Added a --size option to only archive messages over a certain byte size.

CHANGELOG

@@ -4,6 +4,8 @@ Version 0.4.5   - 28 April 2002
     an mbox mailbox you would get a python error.
   * Added a lot more testing to test_archivemail.py (test_archivemail.py is
     now 37% bigger than archivemail -- scary)
+  * Added a new '--size' option which allows you to only archive messages 
+    over a certain size.
 
 Version 0.4.4   - 27 April 2002
   * Fixed a bug where the long --suffix option was not working (although the 

Makefile

@@ -1,5 +1,5 @@
 
-VERSION=0.4.4
+VERSION=0.4.5
 VERSION_TAG=v$(subst .,_,$(VERSION))
 TARFILE=archivemail-$(VERSION).tar.gz
 

TODO

@@ -1,16 +1,14 @@
 
-Goals for next minor release (0.4.5):
+Goals for next minor release (0.4.6):
 -------------------------------------
 * Think about the best way to specify the names of archives created with
   possibly an --archive-name option.
-* Add a lot more tests (see top of test_archivemail.py)
+* Add more tests (see top of test_archivemail.py)
 * We need some better checking to see if we are really looking at a valid
   mbox-format mailbox.
 
 Goals for next major release (0.5.0):
 -------------------------------------
-* Add a --size option to only archive messages greater than a certain byte
-  size.
 * Lock any original .gz files 
   - is this necessary?
 * Check for symlink attacks for tempfiles (although we don't use /var/tmp)

archivemail.py

@@ -22,7 +22,7 @@ Website: http://archivemail.sourceforge.net/
 """
 
 # global administrivia 
-__version__ = "archivemail v0.4.4"
+__version__ = "archivemail v0.4.5"
 __cvs_id__ = "$Id$"
 __copyright__ = """Copyright (C) 2002  Paul Rodger <paul@paulrodger.com>
 This is free software; see the source for copying conditions. There is NO
@@ -147,6 +147,7 @@ class Options:
     quiet                = 0
     read_buffer_size     = 8192
     script_name          = os.path.basename(sys.argv[0])
+    min_size             = None
     verbose              = 0
     warn_duplicates      = 0
 
@@ -162,11 +163,11 @@ class Options:
 
         """
         try:
-            opts, args = getopt.getopt(args, '?D:Vd:hno:qs:uv', 
+            opts, args = getopt.getopt(args, '?D:S:Vd:hno:qs:uv', 
                              ["date=", "days=", "delete", "dry-run", "help",
-                             "include-flagged", "no-compress", "output-dir=", 
-                             "preserve-unread", "quiet", "suffix=", "verbose", 
-                             "version", "warn-duplicate"])
+                             "include-flagged", "no-compress", "output-dir=",
+                             "preserve-unread", "quiet", "size=", "suffix=",
+                             "verbose", "version", "warn-duplicate"])
         except getopt.error, msg:
             user_error(msg)
 
@@ -202,6 +203,8 @@ class Options:
                 self.quiet = 1
             if o in ('-s', '--suffix'):
                 self.archive_suffix = a
+            if o in ('-S', '--size'):
+                self.min_size = string.atoi(a)
             if o in ('-u', '--preserve-unread'):
                 self.preserve_unread = 1
             if o in ('-v', '--verbose'):
@@ -224,9 +227,11 @@ class Options:
                 unexpected_error(("output directory is world-writable: " + \
                     "%s -- I feel nervous!") % self.output_dir)
         if self.days_old_max < 1:
-            user_error("argument to -d must be greater than zero")
+            user_error("--days argument must be greater than zero")
         if self.days_old_max >= 10000:
-            user_error("argument to -d must be less than 10000")
+            user_error("--days argument must be less than 10000")
+        if self.min_size is not None and self.min_size < 1:
+            user_error("--size argument must be greater than zero")
         if self.quiet and self.verbose:
             user_error("you cannot use both the --quiet and --verbose options")
 
@@ -563,6 +568,7 @@ Options are as follows:
   -D, --date=DATE       archive messages older than DATE
   -o, --output-dir=DIR  directory to store archives (default: same as original)
   -s, --suffix=NAME     suffix for archive filename (default: '%s')
+  -S, --size=NUM        only archive messages NUM bytes or larger
   -n, --dry-run         don't write to anything - just show what would be done
   -u, --preserve-unread never archive unread messages
       --delete          delete rather than archive old mail (use with caution!)
@@ -780,6 +786,42 @@ def is_unread(message):
     return 1
 
