version 1.149, 2016/07/10 13:34:30 |
version 1.163, 2017/03/07 20:00:08 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2010-2016 Ingo Schwarze <schwarze@openbsd.org> |
* Copyright (c) 2010-2017 Ingo Schwarze <schwarze@openbsd.org> |
* Copyright (c) 2010, 2012 Joerg Sonnenberger <joerg@netbsd.org> |
* Copyright (c) 2010, 2012 Joerg Sonnenberger <joerg@netbsd.org> |
* |
* |
* Permission to use, copy, modify, and distribute this software for any |
* Permission to use, copy, modify, and distribute this software for any |
|
|
#include "config.h" |
#include "config.h" |
|
|
#include <sys/types.h> |
#include <sys/types.h> |
#if HAVE_MMAP |
|
#include <sys/mman.h> |
#include <sys/mman.h> |
#include <sys/stat.h> |
#include <sys/stat.h> |
#endif |
|
|
|
#include <assert.h> |
#include <assert.h> |
#include <ctype.h> |
#include <ctype.h> |
|
|
#include <errno.h> |
#include <errno.h> |
#include <fcntl.h> |
#include <fcntl.h> |
#include <stdarg.h> |
#include <stdarg.h> |
#include <stdint.h> |
|
#include <stdio.h> |
#include <stdio.h> |
#include <stdlib.h> |
#include <stdlib.h> |
#include <string.h> |
#include <string.h> |
|
|
#define REPARSE_LIMIT 1000 |
#define REPARSE_LIMIT 1000 |
|
|
struct mparse { |
struct mparse { |
struct roff_man *man; /* man parser */ |
|
struct roff *roff; /* roff parser (!NULL) */ |
struct roff *roff; /* roff parser (!NULL) */ |
|
struct roff_man *man; /* man parser */ |
char *sodest; /* filename pointed to by .so */ |
char *sodest; /* filename pointed to by .so */ |
const char *file; /* filename of current input file */ |
const char *file; /* filename of current input file */ |
struct buf *primary; /* buffer currently being parsed */ |
struct buf *primary; /* buffer currently being parsed */ |
|
|
|
|
static void choose_parser(struct mparse *); |
static void choose_parser(struct mparse *); |
static void resize_buf(struct buf *, size_t); |
static void resize_buf(struct buf *, size_t); |
static void mparse_buf_r(struct mparse *, struct buf, size_t, int); |
static int mparse_buf_r(struct mparse *, struct buf, size_t, int); |
static int read_whole_file(struct mparse *, const char *, int, |
static int read_whole_file(struct mparse *, const char *, int, |
struct buf *, int *); |
struct buf *, int *); |
static void mparse_end(struct mparse *); |
static void mparse_end(struct mparse *); |
Line 110 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
Line 107 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
"no document body", |
"no document body", |
"content before first section header", |
"content before first section header", |
"first section is not \"NAME\"", |
"first section is not \"NAME\"", |
"NAME section without name", |
"NAME section without Nm before Nd", |
"NAME section without description", |
"NAME section without description", |
"description not at the end of NAME", |
"description not at the end of NAME", |
"bad NAME section content", |
"bad NAME section content", |
|
"missing comma before name", |
"missing description line, using \"\"", |
"missing description line, using \"\"", |
|
"description line outside NAME section", |
"sections out of conventional order", |
"sections out of conventional order", |
"duplicate section title", |
"duplicate section title", |
"unexpected section", |
"unexpected section", |
Line 143 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
Line 142 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
"empty argument, using 0n", |
"empty argument, using 0n", |
"missing display type, using -ragged", |
"missing display type, using -ragged", |
"list type is not the first argument", |
"list type is not the first argument", |
"missing -width in -tag list, using 8n", |
"missing -width in -tag list, using 6n", |
"missing utility name, using \"\"", |
"missing utility name, using \"\"", |
"missing function name, using \"\"", |
"missing function name, using \"\"", |
"empty head in list item", |
"empty head in list item", |
Line 152 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
Line 151 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
"unknown font type, using \\fR", |
"unknown font type, using \\fR", |
