From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ben Schmidt Date: Mon, 25 Jan 2010 11:08:56 +0000 Subject: Re: PATCH: Not me too Message-Id: <4B5D7BC8.6010201@yahoo.com.au> List-Id: References: <4B5D78F3.30600@yahoo.com.au> In-Reply-To: <4B5D78F3.30600@yahoo.com.au> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: mlmmj@mlmmj.org That bit about testing being successful...might need to change that... It doesn't work for moderated mail; my test of that just got clogged somewhere in the Internet and didn't arrive for a while! Still feel free to offer any comments about the patch, or suggestions regarding implementing this for moderated mail, but also feel free to disregard it until I come back with something a bit more functional. Cheers, Ben. On 25/01/10 9:56 PM, Ben Schmidt wrote: > The attached patch adds a 'not me too' feature to mlmmj, where post > senders are excluded from the post distribution so people don't get > copies of their own mail. A number of people, particularly those who > aren't familiar with traditional email mailing lists, prefer this. It > applies on top of the tunables patch I emailed earlier (which I > neglected to mention applies over 1.2.17). > > Some info and design decisions: > > I decided a simple boolean tunable would suffice for this, at least in > the short term, and quite probably long term, too. For now, people who > really want confirmation their mail has arrived on the list can always > subscribe a nomail email to post from. I will devote my mind to how this > could neatly be extended later if it seems warranted. > > Thanks to Franky for getting me started on this. The from address in > those low level sending functions is the envelope sender, so based on > the list address, not the sender of the mail, though. The sender of the > mail is parsed with the other mail headers in mlmmj-process. This means > I needed to communicate that information from mlmmj-process to > mlmmj-send. Commandline paramaters seemed the way most consistent with > the rest of the program, so I added an option. > > I also had to decide whether the tunable should be tested in > mlmmj-process or mlmmj-send. I opted for mlmmj-process, as that's where > most decisions about whether and where to send mail are made, and as a > general rule it's best if commandline arguments can override tunables > rather than be affected by them. Thus I added an '-o' option to omit a > specified address. The alternative is making the option specify an > originator address, and testing the tunable in mlmmj-send to see whether > or not to omit them from the distribution. > > I decided to do an extra pass through the recipients list in main() to > do the filtering rather than modify the lower level functions, as the > functions are used in numerous locations, but this filtering is only > required once, so an interface change doesn't seem warranted, and > iterating the recipients shouldn't be too slow, even for large lists. > > Comments? Questions? > > I have tested it and it works fine for me. > > Could this be included in mlmmj? > > The patch includes updating the php and perl web admin interfaces' > tunables.pl file. Should I be CCing people of that code specifically? > > Cheers, > > Ben.