[prev in list] [next in list] [prev in thread] [next in thread]
List: majordomo-workers
Subject: majordomo with config file. Now in production slot
From: "John P. Rouillard" <rouilj () terminus ! cs ! umb ! edu>
Date: 1993-11-24 2:50:24
[Download RAW message or body]
Well things seem stable enough for me to inflict my hacks on a larger
audience. I have bitten the bullet and started using my hacked
majordomo in a production slot. From now until otherwise stated,
version 1.62C (C for config) is running at majordomo@cs.umb.edu.
Among the interesting options are:
the bblisa mailing list will accept
subscribe user@foo.baz
if your address is user@grumble.stupid.subdomain.foo.baz
The test list is what I use to test my regression suites
against. The password for the list is test. The
additional commands you may want to look at are:
# to see a copy of the config file
config test test
# to input a config file copy
newconfig test test
[full config file]
EOF
# to rewrite the config file put in with a newconfig
# in normal form
writeconfig test test
Please bang on majordomo. The following is the output of the lists
command:
>>>> lists
Majordomo@cs.umb.edu serves the following lists:
bblisa A list description
bblisa-admin The Back Bay LISA Administrators list.
bblisa-announce The Back Bay LISA announcement list
test This is a test, it is only a test
Use the 'info <list>' command to get more information
about a specific list.
Oops, that reminds me if you have gotten a copy of my changes, I have
an additional patch that you really want. I goofed when I rewrote the
config file in normal form. The previous normal form for array valued
keywords was:
keyword << EOF
stuff
EOF
Sadly I forgot that EOF is the end marker for the config file, so the
config file installation aborted prematurely on the first array entry
leaving an incomplete config file. (The string 'A list description')
is the default value for the description keyword.) I had to use the
trapdoor password in bblisa.passwd to fix my mistake with a
writeconfig, and then I made the config file changes over again. I
have changed the marker to 'END' in the current set of patches. This
shouldn't conflict with any other use of END since the command scanner
is not parsing the contents of the config file. It may cause a problem
if the newconfig commands isn't recognized, but in that case your
request is kind of toast anyway, so better to bail before you cause
too many problems 8-).
Also, this brought up an interesting point. I am thinking about
modifying the way new{config,info} works so that it returns the text
you gave it in case of an error. This would be helpful when say an
info file is trashed because of a disk space crisis etc, just edit out
the quoting text, and resend. Also it helps when you have a number of
changes that were made in the config file, but you screwed up the
syntax for the unpteenth time (you would think I could master the
syntax of here documents by now wouldn't you.) Comments?
-- John
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic