|
Yes, something like this - use DatabaseAccess to locate a database w/o path. Have a look how external tables work now.
But use of current directory to locate database should not be supressed by new feature - legacy applications must work. I think only local connections intentionally use filenames w/o path and the path resolution is handled by the Y-valve in this case, so it will have the precedence automatically. As for the standalone server working via the remote subsystem, filenames w/o path are unlikely to be used. On Windows this means %system% as a working directory :-)
Same logic that was used for external tables is applied to databases.
1. Database file is searched for in all directories, present in restrict list. 2. If not found - first one is used as default directory. This let's create database succeed, attach certainyl fails (like expected). |
||||||||||||||||||||||||||||||||||||||||||||||
I think that database without path should be treated first as alias, than seek on DatabaseAccess path list, and at last at current dir.
Maybe current dir is excess.