version 1.80, 2010/05/17 00:37:26 |
version 1.86, 2010/06/01 11:47:28 |
|
|
#endif |
#endif |
|
|
#include <assert.h> |
#include <assert.h> |
|
#include <ctype.h> |
#include <stdlib.h> |
#include <stdlib.h> |
#include <string.h> |
#include <string.h> |
#include <stdio.h> |
#include <stdio.h> |
|
|
#include "mandoc.h" |
#include "mandoc.h" |
#include "roff.h" |
#include "roff.h" |
|
|
|
#define RSTACK_MAX 128 |
|
|
#define ROFF_CTL(c) \ |
#define ROFF_CTL(c) \ |
('.' == (c) || '\'' == (c)) |
('.' == (c) || '\'' == (c)) |
|
|
|
|
ROFF_de, |
ROFF_de, |
ROFF_dei, |
ROFF_dei, |
ROFF_de1, |
ROFF_de1, |
|
ROFF_ds, |
|
ROFF_el, |
|
ROFF_ie, |
ROFF_if, |
ROFF_if, |
ROFF_ig, |
ROFF_ig, |
|
ROFF_rm, |
|
ROFF_tr, |
ROFF_cblock, |
ROFF_cblock, |
ROFF_ccond, |
ROFF_ccond, |
#if 0 |
|
ROFF_ie, |
|
ROFF_el, |
|
#endif |
|
ROFF_MAX |
ROFF_MAX |
}; |
}; |
|
|
|
enum roffrule { |
|
ROFFRULE_ALLOW, |
|
ROFFRULE_DENY |
|
}; |
|
|
struct roff { |
struct roff { |
struct roffnode *last; /* leaf of stack */ |
struct roffnode *last; /* leaf of stack */ |
mandocmsg msg; /* err/warn/fatal messages */ |
mandocmsg msg; /* err/warn/fatal messages */ |
void *data; /* privdata for messages */ |
void *data; /* privdata for messages */ |
|
enum roffrule rstack[RSTACK_MAX]; /* stack of !`ie' rules */ |
|
int rstackpos; /* position in rstack */ |
}; |
}; |
|
|
enum roffrule { |
|
ROFFRULE_ALLOW, |
|
ROFFRULE_DENY |
|
}; |
|
|
|
struct roffnode { |
struct roffnode { |
enum rofft tok; /* type of node */ |
enum rofft tok; /* type of node */ |
struct roffnode *parent; /* up one in stack */ |
struct roffnode *parent; /* up one in stack */ |
Line 65 struct roffnode { |
|
Line 71 struct roffnode { |
|
int col; /* parse col */ |
int col; /* parse col */ |
char *end; /* end-rules: custom token */ |
char *end; /* end-rules: custom token */ |
int endspan; /* end-rules: next-line or infty */ |
int endspan; /* end-rules: next-line or infty */ |
enum roffrule rule; |
enum roffrule rule; /* current evaluation rule */ |
}; |
}; |
|
|
#define ROFF_ARGS struct roff *r, /* parse ctx */ \ |
#define ROFF_ARGS struct roff *r, /* parse ctx */ \ |
|
|
roffproc sub; /* process as child of macro */ |
roffproc sub; /* process as child of macro */ |
int flags; |
int flags; |
#define ROFFMAC_STRUCT (1 << 0) /* always interpret */ |
#define ROFFMAC_STRUCT (1 << 0) /* always interpret */ |
|
struct roffmac *next; |
}; |
}; |
|
|
static enum rofferr roff_block(ROFF_ARGS); |
static enum rofferr roff_block(ROFF_ARGS); |
Line 93 static enum rofferr roff_block_text(ROFF_ARGS); |
|
Line 100 static enum rofferr roff_block_text(ROFF_ARGS); |
|
static enum rofferr roff_block_sub(ROFF_ARGS); |
static enum rofferr roff_block_sub(ROFF_ARGS); |
static enum rofferr roff_cblock(ROFF_ARGS); |
static enum rofferr roff_cblock(ROFF_ARGS); |
static enum rofferr roff_ccond(ROFF_ARGS); |
static enum rofferr roff_ccond(ROFF_ARGS); |
static enum rofferr roff_if(ROFF_ARGS); |
static enum rofferr roff_cond(ROFF_ARGS); |
static enum rofferr roff_if_text(ROFF_ARGS); |
static enum rofferr roff_cond_text(ROFF_ARGS); |
