version 1.1, 2008/11/22 14:53:29 |
version 1.52, 2009/02/20 11:04:23 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
* Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se> |
* Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se> |
* |
* |
|
|
* TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR |
* TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR |
* PERFORMANCE OF THIS SOFTWARE. |
* PERFORMANCE OF THIS SOFTWARE. |
*/ |
*/ |
#include <sys/param.h> |
|
#include <sys/stat.h> |
#include <sys/stat.h> |
|
#include <sys/param.h> |
|
|
#include <assert.h> |
#include <assert.h> |
#include <err.h> |
|
#include <fcntl.h> |
#include <fcntl.h> |
|
#include <err.h> |
#include <getopt.h> |
#include <getopt.h> |
#include <stdio.h> |
#include <stdio.h> |
#include <stdlib.h> |
#include <stdlib.h> |
#include <string.h> |
#include <string.h> |
#include <unistd.h> |
#include <unistd.h> |
|
|
#include "libmdocml.h" |
#include "mdoc.h" |
|
|
struct md_file { |
#define MD_LINE_SZ (256) /* Max input line size. */ |
int fd; |
|
const char *name; |
|
}; |
|
|
|
struct md_buf { |
struct md_parse { |
struct md_file *file; |
int warn; /* Warning flags. */ |
char *buf; |
#define MD_WARN_SYNTAX (1 << 0) /* Show syntax warnings. */ |
size_t bufsz; |
#define MD_WARN_COMPAT (1 << 1) /* Show compat warnings. */ |
size_t line; |
#define MD_WARN_ALL (0x03) /* Show all warnings. */ |
|
#define MD_WARN_ERR (1 << 2) /* Make warnings->errors. */ |
|
int dbg; /* Debug level. */ |
|
struct mdoc *mdoc; /* Active parser. */ |
|
char *buf; /* Input buffer. */ |
|
u_long bufsz; /* Input buffer size. */ |
|
char *in; /* Input file name. */ |
|
int fdin; /* Input file desc. */ |
}; |
}; |
|
|
static void usage(void); |
extern char *__progname; |
|
|
static int md_begin(const char *, const char *); |
static void usage(void); |
static int md_begin_io(const char *, const char *); |
|
static int md_begin_bufs(struct md_file *, struct md_file *); |
|
static int md_run(struct md_buf *, struct md_buf *); |
|
static int md_line(struct md_buf *, const struct md_buf *, |
|
const char *, size_t); |
|
|
|
static ssize_t md_buf_fill(struct md_buf *); |
static int parse_opts(struct md_parse *, int, char *[]); |
|
static int parse_subopts(struct md_parse *, char *); |
|
|
|
static int parse_begin(struct md_parse *); |
|
static int parse_leave(struct md_parse *, int); |
|
static int io_begin(struct md_parse *); |
|
static int io_leave(struct md_parse *, int); |
|
static int buf_begin(struct md_parse *); |
|
static int buf_leave(struct md_parse *, int); |
|
|
|
static void msg_msg(void *, int, int, const char *); |
|
static int msg_err(void *, int, int, const char *); |
|
static int msg_warn(void *, int, int, |
|
enum mdoc_warn, const char *); |
|
|
|
#ifdef __linux__ |
|
extern int getsubopt(char **, char *const *, char **); |
|
#endif |
|
|
int |
int |
main(int argc, char *argv[]) |
main(int argc, char *argv[]) |
{ |
{ |
int c; |
struct md_parse parser; |
char *out, *in; |
|
|
|
extern char *optarg; |
(void)memset(&parser, 0, sizeof(struct md_parse)); |
