@darius and @raff, setting search.caseSensitiveDatabaseStringComparison to false seems to solve the issue for us.
TRUNK-5071 settles on the conclusion that this GP should be defaulted to false (which is in @raff’s fix commits).
@dkayiwa however both the demo and QA Ref App don’t reflect this, why is that?
On those two servers the GP search.caseSensitiveDatabaseStringComparison is set to true (at least last time I checked).
Default settings work for only those cases where the value was not set or for new installation. They do not automatically overwrite user settings. Is this what you were looking for?
So you mean that the Ref App distro explicitly reverts what is done here?
If that’s what’s happening, I’m just wondering why that choice is being made since TRUNK-5071’s thread settles on the opposite (=GP set to false).
What i mean is that a new installation will have that value of false. But for an update/upgrade, which already had a value of true, that will not overwrite it.
Yes I understand that.
But when one launches a Ref App instance from scratch (such as when the demo instance is re-created based on a released distro and starting from a blank database, is that not a ‘new installation’?), the Core should set the GP to false first, then somewhere else down the line it is set to true again apparently.
Which exact version of the reference application did you try from scratch?
I am not arguing whether setting the GP to false or true does set it to false or true.
I am asking why is it chosen for the demo Ref App to display the opposite behaviour to the conclusion of TRUNK-5071? Is there a good reason for this or is it just accidental?
By accidental I mean, it was set to true for a long time, and the fix didn’t revert it because the fix doesn’t change values that are already set.
By the time the fix was made, the demo refapp had a database with a value of true. So that cannot be automatically overwritten to false with Rafal’s commit. And this is intentional, until an end user explicitly sets it to the opposite value.