version 1.16, 2009/03/25 21:46:24 |
version 1.147, 2011/03/07 01:35:51 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
* Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@openbsd.org> |
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
|
* Copyright (c) 2010, 2011 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 |
* purpose with or without fee is hereby granted, provided that the above |
* above copyright notice and this permission notice appear in all |
* copyright notice and this permission notice appear in all copies. |
* copies. |
|
* |
* |
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL |
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES |
* WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED |
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF |
* WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE |
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR |
* AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL |
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES |
* DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR |
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN |
* PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER |
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF |
* TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR |
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. |
* PERFORMANCE OF THIS SOFTWARE. |
|
*/ |
*/ |
|
#ifdef HAVE_CONFIG_H |
|
#include "config.h" |
|
#endif |
|
|
|
#include <sys/mman.h> |
#include <sys/stat.h> |
#include <sys/stat.h> |
|
|
#include <assert.h> |
#include <assert.h> |
#include <err.h> |
#include <ctype.h> |
#include <fcntl.h> |
#include <fcntl.h> |
#include <stdio.h> |
#include <stdio.h> |
|
#include <stdint.h> |
#include <stdlib.h> |
#include <stdlib.h> |
#include <string.h> |
#include <string.h> |
#include <unistd.h> |
#include <unistd.h> |
|
|
|
#include "mandoc.h" |
|
#include "main.h" |
#include "mdoc.h" |
#include "mdoc.h" |
#include "man.h" |
#include "man.h" |
|
#include "roff.h" |
|
|
/* Account for FreeBSD and Linux in our declarations. */ |
#ifndef MAP_FILE |
|
#define MAP_FILE 0 |
|
#endif |
|
|
#ifdef __linux__ |
#define REPARSE_LIMIT 1000 |
extern int getsubopt(char **, char * const *, char **); |
#define UNCONST(a) ((void *)(uintptr_t)(const void *)(a)) |
# ifndef __dead |
|
# define __dead __attribute__((__noreturn__)) |
/* FIXME: Intel's compiler? LLVM? pcc? */ |
|
|
|
#if !defined(__GNUC__) || (__GNUC__ < 2) |
|
# if !defined(lint) |
|
# define __attribute__(x) |
# endif |
# endif |
#elif defined(__dead2) |
#endif /* !defined(__GNUC__) || (__GNUC__ < 2) */ |
# ifndef __dead |
|
# define __dead __dead2 |
|
# endif |
|
#endif |
|
|
|
|
typedef void (*out_mdoc)(void *, const struct mdoc *); |
|
typedef void (*out_man)(void *, const struct man *); |
|
typedef void (*out_free)(void *); |
|
|
struct buf { |
struct buf { |
char *buf; |
char *buf; |
size_t sz; |
size_t sz; |
}; |
}; |
|
|
struct curparse { |
|
const char *file; |
|
int wflags; |
|
#define WARN_WALL 0x03 /* All-warnings mask. */ |
|
#define WARN_WCOMPAT (1 << 0) /* Compatibility warnings. */ |
|
#define WARN_WSYNTAX (1 << 1) /* Syntax warnings. */ |
|
#define WARN_WERR (1 << 2) /* Warnings->errors. */ |
|
}; |
|
|
|
#define IGN_SCOPE (1 << 0) /* Ignore scope errors. */ |
|
#define IGN_ESCAPE (1 << 1) /* Ignore bad escapes. */ |
|
#define IGN_MACRO (1 << 2) /* Ignore unknown macros. */ |
|
|
|
enum intt { |
enum intt { |
INTT_MDOC = 0, |
INTT_AUTO, |
|
INTT_MDOC, |
INTT_MAN |
INTT_MAN |
}; |
}; |
|
|
enum outt { |
enum outt { |
OUTT_ASCII = 0, |
OUTT_ASCII = 0, |
OUTT_TREE, |
OUTT_TREE, |
OUTT_LINT |
OUTT_HTML, |
|
OUTT_XHTML, |
|
OUTT_LINT, |
|
OUTT_PS, |
|
OUTT_PDF |
}; |
}; |
|
|
typedef int (*out_run)(void *, const struct man *, |
struct curparse { |
const struct mdoc *); |
const char *file; /* Current parse. */ |
typedef void (*out_free)(void *); |
int fd; /* Current parse. */ |
|
int line; /* Line number in the file. */ |
|
enum mandoclevel wlevel; /* Ignore messages below this. */ |
|
int wstop; /* Stop after a file with a warning. */ |
|
enum intt inttype; /* which parser to use */ |
|
struct man *pman; /* persistent man parser */ |
|
struct mdoc *pmdoc; /* persistent mdoc parser */ |
|
struct man *man; /* man parser */ |
|
struct mdoc *mdoc; /* mdoc parser */ |
|
struct roff *roff; /* roff parser (!NULL) */ |
|
struct regset regs; /* roff registers */ |
|
int reparse_count; /* finite interpolation stack */ |
|
enum outt outtype; /* which output to use */ |
|
out_mdoc outmdoc; /* mdoc output ptr */ |