static enum rofferr roff_if_sub(ROFF_ARGS); |
static enum rofferr roff_cond_sub(ROFF_ARGS); |
|
static enum rofferr roff_line(ROFF_ARGS); |
|
|
const struct roffmac roffs[ROFF_MAX] = { |
/* See roff_hash_find() */ |
{ "am", roff_block, roff_block_text, roff_block_sub, 0 }, |
|
{ "ami", roff_block, roff_block_text, roff_block_sub, 0 }, |
#define ASCII_HI 126 |
{ "am1", roff_block, roff_block_text, roff_block_sub, 0 }, |
#define ASCII_LO 33 |
{ "de", roff_block, roff_block_text, roff_block_sub, 0 }, |
#define HASHWIDTH (ASCII_HI - ASCII_LO + 1) |
{ "dei", roff_block, roff_block_text, roff_block_sub, 0 }, |
|
{ "de1", roff_block, roff_block_text, roff_block_sub, 0 }, |
static struct roffmac *hash[HASHWIDTH]; |
{ "if", roff_if, roff_if_text, roff_if_sub, ROFFMAC_STRUCT }, |
|
{ "ig", roff_block, roff_block_text, roff_block_sub, 0 }, |
static struct roffmac roffs[ROFF_MAX] = { |
{ ".", roff_cblock, NULL, NULL, 0 }, |
{ "am", roff_block, roff_block_text, roff_block_sub, 0, NULL }, |
{ "\\}", roff_ccond, NULL, NULL, 0 }, |
{ "ami", roff_block, roff_block_text, roff_block_sub, 0, NULL }, |
|
{ "am1", roff_block, roff_block_text, roff_block_sub, 0, NULL }, |
|
{ "de", roff_block, roff_block_text, roff_block_sub, 0, NULL }, |
|
{ "dei", roff_block, roff_block_text, roff_block_sub, 0, NULL }, |
|
{ "de1", roff_block, roff_block_text, roff_block_sub, 0, NULL }, |
|
{ "ds", roff_line, NULL, NULL, 0, NULL }, |
|
{ "el", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL }, |
|
{ "ie", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL }, |
|
{ "if", roff_cond, roff_cond_text, roff_cond_sub, ROFFMAC_STRUCT, NULL }, |
|
{ "ig", roff_block, roff_block_text, roff_block_sub, 0, NULL }, |
|
{ "rm", roff_line, NULL, NULL, 0, NULL }, |
|
{ "tr", roff_line, NULL, NULL, 0, NULL }, |
|
{ ".", roff_cblock, NULL, NULL, 0, NULL }, |
|
{ "\\}", roff_ccond, NULL, NULL, 0, NULL }, |
}; |
}; |
|
|
static void roff_free1(struct roff *); |
static void roff_free1(struct roff *); |
static enum rofft roff_hash_find(const char *); |
static enum rofft roff_hash_find(const char *); |
|
static void roff_hash_init(void); |
static void roffnode_cleanscope(struct roff *); |
static void roffnode_cleanscope(struct roff *); |
static int roffnode_push(struct roff *, |
static int roffnode_push(struct roff *, |
enum rofft, int, int); |
enum rofft, int, int); |
static void roffnode_pop(struct roff *); |
static void roffnode_pop(struct roff *); |
static enum rofft roff_parse(const char *, int *); |
static enum rofft roff_parse(const char *, int *); |
|
|
|
/* See roff_hash_find() */ |
|
#define ROFF_HASH(p) (p[0] - ASCII_LO) |
|
|
|
static void |
|
roff_hash_init(void) |
|
{ |
|
struct roffmac *n; |
|
int buc, i; |
|
|
|
for (i = 0; i < (int)ROFF_MAX; i++) { |
|
assert(roffs[i].name[0] >= ASCII_LO); |
|
assert(roffs[i].name[0] <= ASCII_HI); |
|
|
|
buc = ROFF_HASH(roffs[i].name); |
|
|
|
if (NULL != (n = hash[buc])) { |
|
for ( ; n->next; n = n->next) |
|
/* Do nothing. */ ; |
|
n->next = &roffs[i]; |
|
} else |
|
hash[buc] = &roffs[i]; |
|
} |
|
} |
|
|
|
|
/* |
/* |
