> > I can test the PPAS stuff, but not the Greenplum objects. But... they > shouldn't be an issue - if an object is derived from pgSchema, you can > (and should use GetSchema(). If it's derived from a pgTable, you > probably want the table name too. If it's derived from pgObject, > pgServer, pgDatabase, then there's no schema available. >
Ok. No problems here. - First, the patch exhibits exactly the problem I was talking about. > See the attached screenshot, which is just a mess. All of the names > (server, object, username etc) are real. None are made up. And yes, > despite what Microsoft seem to be wanting people to do from Windows 8 > onwards, I usually don't have full screen windows, but tend to work > with smaller ones side-by-side. > Really. Your servers have very long names. This should happen only at lower resolutions. And if we put a configuration option? The user decides if he wants this option or not. Or to put this information somewhere else, but where? > - Not only do I see truncated names, but more importantly the "Done." > is not visible. That is *essential*, for obvious reasons. > Yes, I agree. > > - We try to avoid referring to servers by their hostname/IP address in > the UI (anywhere we do, is an oversight). You should be showing the > name. In my case, for remote servers that tends to be the hostname > anyway. For local servers though, it's usually a string like > "PostgreSQL 9.1" or "Postgres Plus Advanced Server 9.1". The reasoning > is that the name is a memorable string that the user can easily > recognise, whilst the hostname or IP address usually isn't (for > example, Amazon EC2 hostnames - which of course, are also very long). > Here, also without problems. I agree too.