|
out_man outman; /* man output ptr */ |
|
out_free outfree; /* free output ptr */ |
|
void *outdata; /* data for output */ |
|
char outopts[BUFSIZ]; /* buf of output opts */ |
|
}; |
|
|
extern char *__progname; |
static const char * const mandoclevels[MANDOCLEVEL_MAX] = { |
|
"SUCCESS", |
|
"RESERVED", |
|
"WARNING", |
|
"ERROR", |
|
"FATAL", |
|
"BADARG", |
|
"SYSERR" |
|
}; |
|
|
extern void *ascii_alloc(void); |
static const enum mandocerr mandoclimits[MANDOCLEVEL_MAX] = { |
extern int terminal_run(void *, const struct man *, |
MANDOCERR_OK, |
const struct mdoc *); |
MANDOCERR_WARNING, |
extern int tree_run(void *, const struct man *, |
MANDOCERR_WARNING, |
const struct mdoc *); |
MANDOCERR_ERROR, |
extern void terminal_free(void *); |
MANDOCERR_FATAL, |
|
MANDOCERR_MAX, |
|
MANDOCERR_MAX |
|
}; |
|
|
static int foptions(int *, char *); |
static const char * const mandocerrs[MANDOCERR_MAX] = { |
static int toptions(enum outt *, char *); |
"ok", |
|
|
|
"generic warning", |
|
|
|
/* related to the prologue */ |
|
"no title in document", |
|
"document title should be all caps", |
|
"unknown manual section", |
|
"date missing, using today's date", |
|
"cannot parse date, using it verbatim", |
|
"prologue macros out of order", |
|
"duplicate prologue macro", |
|
"macro not allowed in prologue", |
|
"macro not allowed in body", |
|
|
|
/* related to document structure */ |
|
".so is fragile, better use ln(1)", |
|
"NAME section must come first", |
|
"bad NAME section contents", |
|
"manual name not yet set", |
|
"sections out of conventional order", |
|
"duplicate section name", |
|
"section not in conventional manual section", |
|
|
|
/* related to macros and nesting */ |
|
"skipping obsolete macro", |
|
"skipping paragraph macro", |
|
"skipping no-space macro", |
|
"blocks badly nested", |
|
"child violates parent syntax", |
|
"nested displays are not portable", |
|
"already in literal mode", |
|
|
|
/* related to missing macro arguments */ |
|
"skipping empty macro", |
|
"argument count wrong", |
|
"missing display type", |
|
"list type must come first", |
|
"tag lists require a width argument", |
|
"missing font type", |
|
"skipping end of block that is not open", |
|
|
|
/* related to bad macro arguments */ |
|
"skipping argument", |
|
"duplicate argument", |
|
"duplicate display type", |
|
"duplicate list type", |
|
"unknown AT&T UNIX version", |
|
"bad Boolean value", |
|
"unknown font", |
|
"unknown standard specifier", |
|
"bad width argument", |
|
|
|
/* related to plain text */ |
|
"blank line in non-literal context", |
|
"tab in non-literal context", |
|
"end of line whitespace", |
|
"bad comment style", |
|
"unknown escape sequence", |
|
"unterminated quoted string", |
|
|
|
"generic error", |
|
|
|
/* related to tables */ |
|
"bad table syntax", |
|
"bad table option", |
|
"bad table layout", |
|
"no table layout cells specified", |
|
"no table data cells specified", |
|
"ignore data in cell", |
|
"data block still open", |
|
"ignoring extra data cells", |
|
|
|
"input stack limit exceeded, infinite loop?", |
|
"skipping bad character", |
|
"escaped character not allowed in a name", |
|
"skipping text before the first section header", |
|
"skipping unknown macro", |
|
"NOT IMPLEMENTED, please use groff: skipping request", |
|
"line scope broken", |
|
"argument count wrong", |
|
"skipping end of block that is not open", |
|
"missing end of block", |
|
"scope open on exit", |
|
"uname(3) system call failed", |
|
"macro requires line argument(s)", |
|
"macro requires body argument(s)", |
|
"macro requires argument(s)", |
|
"missing list type", |
|
"line argument(s) will be lost", |
|
"body argument(s) will be lost", |
|
|
|
"generic fatal error", |
|
|
|
"column syntax is inconsistent", |
|
"NOT IMPLEMENTED: .Bd -file", |
|
"line scope broken, syntax violated", |
|
"argument count wrong, violates syntax", |
|
"child violates parent syntax", |
|
"argument count wrong, violates syntax", |
|
"NOT IMPLEMENTED: .so with absolute path or \"..\"", |
|
"no document body", |
|
"no document prologue", |
|
"static buffer exhausted", |
|
}; |
|
|
|
static void parsebuf(struct curparse *, struct buf, int); |
|
static void pdesc(struct curparse *); |
|
static void fdesc(struct curparse *); |
|
static void ffile(const char *, struct curparse *); |
|
static int pfile(const char *, struct curparse *); |
static int moptions(enum intt *, char *); |
static int moptions(enum intt *, char *); |
static int woptions(int *, char *); |
static int mmsg(enum mandocerr, void *, |
static int merr(void *, int, int, const char *); |
int, int, const char *); |
static int manwarn(void *, int, int, const char *); |
static void pset(const char *, int, struct curparse *); |
static int mdocwarn(void *, int, int, |
static int toptions(struct curparse *, char *); |
enum mdoc_warn, const char *); |
static void usage(void) __attribute__((noreturn)); |
static int file(struct buf *, struct buf *, |
static void version(void) __attribute__((noreturn)); |
const char *, |
static int woptions(struct curparse *, char *); |
struct man *, struct mdoc *); |
|
static int fdesc(struct buf *, struct buf *, |
|
const char *, int, |
|
struct man *, struct mdoc *); |
|
__dead static void version(void); |
|
__dead static void usage(void); |
|
|
|
|
static const char *progname; |
|
static enum mandoclevel file_status = MANDOCLEVEL_OK; |
|
static enum mandoclevel exit_status = MANDOCLEVEL_OK; |
|
|
int |
int |
main(int argc, char *argv[]) |
main(int argc, char *argv[]) |
{ |
{ |
int c, rc, fflags, pflags; |
int c; |
struct mdoc_cb mdoccb; |
|
struct man_cb mancb; |
|
struct man *man; |
|
struct mdoc *mdoc; |
|
void *outdata; |
|
enum outt outtype; |
|
enum intt inttype; |
|
struct buf ln, blk; |
|
out_run outrun; |
|
out_free outfree; |
|
struct curparse curp; |
struct curparse curp; |
|
|
fflags = 0; |
progname = strrchr(argv[0], '/'); |
outtype = OUTT_ASCII; |
if (progname == NULL) |
inttype = INTT_MDOC; |
progname = argv[0]; |
|
else |
|
++progname; |
|
|
bzero(&curp, sizeof(struct curparse)); |
memset(&curp, 0, sizeof(struct curparse)); |
|
|
|
curp.inttype = INTT_AUTO; |
|
curp.outtype = OUTT_ASCII; |
|
curp.wlevel = MANDOCLEVEL_FATAL; |
|
|
/* LINTED */ |
/* LINTED */ |
while (-1 != (c = getopt(argc, argv, "f:m:VW:T:"))) |
while (-1 != (c = getopt(argc, argv, "m:O:T:VW:"))) |
switch (c) { |
switch (c) { |
case ('f'): |
|
if ( ! foptions(&fflags, optarg)) |
|
return(0); |
|
break; |
|
case ('m'): |
case ('m'): |
if ( ! moptions(&inttype, optarg)) |
if ( ! moptions(&curp.inttype, optarg)) |
return(0); |
return((int)MANDOCLEVEL_BADARG); |
break; |
break; |
|
case ('O'): |
|
(void)strlcat(curp.outopts, optarg, BUFSIZ); |
|
(void)strlcat(curp.outopts, ",", BUFSIZ); |
|
break; |
case ('T'): |
case ('T'): |
if ( ! toptions(&outtype, optarg)) |
if ( ! toptions(&curp, optarg)) |
return(0); |
return((int)MANDOCLEVEL_BADARG); |
break; |
break; |
case ('W'): |
case ('W'): |
if ( ! woptions(&curp.wflags, optarg)) |
if ( ! woptions(&curp, optarg)) |
return(0); |
return((int)MANDOCLEVEL_BADARG); |
break; |
break; |
case ('V'): |
case ('V'): |
version(); |
version(); |
Line 154 main(int argc, char *argv[]) |
|
Line 292 main(int argc, char *argv[]) |
|
argc -= optind; |
argc -= optind; |
argv += optind; |
argv += optind; |
|
|
/* |
if (NULL == *argv) { |
* Allocate the appropriate front-end. Note that utf8, latin1 |
curp.file = "<stdin>"; |
* (both not yet implemented) and ascii all resolve to the |
curp.fd = STDIN_FILENO; |
* terminal front-end with different encodings (see terminal.c). |
|
* Not all frontends have cleanup or alloc routines. |
|
*/ |
|
|
|
switch (outtype) { |
fdesc(&curp); |
case (OUTT_TREE): |
|
outdata = NULL; |
|
outrun = tree_run; |
|
outfree = NULL; |
|
break; |
|
case (OUTT_LINT): |
|
outdata = NULL; |
|
outrun = NULL; |
|
outfree = NULL; |
|
break; |
|
default: |
|
outdata = ascii_alloc(); |
|
outrun = terminal_run; |
|
outfree = terminal_free; |
|
break; |
|
} |
} |
|
|
/* |
while (*argv) { |
* All callbacks route into here, where we print them onto the |
ffile(*argv, &curp); |
* screen. XXX - for now, no path for debugging messages. |
if (MANDOCLEVEL_OK != exit_status && curp.wstop) |
*/ |
break; |
|
++argv; |
|
} |
|
|
mdoccb.mdoc_msg = NULL; |
if (curp.outfree) |
mdoccb.mdoc_err = merr; |
(*curp.outfree)(curp.outdata); |
mdoccb.mdoc_warn = mdocwarn; |
if (curp.pmdoc) |
|
mdoc_free(curp.pmdoc); |
|
if (curp.pman) |
|
man_free(curp.pman); |
|
if (curp.roff) |
|
roff_free(curp.roff); |
|
|
mancb.man_err = merr; |
return((int)exit_status); |
mancb.man_warn = manwarn; |
} |
|
|
/* Configure buffers. */ |
|
|
|
bzero(&ln, sizeof(struct buf)); |
static void |
bzero(&blk, sizeof(struct buf)); |
version(void) |
|
{ |
|
|
man = NULL; |
(void)printf("%s %s\n", progname, VERSION); |
mdoc = NULL; |
exit((int)MANDOCLEVEL_OK); |
pflags = 0; |
} |
|
|
/* |
|
* Allocate the parser. There are two kinds of parser: libman |
|
* and libmdoc. We must separately copy over the flags that |
|
* we'll use internally. |
|
*/ |
|
|
|
switch (inttype) { |
static void |
case (INTT_MAN): |
usage(void) |
if (fflags & IGN_MACRO) |
