Re: [hypermail] Strange request for Hypermail!

From: Kent Landfield <kent_at_hypermail-project.org>
Date: Fri, 14 Jun 2002 14:39:43 -0500 (CDT)
Message-Id: <200206141939.OAA02023_at_rkive.landfield.com>


# >Hi. I was wondering if someone has come up with a solution for the following
# >issue:
# >
# >Each time a message is added on a page maintained by Hypermail, I want an e-
# >mail message to be sent to a defined list of users, containing only the
# >Subject and a link to that message, ie.
#
# It sounds like it's time to add a general purpose option to automate
# calling of external commands.
# Here are some possible conditions that could trigger such a call:
# - a new subdirectory is created.
# - any new message is written (this might mean a lot of calls if an archive
# is created from a large mbox).
# - a Hypermail invocation generates one or more new messages (only one call
# per Hypermail invocation).
# - a new thread is created.
# Probably the name of the file or directory which triggered the call should
# be passed as a command line argument, with additional arguments passing
# the archive name and the fields that Hypermail keeps as comments in the
# message .html files (using the info for the first message if there is more
# than one message).
# Any comments?

My concern is the security of such a capability and the system load it may place on a very busy list.

You would not want a message to generate mutiple mail messages. The conditions should be separate configuration items and a single message sent out.

As for the conditions...

# - a new subdirectory is created.

I don't see a need for this but others might...

# - any new message is written (this might mean a lot of calls if an archive
# is created from a large mbox).

If the processing is for a mailbox then it should send one message. The mail alerting fuction could be a part of the cleanup just prior to hypermail terminating.

hypermail could check the configuration for alerting on conditions. If the alerting is to be used then a temp file is created. When a configured alert condition exists, hypermail logs the specific condition's alert message to the temp file. At the end of the processing, the tempfile is sent on stdin to the specified alert "command line" and removes the tempfile.

This type of setup would allow for a single alert message to be generated for all types of invocations. Large batch processing would generate a single message. hypermail processing an inbound message on stdin would generate a single message.

# - a Hypermail invocation generates one or more new messages (only one call
# per Hypermail invocation).

?? Should need only one message per invocation.

# - a new thread is created.

Useful condition...

The command line of the alerting mail command should be a configuration item. Then it might be useful for executing other things than just email. It might be a script that updates a database, for example.

The message that is sent (via logging to the tempfile and later delivery) should use the magic cookie approach (within reason) so there is flexibility in what is sent. That way the archive admin can setup the alert facility to what their needs are.

Make sense ?

-- 
Kent Landfield                        Phone: 1-817-545-2502
Email: kent_at_landfield.com             http://www.landfield.com/
Search the Usenet FAQ Archive at http://www.faqs.org/faqs/
Search the RFC/FYI/STD/BCP Archive at http://www.faqs.org/rfcs/
Received on Fri 14 Jun 2002 09:43:39 PM GMT

This archive was generated by hypermail 2.2.0 : Thu 22 Feb 2007 07:33:54 PM GMT GMT