On Wed, 24 Feb 1999, Bruce Bowler wrote:
> In the version of sendmail we had (sorry don't recall specific version) all
> of the following would get to me...
>
> BBowler@???, bboWler@???, BBoWLeR@??? etc...
>
> With exim, I had to play all sorts of 'games' with locally_caseless (sorry
> if that's not the exact variable, but it's close) and duplicate entries in
> my aliases file in order to get the same functionality to work.
Exim *does* believe (by default) that local local parts should be
caseless. However, you then have the problem of how to implement that.
Exim does it by lower casing the local parts before it starts processing
them.
> This is
> largely because the person I inherited the system from decided by username
> should be BBowler, not bbowler.
Aarrgghh. Traditional Unix user names have always been lower case.
What is Exim to do on receiving mail for any of
BBowler@???, bboWler@???, BBoWLeR@???
? The function getpwnam() which looks up user names is case-sensitive
(as are most things on Unix). What Exim does by default is to look up
getpwnam(bbowler) in all cases. If you set locally_caseless it will look
up what is given, case-independently. However, it cannot possibly look
up all the differently cased versions!
I don't know how sendmail did it. The only way I can think of is that it
actually read the /etc/passwd file serially and did its own caseless
comparisons. That seems to me to be a very heavy-handed way of going
about this.
> Is there an (easy) way to make exim treat *ALL* instances of anything that
> could potentially be a local part, the same way? Assuming exim uses
> getwpent or something similar to check for local users, it would obviously
> mean "rolling your own".
By default, local local parts are lowercased before any processing is
done. Remote local parts are left alone - but don't really take much
part in any processing, except in the spam-checking cases.
In which way would you like Exim to treat local parts? I cannot myself
see any alternative either to forcing the case (one way or another) or
leaving it alone.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
--
*** Exim information can be found at
http://www.exim.org/ ***
