version 1.75, 2009/04/12 19:45:26 |
version 1.229, 2014/11/19 03:08:17 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
* Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@openbsd.org> |
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
|
* Copyright (c) 2010, 2012, 2013, 2014 Ingo Schwarze <schwarze@openbsd.org> |
* |
* |
* Permission to use, copy, modify, and distribute this software for any |
* Permission to use, copy, modify, and distribute this software for any |
* purpose with or without fee is hereby granted, provided that the above |
* purpose with or without fee is hereby granted, provided that the above |
|
|
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF |
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF |
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. |
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. |
*/ |
*/ |
|
#include "config.h" |
|
|
|
#include <sys/types.h> |
|
|
#include <assert.h> |
#include <assert.h> |
#include <ctype.h> |
#include <ctype.h> |
#include <stdarg.h> |
#include <stdarg.h> |
#include <stdio.h> |
#include <stdio.h> |
#include <stdlib.h> |
#include <stdlib.h> |
#include <string.h> |
#include <string.h> |
|
#include <time.h> |
|
|
|
#include "mdoc.h" |
|
#include "mandoc.h" |
|
#include "mandoc_aux.h" |
#include "libmdoc.h" |
#include "libmdoc.h" |
|
#include "libmandoc.h" |
|
|
enum merr { |
const char *const __mdoc_macronames[MDOC_MAX + 1] = { |
ENOCALL, |
"Ap", "Dd", "Dt", "Os", |
EBODYPROL, |
|
EPROLBODY, |
|
ESPACE, |
|
ETEXTPROL, |
|
ENOBLANK, |
|
EMALLOC |
|
}; |
|
|
|
const char *const __mdoc_macronames[MDOC_MAX] = { |
|
"\\\"", "Dd", "Dt", "Os", |
|
"Sh", "Ss", "Pp", "D1", |
"Sh", "Ss", "Pp", "D1", |
"Dl", "Bd", "Ed", "Bl", |
"Dl", "Bd", "Ed", "Bl", |
"El", "It", "Ad", "An", |
"El", "It", "Ad", "An", |
Line 44 const char *const __mdoc_macronames[MDOC_MAX] = { |
|
Line 44 const char *const __mdoc_macronames[MDOC_MAX] = { |
|
"Ic", "In", "Li", "Nd", |
"Ic", "In", "Li", "Nd", |
"Nm", "Op", "Ot", "Pa", |
"Nm", "Op", "Ot", "Pa", |
"Rv", "St", "Va", "Vt", |
"Rv", "St", "Va", "Vt", |
/* LINTED */ |
"Xr", "%A", "%B", "%D", |
"Xr", "\%A", "\%B", "\%D", |
"%I", "%J", "%N", "%O", |
/* LINTED */ |
"%P", "%R", "%T", "%V", |
"\%I", "\%J", "\%N", "\%O", |
|
/* LINTED */ |
|
"\%P", "\%R", "\%T", "\%V", |
|
"Ac", "Ao", "Aq", "At", |
"Ac", "Ao", "Aq", "At", |
"Bc", "Bf", "Bo", "Bq", |
"Bc", "Bf", "Bo", "Bq", |
"Bsx", "Bx", "Db", "Dc", |
"Bsx", "Bx", "Db", "Dc", |
Line 63 const char *const __mdoc_macronames[MDOC_MAX] = { |
|
Line 60 const char *const __mdoc_macronames[MDOC_MAX] = { |
|
"Tn", "Ux", "Xc", "Xo", |
"Tn", "Ux", "Xc", "Xo", |
"Fo", "Fc", "Oo", "Oc", |
"Fo", "Fc", "Oo", "Oc", |
"Bk", "Ek", "Bt", "Hf", |
"Bk", "Ek", "Bt", "Hf", |
"Fr", "Ud", "Lb", "Ap", |
"Fr", "Ud", "Lb", "Lp", |
"Lp", "Lk", "Mt", "Brq", |
"Lk", "Mt", "Brq", "Bro", |
/* LINTED */ |
"Brc", "%C", "Es", "En", |
"Bro", "Brc", "\%C", "Es", |
"Dx", "%Q", "br", "sp", |
/* LINTED */ |
"%U", "Ta", "ll", "text", |
"En", "Dx", "\%Q" |
|
}; |
}; |
|
|
const char *const __mdoc_argnames[MDOC_ARG_MAX] = { |
const char *const __mdoc_argnames[MDOC_ARG_MAX] = { |
"split", "nosplit", "ragged", |
"split", "nosplit", "ragged", |
"unfilled", "literal", "file", |
"unfilled", "literal", "file", |
"offset", "bullet", "dash", |
"offset", "bullet", "dash", |
"hyphen", "item", "enum", |
"hyphen", "item", "enum", |
"tag", "diag", "hang", |
"tag", "diag", "hang", |
"ohang", "inset", "column", |
"ohang", "inset", "column", |
"width", "compact", "std", |
"width", "compact", "std", |
"filled", "words", "emphasis", |
"filled", "words", "emphasis", |
"symbolic", "nested" |
"symbolic", "nested", "centered" |
}; |
}; |
|
|
const char * const *mdoc_macronames = __mdoc_macronames; |
const char * const *mdoc_macronames = __mdoc_macronames; |
const char * const *mdoc_argnames = __mdoc_argnames; |
const char * const *mdoc_argnames = __mdoc_argnames; |
|
|
|
static void mdoc_node_free(struct mdoc_node *); |
|
static void mdoc_node_unlink(struct mdoc *, |
|
struct mdoc_node *); |
static void mdoc_free1(struct mdoc *); |
static void mdoc_free1(struct mdoc *); |
static int mdoc_alloc1(struct mdoc *); |