"nothing follows prefix", |
"nothing follows prefix", |
"empty reference block", |
"empty reference block", |
|
"missing section argument", |
"missing -std argument, adding it", |
"missing -std argument, adding it", |
"missing option string, using \"\"", |
"missing option string, using \"\"", |
"missing resource identifier, using \"\"", |
"missing resource identifier, using \"\"", |
Line 177 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
Line 177 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
"blank line in fill mode, using .sp", |
"blank line in fill mode, using .sp", |
"tab in filled text", |
"tab in filled text", |
"whitespace at end of input line", |
"whitespace at end of input line", |
|
"new sentence, new line", |
"bad comment style", |
"bad comment style", |
"invalid escape sequence", |
"invalid escape sequence", |
"undefined string, using \"\"", |
"undefined string, using \"\"", |
Line 291 choose_parser(struct mparse *curp) |
|
Line 292 choose_parser(struct mparse *curp) |
|
} |
} |
} |
} |
|
|
if (curp->man == NULL) { |
|
curp->man = roff_man_alloc(curp->roff, curp, curp->defos, |
|
curp->options & MPARSE_QUICK ? 1 : 0); |
|
curp->man->macroset = MACROSET_MAN; |
|
curp->man->first->tok = TOKEN_NONE; |
|
} |
|
|
|
if (format == MPARSE_MDOC) { |
if (format == MPARSE_MDOC) { |
mdoc_hash_init(); |
mdoc_hash_init(); |
curp->man->macroset = MACROSET_MDOC; |
curp->man->macroset = MACROSET_MDOC; |
Line 316 choose_parser(struct mparse *curp) |
|
Line 310 choose_parser(struct mparse *curp) |
|
* macros, inline equations, and input line traps) |
* macros, inline equations, and input line traps) |
* and indirectly (for .so file inclusion). |
* and indirectly (for .so file inclusion). |
*/ |
*/ |
static void |
static int |
mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start) |
mparse_buf_r(struct mparse *curp, struct buf blk, size_t i, int start) |
{ |
{ |
const struct tbl_span *span; |
const struct tbl_span *span; |
Line 324 mparse_buf_r(struct mparse *curp, struct buf blk, size |
|
Line 318 mparse_buf_r(struct mparse *curp, struct buf blk, size |
|
const char *save_file; |
const char *save_file; |
char *cp; |
char *cp; |
size_t pos; /* byte number in the ln buffer */ |
size_t pos; /* byte number in the ln buffer */ |
|
size_t j; /* auxiliary byte number in the blk buffer */ |
enum rofferr rr; |
enum rofferr rr; |
int of; |
int of; |
int lnn; /* line number in the real file */ |
int lnn; /* line number in the real file */ |
Line 429 mparse_buf_r(struct mparse *curp, struct buf blk, size |
|
Line 424 mparse_buf_r(struct mparse *curp, struct buf blk, size |
|
} |
} |
|
|
if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) { |
if ('"' == blk.buf[i + 1] || '#' == blk.buf[i + 1]) { |
|
j = i; |
i += 2; |
i += 2; |
/* Comment, skip to end of line */ |
/* Comment, skip to end of line */ |
for (; i < blk.sz; ++i) { |
for (; i < blk.sz; ++i) { |
if ('\n' == blk.buf[i]) { |
if (blk.buf[i] != '\n') |
++i; |
continue; |
++lnn; |
if (blk.buf[i - 1] == ' ' || |
break; |
blk.buf[i - 1] == '\t') |
} |
mandoc_msg( |
|
MANDOCERR_SPACE_EOL, |
|
curp, curp->line, |
|
pos + i-1 - j, NULL); |
|
++i; |
|
++lnn; |
|
break; |
} |
} |
|
|
/* Backout trailing whitespaces */ |
/* Backout trailing whitespaces */ |
|
|
|
|
switch (rr) { |
switch (rr) { |
case ROFF_REPARSE: |
case ROFF_REPARSE: |
if (REPARSE_LIMIT >= ++curp->reparse_count) |
if (++curp->reparse_count > REPARSE_LIMIT) |
mparse_buf_r(curp, ln, of, 0); |
|
else |
|
mandoc_msg(MANDOCERR_ROFFLOOP, curp, |
mandoc_msg(MANDOCERR_ROFFLOOP, curp, |
curp->line, pos, NULL); |
curp->line, pos, NULL); |
pos = 0; |
else if (mparse_buf_r(curp, ln, of, 0) == 1 || |
continue; |
start == 1) { |
|
pos = 0; |
|
continue; |
|
} |
|
free(ln.buf); |
|
return 0; |
case ROFF_APPEND: |
case ROFF_APPEND: |
pos = strlen(ln.buf); |
pos = strlen(ln.buf); |
continue; |
continue; |
|
|
(i >= blk.sz || blk.buf[i] == '\0')) { |
(i >= blk.sz || blk.buf[i] == '\0')) { |
curp->sodest = mandoc_strdup(ln.buf + of); |
curp->sodest = mandoc_strdup(ln.buf + of); |
free(ln.buf); |
free(ln.buf); |
return; |
return 1; |
} |
} |
/* |