{ |
pflags |= MAN_IGN_MACRO; |
|
man = man_alloc(&curp, pflags, &mancb); |
|
break; |
|
default: |
|
if (fflags & IGN_SCOPE) |
|
pflags |= MDOC_IGN_SCOPE; |
|
if (fflags & IGN_ESCAPE) |
|
pflags |= MDOC_IGN_ESCAPE; |
|
if (fflags & IGN_MACRO) |
|
pflags |= MDOC_IGN_MACRO; |
|
mdoc = mdoc_alloc(&curp, pflags, &mdoccb); |
|
break; |
|
} |
|
|
|
|
(void)fprintf(stderr, "usage: %s " |
|
"[-V] " |
|
"[-foption] " |
|
"[-mformat] " |
|
"[-Ooption] " |
|
"[-Toutput] " |
|
"[-Werr] " |
|
"[file...]\n", |
|
progname); |
|
|
|
exit((int)MANDOCLEVEL_BADARG); |
|
} |
|
|
|
static void |
|
ffile(const char *file, struct curparse *curp) |
|
{ |
|
|
/* |
/* |
* Main loop around available files. |
* Called once per input file. Get the file ready for reading, |
|
* pass it through to the parser-driver, then close it out. |
|
* XXX: don't do anything special as this is only called for |
|
* files; stdin goes directly to fdesc(). |
*/ |
*/ |
|
|
if (NULL == *argv) { |
curp->file = file; |
curp.file = "<stdin>"; |
|
rc = 0; |
|
c = fdesc(&blk, &ln, "stdin", STDIN_FILENO, man, mdoc); |
|
|
|
if (c && NULL == outrun) |
if (-1 == (curp->fd = open(curp->file, O_RDONLY, 0))) { |
rc = 1; |
perror(curp->file); |
else if (c && outrun && (*outrun)(outdata, man, mdoc)) |
exit_status = MANDOCLEVEL_SYSERR; |
rc = 1; |
return; |
} else { |
|
while (*argv) { |
|
curp.file = *argv; |
|
c = file(&blk, &ln, *argv, man, mdoc); |
|
if ( ! c) |
|
break; |
|
if (outrun && ! (*outrun)(outdata, man, mdoc)) |
|
break; |
|
if (man) |
|
man_reset(man); |
|
if (mdoc) |
|
mdoc_reset(mdoc); |
|
argv++; |
|
} |
|
rc = NULL == *argv; |
|
} |
} |
|
|
if (blk.buf) |
fdesc(curp); |
free(blk.buf); |
|
if (ln.buf) |
|
free(ln.buf); |
|
if (outfree) |
|
(*outfree)(outdata); |
|
if (mdoc) |
|
mdoc_free(mdoc); |
|
if (man) |
|
man_free(man); |
|
|
|
return(rc ? EXIT_SUCCESS : EXIT_FAILURE); |
if (-1 == close(curp->fd)) |
|
perror(curp->file); |
} |
} |
|
|
|
static int |
__dead static void |
pfile(const char *file, struct curparse *curp) |
version(void) |
|
{ |
{ |
|
const char *savefile; |
|
int fd, savefd; |
|
|
(void)printf("%s %s\n", __progname, VERSION); |
if (-1 == (fd = open(file, O_RDONLY, 0))) { |
exit(0); |
perror(file); |
/* NOTREACHED */ |
file_status = MANDOCLEVEL_SYSERR; |
|
return(0); |
|
} |
|
|
|
savefile = curp->file; |
|
savefd = curp->fd; |
|
|
|
curp->file = file; |
|
curp->fd = fd; |
|
|
|
pdesc(curp); |
|
|
|
curp->file = savefile; |
|
curp->fd = savefd; |
|
|
|
if (-1 == close(fd)) |
|
perror(file); |
|
|
|
return(MANDOCLEVEL_FATAL > file_status ? 1 : 0); |
} |
} |
|
|
|
|
__dead static void |
static void |
usage(void) |
resize_buf(struct buf *buf, size_t initial) |
{ |
{ |
|
|
(void)fprintf(stderr, "usage: %s [-V] [-foption...] " |
buf->sz = buf->sz > initial/2 ? 2 * buf->sz : initial; |
"[-mformat] [-Toutput] [-Werr...]\n", |
buf->buf = realloc(buf->buf, buf->sz); |
__progname); |
if (NULL == buf->buf) { |
exit(1); |
perror(NULL); |
/* NOTREACHED */ |
exit((int)MANDOCLEVEL_SYSERR); |
|
} |
} |
} |
|
|
|
|
static int |
static int |
file(struct buf *blk, struct buf *ln, const char *file, |
read_whole_file(struct curparse *curp, struct buf *fb, int *with_mmap) |
struct man *man, struct mdoc *mdoc) |
|
{ |
{ |
int fd, c; |
struct stat st; |
|
size_t off; |
|
ssize_t ssz; |
|
|
if (-1 == (fd = open(file, O_RDONLY, 0))) { |
if (-1 == fstat(curp->fd, &st)) { |
warn("%s", file); |
perror(curp->file); |
return(0); |
return(0); |
} |
} |
|
|
c = fdesc(blk, ln, file, fd, man, mdoc); |
/* |
|
* If we're a regular file, try just reading in the whole entry |
|
* via mmap(). This is faster than reading it into blocks, and |
|
* since each file is only a few bytes to begin with, I'm not |
|
* concerned that this is going to tank any machines. |
|
*/ |
|
|
if (-1 == close(fd)) |
if (S_ISREG(st.st_mode)) { |
warn("%s", file); |
if (st.st_size >= (1U << 31)) { |
|
fprintf(stderr, "%s: input too large\n", |
|
curp->file); |
|
return(0); |
|
} |
|
*with_mmap = 1; |
|
fb->sz = (size_t)st.st_size; |
|
fb->buf = mmap(NULL, fb->sz, PROT_READ, |
|
MAP_FILE|MAP_SHARED, curp->fd, 0); |
|
if (fb->buf != MAP_FAILED) |
|
return(1); |
|
} |
|
|
return(c); |
/* |
|
* If this isn't a regular file (like, say, stdin), then we must |
|
* go the old way and just read things in bit by bit. |
|
*/ |
|
|
|
*with_mmap = 0; |
|
off = 0; |
|
fb->sz = 0; |
|
fb->buf = NULL; |
|
for (;;) { |
|
if (off == fb->sz) { |
|
if (fb->sz == (1U << 31)) { |
|
fprintf(stderr, "%s: input too large\n", |
|
curp->file); |
|
break; |
|
} |
|
resize_buf(fb, 65536); |
|
} |
