96675 - checksetup should require admin e-mail address satisfy emailregexp.

Status: RESOLVED FIXED

(Bugzilla :: Installation & Upgrading, defect, P1)

Description

[Matthew Tuck [:CodeMachine]]

When checksetup.pl initialises an installation, it asks for the admin address. 
I use a local user, since I don't have sendmail set up to do remote mailing.

Then it doesn't let that user log in because the address doesn't match
emailregexp.  I think that part might not be new.  I don't have a problem with
that behaviour - it shouldn't create the account in the first place.

Instead, if the admin account doesn't satisfy the default emailregexp, we should
 write out a more lax emailregexp and emailregexpdesc.  Optionally, we could
confirm this action out front.

Comment 1

[Matthew Tuck [:CodeMachine]]

Oops wrong component.  Reassigning too.

Comment 2

[Dave Miller [:justdave]]

This is going to be a blocker for the 2.16 release

Comment 3

[Zach Lipton [:zach]]

Attachment: Patch to fix

Comment 4

[Dave Miller [:justdave]]

Comment on attachment 55418 [details] [diff] [review]
Patch to fix

Looks good to me.  I'd give it first-review except that it has a MacBinary header on it so I can't apply it...

Comment 5

[Dave Miller [:justdave]]

Attachment: the above patch without the MacBinary header

Comment 6

[Zach Lipton [:zach]]

Comment on attachment 55419 [details] [diff] [review]
the above patch without the MacBinary header

has dave's 1st review

Comment 7

[Bradley Baetz (:bbaetz)]

Comment on attachment 55419 [details] [diff] [review]
the above patch without the MacBinary header

>
>+      unless ($login =~ /$mailcheckexp/) {
>+        print "\nThe login address must match the following pattern to be valid:\n";
>+        print "$mailcheck\n";
>+        die "Please try again\n";

Is the first line correct? You're not printing the pattern, you're printing a description
of a pattern. What about "The administrator's address is invalid.\n$mailcheck"? That matches
CheckEmailSyntax's text from CGI.pl

r=bbaetz in any case.

Comment 8

[Zach Lipton [:zach]]

fix checked in

Comment 9

[John Vandenberg]

I believe there is a problem with this patch, as it uses $::params rather than
$::param.

Comment 10

[John Vandenberg]

Ive raised the above problem as bug 133210.