[Bug] short or incomple URL's return all bad entries as matched

#1

I have many local domain pages for different services/apps/programs in my house. They all seem to catch each other in the matched login entries field and it drives me nuts.

I’m not really sure why they all are matched logins. Many things are single word machine names (i.e. https://computername)
Those, I guess if it’s resolving out the domain name? maybe that considers a match? But… I have no domain in the entry to match so the behavior should still be exact hostname without a domain…

It also seems to match any entries in my db with a broken or non resolvable link.

Example: If I go to
http://docker:9000/#/auth

I get a dozen+ matched entries. Some of the "matched" links:
http://octopi.local (my DNS fqdn is .local.domainname.com so maybe it's matching the .local?)
https:/booking.com (forgot the second /)
https://.homedepot.com (guessing bad syntax)
chrome://keefox 
chrome://FirefoxAccounts
http://download:80
chrome://weave

This is in the house of a power user. I can only imagine how much of a nightmare this would be if deployed to an organization for their users.
An entry matching a domain shouldn’t have anything to match if no fqdn was entered, but, it seems to return all bad entries instead.

#2

I can’t reproduce the “http://download:80” example and the “.local” “domain” is such a mess in general that I wouldn’t expect that to work correctly but yeah the other examples are due to the URLs for those entries being “bad”, to varying degrees.

I agree that it would be difficult for most users to understand what is wrong and how to fix it so I’ve created a GitHub issue to track the underlying bug and expect to have a new version of KeePassRPC available for testing soon.

#3

I’ve published a new version at https://github.com/kee-org/keepassrpc/releases/tag/v1.10.0

I’ll beta test it for a while before marking it as ready for wider release so it would be great if you could test it out too.