[Bug-tar] [PATCH/RFC] tar: read access to remote archives via HTTP

[Jan Stępień]

This change is an incomplete attempt to allow tar to have read access to
archives published on HTTP servers using wget.
* src/buffer.c (_open_archive): start wget using popen.
* configure.ac: add --with-wget and check for its presence.

---

The main aim of this patch is to reduce the number of keystrokes. Instead
of 'wget http://example.org/foo.tar && tar -xf foo.tar' now you're able to
'tar -xf http://example.org/foo.tar'. I'd be grateful for your comments
regarding this idea.

The patch is far from being complete but it offers some basic functionality.
So far it works only for uncompressed tarballs. I'm a novice when it comes to
tar's code and GNU coding style so please let me know if I'm doing it wrong.

Best regards,
Jan Stępień

---
 configure.ac |    5 +++++
 src/buffer.c |   29 +++++++++++++++++++++++++++++
 2 files changed, 34 insertions(+), 0 deletions(-)

diff --git a/configure.ac b/configure.ac
index c3a3af7..8ff71fe 100644
--- a/configure.ac
+++ b/configure.ac
@@ -243,6 +243,11 @@ AC_ARG_ENABLE(backup-scripts,
                     ;;
                esac])
 
+AC_ARG_WITH([wget],
+            AC_HELP_STRING([--with-wget], [use wget for remote file fetching]))
+AS_IF([test "x$with_wget" = "xyes"],
+      [AC_DEFINE([WITH_WGET], [1], [Should we use wget.])])
+
 AC_SUBST(BACKUP_SED_COND)
 if date +%Y-%m-%d 2>/dev/null >&2; then
        
BACKUP_SED_COND='/^\#ELSE_DATE_FORMAT_OK/,/^\#ENDIF_DATE_FORMAT_OK/d;/^\#IF_DATE_FORMAT_OK/d'
diff --git a/src/buffer.c b/src/buffer.c
index ef4bbac..7f11d0d 100644
--- a/src/buffer.c
+++ b/src/buffer.c
@@ -32,6 +32,7 @@
 
 #include "common.h"
 #include <rmt.h>
+#include <config.h>
 
 /* Number of retries before giving up on read.  */
 #define READ_ERROR_MAX 10
@@ -651,6 +652,34 @@ _open_archive (enum access_mode wanted_access)
           break;
         }
     }
+#ifdef WITH_WGET
+  else if (!strncmp (archive_name_array[0], "http://";, strlen("http://";)) ||
+           !strncmp (archive_name_array[0], "https://";, strlen("https://";)))
+    {
+      if (wanted_access == ACCESS_READ)
+        {
+          bool shortfile;
+          enum compress_type type;
+          char* command = alloca(sizeof(char) * (strlen("wget -qO- ") +
+                strlen(archive_name_array[0]) + 1));
+          sprintf(command, "wget -qO- %s", archive_name_array[0]);
+          archive = fileno(popen(command, "r"));
+          read_full_records = true;
+
+          type = check_compressed_archive (&shortfile);
+          if (type != ct_tar && type != ct_none)
+            FATAL_ERROR ((0, 0,
+                          _("Archive is compressed. Use %s option"),
+                          compress_option (type)));
+          if (shortfile)
+            ERROR ((0, 0, _("This does not look like a tar archive")));
+        }
+      else
+        {
+          ERROR ((0, 0, _("Remote archives are available only for reading")));
+        }
+    }
+#endif /* WITH_WGET */
   else if (verify_option)
     archive = rmtopen (archive_name_array[0], O_RDWR | O_CREAT | O_BINARY,
                        MODE_RW, rsh_command_option);
-- 
1.7.0.4