|
ssz = read(curp->fd, fb->buf + (int)off, fb->sz - off); |
|
if (ssz == 0) { |
|
fb->sz = off; |
|
return(1); |
|
} |
|
if (ssz == -1) { |
|
perror(curp->file); |
|
break; |
|
} |
|
off += (size_t)ssz; |
|
} |
|
|
|
free(fb->buf); |
|
fb->buf = NULL; |
|
return(0); |
} |
} |
|
|
|
|
static int |
static void |
fdesc(struct buf *blk, struct buf *ln, |
fdesc(struct curparse *curp) |
const char *f, int fd, |
|
struct man *man, struct mdoc *mdoc) |
|
{ |
{ |
size_t sz; |
|
ssize_t ssz; |
|
struct stat st; |
|
int j, i, pos, lnn; |
|
|
|
assert( ! (man && mdoc)); |
/* |
|
* Called once per file with an opened file descriptor. All |
|
* pre-file-parse operations (whether stdin or a file) should go |
|
* here. |
|
* |
|
* This calls down into the nested parser, which drills down and |
|
* fully parses a file and all its dependences (i.e., `so'). It |
|
* then runs the cleanup validators and pushes to output. |
|
*/ |
|
|
|
/* Zero the parse type. */ |
|
|
|
curp->mdoc = NULL; |
|
curp->man = NULL; |
|
file_status = MANDOCLEVEL_OK; |
|
|
|
/* Make sure the mandotory roff parser is initialised. */ |
|
|
|
if (NULL == curp->roff) { |
|
curp->roff = roff_alloc(&curp->regs, curp, mmsg); |
|
assert(curp->roff); |
|
} |
|
|
|
/* Fully parse the file. */ |
|
|
|
pdesc(curp); |
|
|
|
if (MANDOCLEVEL_FATAL <= file_status) |
|
goto cleanup; |
|
|
|
/* NOTE a parser may not have been assigned, yet. */ |
|
|
|
if ( ! (curp->man || curp->mdoc)) { |
|
fprintf(stderr, "%s: Not a manual\n", curp->file); |
|
file_status = MANDOCLEVEL_FATAL; |
|
goto cleanup; |
|
} |
|
|
|
/* Clean up the parse routine ASTs. */ |
|
|
|
if (curp->mdoc && ! mdoc_endparse(curp->mdoc)) { |
|
assert(MANDOCLEVEL_FATAL <= file_status); |
|
goto cleanup; |
|
} |
|
|
|
if (curp->man && ! man_endparse(curp->man)) { |
|
assert(MANDOCLEVEL_FATAL <= file_status); |
|
goto cleanup; |
|
} |
|
|
|
assert(curp->roff); |
|
roff_endparse(curp->roff); |
|
|
/* |
/* |
* Two buffers: ln and buf. buf is the input buffer, optimised |
* With -Wstop and warnings or errors of at least |
* for each file's block size. ln is a line buffer. Both |
* the requested level, do not produce output. |
* growable, hence passed in by ptr-ptr. |
|
*/ |
*/ |
|
|
sz = BUFSIZ; |
if (MANDOCLEVEL_OK != file_status && curp->wstop) |
|
goto cleanup; |
|
|
if (-1 == fstat(fd, &st)) |
/* If unset, allocate output dev now (if applicable). */ |
warnx("%s", f); |
|
else if ((size_t)st.st_blksize > sz) |
|
sz = st.st_blksize; |
|
|
|
if (sz > blk->sz) { |
if ( ! (curp->outman && curp->outmdoc)) { |
blk->buf = realloc(blk->buf, sz); |
switch (curp->outtype) { |
if (NULL == blk->buf) |
case (OUTT_XHTML): |
err(1, "realloc"); |
curp->outdata = xhtml_alloc(curp->outopts); |
blk->sz = sz; |
break; |
|
case (OUTT_HTML): |
|
curp->outdata = html_alloc(curp->outopts); |
|
break; |
|
case (OUTT_ASCII): |
|
curp->outdata = ascii_alloc(curp->outopts); |
|
curp->outfree = ascii_free; |
|
break; |
|
case (OUTT_PDF): |
|
curp->outdata = pdf_alloc(curp->outopts); |
|
curp->outfree = pspdf_free; |
|
break; |
|
case (OUTT_PS): |
|
curp->outdata = ps_alloc(curp->outopts); |
|
curp->outfree = pspdf_free; |
|
break; |
|
default: |
|
break; |
|
} |
|
|
|
switch (curp->outtype) { |
|
case (OUTT_HTML): |
|
/* FALLTHROUGH */ |
|
case (OUTT_XHTML): |
|
curp->outman = html_man; |
|
curp->outmdoc = html_mdoc; |
|
curp->outfree = html_free; |
|
break; |
|
case (OUTT_TREE): |
|
curp->outman = tree_man; |
|
curp->outmdoc = tree_mdoc; |
|
break; |
|
case (OUTT_PDF): |
|
/* FALLTHROUGH */ |
|
case (OUTT_ASCII): |
|
/* FALLTHROUGH */ |
|
case (OUTT_PS): |
|
curp->outman = terminal_man; |
|
curp->outmdoc = terminal_mdoc; |
|
break; |
|
default: |
|
break; |
|
} |
} |
} |
|
|
|
/* Execute the out device, if it exists. */ |
|
|
|
if (curp->man && curp->outman) |
|
(*curp->outman)(curp->outdata, curp->man); |
|
if (curp->mdoc && curp->outmdoc) |
|
(*curp->outmdoc)(curp->outdata, curp->mdoc); |
|
|
|
cleanup: |
|
|
|
memset(&curp->regs, 0, sizeof(struct regset)); |
|
|
|
/* Reset the current-parse compilers. */ |
|
|
|
if (curp->mdoc) |
|
mdoc_reset(curp->mdoc); |
|
if (curp->man) |
|
man_reset(curp->man); |
|
|
|
assert(curp->roff); |
|
roff_reset(curp->roff); |
|
|
|
if (exit_status < file_status) |
|
exit_status = file_status; |
|
|
|
return; |
|
} |
|
|
|
static void |
|
pdesc(struct curparse *curp) |
|
{ |
|
struct buf blk; |
|
int with_mmap; |
|
|
/* |
/* |
* Fill buf with file blocksize and parse newlines into ln. |
* Run for each opened file; may be called more than once for |
|
* each full parse sequence if the opened file is nested (i.e., |
