Added new option --dont-mangle-from to turn off From_ mangling.

3 files changed, 37 insertions, 8 deletions

diff --git a/archivemail.1 b/archivemail.1
index 638f798..a2c68cc 100644
--- a/archivemail.1
+++ b/archivemail.1
@@ -3,7 +3,7 @@
 .\" <http://shell.ipoline.com/~elmert/comp/docbook2X/> 
 .\" Please send any bug reports, improvements, comments, patches, 
 .\" etc. to Steve Cheng <steve@ggi-project.org>.
-.TH "ARCHIVEMAIL" "1" "29 September 2006" "SP" ""
+.TH "ARCHIVEMAIL" "1" "12 October 2006" "SP" ""
 
 .SH NAME
 archivemail \- archive and compress your old email
@@ -177,6 +177,13 @@ message has been read. \fBarchivemail\fR determines if a
 been read by looking at the filename. If the filename contains an 'S' after 
 \fI:2,\fR then it assumes the message has been read.
 .TP
+\fB    --dont-mangle-from\fR
+Do not mangle lines in message bodies beginning with "From ".  Since mail user
+agents may interpret such lines as message separators, by default
+\fBarchivemail\fR mangles them by prepending a '>' to these lines (that is,
+the resulting mailbox is in \fBMBOXO\fR-format). See \fBmbox\fR(5) for more
+information.
+.TP
 \fB    --delete\fR
 Delete rather than archive old mail. Use this option with caution!
 .TP
diff --git a/archivemail.py b/archivemail.py
index 72598d7..b0acd04 100755
--- a/archivemail.py
+++ b/archivemail.py
@@ -157,6 +157,7 @@ class Options:
     output_dir           = None
     pwfile               = None
     preserve_unread      = 0
+    mangle_from          = 1
     quiet                = 0
     read_buffer_size     = 8192
     script_name          = os.path.basename(sys.argv[0])
@@ -179,7 +180,7 @@ class Options:
             opts, args = getopt.getopt(args, '?D:S:Vd:hno:F:P:qs:uv', 
                              ["date=", "days=", "delete", "dry-run", "help",
                              "include-flagged", "no-compress", "output-dir=",
-                             "filter-append=", "pwfile=",
+                             "filter-append=", "pwfile=", "dont-mangle-from",
                              "preserve-unread", "quiet", "size=", "suffix=",
                              "verbose", "version", "warn-duplicate"])
         except getopt.error, msg:
@@ -225,6 +226,8 @@ class Options:
                 self.min_size = string.atoi(a)
             if o in ('-u', '--preserve-unread'):
                 self.preserve_unread = 1
+            if o == '--dont-mangle-from':
+                self.mangle_from = 0
             if o in ('-v', '--verbose'):
                 self.verbose = 1
             if o in ('-V', '--version'):
@@ -337,14 +340,15 @@ class Mbox(mailbox.UnixMailbox):
         assert(options.read_buffer_size > 0)
         linebuf = ""
         while 1:
-            chunk = msg.fp.read(options.read_buffer_size)
-            # Be careful not to break pattern matching
-            splitindex = chunk.rfind(os.linesep)
-            body = linebuf + chunk[:splitindex]
-            linebuf = chunk[splitindex:]
+            body = msg.fp.read(options.read_buffer_size)
+            if options.mangle_from:
+                # Be careful not to break pattern matching
+                splitindex = body.rfind(os.linesep)
+                nicebody = linebuf + body[:splitindex]
+                linebuf = body[splitindex:]
+                body = from_re.sub('>From ', nicebody)
             if not body:
                 break
-            body = from_re.sub('>From ', body)
             self.mbox_file.write(body)
         self.mbox_file.write(os.linesep)
 
@@ -610,6 +614,7 @@ Options are as follows:
   -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
+      --dont-mangle-from  do not mangle From_ in message bodies
       --delete          delete rather than archive old mail (use with caution!)
       --include-flagged messages flagged important can also be archived
       --no-compress     do not compress archives with gzip
diff --git a/archivemail.sgml b/archivemail.sgml
index a7991d5..b80be54 100644
--- a/archivemail.sgml
+++ b/archivemail.sgml
@@ -27,6 +27,10 @@
 <!ENTITY crontab "<CiteRefEntry>
 <RefEntryTitle><Command/crontab/</RefEntryTitle>
 <ManVolNum/5/</CiteRefEntry>">
+
+<!ENTITY mbox "<CiteRefEntry>
+<RefEntryTitle><Command/mbox/</RefEntryTitle>
+<ManVolNum/5/</CiteRefEntry>">
 ]>
 
 <RefEntry>
@@ -244,6 +248,19 @@ been read by looking at the filename. If the filename contains an 'S' after
 
 <VarListEntry>
 <Term>
+    <Option>--dont-mangle-from</Option>
+</Term>
+<ListItem><Para>
+Do not mangle lines in message bodies beginning with "From ".  Since mail user
+agents may interpret such lines as message separators, by default
+<command/archivemail/ mangles them by prepending a '>' to these lines (that is,
+the resulting mailbox is in <application/MBOXO/-format). See &mbox; for more
+information.
+</Para></ListItem>
+</VarListEntry>
+
+<VarListEntry>
+<Term>
     <Option>--delete</Option>
 </Term>
 <ListItem><Para>