/* |
* We remove `so' clauses from our lookaside |
* We remove `so' clauses from our lookaside |
|
|
break; |
break; |
} |
} |
|
|
/* |
if (curp->man->macroset == MACROSET_NONE) |
* If input parsers have not been allocated, do so now. |
|
* We keep these instanced between parsers, but set them |
|
* locally per parse routine since we can use different |
|
* parsers with each one. |
|
*/ |
|
|
|
if (curp->man == NULL || |
|
curp->man->macroset == MACROSET_NONE) |
|
choose_parser(curp); |
choose_parser(curp); |
|
|
/* |
/* |
|
|
} |
} |
|
|
free(ln.buf); |
free(ln.buf); |
|
return 1; |
} |
} |
|
|
static int |
static int |
read_whole_file(struct mparse *curp, const char *file, int fd, |
read_whole_file(struct mparse *curp, const char *file, int fd, |
struct buf *fb, int *with_mmap) |
struct buf *fb, int *with_mmap) |
{ |
{ |
|
struct stat st; |
gzFile gz; |
gzFile gz; |
size_t off; |
size_t off; |
ssize_t ssz; |
ssize_t ssz; |
|
|
#if HAVE_MMAP |
|
struct stat st; |
|
|
|
if (fstat(fd, &st) == -1) |
if (fstat(fd, &st) == -1) |
err((int)MANDOCLEVEL_SYSERR, "%s", file); |
err((int)MANDOCLEVEL_SYSERR, "%s", file); |
|
|
Line 637 read_whole_file(struct mparse *curp, const char *file, |
|
Line 633 read_whole_file(struct mparse *curp, const char *file, |
|
if (fb->buf != MAP_FAILED) |
if (fb->buf != MAP_FAILED) |
return 1; |
return 1; |
} |
} |
#endif |
|
|
|
if (curp->gzip) { |
if (curp->gzip) { |
if ((gz = gzdopen(fd, "rb")) == NULL) |
if ((gz = gzdopen(fd, "rb")) == NULL) |
Line 683 read_whole_file(struct mparse *curp, const char *file, |
|
Line 678 read_whole_file(struct mparse *curp, const char *file, |
|
static void |
static void |
mparse_end(struct mparse *curp) |
mparse_end(struct mparse *curp) |
{ |
{ |
|
|
if (curp->man == NULL && curp->sodest == NULL) |
|
curp->man = roff_man_alloc(curp->roff, curp, curp->defos, |
|
curp->options & MPARSE_QUICK ? 1 : 0); |
|
if (curp->man->macroset == MACROSET_NONE) |
if (curp->man->macroset == MACROSET_NONE) |
curp->man->macroset = MACROSET_MAN; |
curp->man->macroset = MACROSET_MAN; |
if (curp->man->macroset == MACROSET_MDOC) |
if (curp->man->macroset == MACROSET_MDOC) |
Line 766 mparse_readfd(struct mparse *curp, int fd, const char |
|
Line 757 mparse_readfd(struct mparse *curp, int fd, const char |
|
(MPARSE_UTF8 | MPARSE_LATIN1); |
(MPARSE_UTF8 | MPARSE_LATIN1); |
mparse_parse_buffer(curp, blk, file); |
mparse_parse_buffer(curp, blk, file); |
curp->filenc = save_filenc; |
curp->filenc = save_filenc; |
#if HAVE_MMAP |
|
if (with_mmap) |
if (with_mmap) |
munmap(blk.buf, blk.sz); |
munmap(blk.buf, blk.sz); |
else |
else |
#endif |
|
free(blk.buf); |
free(blk.buf); |
} |
} |
return curp->file_status; |
return curp->file_status; |
Line 842 mparse_alloc(int options, enum mandoclevel wlevel, man |
|
Line 831 mparse_alloc(int options, enum mandoclevel wlevel, man |
|
void |
void |
mparse_reset(struct mparse *curp) |
mparse_reset(struct mparse *curp) |
{ |
{ |
|
|
roff_reset(curp->roff); |
roff_reset(curp->roff); |
|
roff_man_reset(curp->man); |
|
|
if (curp->man != NULL) |
free(curp->sodest); |
roff_man_reset(curp->man); |
curp->sodest = NULL; |
|
|
if (curp->secondary) |
if (curp->secondary) |
curp->secondary->sz = 0; |
curp->secondary->sz = 0; |
|
|
curp->file_status = MANDOCLEVEL_OK; |
curp->file_status = MANDOCLEVEL_OK; |
|
curp->gzip = 0; |
free(curp->sodest); |
|
curp->sodest = NULL; |
|
} |
} |
|
|
void |
void |
Line 861 mparse_free(struct mparse *curp) |
|
Line 849 mparse_free(struct mparse *curp) |
|
{ |
{ |
|
|
roff_man_free(curp->man); |
roff_man_free(curp->man); |
if (curp->roff) |
roff_free(curp->roff); |
roff_free(curp->roff); |
|
if (curp->secondary) |
if (curp->secondary) |
free(curp->secondary->buf); |
free(curp->secondary->buf); |
|
|
Line 882 mparse_result(struct mparse *curp, struct roff_man **m |
|
Line 869 mparse_result(struct mparse *curp, struct roff_man **m |
|
} |
} |
if (man) |
if (man) |
*man = curp->man; |
*man = curp->man; |
|
} |
|
|
|
void |
|
mparse_updaterc(struct mparse *curp, enum mandoclevel *rc) |
|
{ |
|
if (curp->file_status > *rc) |
|
*rc = curp->file_status; |
} |
} |
|
|
void |
void |