test
, but the forum told me the username already exists. I looked in the database under phpbb_users and that name did not exist. Any ideas?
phpBB 3.2.3
PHP 7.1
test
, but the forum told me the username already exists. I looked in the database under phpbb_users and that name did not exist. Code: Select all
SELECT * FROM phpbb_users WHERE username_clean = 'test'
The odd thing is that it doesn't happen on my other board. I'm wondering if an extension is to blame, but then again, surely a table would exist of aMySQL returned an empty result set (i.e. zero rows). (Query took 0.0008 seconds.)
test
username.test
. Very odd.And since personal webservers are free (at least many of them are) and they are easily configurable (php-version, MySQl version, addons), it's THE way to test all changes beforehand.
This was EXACTLY what the issue was. I had a defined user group called test and apparently that also affects username choice as well. Perhaps phpBB should change that?
I can't think of a scenario off-hand to where "it would be ambiguous whether we're referring to names of groups or names of users", such that the two "must" share the same namespace / not allow duplicates between the two. But I'm guessing some scenario like that must exist, as to why the code would have been added to explicitly test for uniqueness across both tables.John connor wrote: ↑Tue Oct 16, 2018 4:20 am I had a defined user group called test and apparently that also affects username choice as well. Perhaps phpBB should change that?
Private message senders. Upon addressing recipients you enter usernames and click on group names, but when reading a PM (no matter if you're the sender or one of the recipients) the names being listed in one style.EA117 wrote: ↑Tue Oct 16, 2018 5:14 amI can't think of a scenario off-hand to where "it would be ambiguous whether we're referring to names of groups or names of users", such that the two "must" share the same namespace / not allow duplicates between the two. But I'm guessing some scenario like that must exist, as to why the code would have been added to explicitly test for uniqueness across both tables.
I would have to agree that it doesn't seem defensible that we "need" to be able to make group names the same as user names. The real situation is "its not expected that a user name would collide with a group name", or vice-versa, since both types allow arbitrary names to be defined. Accurate and appropriate error messages sounds like the efficient solution.