extern int optind; |
|
|
|
out = NULL; |
if ( ! parse_opts(&parser, argc, argv)) |
|
return(EXIT_FAILURE); |
while (-1 != (c = getopt(argc, argv, "o:"))) |
if ( ! io_begin(&parser)) |
switch (c) { |
return(EXIT_FAILURE); |
case ('o'): |
|
out = optarg; |
|
break; |
|
default: |
|
usage(); |
|
return(1); |
|
} |
|
|
|
argv += optind; |
return(EXIT_SUCCESS); |
if (1 != (argc -= optind)) { |
} |
usage(); |
|
return(1); |
|
} |
|
|
|
argc--; |
|
in = *argv++; |
|
|
|
return(md_begin(out, in)); |
static int |
|
io_leave(struct md_parse *p, int code) |
|
{ |
|
|
|
if (-1 == p->fdin || STDIN_FILENO == p->fdin) |
|
return(code); |
|
|
|
if (-1 == close(p->fdin)) { |
|
warn("%s", p->in); |
|
code = 0; |
|
} |
|
return(code); |
} |
} |
|
|
|
|
static int |
static int |
md_begin(const char *out, const char *in) |
parse_subopts(struct md_parse *p, char *arg) |
{ |
{ |
char buf[MAXPATHLEN]; |
char *v; |
|
char *toks[] = { "all", "compat", |
|
"syntax", "error", NULL }; |
|
|
assert(in); |
/* |
if (out) |
* Future -Wxxx levels and so on should be here. For now we |
return(md_begin_io(out, in)); |
* only recognise syntax and compat warnings as categories, |
|
* beyond the usually "all" and "error" (make warn error out). |
|
*/ |
|
|
if (strlcpy(buf, in, MAXPATHLEN) >= MAXPATHLEN) |
while (*arg) |
warnx("output filename too long"); |
switch (getsubopt(&arg, toks, &v)) { |
else if (strlcat(buf, ".html", MAXPATHLEN) >= MAXPATHLEN) |
case (0): |
warnx("output filename too long"); |
p->warn |= MD_WARN_ALL; |
else |
break; |
return(md_begin_io(buf, in)); |
case (1): |
|
p->warn |= MD_WARN_COMPAT; |
|
break; |
|
case (2): |
|
p->warn |= MD_WARN_SYNTAX; |
|
break; |
|
case (3): |
|
p->warn |= MD_WARN_ERR; |
|
break; |
|
default: |
|
usage(); |
|
return(0); |
|
} |
|
|
return(1); |
return(1); |
} |
} |
|
|
|
|
static int |
static int |
md_begin_io(const char *out, const char *in) |
parse_opts(struct md_parse *p, int argc, char *argv[]) |
{ |
{ |
int c; |
int c; |
struct md_file fin, fout; |
|
|
|
assert(out); |
extern char *optarg; |
assert(in); |
extern int optind; |
|
|
/* XXX: put an output file in TMPDIR and switch it out when the |
p->in = "-"; |
* true file is ready to be written. |
|
*/ |
|
|
|
fin.name = in; |
while (-1 != (c = getopt(argc, argv, "vW:"))) |
|
switch (c) { |
|
case ('v'): |
|
p->dbg++; |
|
break; |
|
case ('W'): |
|
if ( ! parse_subopts(p, optarg)) |
|
return(0); |
|
break; |
|
default: |
|
usage(); |
|
return(0); |
|
} |
|
|
if (-1 == (fin.fd = open(fin.name, O_RDONLY, 0))) { |
argv += optind; |
warn("%s", fin.name); |
if (0 == (argc -= optind)) |
return(1); |
return(1); |
} |
|
|
|
fout.name = out; |
p->in = *argv++; |
|
return(1); |
|
} |
|
|
fout.fd = open(fout.name, O_WRONLY | O_CREAT | O_EXCL, 0644); |
|
if (-1 == fout.fd) { |
|
warn("%s", fout.name); |
|
if (-1 == close(fin.fd)) |
|
warn("%s", fin.name); |
|
return(1); |
|
} |
|
|
|
c = md_begin_bufs(&fout, &fin); |
static int |
|
io_begin(struct md_parse *p) |
|
{ |
|
|
if (-1 == close(fin.fd)) { |
p->fdin = STDIN_FILENO; |
warn("%s", in); |
