Users with multiple e-mail addresses, and their ACL rights

Hi,

We have an RT installation that is configured to allow everyone

to create tickets, but only requesters and CCs to correspond on tickets.
This creates an interesting problem; people with multiple e-mail
addresses often end up having their correspondence rejected, because
they created the ticket from one mail address but responded to the
followup message from another address. has anyone grappled with this
problem?

I can think of two possible solutions to this problem, but both

are structurally heavy in terms of the changes they would entail:

  1. Allow e-mail aliases for a given account; so when a mail comes from
    one of th known aliases, it would be created under the said account.

  2. Allow for a specific type of linking between account that would
    indicate equivalency (or equivalently, add extra roles – ‘member of
    requester’s group’ and ‘member of CC’s group’ where the said
    privilege-propagating group would be selected in the user account
    settings).

    Has anyone a better solution, perhaps one that can be
    implemented in RT as-is and that doesn’t involve letting anyone reply to
    any tickets?

| Victor Danilchenko ±------------------------------------------+
| danilche@cs.umass.edu | With a GUI, what you see is what you get! |
| CSCF | 5-4231 | With a GUI, what you see is ALL you get! |

We have an RT installation that is configured to allow everyone
to create tickets, but only requesters and CCs to correspond on tickets.
This creates an interesting problem; people with multiple e-mail
addresses often end up having their correspondence rejected, because
they created the ticket from one mail address but responded to the
followup message from another address. has anyone grappled with this
problem?

Have you read the config file? or the docs? or even the mailing list
archives? rt supports this through the canonicalize email routine.

seph

We have an RT installation that is configured to allow everyone
to create tickets, but only requesters and CCs to correspond on tickets.
This creates an interesting problem; people with multiple e-mail
addresses often end up having their correspondence rejected, because
they created the ticket from one mail address but responded to the
followup message from another address. has anyone grappled with this
problem?

Have you read the config file? or the docs? or even the mailing list
archives? rt supports this through the canonicalize email routine.

yes, I have read these sources, and I am aware of

canonicalization. We have it enabled; but it’s not enough. For example,
we have multiple aliases for the mail server, and we canonicalize those;
but many people have mail addresses outside the department, and others
have distinct mail addresses within the department (for example, a user
who is also a support person for a research group, with a distinct alias
for that, and who corresponds both from his canonical address and the
said support address).

Basically, as coarse a tool as canonicalization can solve a

limited scope of problems; these problems are the most common ones, of
course, but canonicalization doesn’t help us any when a department chair
initiates a ticket from departmental mail address, and tries to follow
up from his address at a different university.

What we need is some systemic way to treat multiple distinct

e-mails as belonging to the same RT account on a per-account basis.
Sounds like there is currently no such facility in RT. maybe i will look
into it when I understand RT better…

P.S. Frankly, I didn’t even think of canonicalization WRT this problem,
because we already have it enabled. Sorry for not mentioning that
explicitly.

| Victor Danilchenko ±---------------------------------------------+
| danilche@cs.umass.edu | Does Emacs have the buddha nature? |
| CSCF | 5-4231 | Why not, it has bloody well everything else! |

Victor Danilchenko danilche@cs.umass.edu writes:

yes, I have read these sources, and I am aware of
canonicalization. We have it enabled; but it’s not enough. For example,
we have multiple aliases for the mail server, and we canonicalize those;
but many people have mail addresses outside the department, and others
have distinct mail addresses within the department (for example, a user
who is also a support person for a research group, with a distinct alias
for that, and who corresponds both from his canonical address and the
said support address).

It sounds like you’re using rt3. rt3’s CanonicaizeEmail function starts
out as a pretty simple search and replace, however rt3 allows you to
redefine most internal functions. So override it, and have it do what
you want. Use a giant lookup map, use an external db, whatever you can
code you can have.

seph

Victor Danilchenko danilche@cs.umass.edu writes:

yes, I have read these sources, and I am aware of
canonicalization. We have it enabled; but it’s not enough. For example,
we have multiple aliases for the mail server, and we canonicalize those;
but many people have mail addresses outside the department, and others
have distinct mail addresses within the department (for example, a user
who is also a support person for a research group, with a distinct alias
for that, and who corresponds both from his canonical address and the
said support address).

It sounds like you’re using rt3. rt3’s CanonicaizeEmail function starts
out as a pretty simple search and replace, however rt3 allows you to
redefine most internal functions. So override it, and have it do what
you want. Use a giant lookup map, use an external db, whatever you can
code you can have.

Hey, that's a thought -- I could overload it to look up a table

of e-mail aliases… Cool, I will look into it. thanks.

| Victor Danilchenko | You cannot apply a technological |
| danilche@cs.umass.edu | solution to a sociological problem. |
| CSCF | 5-4231 | Edwards’ Law |