static void mdoc_alloc1(struct mdoc *); |
static struct mdoc_node *node_alloc(struct mdoc *, int, int, |
static struct mdoc_node *node_alloc(struct mdoc *, int, int, |
int, enum mdoc_type); |
enum mdoct, enum mdoc_type); |
static int node_append(struct mdoc *, |
static int node_append(struct mdoc *, |
struct mdoc_node *); |
struct mdoc_node *); |
static int parsetext(struct mdoc *, int, char *); |
static int mdoc_ptext(struct mdoc *, int, char *, int); |
static int parsemacro(struct mdoc *, int, char *); |
static int mdoc_pmacro(struct mdoc *, int, char *, int); |
static int macrowarn(struct mdoc *, int, const char *); |
|
static int perr(struct mdoc *, int, int, enum merr); |
|
|
|
#define verr(m, t) perr((m), (m)->last->line, (m)->last->pos, (t)) |
|
|
|
/* |
|
* Get the first (root) node of the parse tree. |
|
*/ |
|
const struct mdoc_node * |
const struct mdoc_node * |
mdoc_node(const struct mdoc *m) |
mdoc_node(const struct mdoc *mdoc) |
{ |
{ |
|
|
return(MDOC_HALT & m->flags ? NULL : m->first); |
return(mdoc->first); |
} |
} |
|
|
|
|
const struct mdoc_meta * |
const struct mdoc_meta * |
mdoc_meta(const struct mdoc *m) |
mdoc_meta(const struct mdoc *mdoc) |
{ |
{ |
|
|
return(MDOC_HALT & m->flags ? NULL : &m->meta); |
return(&mdoc->meta); |
} |
} |
|
|
|
/* |
|
* Frees volatile resources (parse tree, meta-data, fields). |
|
*/ |
static void |
static void |
mdoc_free1(struct mdoc *mdoc) |
mdoc_free1(struct mdoc *mdoc) |
{ |
{ |
|
|
if (mdoc->first) |
if (mdoc->first) |
mdoc_node_freelist(mdoc->first); |
mdoc_node_delete(mdoc, mdoc->first); |
if (mdoc->meta.title) |
free(mdoc->meta.msec); |
free(mdoc->meta.title); |
free(mdoc->meta.vol); |
if (mdoc->meta.os) |
free(mdoc->meta.arch); |
free(mdoc->meta.os); |
free(mdoc->meta.date); |
if (mdoc->meta.name) |
free(mdoc->meta.title); |
free(mdoc->meta.name); |
free(mdoc->meta.os); |
if (mdoc->meta.arch) |
free(mdoc->meta.name); |
free(mdoc->meta.arch); |
|
if (mdoc->meta.vol) |
|
free(mdoc->meta.vol); |
|
} |
} |
|
|
|
/* |
static int |
* Allocate all volatile resources (parse tree, meta-data, fields). |
|
*/ |
|
static void |
mdoc_alloc1(struct mdoc *mdoc) |
mdoc_alloc1(struct mdoc *mdoc) |
{ |
{ |
|
|
bzero(&mdoc->meta, sizeof(struct mdoc_meta)); |
memset(&mdoc->meta, 0, sizeof(struct mdoc_meta)); |
mdoc->flags = 0; |
mdoc->flags = 0; |
mdoc->lastnamed = mdoc->lastsec = 0; |
mdoc->lastnamed = mdoc->lastsec = SEC_NONE; |
mdoc->last = calloc(1, sizeof(struct mdoc_node)); |
mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node)); |
if (NULL == mdoc->last) |
|
return(0); |
|
|
|
mdoc->first = mdoc->last; |
mdoc->first = mdoc->last; |
mdoc->last->type = MDOC_ROOT; |
mdoc->last->type = MDOC_ROOT; |
|
mdoc->last->tok = MDOC_MAX; |
mdoc->next = MDOC_NEXT_CHILD; |
mdoc->next = MDOC_NEXT_CHILD; |
return(1); |
|
} |
} |
|
|
|
|
/* |
/* |
* Free up all resources contributed by a parse: the node tree, |
* Free up volatile resources (see mdoc_free1()) then re-initialises the |
* meta-data and so on. Then reallocate the root node for another |
* data with mdoc_alloc1(). After invocation, parse data has been reset |
* parse. |
* and the parser is ready for re-invocation on a new tree; however, |
|
* cross-parse non-volatile data is kept intact. |
*/ |
*/ |
int |
void |
mdoc_reset(struct mdoc *mdoc) |
mdoc_reset(struct mdoc *mdoc) |
{ |
{ |
|
|
mdoc_free1(mdoc); |
mdoc_free1(mdoc); |
return(mdoc_alloc1(mdoc)); |
mdoc_alloc1(mdoc); |
} |
} |
|
|
|
|
/* |
/* |
* Completely free up all resources. |
* Completely free up all volatile and non-volatile parse resources. |
|
* After invocation, the pointer is no longer usable. |
*/ |
*/ |
void |
void |
mdoc_free(struct mdoc *mdoc) |
mdoc_free(struct mdoc *mdoc) |
{ |
{ |
|
|
mdoc_free1(mdoc); |
mdoc_free1(mdoc); |
if (mdoc->htab) |
|
mdoc_hash_free(mdoc->htab); |
|
free(mdoc); |
free(mdoc); |
} |
} |
|
|
|
/* |
|
* Allocate volatile and non-volatile parse resources. |
|
*/ |
struct mdoc * |
struct mdoc * |
mdoc_alloc(void *data, int pflags, const struct mdoc_cb *cb) |
mdoc_alloc(struct roff *roff, struct mparse *parse, |
|
const char *defos, int quick) |
{ |
{ |
struct mdoc *p; |
struct mdoc *p; |
|
|
if (NULL == (p = calloc(1, sizeof(struct mdoc)))) |
p = mandoc_calloc(1, sizeof(struct mdoc)); |
return(NULL); |
|
if (cb) |
|
(void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb)); |
|
|
|
p->data = data; |
p->parse = parse; |
p->pflags = pflags; |
p->defos = defos; |
|
p->quick = quick; |
|
p->roff = roff; |
|
|
if (NULL == (p->htab = mdoc_hash_alloc())) { |
mdoc_hash_init(); |
free(p); |
mdoc_alloc1(p); |
return(NULL); |
return(p); |
} else if (mdoc_alloc1(p)) |
|
return(p); |
|
|
|
free(p); |
|
return(NULL); |
|
} |
} |
|
|
|
|
/* |
|
* Climb back up the parse tree, validating open scopes. Mostly calls |
|
* through to macro_end in macro.c. |
|
*/ |
|
int |
int |
mdoc_endparse(struct mdoc *m) |
mdoc_endparse(struct mdoc *mdoc) |
{ |
{ |
|
|
if (MDOC_HALT & m->flags) |
return(mdoc_macroend(mdoc)); |
return(0); |
|
else if (mdoc_macroend(m)) |
|
return(1); |
|
m->flags |= MDOC_HALT; |
|
return(0); |
|
} |
} |
|
|
|
|
/* |
|
* Main parse routine. Parses a single line -- really just hands off to |
|
* the macro or text parser. |
|
*/ |
|
int |
int |
mdoc_parseln(struct mdoc *m, int ln, char *buf) |
mdoc_addeqn(struct mdoc *mdoc, const struct eqn *ep) |
{ |
{ |
|
struct mdoc_node *n; |
|
|
/* If in error-mode, then we parse no more. */ |
n = node_alloc(mdoc, ep->ln, ep->pos, MDOC_MAX, MDOC_EQN); |
|
n->eqn = ep; |
|
if (ep->ln > mdoc->last->line) |
|
n->flags |= MDOC_LINE; |
|
|
if (MDOC_HALT & m->flags) |
if ( ! node_append(mdoc, n)) |
return(0); |
return(0); |
|
|
return('.' == *buf ? parsemacro(m, ln, buf) : |
mdoc->next = MDOC_NEXT_SIBLING; |
parsetext(m, ln, buf)); |
return(1); |
} |
} |
|
|
|
|
void |
|
mdoc_vmsg(struct mdoc *mdoc, int ln, int pos, const char *fmt, ...) |
|
{ |
|
char buf[256]; |
|
va_list ap; |
|
|
|
if (NULL == mdoc->cb.mdoc_msg) |
|
return; |
|
|
|
va_start(ap, fmt); |
|
(void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap); |
|
va_end(ap); |
|
(*mdoc->cb.mdoc_msg)(mdoc->data, ln, pos, buf); |
|
} |
|
|
|
|
|
int |
int |
mdoc_verr(struct mdoc *mdoc, int ln, int pos, |
mdoc_addspan(struct mdoc *mdoc, const struct tbl_span *sp) |
const char *fmt, ...) |
|
{ |
{ |
char buf[256]; |
struct mdoc_node *n; |
va_list ap; |
|
|
|
if (NULL == mdoc->cb.mdoc_err) |
n = node_alloc(mdoc, sp->line, 0, MDOC_MAX, MDOC_TBL); |
return(0); |
n->span = sp; |
|
|
va_start(ap, fmt); |
if ( ! node_append(mdoc, n)) |
(void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap); |
|
va_end(ap); |
|
return((*mdoc->cb.mdoc_err)(mdoc->data, ln, pos, buf)); |
|
} |
|
|
|
|
|
int |
|
mdoc_vwarn(struct mdoc *mdoc, int ln, int pos, |
|
enum mdoc_warn type, const char *fmt, ...) |
|
{ |
|
char buf[256]; |
|
va_list ap; |
|
|
|
if (NULL == mdoc->cb.mdoc_warn) |
|
return(0); |
return(0); |
|
|
va_start(ap, fmt); |
mdoc->next = MDOC_NEXT_SIBLING; |
(void)vsnprintf(buf, sizeof(buf) - 1, fmt, ap); |
return(1); |
va_end(ap); |
|
return((*mdoc->cb.mdoc_warn)(mdoc->data, ln, pos, type, buf)); |
|
} |
} |
|
|
|
/* |
|
* Main parse routine. Parses a single line -- really just hands off to |
|
* the macro (mdoc_pmacro()) or text parser (mdoc_ptext()). |
|
*/ |
int |
int |
mdoc_macro(struct mdoc *m, int tok, |
mdoc_parseln(struct mdoc *mdoc, int ln, char *buf, int offs) |
int ln, int pp, int *pos, char *buf) |
|
{ |
{ |
|
|
/* FIXME - these should happen during validation. */ |
if (mdoc->last->type != MDOC_EQN || ln > mdoc->last->line) |
|
mdoc->flags |= MDOC_NEWLINE; |
|
|
if (MDOC_PROLOGUE & mdoc_macros[tok].flags && |
/* |
SEC_PROLOGUE != m->lastnamed) |
* Let the roff nS register switch SYNOPSIS mode early, |
return(perr(m, ln, pp, EPROLBODY)); |
* such that the parser knows at all times |
|
* whether this mode is on or off. |
|
* Note that this mode is also switched by the Sh macro. |
|
*/ |
|
if (roff_getreg(mdoc->roff, "nS")) |
|
mdoc->flags |= MDOC_SYNOPSIS; |
|
else |
|
mdoc->flags &= ~MDOC_SYNOPSIS; |
|
|
if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && |
return(roff_getcontrol(mdoc->roff, buf, &offs) ? |
SEC_PROLOGUE == m->lastnamed) |
mdoc_pmacro(mdoc, ln, buf, offs) : |
return(perr(m, ln, pp, EBODYPROL)); |
mdoc_ptext(mdoc, ln, buf, offs)); |
|
|
if (1 != pp && ! (MDOC_CALLABLE & mdoc_macros[tok].flags)) |
|
return(perr(m, ln, pp, ENOCALL)); |