if (0 != strncmp(p->in, "-", 1)) |
c = 1; |
if (-1 == (p->fdin = open(p->in, O_RDONLY, 0))) { |
} |
warn("%s", p->in); |
if (-1 == close(fout.fd)) { |
return(io_leave(p, 0)); |
warn("%s", out); |
} |
c = 1; |
|
} |
|
|
|
return(c); |
return(io_leave(p, buf_begin(p))); |
} |
} |
|
|
|
|
static int |
static int |
md_begin_bufs(struct md_file *out, struct md_file *in) |
buf_leave(struct md_parse *p, int code) |
{ |
{ |
struct stat stin, stout; |
|
struct md_buf inbuf, outbuf; |
|
int c; |
|
|
|
assert(in); |
if (p->buf) |
assert(out); |
free(p->buf); |
|
return(code); |
|
} |
|
|
if (-1 == fstat(in->fd, &stin)) { |
|
warn("fstat: %s", in->name); |
|
return(1); |
|
} else if (-1 == fstat(out->fd, &stout)) { |
|
warn("fstat: %s", out->name); |
|
return(1); |
|
} |
|
|
|
inbuf.file = in; |
static int |
inbuf.line = 1; |
buf_begin(struct md_parse *p) |
inbuf.bufsz = MAX(stin.st_blksize, BUFSIZ); |
{ |
|
struct stat st; |
|
|
outbuf.file = out; |
if (-1 == fstat(p->fdin, &st)) { |
outbuf.line = 1; |
warn("%s", p->in); |
outbuf.bufsz = MAX(stout.st_blksize, BUFSIZ); |
return(0); |
|
} |
|
|
if (NULL == (inbuf.buf = malloc(inbuf.bufsz))) { |
/* |
|
* Try to intuit the fastest way of sucking down buffered data |
|
* by using either the block buffer size or the hard-coded one. |
|
* This is inspired by bin/cat.c. |
|
*/ |
|
|
|
p->bufsz = MAX(st.st_blksize, BUFSIZ); |
|
|
|
if (NULL == (p->buf = malloc(p->bufsz))) { |
warn("malloc"); |
warn("malloc"); |
return(1); |
return(buf_leave(p, 0)); |
} else if (NULL == (outbuf.buf = malloc(outbuf.bufsz))) { |
|
warn("malloc"); |
|
free(inbuf.buf); |
|
return(1); |
|
} |
} |
|
|
c = md_run(&outbuf, &inbuf); |
return(buf_leave(p, parse_begin(p))); |
|
|
free(inbuf.buf); |
|
free(outbuf.buf); |
|
|
|
return(c); |
|
} |
} |
|
|
|
|
static ssize_t |
static int |
md_buf_fill(struct md_buf *in) |
parse_leave(struct md_parse *p, int code) |
{ |
{ |
ssize_t ssz; |
extern int termprint(const struct mdoc_node *, |
|
const struct mdoc_meta *); |
|
|
assert(in); |
if (NULL == p->mdoc) |
assert(in->file); |
return(code); |
assert(in->buf); |
|
assert(in->bufsz > 0); |
|
assert(in->file->name); |
|
|
|
if (-1 == (ssz = read(in->file->fd, in->buf, in->bufsz))) |
if ( ! mdoc_endparse(p->mdoc)) |
warn("%s", in->file->name); |
code = 0; |
|
|
return(ssz); |
/* TODO */ |
|
if (code && ! termprint(mdoc_node(p->mdoc), mdoc_meta(p->mdoc))) |
|
code = 0; |
|
|
|
mdoc_free(p->mdoc); |
|
return(code); |
} |
} |
|
|
|
|
static int |
static int |
md_run(struct md_buf *out, struct md_buf *in) |
parse_begin(struct md_parse *p) |
{ |
{ |
ssize_t sz, i; |
ssize_t sz, i; |
char line[BUFSIZ]; |
|
size_t pos; |
size_t pos; |
|
char line[MD_LINE_SZ]; |
|
struct mdoc_cb cb; |
|
int lnn; |
|
|
assert(in); |
cb.mdoc_err = msg_err; |
assert(out); |
cb.mdoc_warn = msg_warn; |
|
cb.mdoc_msg = msg_msg; |
|
|
/* LINTED */ |
if (NULL == (p->mdoc = mdoc_alloc(p, &cb))) |
for (pos = 0; ; ) { |
return(parse_leave(p, 0)); |
if (-1 == (sz = md_buf_fill(in))) |
|
return(1); |
/* |
else if (0 == sz) |
* This is a little more complicated than fgets. TODO: have |
