' has been invoked will cause the
                   1887:  * non-breaking macro contents to be printed verbatim.
                   1888:  */
                   1889: int
                   1890: roff_getcontrol(const struct roff *r, const char *cp, int *ppos)
                   1891: {
                   1892:        int             pos;
                   1893:
                   1894:        pos = *ppos;
                   1895:
                   1896:        if (0 != r->control && cp[pos] == r->control)
                   1897:                pos++;
                   1898:        else if (0 != r->control)
                   1899:                return(0);
                   1900:        else if ('\\' == cp[pos] && '.' == cp[pos + 1])
                   1901:                pos += 2;
                   1902:        else if ('.' == cp[pos] || '\'' == cp[pos])
                   1903:                pos++;
                   1904:        else
                   1905:                return(0);
                   1906:
                   1907:        while (' ' == cp[pos] || '\t' == cp[pos])
                   1908:                pos++;
                   1909:
                   1910:        *ppos = pos;
                   1911:        return(1);
1.1       schwarze 1912: }