|
|
|
return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf)); |
|
} |
} |
|
|
|
int |
static int |
mdoc_macro(MACRO_PROT_ARGS) |
perr(struct mdoc *m, int line, int pos, enum merr type) |
|
{ |
{ |
char *p; |
assert(tok < MDOC_MAX); |
|
|
p = NULL; |
if (mdoc->flags & MDOC_PBODY) { |
switch (type) { |
if (tok == MDOC_Dt) { |
case (ENOCALL): |
mandoc_vmsg(MANDOCERR_DT_LATE, |
p = "not callable"; |
mdoc->parse, line, ppos, |
break; |
"Dt %s", buf + *pos); |
case (EPROLBODY): |
return(1); |
p = "macro disallowed in document body"; |
} |
break; |
} else if ( ! (mdoc_macros[tok].flags & MDOC_PROLOGUE)) { |
case (EBODYPROL): |
if (mdoc->meta.title == NULL) { |
p = "macro disallowed in document prologue"; |
mandoc_vmsg(MANDOCERR_DT_NOTITLE, |
break; |
mdoc->parse, line, ppos, "%s %s", |
case (EMALLOC): |
mdoc_macronames[tok], buf + *pos); |
p = "memory exhausted"; |
mdoc->meta.title = mandoc_strdup("UNTITLED"); |
break; |
} |
case (ETEXTPROL): |
if (NULL == mdoc->meta.vol) |
p = "text disallowed in document prologue"; |
mdoc->meta.vol = mandoc_strdup("LOCAL"); |
break; |
mdoc->flags |= MDOC_PBODY; |
case (ENOBLANK): |
|
p = "blank lines disallowed in non-literal contexts"; |
|
break; |
|
case (ESPACE): |
|
p = "whitespace disallowed after delimiter"; |
|
break; |
|
} |
} |
assert(p); |
|
return(mdoc_perr(m, line, pos, p)); |
return((*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf)); |
} |
} |
|
|
|
|
Line 357 node_append(struct mdoc *mdoc, struct mdoc_node *p) |
|
Line 294 node_append(struct mdoc *mdoc, struct mdoc_node *p) |
|
assert(MDOC_ROOT != p->type); |
assert(MDOC_ROOT != p->type); |
|
|
switch (mdoc->next) { |
switch (mdoc->next) { |
case (MDOC_NEXT_SIBLING): |
case MDOC_NEXT_SIBLING: |
mdoc->last->next = p; |
mdoc->last->next = p; |
p->prev = mdoc->last; |
p->prev = mdoc->last; |
p->parent = mdoc->last->parent; |
p->parent = mdoc->last->parent; |
break; |
break; |
case (MDOC_NEXT_CHILD): |
case MDOC_NEXT_CHILD: |
mdoc->last->child = p; |
mdoc->last->child = p; |
p->parent = mdoc->last; |
p->parent = mdoc->last; |
break; |
break; |
Line 371 node_append(struct mdoc *mdoc, struct mdoc_node *p) |
|
Line 308 node_append(struct mdoc *mdoc, struct mdoc_node *p) |
|
/* NOTREACHED */ |
/* NOTREACHED */ |
} |
} |
|
|
|
p->parent->nchild++; |
|
|
|
/* |
|
* Copy over the normalised-data pointer of our parent. Not |
|
* everybody has one, but copying a null pointer is fine. |
|
*/ |
|
|
|
switch (p->type) { |
|
case MDOC_BODY: |
|
if (ENDBODY_NOT != p->end) |
|
break; |
|
/* FALLTHROUGH */ |
|
case MDOC_TAIL: |
|
/* FALLTHROUGH */ |
|
case MDOC_HEAD: |
|
p->norm = p->parent->norm; |
|
break; |
|
default: |
|
break; |
|
} |
|
|
if ( ! mdoc_valid_pre(mdoc, p)) |
if ( ! mdoc_valid_pre(mdoc, p)) |
return(0); |
return(0); |
if ( ! mdoc_action_pre(mdoc, p)) |
|
return(0); |
|
|
|
switch (p->type) { |
switch (p->type) { |
case (MDOC_HEAD): |
case MDOC_HEAD: |
assert(MDOC_BLOCK == p->parent->type); |
assert(MDOC_BLOCK == p->parent->type); |
p->parent->head = p; |
p->parent->head = p; |
break; |
break; |
case (MDOC_TAIL): |
case MDOC_TAIL: |
assert(MDOC_BLOCK == p->parent->type); |
assert(MDOC_BLOCK == p->parent->type); |
p->parent->tail = p; |
p->parent->tail = p; |
break; |
break; |
case (MDOC_BODY): |
case MDOC_BODY: |
|
if (p->end) |
|
break; |
assert(MDOC_BLOCK == p->parent->type); |
assert(MDOC_BLOCK == p->parent->type); |
p->parent->body = p; |
p->parent->body = p; |
break; |
break; |
Line 396 node_append(struct mdoc *mdoc, struct mdoc_node *p) |
|
Line 354 node_append(struct mdoc *mdoc, struct mdoc_node *p) |
|
mdoc->last = p; |
mdoc->last = p; |
|
|
switch (p->type) { |
switch (p->type) { |
case (MDOC_TEXT): |
case MDOC_TBL: |
|
/* FALLTHROUGH */ |
|
case MDOC_TEXT: |
if ( ! mdoc_valid_post(mdoc)) |
if ( ! mdoc_valid_post(mdoc)) |
return(0); |
return(0); |
if ( ! mdoc_action_post(mdoc)) |
|
return(0); |
|
break; |
break; |
default: |
default: |
break; |
break; |
Line 409 node_append(struct mdoc *mdoc, struct mdoc_node *p) |
|
Line 367 node_append(struct mdoc *mdoc, struct mdoc_node *p) |
|
return(1); |
return(1); |
} |
} |
|
|
|
|
static struct mdoc_node * |
static struct mdoc_node * |
