version 1.174, 2017/06/10 01:48:53 |
version 1.196, 2018/07/28 18:34:15 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2010-2017 Ingo Schwarze <schwarze@openbsd.org> |
* Copyright (c) 2010-2018 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 <assert.h> |
#include <assert.h> |
#include <ctype.h> |
#include <ctype.h> |
#if HAVE_ERR |
|
#include <err.h> |
|
#endif |
|
#include <errno.h> |
#include <errno.h> |
#include <fcntl.h> |
#include <fcntl.h> |
#include <stdarg.h> |
#include <stdarg.h> |
|
|
#include "mdoc.h" |
#include "mdoc.h" |
#include "man.h" |
#include "man.h" |
#include "libmandoc.h" |
#include "libmandoc.h" |
#include "roff_int.h" |
|
|
|
#define REPARSE_LIMIT 1000 |
#define REPARSE_LIMIT 1000 |
|
|
|
|
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 */ |
struct buf *secondary; /* preprocessed copy of input */ |
struct buf *secondary; /* preprocessed copy of input */ |
const char *defos; /* default operating system */ |
const char *os_s; /* default operating system */ |
mandocmsg mmsg; /* warning/error message handler */ |
mandocmsg mmsg; /* warning/error message handler */ |
enum mandoclevel file_status; /* status of current parse */ |
enum mandoclevel file_status; /* status of current parse */ |
enum mandoclevel wlevel; /* ignore messages below this */ |
enum mandocerr mmin; /* ignore messages below this */ |
int options; /* parser options */ |
int options; /* parser options */ |
int gzip; /* current input file is gzipped */ |
int gzip; /* current input file is gzipped */ |
int filenc; /* encoding of the current file */ |
int filenc; /* encoding of the current file */ |
Line 75 static void mparse_parse_buffer(struct mparse *, str |
|
Line 71 static void mparse_parse_buffer(struct mparse *, str |
|
|
|
static const enum mandocerr mandoclimits[MANDOCLEVEL_MAX] = { |
static const enum mandocerr mandoclimits[MANDOCLEVEL_MAX] = { |
MANDOCERR_OK, |
MANDOCERR_OK, |
MANDOCERR_STYLE, |
MANDOCERR_OK, |
MANDOCERR_WARNING, |
MANDOCERR_WARNING, |
MANDOCERR_ERROR, |
MANDOCERR_ERROR, |
MANDOCERR_UNSUPP, |
MANDOCERR_UNSUPP, |
Line 86 static const enum mandocerr mandoclimits[MANDOCLEVEL_M |
|
Line 82 static const enum mandocerr mandoclimits[MANDOCLEVEL_M |
|
static const char * const mandocerrs[MANDOCERR_MAX] = { |
static const char * const mandocerrs[MANDOCERR_MAX] = { |
"ok", |
"ok", |
|
|
|
"base system convention", |
|
|
|
"Mdocdate found", |
|
"Mdocdate missing", |
|
"unknown architecture", |
|
"operating system explicitly specified", |
|
"RCS id missing", |
|
"referenced manual not found", |
|
|
"generic style suggestion", |
"generic style suggestion", |
|
|
|
"legacy man(7) date format", |
|
"normalizing date format to", |
|
"lower case character in document title", |
|
"duplicate RCS id", |
|
"possible typo in section name", |
|
"unterminated quoted argument", |
"useless macro", |
"useless macro", |
"consider using OS macro", |
"consider using OS macro", |
"errnos out of order", |
"errnos out of order", |
"duplicate errno", |
"duplicate errno", |
"description line ends with a full stop", |
"trailing delimiter", |
"no blank before trailing delimiter", |
"no blank before trailing delimiter", |
|
"fill mode already enabled, skipping", |
|
"fill mode already disabled, skipping", |
|
"verbatim \"--\", maybe consider using \\(em", |
|
"function name without markup", |
|
"whitespace at end of input line", |
|
"bad comment style", |
|
|
