From e8e87978737f88a536f6606b9029ece7b1001b4e Mon Sep 17 00:00:00 2001
From: hniksic <devnull@localhost>
Date: Sun, 18 Nov 2001 17:14:14 -0800
Subject: [PATCH] [svn] Rewrite shorthand URLs in a step separate from parsing.
Published in <sxspu6f7ecz.fsf@florida.arsdigita.de>.
---
src/ChangeLog | 6 ++++++
src/main.c | 14 +++++++++---
src/url.c | 60 +++++++++++++++++++++++++++++++++++++++++++++++++++
src/url.h | 2 ++
4 files changed, 79 insertions(+), 3 deletions(-)
diff --git a/src/ChangeLog b/src/ChangeLog
index c37a399c..f375f50f 100644
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,9 @@
+2001-11-19 Hrvoje Niksic <hniksic@arsdigita.com>
+
+ * main.c (main): Use it.
+
+ * url.c (rewrite_url_maybe): New function.
+
2001-11-19 Hrvoje Niksic <hniksic@arsdigita.com>
* url.c: Clean up handling of URL schemes.
diff --git a/src/main.c b/src/main.c
index 6a79e0b0..092a455f 100644
--- a/src/main.c
+++ b/src/main.c
@@ -50,6 +50,7 @@ Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
#include "recur.h"
#include "host.h"
#include "cookies.h"
+#include "url.h"
/* On GNU system this will include system-wide getopt.h. */
#include "getopt.h"
@@ -739,9 +740,14 @@ Can't timestamp and not clobber old files at the same time.\n"));
/* Fill in the arguments. */
for (i = 0; i < nurl; i++, optind++)
{
- char *irix4_cc_needs_this;
- STRDUP_ALLOCA (irix4_cc_needs_this, argv[optind]);
- url[i] = irix4_cc_needs_this;
+ char *rewritten = rewrite_url_maybe (argv[optind]);
+ if (rewritten)
+ {
+ printf ("Converted %s to %s\n", argv[optind], rewritten);
+ url[i] = rewritten;
+ }
+ else
+ url[i] = xstrdup (argv[optind]);
}
url[i] = NULL;
@@ -853,6 +859,8 @@ Can't timestamp and not clobber old files at the same time.\n"));
convert_all_links ();
}
log_close ();
+ for (i = 0; i < nurl; i++)
+ free (url[i]);
cleanup ();
#ifdef DEBUG_MALLOC
print_malloc_debug_stats ();
diff --git a/src/url.c b/src/url.c
index c1c2b596..2159bb91 100644
--- a/src/url.c
+++ b/src/url.c
@@ -296,6 +296,66 @@ url_skip_uname (const char *url)
else
return 0;
}
+
+/* Used by main.c: detect URLs written using the "shorthand" URL forms
+ popularized by Netscape and NcFTP. HTTP shorthands look like this:
+
+ www.foo.com[:port]/dir/file -> http://www.foo.com[:port]/dir/file
+ www.foo.com[:port] -> http://www.foo.com[:port]
+
+ FTP shorthands look like this:
+
+ foo.bar.com:dir/file -> ftp://foo.bar.com/dir/file
+ foo.bar.com:/absdir/file -> ftp://foo.bar.com//absdir/file
+
+ If the URL needs not or cannot be rewritten, return NULL. */
+char *
+rewrite_url_maybe (const char *url)
+{
+ const char *p;
+
+ if (url_has_scheme (url))
+ return NULL;
+
+ /* Look for a ':' or '/'. The former signifies NcFTP syntax, the
+ latter Netscape. */
+ for (p = url; *p && *p != ':' && *p != '/'; p++)
+ ;
+
+ if (p == url)
+ return NULL;
+
+ if (*p == ':')
+ {
+ const char *pp, *path;
+ char *res;
+ /* If the characters after the colon and before the next slash
+ or end of string are all digits, it's HTTP. */
+ int digits = 0;
+ for (pp = p + 1; ISDIGIT (*pp); pp++)
+ ++digits;
+ if (digits > 0
+ && (*pp == '/' || *pp == '\0'))
+ goto http;
+
+ /* Prepend "ftp://" to the entire URL... */
+ path = p + 1;
+ res = xmalloc (6 + strlen (url) + 1);
+ sprintf (res, "ftp://%s", url);
+ /* ...and replace ':' with '/'. */
+ res[6 + (p - url)] = '/';
+ return res;
+ }
+ else
+ {
+ char *res;
+ http:
+ /* Just prepend "http://" to what we have. */
+ res = xmalloc (7 + strlen (url) + 1);
+ sprintf (res, "http://%s", url);
+ return res;
+ }
+}
/* Allocate a new urlinfo structure, fill it with default values and
return a pointer to it. */
diff --git a/src/url.h b/src/url.h
index 3c74a0e2..e9bcc379 100644
--- a/src/url.h
+++ b/src/url.h
@@ -137,4 +137,6 @@ urlpos *add_url PARAMS ((urlpos *, const char *, const char *));
downloaded_file_t downloaded_file PARAMS ((downloaded_file_t, const char *));
+char *rewrite_url_maybe PARAMS ((const char *));
+
#endif /* URL_H */