node_alloc(struct mdoc *mdoc, int line, |
node_alloc(struct mdoc *mdoc, int line, int pos, |
int pos, int tok, enum mdoc_type type) |
enum mdoct tok, enum mdoc_type type) |
{ |
{ |
struct mdoc_node *p; |
struct mdoc_node *p; |
|
|
if (NULL == (p = calloc(1, sizeof(struct mdoc_node)))) { |
p = mandoc_calloc(1, sizeof(struct mdoc_node)); |
(void)verr(mdoc, EMALLOC); |
|
return(NULL); |
|
} |
|
|
|
p->sec = mdoc->lastsec; |
p->sec = mdoc->lastsec; |
p->line = line; |
p->line = line; |
p->pos = pos; |
p->pos = pos; |
|
p->lastline = line; |
p->tok = tok; |
p->tok = tok; |
if (MDOC_TEXT != (p->type = type)) |
p->type = type; |
assert(p->tok >= 0); |
|
|
|
|
/* Flag analysis. */ |
|
|
|
if (MDOC_SYNOPSIS & mdoc->flags) |
|
p->flags |= MDOC_SYNPRETTY; |
|
else |
|
p->flags &= ~MDOC_SYNPRETTY; |
|
if (MDOC_NEWLINE & mdoc->flags) |
|
p->flags |= MDOC_LINE; |
|
mdoc->flags &= ~MDOC_NEWLINE; |
|
|
return(p); |
return(p); |
} |
} |
|
|
|
|
int |
int |
mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, int tok) |
mdoc_tail_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok) |
{ |
{ |
struct mdoc_node *p; |
struct mdoc_node *p; |
|
|
p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL); |
p = node_alloc(mdoc, line, pos, tok, MDOC_TAIL); |
if (NULL == p) |
if ( ! node_append(mdoc, p)) |
return(0); |
return(0); |
return(node_append(mdoc, p)); |
mdoc->next = MDOC_NEXT_CHILD; |
|
return(1); |
} |
} |
|
|
|
|
int |
int |
mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, int tok) |
mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok) |
{ |
{ |
struct mdoc_node *p; |
struct mdoc_node *p; |
|
|
Line 453 mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, |
|
Line 415 mdoc_head_alloc(struct mdoc *mdoc, int line, int pos, |
|
assert(mdoc->last); |
assert(mdoc->last); |
|
|
p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD); |
p = node_alloc(mdoc, line, pos, tok, MDOC_HEAD); |
if (NULL == p) |
if ( ! node_append(mdoc, p)) |
return(0); |
return(0); |
return(node_append(mdoc, p)); |
mdoc->next = MDOC_NEXT_CHILD; |
|
return(1); |
} |
} |
|
|
|
|
int |
int |
mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, int tok) |
mdoc_body_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok) |
{ |
{ |
struct mdoc_node *p; |
struct mdoc_node *p; |
|
|
p = node_alloc(mdoc, line, pos, tok, MDOC_BODY); |
p = node_alloc(mdoc, line, pos, tok, MDOC_BODY); |
if (NULL == p) |
if ( ! node_append(mdoc, p)) |
return(0); |
return(0); |
return(node_append(mdoc, p)); |
mdoc->next = MDOC_NEXT_CHILD; |
|
return(1); |
} |
} |
|
|
|
int |
|
mdoc_endbody_alloc(struct mdoc *mdoc, int line, int pos, enum mdoct tok, |
|
struct mdoc_node *body, enum mdoc_endbody end) |
|
{ |
|
struct mdoc_node *p; |
|
|
|
p = node_alloc(mdoc, line, pos, tok, MDOC_BODY); |
|
p->pending = body; |
|
p->norm = body->norm; |
|
p->end = end; |
|
if ( ! node_append(mdoc, p)) |
|
return(0); |
|
mdoc->next = MDOC_NEXT_SIBLING; |
|
return(1); |
|
} |
|
|
int |
int |
mdoc_block_alloc(struct mdoc *mdoc, int line, int pos, |
mdoc_block_alloc(struct mdoc *mdoc, int line, int pos, |
int tok, struct mdoc_arg *args) |
enum mdoct tok, struct mdoc_arg *args) |
{ |
{ |
struct mdoc_node *p; |
struct mdoc_node *p; |
|
|
p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK); |
p = node_alloc(mdoc, line, pos, tok, MDOC_BLOCK); |
if (NULL == p) |
p->args = args; |
return(0); |
if (p->args) |
if ((p->args = args)) |
|
(args->refcnt)++; |
(args->refcnt)++; |
return(node_append(mdoc, p)); |
|
} |
|
|
|
|
switch (tok) { |
|
case MDOC_Bd: |
|
/* FALLTHROUGH */ |
|
case MDOC_Bf: |
|
/* FALLTHROUGH */ |
|
case MDOC_Bl: |
|
/* FALLTHROUGH */ |
|
case MDOC_En: |
|
/* FALLTHROUGH */ |
|
case MDOC_Rs: |
|
p->norm = mandoc_calloc(1, sizeof(union mdoc_data)); |
|
break; |
|
default: |
|
break; |
|
} |
|
|
|
if ( ! node_append(mdoc, p)) |
|
return(0); |
|
mdoc->next = MDOC_NEXT_CHILD; |
|
return(1); |
|
} |
|
|
int |
int |
mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos, |
mdoc_elem_alloc(struct mdoc *mdoc, int line, int pos, |
int tok, struct mdoc_arg *args) |
enum mdoct tok, struct mdoc_arg *args) |
{ |
{ |
struct mdoc_node *p; |
struct mdoc_node *p; |
|
|