* Look up a roff token by its name. Returns ROFF_MAX if no macro by |
* Look up a roff token by its name. Returns ROFF_MAX if no macro by |
* the nil-terminated string name could be found. |
* the nil-terminated string name could be found. |
Line 126 static enum rofft roff_parse(const char *, int *); |
|
Line 172 static enum rofft roff_parse(const char *, int *); |
|
static enum rofft |
static enum rofft |
roff_hash_find(const char *p) |
roff_hash_find(const char *p) |
{ |
{ |
int i; |
int buc; |
|
struct roffmac *n; |
|
|
/* FIXME: make this be fast and efficient. */ |
/* |
|
* libroff has an extremely simple hashtable, for the time |
|
* being, which simply keys on the first character, which must |
|
* be printable, then walks a chain. It works well enough until |
|
* optimised. |
|
*/ |
|
|
for (i = 0; i < (int)ROFF_MAX; i++) |
if (p[0] < ASCII_LO || p[0] > ASCII_HI) |
if (0 == strcmp(roffs[i].name, p)) |
return(ROFF_MAX); |
return((enum rofft)i); |
|
|
|
|
buc = ROFF_HASH(p); |
|
|
|
if (NULL == (n = hash[buc])) |
|
return(ROFF_MAX); |
|
for ( ; n; n = n->next) |
|
if (0 == strcmp(n->name, p)) |
|
return((enum rofft)(n - roffs)); |
|
|
return(ROFF_MAX); |
return(ROFF_MAX); |
} |
} |
|
|
Line 149 roffnode_pop(struct roff *r) |
|
Line 208 roffnode_pop(struct roff *r) |
|
|
|
assert(r->last); |
assert(r->last); |
p = r->last; |
p = r->last; |
|
|
|
if (ROFF_el == p->tok) |
|
if (r->rstackpos > -1) |
|
r->rstackpos--; |
|
|
r->last = r->last->parent; |
r->last = r->last->parent; |
if (p->end) |
if (p->end) |
free(p->end); |
free(p->end); |
Line 219 roff_alloc(const mandocmsg msg, void *data) |
|
Line 283 roff_alloc(const mandocmsg msg, void *data) |
|
|
|
r->msg = msg; |
r->msg = msg; |
r->data = data; |
r->data = data; |
|
r->rstackpos = -1; |
|
|
|
roff_hash_init(); |
return(r); |
return(r); |
} |
} |
|
|
Line 306 roff_parse(const char *buf, int *pos) |
|
Line 373 roff_parse(const char *buf, int *pos) |
|
for (j = 0; j < 4; j++, (*pos)++) |
for (j = 0; j < 4; j++, (*pos)++) |
if ('\0' == (mac[j] = buf[*pos])) |
if ('\0' == (mac[j] = buf[*pos])) |
break; |
break; |
else if (' ' == buf[*pos]) |
else if (' ' == buf[*pos] || (j && '\\' == buf[*pos])) |
break; |
break; |
|
|
if (j == 4 || j < 1) |
if (j == 4 || j < 1) |
Line 340 roff_cblock(ROFF_ARGS) |
|
Line 407 roff_cblock(ROFF_ARGS) |
|
return(ROFF_IGN); |
return(ROFF_IGN); |
} |
} |
|
|
if (ROFF_ig != r->last->tok) { |
switch (r->last->tok) { |
|
case (ROFF_am): |
|
/* FALLTHROUGH */ |
|
case (ROFF_ami): |
|
/* FALLTHROUGH */ |
|
case (ROFF_am1): |
|
/* FALLTHROUGH */ |
|
case (ROFF_de): |
|
/* FALLTHROUGH */ |
|
case (ROFF_dei): |
|
/* FALLTHROUGH */ |
|
case (ROFF_de1): |
|
/* FALLTHROUGH */ |
|
case (ROFF_ig): |
|
break; |
|
default: |
if ( ! (*r->msg)(MANDOCERR_NOSCOPE, r->data, ln, ppos, NULL)) |
if ( ! (*r->msg)(MANDOCERR_NOSCOPE, r->data, ln, ppos, NULL)) |
return(ROFF_ERR); |
return(ROFF_ERR); |
return(ROFF_IGN); |
return(ROFF_IGN); |
Line 380 roff_ccond(ROFF_ARGS) |
|
Line 462 roff_ccond(ROFF_ARGS) |
|
return(ROFF_IGN); |
return(ROFF_IGN); |
} |
} |
|
|
if (ROFF_if != r->last->tok) { |
