Re: [mlmmj] RFE: improving access rejection messages
Baptiste Daroussin <[email protected]> Thu, 05 Oct 2023 13:51:59 +0000
| Newsgroups | org.mlmmj.mlmmj |
|---|---|
| Message-ID | <7ffoiewbvfs7jstloajinn7vbnze67vf2mntknu2v5tpz7hdgu@oqfirw2oqy4j> |
On Thu, Oct 05, 2023 at 09:30:18AM -0400, Konstantin Ryabitsev wrote: > On Thu, Oct 05, 2023 at 07:50:21AM +0200, Baptiste Daroussin wrote: > > >One of the problems I'm having is not being able to tell submitters exactly > > >why their message was denied. For example, I may have the following rules in > > >my access file: > > > > > > deny ^Content-Type: text/html > > > deny ^Authentication-Results:.*dmarcĂșil > > > deny ^Subject:.*forbidden-topic > > > > > >Currently, there is only one message generated on deny, and I have to include > > >all possible reasons why didn't allow the message through. This can be greatly > > >improved if we can tell the user exactly why their message was rejected, > > >instead of making them guess. > > > > > >I propose extending the functionality to additionally include the reason that > > >can point at a different file, e.g.: > > > > > > deny-html ^Content-Type: text/html > > > deny-dmarc ^Authentication-Results:.*dmarcĂșil > > > deny-topic ^Subject:.*forbidden-topic > > > > > >The dash-extension is used to identify the listtext to use, if present. For > > >example, if "deny-html" is triggered, the rejection message is loaded from > > >listtexts/deny-post-access-html. If "deny-dmarc" then from > > >deny-post-access-dmarc. If there is no matching file present, the usual > > >fallback path is followed (deny-post-access, then deny-post). > > > > I do like the idea, do you plan to provide a patch? > > I'm no good at writing C, so I am not comfortable producing a patch. That is not a problem, I will append this to my todo list > > > Otherwise please fill an issue with the proposal on codeberg so I keep track > > of the idea. > > Will do. Is that the recommended way for development discussions? I wasn't > sure if the list should be used for that, or if it is only for user support. > The list is an excellent place to discuss this and the issue are a good place to store the information once it has been discussed, so it is easy for me or anyone to find it and actually write the patch. Best regards, Bapt