|
* some benchmarks that show it's faster (note that I want to |
|
* check many, many manuals simultaneously, so speed is |
|
* important). Fill a buffer (sized to the block size) with a |
|
* single read, then parse \n-terminated lines into a line |
|
* buffer, which is passed to the parser. Hard-code the line |
|
* buffer to a particular size -- a reasonable assumption. |
|
*/ |
|
|
|
for (lnn = 1, pos = 0; ; ) { |
|
if (-1 == (sz = read(p->fdin, p->buf, p->bufsz))) { |
|
warn("%s", p->in); |
|
return(parse_leave(p, 0)); |
|
} else if (0 == sz) |
break; |
break; |
|
|
for (i = 0; i < sz; i++) { |
for (i = 0; i < sz; i++) { |
if ('\n' == in->buf[i]) { |
if ('\n' != p->buf[i]) { |
if (md_line(out, in, line, pos)) |
if (pos < sizeof(line)) { |
return(1); |
line[(int)pos++] = p->buf[(int)i]; |
in->line++; |
continue; |
pos = 0; |
} |
continue; |
warnx("%s: line %d too long", p->in, lnn); |
|
return(parse_leave(p, 0)); |
} |
} |
|
|
|
line[(int)pos] = 0; |
|
if ( ! mdoc_parseln(p->mdoc, lnn, line)) |
|
return(parse_leave(p, 0)); |
|
|
if (pos < BUFSIZ) { |
lnn++; |
/* LINTED */ |
pos = 0; |
line[pos++] = in->buf[i]; |
|
continue; |
|
} |
|
|
|
warnx("%s: line %zu too long", |
|
in->file->name, in->line); |
|
return(1); |
|
} |
} |
} |
} |
|
|
if (0 != pos) |
return(parse_leave(p, 1)); |
return(md_line(out, in, line, pos)); |
} |
|
|
|
|
|
static int |
|
msg_err(void *arg, int line, int col, const char *msg) |
|
{ |
|
struct md_parse *p; |
|
|
|
p = (struct md_parse *)arg; |
|
|
|
warnx("%s:%d: error: %s (column %d)", |
|
p->in, line, msg, col); |
return(0); |
return(0); |
} |
} |
|
|
|
|
|
static void |
|
msg_msg(void *arg, int line, int col, const char *msg) |
|
{ |
|
struct md_parse *p; |
|
|
|
p = (struct md_parse *)arg; |
|
|
|
if (0 == p->dbg) |
|
return; |
|
|
|
warnx("%s:%d: debug: %s (column %d)", |
|
p->in, line, msg, col); |
|
} |
|
|
|
|
static int |
static int |
md_line(struct md_buf *out, const struct md_buf *in, |
msg_warn(void *arg, int line, int col, |
const char *buf, size_t sz) |
enum mdoc_warn type, const char *msg) |
{ |
{ |
size_t i; |
struct md_parse *p; |
|
|
assert(buf); |
p = (struct md_parse *)arg; |
assert(out); |
|
assert(in); |
|
|
|
for (i = 0; i < sz; i++) |
switch (type) { |
(void)putchar(buf[i]); |
case (WARN_COMPAT): |
|
if (p->warn & MD_WARN_COMPAT) |
|
break; |
|
return(1); |
|
case (WARN_SYNTAX): |
|
if (p->warn & MD_WARN_SYNTAX) |
|
break; |
|
return(1); |
|
} |
|
|
(void)putchar('\n'); |
warnx("%s:%d: warning: %s (column %d)", |
|
p->in, line, msg, col); |
|
|
|
if ( ! (p->warn & MD_WARN_ERR)) |
|
return(1); |
|
|
|
warnx("%s: considering warnings as errors", __progname); |
return(0); |
return(0); |
} |
} |
|
|
Line 281 md_line(struct md_buf *out, const struct md_buf *in, |
|
Line 359 md_line(struct md_buf *out, const struct md_buf *in, |
|
static void |
static void |
usage(void) |
usage(void) |
{ |
{ |
extern char *__progname; |
|
|
|
(void)printf("usage: %s [-o outfile] infile\n", __progname); |
warnx("usage: %s [-v] [-Wwarn...] [infile]", __progname); |
} |
} |
|
|