Boolean validation.

[mandoc.git] / mdocml.c

diff --git a/mdocml.c b/mdocml.c
index df834d90d0e0ea73c6ad7960bf1cda4ee33febd7..2b7c4388fb4f9bde9142d0a3f9ddaddeae43d20f 100644 (file)
--- a/mdocml.c
+++ b/mdocml.c
@@ -1,4 +1,4 @@
-/* $Id: mdocml.c,v 1.8 2008/11/23 22:30:53 kristaps Exp $ */
+/* $Id: mdocml.c,v 1.46 2009/01/16 14:15:12 kristaps Exp $ */

 /*
  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
  *
 /*
  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
  *


@@ -16,49 +16,103 @@
  * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
  * PERFORMANCE OF THIS SOFTWARE.
  */
  * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
  * 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 <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <unistd.h>
 
 #include <getopt.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <unistd.h>
 

-#include "libmdocml.h"
+#include "mdoc.h"
+
+#define        xfprintf        (void)fprintf
+#define        xprintf         (void)printf
+#define        xvfprintf       (void)fvprintf
+
+#define        MD_LINE_SZ      (256)           /* Max input line size. */
+
+struct md_parse {
+       int              warn;          /* Warning flags. */
+#define        MD_WARN_SYNTAX  (1 << 0)        /* Show syntax warnings. */
+#define        MD_WARN_COMPAT  (1 << 1)        /* Show compat warnings. */
+#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            *name;          /* Input file name. */
+       int              fd;            /* Input file desc. */
+};
+
+extern char            *__progname;
+
+static void             usage(void);
+
+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);

 
 

-#define        BUFFER_IN_DEF   BUFSIZ
-#define        BUFFER_OUT_DEF  BUFSIZ
+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 *);

 
 

-static void             usage(void);
-static int              begin_io(const struct md_args *, 
-                               char *, char *);
-static int              leave_io(const struct md_buf *, 
-                               const struct md_buf *, int);
-static int              begin_bufs(const struct md_args *,
-                               struct md_buf *, struct md_buf *);
-static int              leave_bufs(const struct md_buf *, 
-                               const struct md_buf *, int);
+#ifdef __linux__
+extern int              getsubopt(char **, char *const *, char **);
+#endif

 
 int
 main(int argc, char *argv[])
 {
        int              c;
 
 int
 main(int argc, char *argv[])
 {
        int              c;

-       char            *out, *in;
-       struct md_args   args;
+       struct md_parse  parser;
+       char            *opts, *v;
+#define ALL             0
+#define COMPAT          1
+#define SYNTAX          2
+#define ERROR           3
+       char            *toks[] = { "all", "compat", "syntax", 
+                                   "error", NULL };

 
        extern char     *optarg;
        extern int       optind;
 
 
        extern char     *optarg;
        extern int       optind;
 

-       out = in = NULL;
-       
-       while (-1 != (c = getopt(argc, argv, "o:")))
+       (void)memset(&parser, 0, sizeof(struct md_parse));
+
+       while (-1 != (c = getopt(argc, argv, "vW:")))

                switch (c) {
                switch (c) {

-               case ('o'):
-                       out = optarg;
+               case ('v'):
+                       parser.dbg++;
+                       break;
+               case ('W'):
+                       opts = optarg;
+                       while (*opts) 
+                               switch (getsubopt(&opts, toks, &v)) {
+                               case (ALL):
+                                       parser.warn |= MD_WARN_ALL;
+                                       break;
+                               case (COMPAT):
+                                       parser.warn |= MD_WARN_COMPAT;
+                                       break;
+                               case (SYNTAX):
+                                       parser.warn |= MD_WARN_SYNTAX;
+                                       break;
+                               case (ERROR):
+                                       parser.warn |= MD_WARN_ERR;
+                                       break;
+                               default:
+                                       usage();
+                                       return(1);
+                               }

                        break;
                default:
                        usage();
                        break;
                default:
                        usage();


@@ -68,133 +122,281 @@ main(int argc, char *argv[])
        argv += optind;
        argc -= optind;
 
        argv += optind;
        argc -= optind;
 

+       parser.name = "-";

        if (1 == argc)
        if (1 == argc)

-               in = *argv++;
+               parser.name = *argv++;
+
+       if ( ! io_begin(&parser))
+               return(EXIT_FAILURE);
+
+       return(EXIT_SUCCESS);
+}
+
+
+static int
+io_leave(struct md_parse *p, int code)
+{
+
+       if (-1 == p->fd || STDIN_FILENO == p->fd)
+               return(code);
+
+       if (-1 == close(p->fd)) {
+               warn("%s", p->name);
+               code = 0;
+       }
+       return(code);
+}

 
 

-       args.type = MD_HTML4_STRICT;
-       args.dbg = MD_DBG_TREE;

 
 

