User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1.1) Gecko/20061204 Firefox/2.0.0.1 Build Identifier: email_in.pl properly sends a notification when it doesn't recognize a product name. However, if the product name matches from a case-insensitive point of view, I get a rather obtuse error message: can_enter_product reached an unreachable location. Reproducible: Always

Wow, that's really weird. Are you on PostgreSQL or MySQL? Could you tell me what the two products were that you were using, and if they have any group controls on them?

I'm using MySQL. The product names only differed by the case sensitivity, such as "FireFox" vs. "Firefox". But only "FireFox" exists, not "Firefox". I don't use groups right now.

Here's the code that's at fault, I'm pretty sure: my $can_enter = grep($_->name eq $product_name, @{$self->get_enterable_products});

Okay, this should fix it, and it makes sense to me. Product names are case-insensitive on both MySQL and PostgreSQL.

Comment on attachment 250394 [details] [diff] [review] v1 If we want can_enter_product() to be case-insensitive, then we should also make can_see_product() case-insensitive. Also, you have to make sure that you don't introduce any regression, neither in enter/post/process_bug.cgi, nor in editproducts/components/milestones/versions.cgi. This makes me think: how could you change the case of a product with this change, i.e. if you have a product FireFox, and you want to rename it as Firefox? I know we have a bug about that already, and we should think about this first.

> This makes me think: how could you change the case of a product with this > change, i.e. if you have a product FireFox, and you want to rename it as > Firefox? We have the same problem with email addresses, although I can't find a ticket for that.

This is IMO cleaner, especially now that we know that lc() has some problems with some locales.

Comment on attachment 327716 [details] [diff] [review] patch for 3.2 and HEAD, v2 Why not just call Product->check?

(In reply to comment #9) > Why not just call Product->check? Because ->check will inform the user that the product doesn't exist instead of the current message which says it either doesn't exist or you don't have access to it. This is the kind of information we don't want to leak.

Comment on attachment 327716 [details] [diff] [review] patch for 3.2 and HEAD, v2 Okay, sure, this looks fine then! :-)

I'm not sure of the performance impact of that change, so I don't really want to approve it for 3.0.

Minimal change for 3.0.5: I use lc()

a3.0=mkanat for the 3.0.5 patch.

tip: Checking in Bugzilla/User.pm; /cvsroot/mozilla/webtools/bugzilla/Bugzilla/User.pm,v <-- User.pm new revision: 1.166; previous revision: 1.165 done 3.1.4: Checking in Bugzilla/User.pm; /cvsroot/mozilla/webtools/bugzilla/Bugzilla/User.pm,v <-- User.pm new revision: 1.164.2.1; previous revision: 1.164 done 3.0.4: Checking in Bugzilla/User.pm; /cvsroot/mozilla/webtools/bugzilla/Bugzilla/User.pm,v <-- User.pm new revision: 1.148.2.7; previous revision: 1.148.2.6 done