switch (r->last->tok) { |
|
case (ROFF_el): |
|
/* FALLTHROUGH */ |
|
case (ROFF_ie): |
|
/* FALLTHROUGH */ |
|
case (ROFF_if): |
|
break; |
|
default: |
if ( ! (*r->msg)(MANDOCERR_NOSCOPE, r->data, ln, ppos, NULL)) |
if ( ! (*r->msg)(MANDOCERR_NOSCOPE, r->data, ln, ppos, NULL)) |
return(ROFF_ERR); |
return(ROFF_ERR); |
return(ROFF_IGN); |
return(ROFF_IGN); |
Line 463 roff_block(ROFF_ARGS) |
|
Line 552 roff_block(ROFF_ARGS) |
|
|
|
/* ARGSUSED */ |
/* ARGSUSED */ |
static enum rofferr |
static enum rofferr |
roff_if_sub(ROFF_ARGS) |
|
{ |
|
enum rofft t; |
|
enum roffrule rr; |
|
|
|
ppos = pos; |
|
rr = r->last->rule; |
|
roffnode_cleanscope(r); |
|
|
|
if (ROFF_MAX == (t = roff_parse(*bufp, &pos))) |
|
return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT); |
|
|
|
/* |
|
* A denied conditional must evaluate its children if and only |
|
* if they're either structurally required (such as loops and |
|
* conditionals) or a closing macro. |
|
*/ |
|
if (ROFFRULE_DENY == rr) |
|
if ( ! (ROFFMAC_STRUCT & roffs[t].flags)) |
|
if (ROFF_ccond != t) |
|
return(ROFF_IGN); |
|
|
|
assert(roffs[t].proc); |
|
return((*roffs[t].proc) |
|
(r, t, bufp, szp, ln, ppos, pos, offs)); |
|
} |
|
|
|
|
|
/* ARGSUSED */ |
|
static enum rofferr |
|
roff_block_sub(ROFF_ARGS) |
roff_block_sub(ROFF_ARGS) |
{ |
{ |
enum rofft t; |
enum rofft t; |
Line 558 roff_block_text(ROFF_ARGS) |
|
Line 617 roff_block_text(ROFF_ARGS) |
|
|
|
/* ARGSUSED */ |
/* ARGSUSED */ |
static enum rofferr |
static enum rofferr |
roff_if_text(ROFF_ARGS) |
roff_cond_sub(ROFF_ARGS) |
{ |
{ |
|
enum rofft t; |
|
enum roffrule rr; |
|
|
|
ppos = pos; |
|
rr = r->last->rule; |
|
|
|
roff_cond_text(r, tok, bufp, szp, ln, ppos, pos, offs); |
|
|
|
if (ROFF_MAX == (t = roff_parse(*bufp, &pos))) |
|
return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT); |
|
|
|
/* |
|
* A denied conditional must evaluate its children if and only |
|
* if they're either structurally required (such as loops and |
|
* conditionals) or a closing macro. |
|
*/ |
|
if (ROFFRULE_DENY == rr) |
|
if ( ! (ROFFMAC_STRUCT & roffs[t].flags)) |
|
if (ROFF_ccond != t) |
|
return(ROFF_IGN); |
|
|
|
assert(roffs[t].proc); |
|
return((*roffs[t].proc) |
|
(r, t, bufp, szp, ln, ppos, pos, offs)); |
|
} |
|
|
|
|
|
/* ARGSUSED */ |
|
static enum rofferr |
|
roff_cond_text(ROFF_ARGS) |
|
{ |
char *ep, *st; |
char *ep, *st; |
|
enum roffrule rr; |
|
|
|
rr = r->last->rule; |
|
|
|
/* |
|
* We display the value of the text if out current evaluation |
|
* scope permits us to do so. |
|
*/ |
|
|
st = &(*bufp)[pos]; |
st = &(*bufp)[pos]; |
if (NULL == (ep = strstr(st, "\\}"))) { |
if (NULL == (ep = strstr(st, "\\}"))) { |
roffnode_cleanscope(r); |
roffnode_cleanscope(r); |
return(ROFF_IGN); |
return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT); |
} |
} |
|
|
if (ep > st && '\\' != *(ep - 1)) |
if (ep == st || (ep > st && '\\' != *(ep - 1))) |
roffnode_pop(r); |
roffnode_pop(r); |
|
|
roffnode_cleanscope(r); |
roffnode_cleanscope(r); |
return(ROFF_IGN); |
return(ROFFRULE_DENY == rr ? ROFF_IGN : ROFF_CONT); |
} |
} |
|