-       return(begin_io(&args, out ? out : "-", in ? in : "-"));
+static int
+io_begin(struct md_parse *p)
+{
+
+       p->fd = STDIN_FILENO;
+       if (0 != strncmp(p->name, "-", 1))
+               if (-1 == (p->fd = open(p->name, O_RDONLY, 0))) {
+                       warn("%s", p->name);
+                       return(io_leave(p, 0));
+               }
+
+       return(io_leave(p, buf_begin(p)));

 }
 
 
 static int
 }
 
 
 static int

-leave_io(const struct md_buf *out, 
-               const struct md_buf *in, int c)
+buf_leave(struct md_parse *p, int code)

 {
 {

-       assert(out);
-       assert(in);

 
 

-       if (-1 != in->fd && -1 == close(in->fd)) {
-               assert(in->name);
-               warn("%s", in->name);
-               c = 1;
+       if (p->buf)
+               free(p->buf);
+       return(code);
+}
+
+
+static int
+buf_begin(struct md_parse *p)
+{
+       struct stat      st;
+
+       if (-1 == fstat(p->fd, &st)) {
+               warn("%s", p->name);
+               return(1);
+       } 
+
+       p->bufsz = MAX(st.st_blksize, BUFSIZ);
+
+       if (NULL == (p->buf = malloc(p->bufsz))) {
+               warn("malloc");
+               return(buf_leave(p, 0));
+       }
+
+       return(buf_leave(p, parse_begin(p)));
+}
+
+
+/* TODO: remove this to a print-tree output filter. */
+static void
+print_node(const struct mdoc_node *n, int indent)
+{
+       const char       *p, *t;
+       int               i, j;
+       size_t            argc, sz;
+       char            **params;
+       struct mdoc_arg  *argv;
+
+       argv = NULL;
+       argc = sz = 0;
+       params = NULL;
+
+       t = mdoc_type2a(n->type);
+
+       switch (n->type) {
+       case (MDOC_TEXT):
+               p = n->data.text.string;
+               break;
+       case (MDOC_BODY):
+               p = mdoc_macronames[n->tok];
+               break;
+       case (MDOC_HEAD):
+               p = mdoc_macronames[n->tok];
+               break;
+       case (MDOC_TAIL):
+               p = mdoc_macronames[n->tok];
+               break;
+       case (MDOC_ELEM):
+               p = mdoc_macronames[n->tok];
+               argv = n->data.elem.argv;
+               argc = n->data.elem.argc;
+               break;
+       case (MDOC_BLOCK):
+               p = mdoc_macronames[n->tok];
+               argv = n->data.block.argv;
+               argc = n->data.block.argc;
+               break;
+       case (MDOC_ROOT):
+               p = "root";
+               break;
+       default:
+               abort();
+               /* NOTREACHED */

        }
        }

-       if (-1 != out->fd && STDOUT_FILENO != out->fd &&
-                       -1 == close(out->fd)) {
-               assert(out->name);
-               warn("%s", out->name);
-               c = 1;
+
+       for (i = 0; i < indent; i++)
+               xprintf("    ");
+       xprintf("%s (%s)", p, t);
+
+       for (i = 0; i < (int)argc; i++) {
+               xprintf(" -%s", mdoc_argnames[argv[i].arg]);
+               if (argv[i].sz > 0)
+                       xprintf(" [");
+               for (j = 0; j < (int)argv[i].sz; j++)
+                       xprintf(" [%s]", argv[i].value[j]);
+               if (argv[i].sz > 0)
+                       xprintf(" ]");

        }
 
        }
 

-       return(c);
+       for (i = 0; i < (int)sz; i++)
+               xprintf(" [%s]", params[i]);
+
+       xprintf(" %d:%d\n", n->line, n->pos);
+
+       if (n->child)
+               print_node(n->child, indent + 1);
+       if (n->next)
+               print_node(n->next, indent);

 }
 
 
 static int
 }
 
 
 static int

-begin_io(const struct md_args *args, char *out, char *in)
+parse_leave(struct md_parse *p, int code)

 {
 {

-       struct md_buf    fi;
-       struct md_buf    fo;
+       const struct mdoc_node *n;

 
 

-#define        FI_FL   O_RDONLY
-#define        FO_FL   O_WRONLY|O_CREAT|O_TRUNC
+       if (NULL == p->mdoc)
+               return(code);

 
 

-       assert(args);
-       assert(out);
-       assert(in);
+       if ( ! mdoc_endparse(p->mdoc))
+               code = 0;
+       if ((n = mdoc_result(p->mdoc)))
+               print_node(n, 0);

 
 

-       bzero(&fi, sizeof(struct md_buf));
-       bzero(&fo, sizeof(struct md_buf));
+       mdoc_free(p->mdoc);

 
 

-       fi.fd = STDIN_FILENO;
-       fo.fd = STDOUT_FILENO;
+       return(code);
+}

 
 

-       fi.name = in;
-       fo.name = out;

 
 