p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM); |
p = node_alloc(mdoc, line, pos, tok, MDOC_ELEM); |
if (NULL == p) |
p->args = args; |
return(0); |
if (p->args) |
if ((p->args = args)) |
|
(args->refcnt)++; |
(args->refcnt)++; |
return(node_append(mdoc, p)); |
|
} |
|
|
|
|
switch (tok) { |
|
case MDOC_An: |
|
p->norm = mandoc_calloc(1, sizeof(union mdoc_data)); |
|
break; |
|
default: |
|
break; |
|
} |
|
|
|
if ( ! node_append(mdoc, p)) |
|
return(0); |
|
mdoc->next = MDOC_NEXT_CHILD; |
|
return(1); |
|
} |
|
|
int |
int |
mdoc_word_alloc(struct mdoc *mdoc, |
mdoc_word_alloc(struct mdoc *mdoc, int line, int pos, const char *p) |
int line, int pos, const char *word) |
|
{ |
{ |
struct mdoc_node *p; |
struct mdoc_node *n; |
|
|
p = node_alloc(mdoc, line, pos, -1, MDOC_TEXT); |
n = node_alloc(mdoc, line, pos, MDOC_MAX, MDOC_TEXT); |
if (NULL == p) |
n->string = roff_strdup(mdoc->roff, p); |
|
|
|
if ( ! node_append(mdoc, n)) |
return(0); |
return(0); |
if (NULL == (p->string = strdup(word))) { |
|
(void)verr(mdoc, EMALLOC); |
mdoc->next = MDOC_NEXT_SIBLING; |
return(0); |
return(1); |
} |
|
return(node_append(mdoc, p)); |
|
} |
} |
|
|
|
|
void |
void |
|
mdoc_word_append(struct mdoc *mdoc, const char *p) |
|
{ |
|
struct mdoc_node *n; |
|
char *addstr, *newstr; |
|
|
|
n = mdoc->last; |
|
addstr = roff_strdup(mdoc->roff, p); |
|
mandoc_asprintf(&newstr, "%s %s", n->string, addstr); |
|
free(addstr); |
|
free(n->string); |
|
n->string = newstr; |
|
mdoc->next = MDOC_NEXT_SIBLING; |
|
} |
|
|
|
static void |
mdoc_node_free(struct mdoc_node *p) |
mdoc_node_free(struct mdoc_node *p) |
{ |
{ |
|
|
|
if (MDOC_BLOCK == p->type || MDOC_ELEM == p->type) |
|
free(p->norm); |
if (p->string) |
if (p->string) |
free(p->string); |
free(p->string); |
if (p->args) |
if (p->args) |
Line 529 mdoc_node_free(struct mdoc_node *p) |
|
Line 550 mdoc_node_free(struct mdoc_node *p) |
|
free(p); |
free(p); |
} |
} |
|
|
|
static void |
|
mdoc_node_unlink(struct mdoc *mdoc, struct mdoc_node *n) |
|
{ |
|
|
|
/* Adjust siblings. */ |
|
|
|
if (n->prev) |
|
n->prev->next = n->next; |
|
if (n->next) |
|
n->next->prev = n->prev; |
|
|
|
/* Adjust parent. */ |
|
|
|
if (n->parent) { |
|
n->parent->nchild--; |
|
if (n->parent->child == n) |
|
n->parent->child = n->prev ? n->prev : n->next; |
|
if (n->parent->last == n) |
|
n->parent->last = n->prev ? n->prev : NULL; |
|
} |
|
|
|
/* Adjust parse point, if applicable. */ |
|
|
|
if (mdoc && mdoc->last == n) { |
|
if (n->prev) { |
|
mdoc->last = n->prev; |
|
mdoc->next = MDOC_NEXT_SIBLING; |
|
} else { |
|
mdoc->last = n->parent; |
|
mdoc->next = MDOC_NEXT_CHILD; |
|
} |
|
} |
|
|
|
if (mdoc && mdoc->first == n) |
|
mdoc->first = NULL; |
|
} |
|
|
void |
void |
mdoc_node_freelist(struct mdoc_node *p) |
mdoc_node_delete(struct mdoc *mdoc, struct mdoc_node *p) |
{ |
{ |
|
|
if (p->child) |
while (p->child) { |
mdoc_node_freelist(p->child); |
assert(p->nchild); |
if (p->next) |
mdoc_node_delete(mdoc, p->child); |
mdoc_node_freelist(p->next); |
} |
|
assert(0 == p->nchild); |
|
|
|
mdoc_node_unlink(mdoc, p); |
mdoc_node_free(p); |
mdoc_node_free(p); |
} |
} |
|
|
|
int |
|
mdoc_node_relink(struct mdoc *mdoc, struct mdoc_node *p) |
|
{ |
|
|
|
mdoc_node_unlink(mdoc, p); |
|
return(node_append(mdoc, p)); |
|
} |
|
|
/* |
/* |
* Parse free-form text, that is, a line that does not begin with the |
* Parse free-form text, that is, a line that does not begin with the |
* control character. |
* control character. |
*/ |
*/ |
static int |
static int |
parsetext(struct mdoc *m, int line, char *buf) |
mdoc_ptext(struct mdoc *mdoc, int line, char *buf, int offs) |
{ |
{ |
|
char *c, *ws, *end; |
|
struct mdoc_node *n; |
|
|
if (SEC_PROLOGUE == m->lastnamed) |
assert(mdoc->last); |
return(perr(m, line, 0, ETEXTPROL)); |
n = mdoc->last; |
|
|
if (0 == buf[0] && ! (MDOC_LITERAL & m->flags)) |
/* |
return(perr(m, line, 0, ENOBLANK)); |
* Divert directly to list processing if we're encountering a |
|
* columnar MDOC_BLOCK with or without a prior MDOC_BLOCK entry |
|
* (a MDOC_BODY means it's already open, in which case we should |
|
* process within its context in the normal way). |
|
*/ |
|
|
if ( ! mdoc_word_alloc(m, line, 0, buf)) |
