Trac: Wizard autoupgrade to unhardcode scripts3 and scripts4

[andersk]

From /mit/scripts/deploy/trac/trac.fcgi (which BTW should be in svn):

From an Athena machine, type
  ssh %s@scripts trac-admin %s upgrade --no-backup
  ssh %s@scripts trac-admin %s wiki upgrade
to upgrade, and then
  ssh %s@scripts3 pkill trac.fcgi
  ssh %s@scripts4 pkill trac.fcgi
to get this message out of the way.

[andersk]

--- trac.fcgi	2008-03-24 22:56:32.000000000 -0400
+++ trac.fcgi	2009-11-23 03:30:13.000000000 -0500
@@ -16,13 +16,13 @@
   ssh %s@scripts trac-admin %s upgrade --no-backup
   ssh %s@scripts trac-admin %s wiki upgrade
 to upgrade, and then
-  ssh %s@scripts3 pkill trac.fcgi
-  ssh %s@scripts4 pkill trac.fcgi
+  add scripts
+  for-each-server -l %s pkill trac.fcgi
 to get this message out of the way.
 
 Please ask the scripts.mit.edu maintainers for help
 if you have any trouble, at scripts@mit.edu.
-''' % (locker, env_path, locker, env_path, locker, locker)]
+''' % (locker, env_path, locker, env_path, locker)]
 
 def setup_env():
     '''Obtain the environment, handling the needs-upgrade check, and cache it.

[andersk]

Oops, /mit/scripts/deploy/trac/trac.fcgi is not the only place this needed to be fixed. We should fix the tarball and deploy an autoupgrade.

[adehnert]

Is this still an issue? I don't find any instances of scripts{3,4} in /mit/scripts/deploy.

alex@kronborg deploy [02:47] $ grep -Er "scripts[0-9]" . alex@kronborg deploy [03:01] $ pwd /afs/athena.mit.edu/contrib/scripts/deploy

[xavid]

The autoupgrade to fix this in autoinstalled tracs still needs to happen; see, e.g., ​http://xavecode.mit.edu/tailspin/ (which is /mit/xavid/web_scripts/xavecode/tailspin/trac.fcgi).

Also, we should migrate Trac to Wizard in such a way that it doesn't need this manual step at all.