 
+def is_smaller(message, size):
+    """Return true if the message is smaller than size bytes, false otherwise"""
+    assert(message)
+    assert(size > 0)
+    file_name = None
+    message_size = None
+    try:
+        file_name = message.fp.name
+    except AttributeError:
+        pass
+    if file_name:
+        # with maildir and MH mailboxes, we can just use the file size
+        message_size = os.path.getsize(file_name)
+    else:
+        # with mbox mailboxes, not so easy
+        message_size = 0
+        if message.unixfrom:
+            message_size = message_size + len(message.unixfrom)
+        for header in message.headers:
+            message_size = message_size + len(header)
+        message_size = message_size + 1 # the blank line after the headers
+        start_offset = message.fp.tell()
+        message.fp.seek(0, 2) # seek to the end of the message
+        end_offset = message.fp.tell()
+        message.rewindbody()
+        message_size = message_size + (end_offset - start_offset)
+    if message_size < size:
+        vprint("message is too small (%d bytes), minimum bytes : %d" % \
+            (message_size, size))
+        return 1
+    else:
+        vprint("message is not too small (%d bytes), minimum bytes: %d" % \
+            (message_size, size))
+        return 0
+
+
 def should_archive(message):
     """Return true if we should archive the message, false otherwise"""
     old = 0
@@ -795,6 +837,8 @@ def should_archive(message):
         return 0
     if not options.include_flagged and is_flagged(message):
         return 0
+    if options.min_size and is_smaller(message, options.min_size):
+        return 0
     if options.preserve_unread and is_unread(message):
         return 0
     return 1

setup.py

@@ -19,7 +19,7 @@ check_python_version()  # define & run this early - 'distutils.core' is new
 from distutils.core import setup
 
 setup(name="archivemail",
-      version="0.4.4",
+      version="0.4.5",
       description="archive and compress old email",
       license="GNU GPL",
       url="http://archivemail.sourceforge.net/",

test_archivemail.py

@@ -1216,6 +1216,75 @@ class TestArchiveMboxUncompressed(unittest.TestCase):
                 os.remove(name)
 
 
+class TestArchiveSize(unittest.TestCase):
+    """check that the 'size' argument works"""
+    def testSmaller(self):
+        """giving a size argument smaller than the message"""
+        for execute in ("package", "system_long", "system_short"):
+            self.mbox_name = make_mbox(messages=1, hours_old=(24 * 181))
+            size_arg = os.path.getsize(self.mbox_name) - 1
+            self.copy_name = tempfile.mktemp()
+            shutil.copyfile(self.mbox_name, self.copy_name)
+            if execute == "package":
+                archivemail.options.quiet = 1
+                archivemail.options.min_size = size_arg
+                archivemail.archive(self.mbox_name)
+            elif execute == "system_long":
+                run = "./archivemail.py --size=%d --quiet %s" % \
+                    (size_arg, self.mbox_name)
+                self.assertEqual(os.system(run), 0)
+            elif execute == "system_short":
+                run = "./archivemail.py -S%d --quiet %s" % \
+                    (size_arg, self.mbox_name)
+                self.assertEqual(os.system(run), 0)
+            else:
+                sys.exit(1)
+            assert(os.path.exists(self.mbox_name))
+            self.assertEqual(os.path.getsize(self.mbox_name), 0)
+            archive_name = self.mbox_name + "_archive.gz"
+            assert(os.path.exists(archive_name))
+            self.assertEqual(os.system("gzip -d %s" % archive_name), 0)
+            archive_name = self.mbox_name + "_archive"
+            assert(os.path.exists(archive_name))
+            assert(filecmp.cmp(archive_name, self.copy_name, shallow=0))
+            self.tearDown()
+
+    def testBigger(self):
+        """giving a size argument bigger than the message"""
+        for execute in ("package", "system_long", "system_short"):
+            self.mbox_name = make_mbox(messages=1, hours_old=(24 * 181))
+            size_arg = os.path.getsize(self.mbox_name) + 1
+            self.copy_name = tempfile.mktemp()
+            shutil.copyfile(self.mbox_name, self.copy_name)
+            if execute == "package":
+                archivemail.options.quiet = 1
+                archivemail.options.min_size = size_arg
+                archivemail.archive(self.mbox_name)
+            elif execute == "system_long":
+                run = "./archivemail.py --size=%d --quiet %s" % \
+                    (size_arg, self.mbox_name)
+                self.assertEqual(os.system(run), 0)
+            elif execute == "system_short":
+                run = "./archivemail.py -S%d --quiet %s" % \
+                    (size_arg, self.mbox_name)
+                self.assertEqual(os.system(run), 0)
+            else:
+                sys.exit(1)
+            assert(os.path.exists(self.mbox_name))
+            assert(filecmp.cmp(self.mbox_name, self.copy_name, shallow=0))
+            archive_name = self.mbox_name + "_archive.gz"
+            assert(not os.path.exists(archive_name))
+            self.tearDown()
+
+    def tearDown(self):
+        archivemail.options.quiet = 0
+        archivemail.options.min_size = None
+        archive = self.mbox_name + "_archive"
+        for name in (self.mbox_name, self.copy_name, archive, archive + ".gz"):
+            if name and os.path.exists(name):
+                os.remove(name)
+
+
 class TestArchiveMboxMode(unittest.TestCase):
     """file mode (permissions) of the original mbox should be preserved"""
     def testOld(self):