-       if (0 != strncmp(fi.name, "-", 1))
-               if (-1 == (fi.fd = open(fi.name, FI_FL, 0))) {
-                       warn("%s", fi.name);
-                       return(leave_io(&fo, &fi, 1));
-               }
+static int
+parse_begin(struct md_parse *p)
+{
+       ssize_t          sz, i;
+       size_t           pos;
+       char             line[MD_LINE_SZ];
+       struct mdoc_cb   cb;
+       int              lnn;
+
+       cb.mdoc_err = msg_err;
+       cb.mdoc_warn = msg_warn;
+       cb.mdoc_msg = msg_msg;
+
+       if (NULL == (p->mdoc = mdoc_alloc(p, &cb)))
+               return(parse_leave(p, 0));
+
+       for (lnn = 1, pos = 0; ; ) {
+               if (-1 == (sz = read(p->fd, p->buf, p->bufsz))) {
+                       warn("%s", p->name);
+                       return(parse_leave(p, 0));
+               } else if (0 == sz) 
+                       break;
+
+               for (i = 0; i < sz; i++) {
+                       if ('\n' != p->buf[i]) {
+                               if (pos < sizeof(line)) {
+                                       line[(int)pos++] = p->buf[(int)i];
+                                       continue;
+                               }
+                               warnx("%s: line %d too long", 
+                                               p->name, lnn);
+                               return(parse_leave(p, 0));
+                       }
+       
+                       line[(int)pos] = 0;
+                       if ( ! mdoc_parseln(p->mdoc, lnn, line))
+                               return(parse_leave(p, 0));

 
 

-       if (0 != strncmp(fo.name, "-", 1)) 
-               if (-1 == (fo.fd = open(fo.name, FO_FL, 0644))) {
-                       warn("%s", fo.name);
-                       return(leave_io(&fo, &fi, 1));
+                       lnn++;
+                       pos = 0;

                }
                }

+       }

 
 

-       return(leave_io(&fo, &fi, begin_bufs(args, &fo, &fi)));
+       return(parse_leave(p, 1));

 }
 
 
 static int
 }
 
 
 static int

-leave_bufs(const struct md_buf *out, 
-               const struct md_buf *in, int c)
+msg_err(void *arg, int line, int col, const char *msg)

 {
 {

-       assert(out);
-       assert(in);
-       if (out->buf)
-               free(out->buf);
-       if (in->buf)
-               free(in->buf);
-       return(c);
+       struct md_parse  *p;
+
+       p = (struct md_parse *)arg;
+
+       xfprintf(stderr, "%s:%d: error: %s (column %d)\n", 
+                       p->name, line, msg, col);
+       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;
+
+       xfprintf(stderr, "%s:%d: debug: %s (column %d)\n", 
+                       p->name, line, msg, col);

 }
 
 
 static int
 }
 
 
 static int

-begin_bufs(const struct md_args *args, 
-               struct md_buf *out, struct md_buf *in)
+msg_warn(void *arg, int line, int col, 
+               enum mdoc_warn type, const char *msg)

 {
 {

-       struct stat      stin, stout;
-       int              c;
+       struct md_parse  *p;

 
 

-       assert(args);
-       assert(in);
-       assert(out);
+       p = (struct md_parse *)arg;

 
 

-       if (-1 == fstat(in->fd, &stin)) {
-               warn("%s", in->name);
-               return(1);
-       } else if (0 == stin.st_size) {
-               warnx("%s: empty file", in->name);
+       switch (type) {
+       case (WARN_COMPAT):
+               if (p->warn & MD_WARN_COMPAT)
+                       break;

                return(1);
                return(1);

-       } else if (-1 == fstat(out->fd, &stout)) {
-               warn("%s", out->name);
+       case (WARN_SYNTAX):
+               if (p->warn & MD_WARN_SYNTAX)
+                       break;

                return(1);
        }
 
                return(1);
        }
 

-       in->bufsz = MAX(stin.st_blksize, BUFFER_IN_DEF);
-       out->bufsz = MAX(stout.st_blksize, BUFFER_OUT_DEF);
+       xfprintf(stderr, "%s:%d: warning: %s (column %d)\n", 
+                       p->name, line, msg, col);

 
 

-       if (NULL == (in->buf = malloc(in->bufsz))) {
-               warn("malloc");
-               return(leave_bufs(out, in, 1));
-       } else if (NULL == (out->buf = malloc(out->bufsz))) {
-               warn("malloc");
-               return(leave_bufs(out, in, 1));
-       }
+       if ( ! (p->warn & MD_WARN_ERR))
+               return(1);

 
 

-       c = md_run(args, out, in);
-       return(leave_bufs(out, in, -1 == c ? 1 : 0));
+       xfprintf(stderr, "%s: considering warnings as errors\n", 
+                       __progname);
+       return(0);

 }
 
 
 static void
 usage(void)
 {
 }
 
 
 static void
 usage(void)
 {

-       extern char     *__progname;

 
 

-       (void)printf("usage: %s [-o outfile] [infile]\n", __progname);
+       xfprintf(stderr, "usage: %s [-v] [-Wwarn...] [infile]\n",
+                       __progname);

 }
 }

+