|
|
|
/* ARGSUSED */ |
/* ARGSUSED */ |
static enum rofferr |
static enum rofferr |
roff_if(ROFF_ARGS) |
roff_cond(ROFF_ARGS) |
{ |
{ |
|
int cpos; /* position of the condition */ |
int sv; |
int sv; |
|
|
/* |
/* Stack overflow! */ |
* Read ahead past the conditional. |
|
* FIXME: this does not work, as conditionals don't end on |
|
* whitespace, but are parsed according to a formal grammar. |
|
* It's good enough for now, however. |
|
*/ |
|
|
|
while ((*bufp)[pos] && ' ' != (*bufp)[pos]) |
if (ROFF_ie == tok && r->rstackpos == RSTACK_MAX - 1) { |
pos++; |
(*r->msg)(MANDOCERR_MEM, r->data, ln, ppos, NULL); |
|
return(ROFF_ERR); |
|
} |
|
|
|
cpos = pos; |
|
|
|
if (ROFF_if == tok || ROFF_ie == tok) { |
|
/* |
|
* Read ahead past the conditional. FIXME: this does |
|
* not work, as conditionals don't end on whitespace, |
|
* but are parsed according to a formal grammar. It's |
|
* good enough for now, however. |
|
*/ |
|
while ((*bufp)[pos] && ' ' != (*bufp)[pos]) |
|
pos++; |
|
} |
|
|
sv = pos; |
sv = pos; |
while (' ' == (*bufp)[pos]) |
while (' ' == (*bufp)[pos]) |
pos++; |
pos++; |
Line 602 roff_if(ROFF_ARGS) |
|
Line 711 roff_if(ROFF_ARGS) |
|
* really doing anything. Warn about this. It's probably |
* really doing anything. Warn about this. It's probably |
* wrong. |
* wrong. |
*/ |
*/ |
|
|
if ('\0' == (*bufp)[pos] && sv != pos) { |
if ('\0' == (*bufp)[pos] && sv != pos) { |
if ( ! (*r->msg)(MANDOCERR_NOARGS, r->data, ln, ppos, NULL)) |
if ( ! (*r->msg)(MANDOCERR_NOARGS, r->data, ln, ppos, NULL)) |
return(ROFF_ERR); |
return(ROFF_ERR); |
Line 612 roff_if(ROFF_ARGS) |
|
Line 720 roff_if(ROFF_ARGS) |
|
if ( ! roffnode_push(r, tok, ln, ppos)) |
if ( ! roffnode_push(r, tok, ln, ppos)) |
return(ROFF_ERR); |
return(ROFF_ERR); |
|
|
/* Don't evaluate: just assume NO. */ |
/* XXX: Implement more conditionals. */ |
|
|
|
if (ROFF_if == tok || ROFF_ie == tok) |
|
r->last->rule = 'n' == (*bufp)[cpos] ? |
|
ROFFRULE_ALLOW : ROFFRULE_DENY; |
|
else if (ROFF_el == tok) { |
|
/* |
|
* An `.el' will get the value of the current rstack |
|
* entry set in prior `ie' calls or defaults to DENY. |
|
*/ |
|
if (r->rstackpos < 0) |
|
r->last->rule = ROFFRULE_DENY; |
|
else |
|
r->last->rule = r->rstack[r->rstackpos]; |
|
} |
|
if (ROFF_ie == tok) { |
|
/* |
|
* An if-else will put the NEGATION of the current |
|
* evaluated conditional into the stack. |
|
*/ |
|
r->rstackpos++; |
|
if (ROFFRULE_DENY == r->last->rule) |
|
r->rstack[r->rstackpos] = ROFFRULE_ALLOW; |
|
else |
|
r->rstack[r->rstackpos] = ROFFRULE_DENY; |
|
} |
|
if (r->last->parent && ROFFRULE_DENY == r->last->parent->rule) |
|
r->last->rule = ROFFRULE_DENY; |
|
|
r->last->endspan = 1; |
r->last->endspan = 1; |
|
|
if ('\\' == (*bufp)[pos] && '{' == (*bufp)[pos + 1]) { |
if ('\\' == (*bufp)[pos] && '{' == (*bufp)[pos + 1]) { |
Line 633 roff_if(ROFF_ARGS) |
|
Line 768 roff_if(ROFF_ARGS) |
|
|
|
*offs = pos; |
*offs = pos; |
return(ROFF_RERUN); |
return(ROFF_RERUN); |
|
} |
|
|
|
|
|
/* ARGSUSED */ |
|
static enum rofferr |
|
roff_line(ROFF_ARGS) |
|
{ |
|
|
|
return(ROFF_IGN); |
} |
} |