"generic warning", |
"generic warning", |
|
|
/* related to the prologue */ |
/* related to the prologue */ |
"missing manual title, using UNTITLED", |
"missing manual title, using UNTITLED", |
"missing manual title, using \"\"", |
"missing manual title, using \"\"", |
"lower case character in document title", |
|
"missing manual section, using \"\"", |
"missing manual section, using \"\"", |
"unknown manual section", |
"unknown manual section", |
"missing date, using today's date", |
"missing date, using today's date", |
"cannot parse date, using it verbatim", |
"cannot parse date, using it verbatim", |
|
"date in the future, using it anyway", |
"missing Os macro, using \"\"", |
"missing Os macro, using \"\"", |
"duplicate prologue macro", |
|
"late prologue macro", |
"late prologue macro", |
"skipping late title macro", |
|
"prologue macros out of order", |
"prologue macros out of order", |
|
|
/* related to document structure */ |
/* related to document structure */ |
Line 126 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
Line 141 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
"sections out of conventional order", |
"sections out of conventional order", |
"duplicate section title", |
"duplicate section title", |
"unexpected section", |
"unexpected section", |
|
"cross reference to self", |
"unusual Xr order", |
"unusual Xr order", |
"unusual Xr punctuation", |
"unusual Xr punctuation", |
"AUTHORS section without An macro", |
"AUTHORS section without An macro", |
Line 139 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
Line 155 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
"blocks badly nested", |
"blocks badly nested", |
"nested displays are not portable", |
"nested displays are not portable", |
"moving content out of list", |
"moving content out of list", |
"fill mode already enabled, skipping", |
"first macro on line", |
"fill mode already disabled, skipping", |
|
"line scope broken", |
"line scope broken", |
"skipping blank line in line scope", |
"skipping blank line in line scope", |
|
|
Line 157 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
Line 172 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
"missing function name, using \"\"", |
"missing function name, using \"\"", |
"empty head in list item", |
"empty head in list item", |
"empty list item", |
"empty list item", |
|
"missing argument, using next line", |
"missing font type, using \\fR", |
"missing font type, using \\fR", |
"unknown font type, using \\fR", |
"unknown font type, using \\fR", |
"nothing follows prefix", |
"nothing follows prefix", |
Line 168 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
Line 184 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
"missing eqn box, using \"\"", |
"missing eqn box, using \"\"", |
|
|
/* related to bad macro arguments */ |
/* related to bad macro arguments */ |
"unterminated quoted argument", |
|
"duplicate argument", |
"duplicate argument", |
"skipping duplicate argument", |
"skipping duplicate argument", |
"skipping duplicate display type", |
"skipping duplicate display type", |
Line 187 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
Line 202 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
/* related to plain text */ |
/* related to plain text */ |
"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", |
|
"new sentence, new line", |
"new sentence, new line", |
"bad comment style", |
|
