[exim-dev] [Bug 2251] Segfault when LDAP lookup does not ret…

https://bugs.exim.org/show_bug.cgi?id=2251

--- Comment #12 from Matthew Slowe <M.Slowe@???> ---
Jeremy,

I brought the new binary into service for a little while and noticed that, for
addresses which would have triggered the segfault before, we're now getting:

Rewrite of <local>@kent.ac.uk yielded unparseable address: no local part in
address @kent.ac.uk

This wasn't happening before.

# exim -d+all -bt someone@???
...
08:45:36 23783 ldap_parse_result: 0
08:45:36 23783 ldap_parse_result yielded 0: Success
08:45:36 23783 LDAP search: returning:
08:45:36 23783 lookup yielded:
08:45:36 23783 ┌considering: $value}fail}@???
08:45:36 23783 ├──expanding: $value
08:45:36 23783 └─────result:
08:45:36 23783 ├──expanding: ${lookup ldap {user="<user>" pass="<pass>"
ldap:///o=kent.ac.uk,o=uni?unikentmailid?sub?(&(objectClass=mailRecipient)(!(inetUserStatus=deleted))(mailAlternateAddress=smtp:${quote_ldap:${local_part}@${domain}})(!(&(unikentmailOptinActive=*)(!(unikentmailOptinChoice=*)))))}{$value}fail}@???
08:45:36 23783 └─────result: @kent.ac.uk
08:45:36 23783 LOG: MAIN PANIC

An untouched one looks like:

08:50:36 21358 ldap_parse_result: 0
08:50:36 21358 ldap_parse_result yielded 0: Success
08:50:36 21358 lookup failed
08:50:36 21358 ┌───scanning: $value}fail}@???
08:50:36 21358 ├──expanding: $value
08:50:36 21358 ├─────result:
08:50:36 21358 └───skipping: result is not used
08:50:36 21358 ├failed to expand: ${lookup ldap {user="<user>" pass="<pass>"
ldap:///o=kent.ac.uk,o=uni?unikentmailid?sub?(&(objectClass=mailRecipient)(!(inetUserStatus=deleted))(mailAlternateAddress=smtp:${quote_ldap:${local_part}@${domain}})(!(&(unikentmailOptinActive=*)(!(unikentmailOptinChoice=*)))))}{$value}fail}@???
08:50:36 21358 ├───error message: "lookup" failed and "fail" requested
08:50:36 21358 └failure was forced

This one then continues to other rewrites rather than bailing out.

This strikes me as the sort of failure that might be solvable with a config
change on the rewrite - if you think so, I'm happy to try...

Thanks

--
You are receiving this mail because:
You are on the CC list for the bug.