With Aero off it took roughly 5 seconds as well. My buffers, however, were empty. Of course it wouldn't surprise me if minimizing full buffers would take longer than 5 seconds.
Why don't we isolate variables here: Do not use full buffers to test in either XP or Vista.
//var %i = 1 | while (%i <= 50) { query x $+ %i | inc %i }Use the above code, cascade the windows and press escape. Time the results under both OS's and report.
Then, if you're curious, you can repeat for full buffers, but of course it will be slower. What we're interested in is the empty window baseline times.
Now, given that GDI under Vista is NOT hardware accelerated (in XP it was), you should expect it to be slower. Slower by how much depends on the scenario, but if there's more content to redraw (in the buffers) then it seems perfectly reasonable to me that 5 seconds can turn into 20 if every one of those 50 windows is full of data.
Windows 7, however, should be hardware accelerating GDI, so there shouldn't be an issue in Win7 (of course it should still take longer than 5 seconds). How are you testing Windows 7?
On a sidenote, I'm just wondering, do you think mIRC is responsible for any of this behaviour? If not, why file a bug report? It seems quite obvious (at least to me) that this is unlikely an issue with mIRC but rather one with GDI.
