* added some points to the roadmap.

* check for citadel.sock before we try to shutdown citadel on upgrade.

diff --git a/citadel/debian/citadel-server.postinst b/citadel/debian/citadel-server.postinst
index 8c9b1de97b4821d4463dd1c18efb3f948b84e7bc..f75e77da014894a783b3c4ac57181552ddc045c1 100644 (file)
--- a/citadel/debian/citadel-server.postinst
+++ b/citadel/debian/citadel-server.postinst
@@ -182,7 +182,9 @@ case "$1" in
 #no
 #" | \
        /usr/lib/citadel-server/setup -q
-       /usr/sbin/sendcommand "DOWN"
+       if test -S /var/run/citadel/citadel.socket; then 
+           /usr/sbin/sendcommand "DOWN"
+       fi
        killall -HUP init
     #### There are three sub-cases:
     ###if test "${2+set}" != set; then

diff --git a/citadel/debian/citadel-server.substvars b/citadel/debian/citadel-server.substvars
index 7dbe5f8efae722248df006b69f186b51c00e512d..a8fe3cff5d20eaabf0bae27f6105caf01ec07c70 100644 (file)
--- a/citadel/debian/citadel-server.substvars
+++ b/citadel/debian/citadel-server.substvars
@@ -1,2 +1 @@
 misc:Depends=debconf (>= 0.5) | debconf-2.0
-shlibs:Depends=libc6 (>= 2.3.6-6), libdb4.3 (>= 4.3.28-1), libldap2 (>= 2.1.17-1), libncurses5 (>= 5.4-5), libssl0.9.8 (>= 0.9.8b-1), zlib1g (>= 1:1.2.1)

diff --git a/citadel/techdoc/roadmap.txt b/citadel/techdoc/roadmap.txt
index e1adf4e334d284f1e1f33d2b0a25c0e612176244..7408fa02853a5463bc0e5d7e593dbf20480763ce 100644 (file)
--- a/citadel/techdoc/roadmap.txt
+++ b/citadel/techdoc/roadmap.txt
@@ -56,3 +56,25 @@ Various Whishlist
 
 * backtraces in the log in case of panics, etc.
 
+* auto-login should move from compile to installtime.
+
+* scheduled restarts there should be a way to notify all users 
+   (whether via imap or webcit) that citadel is going to be unavailable.
+   
+* there should be a way to set a 'service unavailable because of' message, 
+  for people connecting via webcit, so they can see this is a scheduled 
+  downtime (so they don't take further actions like anoying the admin
+  (you) via phone or restart citadel...
+
+* there should be a way to keep citadel from starting while you're db recovering etc.
+  (perhaps something like the pre-shutdown state would be cool, with a 
+   message? so you tell it "sysadmin is reworking the database" as param
+   to the server, it will close the db, but keep open the port, and 
+   print out your message as error message. this should last until citserver
+   receives a kill signal, and then terminates.)
+
+* citserver should bypass some bdb errors and symptoms. it should remember 
+  the actual berkeley db version in its state file, and in case of an up/downgrade
+  moan this to the user / logfile. bdb upgrades should trigger a mail to aide.
+
+