"invalid escape sequence", |
"invalid escape sequence", |
"undefined string, using \"\"", |
"undefined string, using \"\"", |
|
|
Line 215 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
Line 228 static const char * const mandocerrs[MANDOCERR_MAX] = |
|
|
|
/* related to document structure and macros */ |
/* related to document structure and macros */ |
NULL, |
NULL, |
|
"duplicate prologue macro", |
|
"skipping late title macro", |
"input stack limit exceeded, infinite loop?", |
"input stack limit exceeded, infinite loop?", |
"skipping bad character", |
"skipping bad character", |
"skipping unknown macro", |
"skipping unknown macro", |
Line 326 choose_parser(struct mparse *curp) |
|
Line 341 choose_parser(struct mparse *curp) |
|
static int |
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; |
|
struct buf ln; |
struct buf ln; |
const char *save_file; |
const char *save_file; |
char *cp; |
char *cp; |
|
|
if (curp->man->macroset == MACROSET_NONE) |
if (curp->man->macroset == MACROSET_NONE) |
choose_parser(curp); |
choose_parser(curp); |
|
|
/* |
if ((curp->man->macroset == MACROSET_MDOC ? |
* Lastly, push down into the parsers themselves. |
|
* If libroff returns ROFF_TBL, then add it to the |
|
* currently open parse. Since we only get here if |
|
* there does exist data (see tbl_data.c), we're |
|
* guaranteed that something's been allocated. |
|
* Do the same for ROFF_EQN. |
|
*/ |
|
|
|
if (rr == ROFF_TBL) |
|
while ((span = roff_span(curp->roff)) != NULL) |
|
roff_addtbl(curp->man, span); |
|
else if (rr == ROFF_EQN) |
|
roff_addeqn(curp->man, roff_eqn(curp->roff)); |
|
else if ((curp->man->macroset == MACROSET_MDOC ? |
|
mdoc_parseln(curp->man, curp->line, ln.buf, of) : |
mdoc_parseln(curp->man, curp->line, ln.buf, of) : |
man_parseln(curp->man, curp->line, ln.buf, of)) == 2) |
man_parseln(curp->man, curp->line, ln.buf, of)) == 2) |
break; |
break; |
Line 558 read_whole_file(struct mparse *curp, const char *file, |
|
Line 558 read_whole_file(struct mparse *curp, const char *file, |
|
gzFile gz; |
gzFile gz; |
size_t off; |
size_t off; |
ssize_t ssz; |
ssize_t ssz; |
|
int gzerrnum, retval; |
|
|
if (fstat(fd, &st) == -1) |
if (fstat(fd, &st) == -1) { |
err((int)MANDOCLEVEL_SYSERR, "%s", file); |
mandoc_vmsg(MANDOCERR_FILE, curp, 0, 0, |
|
"fstat: %s", strerror(errno)); |
|
return 0; |
|
} |
|
|
/* |
/* |
* If we're a regular file, try just reading in the whole entry |
* If we're a regular file, try just reading in the whole entry |
Line 582 read_whole_file(struct mparse *curp, const char *file, |
|
Line 586 read_whole_file(struct mparse *curp, const char *file, |
|
} |
} |
|
|
if (curp->gzip) { |
if (curp->gzip) { |
if ((gz = gzdopen(fd, "rb")) == NULL) |
/* |
err((int)MANDOCLEVEL_SYSERR, "%s", file); |
* Duplicating the file descriptor is required |
|
* because we will have to call gzclose(3) |
|
* to free memory used internally by zlib, |
|
* but that will also close the file descriptor, |
|
* which this function must not do. |
|
*/ |
|
if ((fd = dup(fd)) == -1) { |
|
mandoc_vmsg(MANDOCERR_FILE, curp, 0, 0, |
|
"dup: %s", strerror(errno)); |
|
return 0; |
|
} |
|
if ((gz = gzdopen(fd, "rb")) == NULL) { |
|
mandoc_vmsg(MANDOCERR_FILE, curp, 0, 0, |
|
"gzdopen: %s", strerror(errno)); |
|
close(fd); |
|
return 0; |
|
} |
} else |
} else |
gz = NULL; |
gz = NULL; |
|
|
Line 594 read_whole_file(struct mparse *curp, const char *file, |
|
