Miroslav Trmac writes:

The 'port' setting used to be an integer, now it can be an integer
or a string, depending on the presence of IPv6 support.  fetchmailconf
assumes 'port' is an integer, though.

When 'port' is a string, it defaults to None; fetchmailconf can't
handle that and writes 'port None' to the output config file if the
server entry is not edited, see
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=55623
The attached patch fixes this.

Note that fetchmailconf still doesn't support other string values for
'port', even with this patch.
    Mirek

svn path=/trunk/; revision=4058

2 files changed, 4 insertions, 1 deletions

diff --git a/NEWS b/NEWS
index 2ae334d3..d60cbfe2 100644
--- a/NEWS
+++ b/NEWS
@@ -65,6 +65,9 @@
   manual page. UIDL isn't flaky as the man page suggested, but a
   reliability feature. In fact, IMAP4 code is flaky in that it relies on
   the upstream seen flags. (Matthias Andree)
+* Miloslav Trmac's patch for fetchmailconf to support string-type values
+  of the "port" variable, avoiding "port None" corruption in .fetchmailrc.
+  To fix Redhat Bug #55623 (Matthias Andree)
 
 fetchmail-6.2.5 (Wed Oct 15 18:39:22 EDT 2003), 23079 lines:
 
diff --git a/fetchmailconf b/fetchmailconf
index ef41a5df..b28a5d96 100755
--- a/fetchmailconf
+++ b/fetchmailconf
@@ -133,7 +133,7 @@ class Server:
 	    res = res + (" via " + str(self.via) + "\n");
 	if self.protocol != ServerDefaults.protocol:
 	    res = res + " with proto " + self.protocol 
-	if self.port != defaultports[self.protocol] and self.port != 0:
+	if self.port != defaultports[self.protocol] and self.port:
 	    res = res + " port " + `self.port`
 	if self.timeout != ServerDefaults.timeout:
 	    res = res + " timeout " + `self.timeout`