|
* from `so'). Simply sucks in the whole file and moves into |
|
* the parse phase for the file. |
*/ |
*/ |
|
|
for (lnn = 1, pos = 0; ; ) { |
if ( ! read_whole_file(curp, &blk, &with_mmap)) { |
if (-1 == (ssz = read(fd, blk->buf, sz))) { |
file_status = MANDOCLEVEL_SYSERR; |
warn("%s", f); |
return; |
return(0); |
} |
} else if (0 == ssz) |
|
|
/* Line number is per-file. */ |
|
|
|
curp->line = 1; |
|
|
|
parsebuf(curp, blk, 1); |
|
|
|
if (with_mmap) |
|
munmap(blk.buf, blk.sz); |
|
else |
|
free(blk.buf); |
|
} |
|
|
|
static void |
|
parsebuf(struct curparse *curp, struct buf blk, int start) |
|
{ |
|
const struct tbl_span *span; |
|
struct buf ln; |
|
enum rofferr rr; |
|
int i, of, rc; |
|
int pos; /* byte number in the ln buffer */ |
|
int lnn; /* line number in the real file */ |
|
unsigned char c; |
|
|
|
/* |
|
* Main parse routine for an opened file. This is called for |
|
* each opened file and simply loops around the full input file, |
|
* possibly nesting (i.e., with `so'). |
|
*/ |
|
|
|
memset(&ln, 0, sizeof(struct buf)); |
|
|
|
lnn = curp->line; |
|
pos = 0; |
|
|
|
for (i = 0; i < (int)blk.sz; ) { |
|
if (0 == pos && '\0' == blk.buf[i]) |
break; |
break; |
|
|
for (i = 0; i < (int)ssz; i++) { |
if (start) { |
if (pos >= (int)ln->sz) { |
curp->line = lnn; |
ln->sz += 256; /* Step-size. */ |
curp->reparse_count = 0; |
ln->buf = realloc(ln->buf, ln->sz); |
} |
if (NULL == ln->buf) |
|
err(1, "realloc"); |
while (i < (int)blk.sz && (start || '\0' != blk.buf[i])) { |
|
|
|
/* |
|
* When finding an unescaped newline character, |
|
* leave the character loop to process the line. |
|
* Skip a preceding carriage return, if any. |
|
*/ |
|
|
|
if ('\r' == blk.buf[i] && i + 1 < (int)blk.sz && |
|
'\n' == blk.buf[i + 1]) |
|
++i; |
|
if ('\n' == blk.buf[i]) { |
|
++i; |
|
++lnn; |
|
break; |
} |
} |
|
|
if ('\n' != blk->buf[i]) { |
/* |
ln->buf[pos++] = blk->buf[i]; |
* Warn about bogus characters. If you're using |
|
* non-ASCII encoding, you're screwing your |
|
* readers. Since I'd rather this not happen, |
|
* I'll be helpful and drop these characters so |
|
* we don't display gibberish. Note to manual |
|
* writers: use special characters. |
|
*/ |
|
|
|
c = (unsigned char) blk.buf[i]; |
|
|
|
if ( ! (isascii(c) && |
|
(isgraph(c) || isblank(c)))) { |
|
mmsg(MANDOCERR_BADCHAR, curp, |
|
curp->line, pos, "ignoring byte"); |
|
i++; |
continue; |
continue; |
} |
} |
|
|
/* Check for CPP-escaped newline. */ |
/* Trailing backslash = a plain char. */ |
|
|
if (pos > 0 && '\\' == ln->buf[pos - 1]) { |
if ('\\' != blk.buf[i] || i + 1 == (int)blk.sz) { |
for (j = pos - 1; j >= 0; j--) |
if (pos >= (int)ln.sz) |
if ('\\' != ln->buf[j]) |
resize_buf(&ln, 256); |
|
ln.buf[pos++] = blk.buf[i++]; |
|
continue; |
|
} |
|
|
|
/* |
|
* Found escape and at least one other character. |
|
* When it's a newline character, skip it. |
|
* When there is a carriage return in between, |
|
* skip that one as well. |
|
*/ |
|
|
|
if ('\r' == blk.buf[i + 1] && i + 2 < (int)blk.sz && |
|
'\n' == blk.buf[i + 2]) |
|
++i; |
|
if ('\n' == blk.buf[i + 1]) { |
|
i += 2; |
|
++lnn; |
|
continue; |
|
} |
|
|
|
if ('"' == blk.buf[i + 1]) { |
|
i += 2; |
|
/* Comment, skip to end of line */ |
|
for (; i < (int)blk.sz; ++i) { |
|
if ('\n' == blk.buf[i]) { |
|
++i; |
|
++lnn; |
break; |
break; |
|
} |
|
} |
|
|
if ( ! ((pos - j) % 2)) { |
/* Backout trailing whitespaces */ |
pos--; |
for (; pos > 0; --pos) { |
lnn++; |
if (ln.buf[pos - 1] != ' ') |
continue; |
break; |
|
if (pos > 2 && ln.buf[pos - 2] == '\\') |
|
break; |
} |
} |
|
break; |
} |
} |
|
|
ln->buf[pos] = 0; |
/* Some other escape sequence, copy & cont. */ |
if (mdoc && ! mdoc_parseln(mdoc, lnn, ln->buf)) |
|
return(0); |
if (pos + 1 >= (int)ln.sz) |
if (man && ! man_parseln(man, lnn, ln->buf)) |
resize_buf(&ln, 256); |
return(0); |
|
lnn++; |
ln.buf[pos++] = blk.buf[i++]; |
|
ln.buf[pos++] = blk.buf[i++]; |
|
} |
|
|
|
if (pos >= (int)ln.sz) |
|
resize_buf(&ln, 256); |
|
|
|
ln.buf[pos] = '\0'; |
|
|
|
/* |
|
* A significant amount of complexity is contained by |
|
* the roff preprocessor. It's line-oriented but can be |
|
* expressed on one line, so we need at times to |
|
* readjust our starting point and re-run it. The roff |
|
* preprocessor can also readjust the buffers with new |
|
* data, so we pass them in wholesale. |
|
*/ |
|
|
|
of = 0; |
|
|
|
rerun: |
|
rr = roff_parseln |
|
(curp->roff, curp->line, |
|
&ln.buf, &ln.sz, of, &of); |
|
|
|
switch (rr) { |
|
case (ROFF_REPARSE): |
|
if (REPARSE_LIMIT >= ++curp->reparse_count) |
