On Fri, 25 Aug 2000, David T-G wrote:
> pub 1024D/AFEFC23B 2000-06-29 jimh <[EMAIL PROTECTED]>
> pub 1024D/401A068F 1998-09-03 PHXMGNT <[EMAIL PROTECTED]>
> I did my same trick of commenting out the greps and seds, and this
> time the "problem" showed up at only the second line.
> In fact, it seems on second look that m_gpg *does* give me all of my
> entries. Further digging reveals that both key addresses end up in
> $collection after lbdbq calls m_gpg_query and that they get stripped
> by munge and munge-keeporder.
That's the point.
> I'm not exactly sure what's going on in munge-keeporder, but munge
> appears to be building an array (named "line") indexed by email
> address -- only.
Yes. This is a feature :-)
My understanding of a mail address is that it is unique to a user
(with a user name). So deduping means to reduce multiple lines with
the same mail address but different user names to one line. Otherwise
you usually have multiple lines like this:
foo@bar User Name
foo@bar "User Name"
foo@bar User M. Name
foo@bar "User M. Name"
foo@bar user name
foo@bar User M . Name
foo@bar "User M . Name"
Or something like this, which doesn't make much sense. For this the
munge algorithm reduces all lines for one mail address to one line
(with a randomly chosen real name).
> Have I stumbled on a limitation of lbdbq and munge, or has my habit
> of keeping keys to decrypt old messages messed me up unlike anyone
> else?
Don't forget that lbdb isn't used for decrypting mail but only for
collecting mail addresses and attached user names. It is intended to
find hopefully all matching addresses/user names and to offer not too
many duplicates, which would make it useless.
But I added a line to the TODO list to think about changing this
behavior...
Tschoeeee
Roland
--
* [EMAIL PROTECTED] * http://www.spinnaker.de/ *