if (MDOC_Bl == n->tok && MDOC_BODY == n->type && |
|
LIST_column == n->norm->Bl.type) { |
|
/* `Bl' is open without any children. */ |
|
mdoc->flags |= MDOC_FREECOL; |
|
return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf)); |
|
} |
|
|
|
if (MDOC_It == n->tok && MDOC_BLOCK == n->type && |
|
NULL != n->parent && |
|
MDOC_Bl == n->parent->tok && |
|
LIST_column == n->parent->norm->Bl.type) { |
|
/* `Bl' has block-level `It' children. */ |
|
mdoc->flags |= MDOC_FREECOL; |
|
return(mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf)); |
|
} |
|
|
|
/* |
|
* Search for the beginning of unescaped trailing whitespace (ws) |
|
* and for the first character not to be output (end). |
|
*/ |
|
|
|
/* FIXME: replace with strcspn(). */ |
|
ws = NULL; |
|
for (c = end = buf + offs; *c; c++) { |
|
switch (*c) { |
|
case ' ': |
|
if (NULL == ws) |
|
ws = c; |
|
continue; |
|
case '\t': |
|
/* |
|
* Always warn about trailing tabs, |
|
* even outside literal context, |
|
* where they should be put on the next line. |
|
*/ |
|
if (NULL == ws) |
|
ws = c; |
|
/* |
|
* Strip trailing tabs in literal context only; |
|
* outside, they affect the next line. |
|
*/ |
|
if (MDOC_LITERAL & mdoc->flags) |
|
continue; |
|
break; |
|
case '\\': |
|
/* Skip the escaped character, too, if any. */ |
|
if (c[1]) |
|
c++; |
|
/* FALLTHROUGH */ |
|
default: |
|
ws = NULL; |
|
break; |
|
} |
|
end = c + 1; |
|
} |
|
*end = '\0'; |
|
|
|
if (ws) |
|
mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse, |
|
line, (int)(ws-buf), NULL); |
|
|
|
if ('\0' == buf[offs] && ! (MDOC_LITERAL & mdoc->flags)) { |
|
mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse, |
|
line, (int)(c - buf), NULL); |
|
|
|
/* |
|
* Insert a `sp' in the case of a blank line. Technically, |
|
* blank lines aren't allowed, but enough manuals assume this |
|
* behaviour that we want to work around it. |
|
*/ |
|
if ( ! mdoc_elem_alloc(mdoc, line, offs, MDOC_sp, NULL)) |
|
return(0); |
|
|
|
mdoc->next = MDOC_NEXT_SIBLING; |
|
|
|
return(mdoc_valid_post(mdoc)); |
|
} |
|
|
|
if ( ! mdoc_word_alloc(mdoc, line, offs, buf+offs)) |
return(0); |
return(0); |
|
|
m->next = MDOC_NEXT_SIBLING; |
if (MDOC_LITERAL & mdoc->flags) |
return(1); |
return(1); |
} |
|
|
|
|
/* |
|
* End-of-sentence check. If the last character is an unescaped |
|
* EOS character, then flag the node as being the end of a |
|
* sentence. The front-end will know how to interpret this. |
|
*/ |
|
|
static int |
assert(buf < end); |
macrowarn(struct mdoc *m, int ln, const char *buf) |
|
{ |
|
if ( ! (MDOC_IGN_MACRO & m->pflags)) |
|
return(mdoc_perr(m, ln, 1, |
|
"unknown macro: %s%s", |
|
buf, strlen(buf) > 3 ? "..." : "")); |
|
return(mdoc_pwarn(m, ln, 1, WARN_SYNTAX, |
|
"unknown macro: %s%s", |
|
buf, strlen(buf) > 3 ? "..." : "")); |
|
} |
|
|
|
|
if (mandoc_eos(buf+offs, (size_t)(end-buf-offs))) |
|
mdoc->last->flags |= MDOC_EOS; |
|
|
|
return(1); |
|
} |
|
|
/* |
/* |
* Parse a macro line, that is, a line beginning with the control |
* Parse a macro line, that is, a line beginning with the control |
* character. |
* character. |
*/ |
*/ |
int |
static int |
parsemacro(struct mdoc *m, int ln, char *buf) |
mdoc_pmacro(struct mdoc *mdoc, int ln, char *buf, int offs) |
{ |
{ |
int i, c; |
struct mdoc_node *n; |
|
const char *cp; |
|
enum mdoct tok; |
|
int i, sv; |
char mac[5]; |
char mac[5]; |
|
|
/* Comments and empties are quickly ignored. */ |
sv = offs; |
|
|
if (0 == buf[1]) |
/* |
|
* Copy the first word into a nil-terminated buffer. |
|
* Stop when a space, tab, escape, or eoln is encountered. |
|
*/ |
|
|
|
i = 0; |
|
while (i < 4 && strchr(" \t\\", buf[offs]) == NULL) |
|
mac[i++] = buf[offs++]; |
|
|
|
mac[i] = '\0'; |
|
|
|
tok = (i > 1 && i < 4) ? mdoc_hash_find(mac) : MDOC_MAX; |
|
|
|
if (tok == MDOC_MAX) { |
|
mandoc_msg(MANDOCERR_MACRO, mdoc->parse, |
|
ln, sv, buf + sv - 1); |
return(1); |
return(1); |
|
} |
|
|
if (' ' == buf[1]) { |
/* Skip a leading escape sequence or tab. */ |
i = 2; |
|
while (buf[i] && ' ' == buf[i]) |
switch (buf[offs]) { |
i++; |
case '\\': |
if (0 == buf[i]) |
cp = buf + offs + 1; |
return(1); |
mandoc_escape(&cp, NULL, NULL); |