|
parsebuf(curp, ln, 0); |
|
else |
|
mmsg(MANDOCERR_ROFFLOOP, curp, |
|
curp->line, pos, NULL); |
pos = 0; |
pos = 0; |
|
continue; |
|
case (ROFF_APPEND): |
|
pos = strlen(ln.buf); |
|
continue; |
|
case (ROFF_RERUN): |
|
goto rerun; |
|
case (ROFF_IGN): |
|
pos = 0; |
|
continue; |
|
case (ROFF_ERR): |
|
assert(MANDOCLEVEL_FATAL <= file_status); |
|
break; |
|
case (ROFF_SO): |
|
if (pfile(ln.buf + of, curp)) { |
|
pos = 0; |
|
continue; |
|
} else |
|
break; |
|
default: |
|
break; |
} |
} |
} |
|
|
|
if (mdoc) |
/* |
return(mdoc_endparse(mdoc)); |
* If we encounter errors in the recursive parsebuf() |
|
* call, make sure we don't continue parsing. |
|
*/ |
|
|
return(man_endparse(man)); |
if (MANDOCLEVEL_FATAL <= file_status) |
|
break; |
|
|
|
/* |
|
* If input parsers have not been allocated, do so now. |
|
* We keep these instanced betwen parsers, but set them |
|
* locally per parse routine since we can use different |
|
* parsers with each one. |
|
*/ |
|
|
|
if ( ! (curp->man || curp->mdoc)) |
|
pset(ln.buf + of, pos - of, curp); |
|
|
|
/* |
|
* Lastly, push down into the parsers themselves. One |
|
* of these will have already been set in the pset() |
|
* routine. |
|
* 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. |
|
*/ |
|
|
|
rc = -1; |
|
|
|
if (ROFF_TBL == rr) |
|
while (NULL != (span = roff_span(curp->roff))) { |
|
rc = curp->man ? |
|
man_addspan(curp->man, span) : |
|
mdoc_addspan(curp->mdoc, span); |
|
if (0 == rc) |
|
break; |
|
} |
|
else if (ROFF_EQN == rr) |
|
rc = curp->mdoc ? |
|
mdoc_addeqn(curp->mdoc, |
|
roff_eqn(curp->roff)) : |
|
man_addeqn(curp->man, |
|
roff_eqn(curp->roff)); |
|
else if (curp->man || curp->mdoc) |
|
rc = curp->man ? |
|
man_parseln(curp->man, |
|
curp->line, ln.buf, of) : |
|
mdoc_parseln(curp->mdoc, |
|
curp->line, ln.buf, of); |
|
|
|
if (0 == rc) { |
|
assert(MANDOCLEVEL_FATAL <= file_status); |
|
break; |
|
} |
|
|
|
/* Temporary buffers typically are not full. */ |
|
|
|
if (0 == start && '\0' == blk.buf[i]) |
|
break; |
|
|
|
/* Start the next input line. */ |
|
|
|
pos = 0; |
|
} |
|
|
|
free(ln.buf); |
} |
} |
|
|
|
static void |
|
pset(const char *buf, int pos, struct curparse *curp) |
|
{ |
|
int i; |
|
|
|
/* |
|
* Try to intuit which kind of manual parser should be used. If |
|
* passed in by command-line (-man, -mdoc), then use that |
|
* explicitly. If passed as -mandoc, then try to guess from the |
|
* line: either skip dot-lines, use -mdoc when finding `.Dt', or |
|
* default to -man, which is more lenient. |
|
* |
|
* Separate out pmdoc/pman from mdoc/man: the first persists |
|
* through all parsers, while the latter is used per-parse. |
|
*/ |
|
|
|
if ('.' == buf[0] || '\'' == buf[0]) { |
|
for (i = 1; buf[i]; i++) |
|
if (' ' != buf[i] && '\t' != buf[i]) |
|
break; |
|
if ('\0' == buf[i]) |
|
return; |
|
} |
|
|
|
switch (curp->inttype) { |
|
case (INTT_MDOC): |
|
if (NULL == curp->pmdoc) |
|
curp->pmdoc = mdoc_alloc |
|
(&curp->regs, curp, mmsg); |
|
assert(curp->pmdoc); |
|
curp->mdoc = curp->pmdoc; |
|
return; |
|
case (INTT_MAN): |
|
if (NULL == curp->pman) |
|
curp->pman = man_alloc |
|
(&curp->regs, curp, mmsg); |
|
assert(curp->pman); |
|
curp->man = curp->pman; |
|
return; |
|
default: |
|
break; |
|
} |
|
|
|
if (pos >= 3 && 0 == memcmp(buf, ".Dd", 3)) { |
|
if (NULL == curp->pmdoc) |
|
curp->pmdoc = mdoc_alloc |
|
(&curp->regs, curp, mmsg); |
|
assert(curp->pmdoc); |
|
curp->mdoc = curp->pmdoc; |
|
return; |
|
} |
|
|
|
if (NULL == curp->pman) |
|
curp->pman = man_alloc(&curp->regs, curp, mmsg); |
|
assert(curp->pman); |
|
curp->man = curp->pman; |
|
} |
|
|
static int |
static int |
moptions(enum intt *tflags, char *arg) |
moptions(enum intt *tflags, char *arg) |
{ |
{ |
|
|
if (0 == strcmp(arg, "mdoc")) |
if (0 == strcmp(arg, "doc")) |
*tflags = INTT_MDOC; |
*tflags = INTT_MDOC; |
else if (0 == strcmp(arg, "man")) |
else if (0 == strcmp(arg, "andoc")) |
|
*tflags = INTT_AUTO; |
|
else if (0 == strcmp(arg, "an")) |
*tflags = INTT_MAN; |
*tflags = INTT_MAN; |
else { |
else { |
warnx("bad argument: -m%s", arg); |
fprintf(stderr, "%s: Bad argument\n", arg); |
return(0); |
return(0); |
} |
} |
|
|
return(1); |
return(1); |
} |
} |
|
|
|
|
static int |
static int |
toptions(enum outt *tflags, char *arg) |
toptions(struct curparse *curp, char *arg) |
{ |
{ |
|
|
if (0 == strcmp(arg, "ascii")) |
if (0 == strcmp(arg, "ascii")) |
*tflags = OUTT_ASCII; |
curp->outtype = OUTT_ASCII; |
else if (0 == strcmp(arg, "lint")) |
else if (0 == strcmp(arg, "lint")) { |
*tflags = OUTT_LINT; |
