In majordomo.pl, in the main'valid_addr routine, there's a stretch around
line 556 which goes:
# This mess turns "jason ti bb i tt s"@hpc.uh.edu into
# "jasontibbitts"@hpc.uh.edu
1 while $temp =~ s/\"(.*)\s(.*)\"/\"$1$2\"/g;
# This compresses space before dots or `@'s. " <- placate emacs'
highlighting
1 while $a =~ s/\s(\.|@)/$1/g;
# This compresses space after dots or `@'s.
1 while $a =~ s/(\.|@)\s/$1/g;
Correct me if I'm wrong, but shouldn't "$a" be "$temp" in the second and
third while line?
Also, I "get" the point of a rule further down:
# Look for addresses of the form rouilj@cs
if (/\@/ && !/\./) {
However, it appears this routine allows through email addresses without any
"@" sign at all. For example, if someone sends in the following request to
majordomo:
subscribe sfraves digest
majordomo will attempt to send a confirmation to the address "digest",
expanded locally to digest@hyperreal.org, which doesn't exist. This
happens in one form or another about 5 times a day for me, sometimes
"hitting" an actual person. Would there be any good reason against adding
a check for an "@" sign? I don't see a point in supporting non-domained
local email addresses.
Brian
--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
"it's a big world, with lots of records to play."-sig brian@hyperreal.org
Follow-Ups:
-
Re: hmmm..
From: Jason L Tibbitts III <tibbs@hpc.uh.edu>
|
|