return(perr(m, ln, 1, ESPACE)); |
offs = cp - buf; |
|
break; |
|
case '\t': |
|
offs++; |
|
break; |
|
default: |
|
break; |
} |
} |
|
|
if (buf[1] && '\\' == buf[1]) |
/* Jump to the next non-whitespace word. */ |
if (buf[2] && '\"' == buf[2]) |
|
return(1); |
|
|
|
/* Copy the first word into a nil-terminated buffer. */ |
while (buf[offs] && ' ' == buf[offs]) |
|
offs++; |
|
|
for (i = 1; i < 5; i++) { |
/* |
if (0 == (mac[i - 1] = buf[i])) |
* Trailing whitespace. Note that tabs are allowed to be passed |
break; |
* into the parser as "text", so we only warn about spaces here. |
else if (' ' == buf[i]) |
*/ |
break; |
|
|
if ('\0' == buf[offs] && ' ' == buf[offs - 1]) |
|
mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse, |
|
ln, offs - 1, NULL); |
|
|
|
/* |
|
* If an initial macro or a list invocation, divert directly |
|
* into macro processing. |
|
*/ |
|
|
|
if (NULL == mdoc->last || MDOC_It == tok || MDOC_El == tok) |
|
return(mdoc_macro(mdoc, tok, ln, sv, &offs, buf)); |
|
|
|
n = mdoc->last; |
|
assert(mdoc->last); |
|
|
|
/* |
|
* If the first macro of a `Bl -column', open an `It' block |
|
* context around the parsed macro. |
|
*/ |
|
|
|
if (MDOC_Bl == n->tok && MDOC_BODY == n->type && |
|
LIST_column == n->norm->Bl.type) { |
|
mdoc->flags |= MDOC_FREECOL; |
|
return(mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf)); |
} |
} |
|
|
mac[i - 1] = 0; |
/* |
|
* If we're following a block-level `It' within a `Bl -column' |
|
* context (perhaps opened in the above block or in ptext()), |
|
* then open an `It' block context around the parsed macro. |
|
*/ |
|
|
if (i == 5 || i <= 2) { |
if (MDOC_It == n->tok && MDOC_BLOCK == n->type && |
if ( ! macrowarn(m, ln, mac)) |
NULL != n->parent && |
goto err; |
MDOC_Bl == n->parent->tok && |
return(1); |
LIST_column == n->parent->norm->Bl.type) { |
} |
mdoc->flags |= MDOC_FREECOL; |
|
return(mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf)); |
if (MDOC_MAX == (c = mdoc_hash_find(m->htab, mac))) { |
|
if ( ! macrowarn(m, ln, mac)) |
|
goto err; |
|
return(1); |
|
} |
} |
|
|
/* The macro is sane. Jump to the next word. */ |
/* Normal processing of a macro. */ |
|
|
while (buf[i] && ' ' == buf[i]) |
if ( ! mdoc_macro(mdoc, tok, ln, sv, &offs, buf)) |
i++; |
return(0); |
|
|
/* Begin recursive parse sequence. */ |
/* In quick mode (for mandocdb), abort after the NAME section. */ |
|
|
if ( ! mdoc_macro(m, c, ln, 1, &i, buf)) |
if (mdoc->quick && MDOC_Sh == tok && |
goto err; |
SEC_NAME != mdoc->last->sec) |
|
return(2); |
|
|
return(1); |
return(1); |
|
} |
|
|
err: /* Error out. */ |
enum mdelim |
|
mdoc_isdelim(const char *p) |
|
{ |
|
|
m->flags |= MDOC_HALT; |
if ('\0' == p[0]) |
return(0); |
return(DELIM_NONE); |
|
|
|
if ('\0' == p[1]) |
|
switch (p[0]) { |
|
case '(': |
|
/* FALLTHROUGH */ |
|
case '[': |
|
return(DELIM_OPEN); |
|
case '|': |
|
return(DELIM_MIDDLE); |
|
case '.': |
|
/* FALLTHROUGH */ |
|
case ',': |
|
/* FALLTHROUGH */ |
|
case ';': |
|
/* FALLTHROUGH */ |
|
case ':': |
|
/* FALLTHROUGH */ |
|
case '?': |
|
/* FALLTHROUGH */ |
|
case '!': |
|
/* FALLTHROUGH */ |
|
case ')': |
|
/* FALLTHROUGH */ |
|
case ']': |
|
return(DELIM_CLOSE); |
|
default: |
|
return(DELIM_NONE); |
|
} |
|
|
|
if ('\\' != p[0]) |
|
return(DELIM_NONE); |
|
|
|
if (0 == strcmp(p + 1, ".")) |
|
return(DELIM_CLOSE); |
|
if (0 == strcmp(p + 1, "fR|\\fP")) |
|
return(DELIM_MIDDLE); |
|
|
|
return(DELIM_NONE); |
|
} |
|
|
|
void |
|
mdoc_deroff(char **dest, const struct mdoc_node *n) |
|
{ |
|
char *cp; |
|
size_t sz; |
|
|
|
if (MDOC_TEXT != n->type) { |
|
for (n = n->child; n; n = n->next) |
|
mdoc_deroff(dest, n); |
|
return; |
|
} |
|
|
|
/* Skip leading whitespace. */ |
|
|
|
for (cp = n->string; '\0' != *cp; cp++) |
|
if (0 == isspace((unsigned char)*cp)) |
|
break; |
|
|
|
/* Skip trailing whitespace. */ |
|
|
|
for (sz = strlen(cp); sz; sz--) |
|
if (0 == isspace((unsigned char)cp[sz-1])) |
|
break; |
|
|
|
/* Skip empty strings. */ |
|
|
|
if (0 == sz) |
|
return; |
|
|
|
if (NULL == *dest) { |
|
*dest = mandoc_strndup(cp, sz); |
|
return; |
|
} |
|
|
|
mandoc_asprintf(&cp, "%s %*s", *dest, (int)sz, cp); |
|
free(*dest); |
|
*dest = cp; |
} |
} |