curp->outtype = OUTT_LINT; |
|
curp->wlevel = MANDOCLEVEL_WARNING; |
|
} |
else if (0 == strcmp(arg, "tree")) |
else if (0 == strcmp(arg, "tree")) |
*tflags = OUTT_TREE; |
curp->outtype = OUTT_TREE; |
|
else if (0 == strcmp(arg, "html")) |
|
curp->outtype = OUTT_HTML; |
|
else if (0 == strcmp(arg, "xhtml")) |
|
curp->outtype = OUTT_XHTML; |
|
else if (0 == strcmp(arg, "ps")) |
|
curp->outtype = OUTT_PS; |
|
else if (0 == strcmp(arg, "pdf")) |
|
curp->outtype = OUTT_PDF; |
else { |
else { |
warnx("bad argument: -T%s", arg); |
fprintf(stderr, "%s: Bad argument\n", arg); |
return(0); |
return(0); |
} |
} |
|
|
return(1); |
return(1); |
} |
} |
|
|
|
|
/* |
|
* Parse out the options for [-fopt...] setting compiler options. These |
|
* can be comma-delimited or called again. |
|
*/ |
|
static int |
static int |
foptions(int *fflags, char *arg) |
woptions(struct curparse *curp, char *arg) |
{ |
{ |
char *v; |
char *v, *o; |
char *toks[4]; |
const char *toks[6]; |
|
|
toks[0] = "ign-scope"; |
toks[0] = "stop"; |
toks[1] = "ign-escape"; |
toks[1] = "all"; |
toks[2] = "ign-macro"; |
toks[2] = "warning"; |
toks[3] = NULL; |
|
|
|
while (*arg) |
|
switch (getsubopt(&arg, toks, &v)) { |
|
case (0): |
|
*fflags |= IGN_SCOPE; |
|
break; |
|
case (1): |
|
*fflags |= IGN_ESCAPE; |
|
break; |
|
case (2): |
|
*fflags |= IGN_MACRO; |
|
break; |
|
default: |
|
warnx("bad argument: -f%s", arg); |
|
return(0); |
|
} |
|
|
|
return(1); |
|
} |
|
|
|
|
|
/* |
|
* Parse out the options for [-Werr...], which sets warning modes. |
|
* These can be comma-delimited or called again. |
|
*/ |
|
static int |
|
woptions(int *wflags, char *arg) |
|
{ |
|
char *v; |
|
char *toks[5]; |
|
|
|
toks[0] = "all"; |
|
toks[1] = "compat"; |
|
toks[2] = "syntax"; |
|
toks[3] = "error"; |
toks[3] = "error"; |
toks[4] = NULL; |
toks[4] = "fatal"; |
|
toks[5] = NULL; |
|
|
while (*arg) |
while (*arg) { |
switch (getsubopt(&arg, toks, &v)) { |
o = arg; |
|
switch (getsubopt(&arg, UNCONST(toks), &v)) { |
case (0): |
case (0): |
*wflags |= WARN_WALL; |
curp->wstop = 1; |
break; |
break; |
case (1): |
case (1): |
*wflags |= WARN_WCOMPAT; |
/* FALLTHROUGH */ |
break; |
|
case (2): |
case (2): |
*wflags |= WARN_WSYNTAX; |
curp->wlevel = MANDOCLEVEL_WARNING; |
break; |
break; |
case (3): |
case (3): |
*wflags |= WARN_WERR; |
curp->wlevel = MANDOCLEVEL_ERROR; |
break; |
break; |
|
case (4): |
|
curp->wlevel = MANDOCLEVEL_FATAL; |
|
break; |
default: |
default: |
warnx("bad argument: -W%s", arg); |
fprintf(stderr, "-W%s: Bad argument\n", o); |
return(0); |
return(0); |
} |
} |
|
} |
|
|
return(1); |
return(1); |
} |
} |
|
|
|
|
/* ARGSUSED */ |
|
static int |
static int |
merr(void *arg, int line, int col, const char *msg) |
mmsg(enum mandocerr t, void *arg, int ln, int col, const char *msg) |
{ |
{ |
struct curparse *curp; |
struct curparse *cp; |
|
enum mandoclevel level; |
|
|
curp = (struct curparse *)arg; |
level = MANDOCLEVEL_FATAL; |
|
while (t < mandoclimits[level]) |
|
/* LINTED */ |
|
level--; |
|
|
warnx("%s:%d: error: %s (column %d)", |
cp = (struct curparse *)arg; |
curp->file, line, msg, col); |
if (level < cp->wlevel) |
return(0); |
|
} |
|
|
|
|
|
static int |
|
mdocwarn(void *arg, int line, int col, |
|
enum mdoc_warn type, const char *msg) |
|
{ |
|
struct curparse *curp; |
|
char *wtype; |
|
|
|
curp = (struct curparse *)arg; |
|
wtype = NULL; |
|
|
|
switch (type) { |
|
case (WARN_COMPAT): |
|
wtype = "compat"; |
|
if (curp->wflags & WARN_WCOMPAT) |
|
break; |
|
return(1); |
return(1); |
case (WARN_SYNTAX): |
|
wtype = "syntax"; |
|
if (curp->wflags & WARN_WSYNTAX) |
|
break; |
|
return(1); |
|
} |
|
|
|
assert(wtype); |
fprintf(stderr, "%s:%d:%d: %s: %s", |
warnx("%s:%d: %s warning: %s (column %d)", |
cp->file, ln, col + 1, mandoclevels[level], mandocerrs[t]); |
curp->file, line, wtype, msg, col); |
if (msg) |
|
fprintf(stderr, ": %s", msg); |
|
fputc('\n', stderr); |
|
|
if ( ! (curp->wflags & WARN_WERR)) |
if (file_status < level) |
return(1); |
file_status = level; |
|
|
warnx("%s: considering warnings as errors", |
return(level < MANDOCLEVEL_FATAL); |
__progname); |
|
return(0); |
|
} |
|
|
|
|
|
static int |
|
manwarn(void *arg, int line, int col, const char *msg) |
|
{ |
|
struct curparse *curp; |
|
|
|
curp = (struct curparse *)arg; |
|
|
|
if ( ! (curp->wflags & WARN_WSYNTAX)) |
|
return(1); |
|
|
|
warnx("%s:%d: syntax warning: %s (column %d)", |
|
curp->file, line, msg, col); |
|
|
|
if ( ! (curp->wflags & WARN_WERR)) |
|
return(1); |
|
|
|
warnx("%s: considering warnings as errors", |
|
__progname); |
|
return(0); |
|
} |
} |