Line 614 read_whole_file(struct mparse *curp, const char *file, |
|
|
|
*with_mmap = 0; |
*with_mmap = 0; |
off = 0; |
off = 0; |
|
retval = 0; |
fb->sz = 0; |
fb->sz = 0; |
fb->buf = NULL; |
fb->buf = NULL; |
for (;;) { |
for (;;) { |
Line 610 read_whole_file(struct mparse *curp, const char *file, |
|
Line 631 read_whole_file(struct mparse *curp, const char *file, |
|
read(fd, fb->buf + (int)off, fb->sz - off); |
read(fd, fb->buf + (int)off, fb->sz - off); |
if (ssz == 0) { |
if (ssz == 0) { |
fb->sz = off; |
fb->sz = off; |
return 1; |
retval = 1; |
|
break; |
} |
} |
if (ssz == -1) |
if (ssz == -1) { |
err((int)MANDOCLEVEL_SYSERR, "%s", file); |
if (curp->gzip) |
|
(void)gzerror(gz, &gzerrnum); |
|
mandoc_vmsg(MANDOCERR_FILE, curp, 0, 0, "read: %s", |
|
curp->gzip && gzerrnum != Z_ERRNO ? |
|
zError(gzerrnum) : strerror(errno)); |
|
break; |
|
} |
off += (size_t)ssz; |
off += (size_t)ssz; |
} |
} |
|
|
free(fb->buf); |
if (curp->gzip && (gzerrnum = gzclose(gz)) != Z_OK) |
fb->buf = NULL; |
mandoc_vmsg(MANDOCERR_FILE, curp, 0, 0, "gzclose: %s", |
return 0; |
gzerrnum == Z_ERRNO ? strerror(errno) : |
|
zError(gzerrnum)); |
|
if (retval == 0) { |
|
free(fb->buf); |
|
fb->buf = NULL; |
|
} |
|
return retval; |
} |
} |
|
|
static void |
static void |
Line 749 mparse_open(struct mparse *curp, const char *file) |
|
Line 783 mparse_open(struct mparse *curp, const char *file) |
|
} |
} |
|
|
struct mparse * |
struct mparse * |
mparse_alloc(int options, enum mandoclevel wlevel, mandocmsg mmsg, |
mparse_alloc(int options, enum mandocerr mmin, mandocmsg mmsg, |
const char *defos) |
enum mandoc_os os_e, const char *os_s) |
{ |
{ |
struct mparse *curp; |
struct mparse *curp; |
|
|
curp = mandoc_calloc(1, sizeof(struct mparse)); |
curp = mandoc_calloc(1, sizeof(struct mparse)); |
|
|
curp->options = options; |
curp->options = options; |
curp->wlevel = wlevel; |
curp->mmin = mmin; |
curp->mmsg = mmsg; |
curp->mmsg = mmsg; |
curp->defos = defos; |
curp->os_s = os_s; |
|
|
curp->roff = roff_alloc(curp, options); |
curp->roff = roff_alloc(curp, options); |
curp->man = roff_man_alloc( curp->roff, curp, curp->defos, |
curp->man = roff_man_alloc(curp->roff, curp, curp->os_s, |
curp->options & MPARSE_QUICK ? 1 : 0); |
curp->options & MPARSE_QUICK ? 1 : 0); |
if (curp->options & MPARSE_MDOC) { |
if (curp->options & MPARSE_MDOC) { |
curp->man->macroset = MACROSET_MDOC; |
curp->man->macroset = MACROSET_MDOC; |
Line 774 mparse_alloc(int options, enum mandoclevel wlevel, man |
|
Line 808 mparse_alloc(int options, enum mandoclevel wlevel, man |
|
curp->man->manmac = roffhash_alloc(MAN_TH, MAN_MAX); |
curp->man->manmac = roffhash_alloc(MAN_TH, MAN_MAX); |
} |
} |
curp->man->first->tok = TOKEN_NONE; |
curp->man->first->tok = TOKEN_NONE; |
|
curp->man->meta.os_e = os_e; |
return curp; |
return curp; |
} |
} |
|
|
Line 849 mandoc_msg(enum mandocerr er, struct mparse *m, |
|
Line 884 mandoc_msg(enum mandocerr er, struct mparse *m, |
|
{ |
{ |
enum mandoclevel level; |
enum mandoclevel level; |
|
|
|
if (er < m->mmin && er != MANDOCERR_FILE) |
|
return; |
|
|
level = MANDOCLEVEL_UNSUPP; |
level = MANDOCLEVEL_UNSUPP; |
while (er < mandoclimits[level]) |
while (er < mandoclimits[level]) |
level--; |
level--; |
|
|
if (level < m->wlevel && er != MANDOCERR_FILE) |
|
return; |
|
|
|
if (m->mmsg) |
if (m->mmsg) |
(*m->mmsg)(er, level, m->file, ln, col, msg); |
(*m->mmsg)(er, level, m->file, ln, col, msg); |