Changes in snapshot over versions

[windowsLC]

I have an old standalone (maybe version 2.9) that uses the import snapshot command (with no object or rec) and you get the cross icon to click and drag. Well If you do not move the cursor and double click an image of the whole desktop is imported. If I do a snapshot with version 9 dp11(windows) and just double click it just fails silently.

Does anyone know if this is a bug or was it changed at some point?

I liked the old way better plus it seems in a standalone it turns the icon into a permanent ring/clock though it doesn't freeze the program so you have to click with the clock cursor. in the IDE it just fails.

[bogs]

In 2.2.1 (only older version I have) it works as you describe, no matter where you just click, it shoots the whole desktop.

In 6.5.2, it will shoot the whole desktop if you click outside of the stack, otherwise it shoots the stack.

in 7.1.4 and 8.1.6 it actually froze the desktop screen and the IDE, forcing a reboot (go figure). I stopped testing at that point, so haven't narrowed down exactly when it changed, but it did change at least twice.

I think I liked it the way it worked better in previous editions myself.

[bn]

Hi windowsLC,

I tried double-clicking on a Mac using LC9 DP11 and it imported a snapshot of the whole screen.

Using the crosshair cursor it imported the selection.

If accidentally your stack is located at the screenloc you would not notice that an image was imported.

If that is not the case you probably should do a bug report.

As bogs noted in its testing this feature was broken on and off over time. But now it works for me.

Kind regards

Bernd

[windowsLC]

thanks for the testing.

I tested a bit also on Windows.

On version v6.0.2 it works and I get full screenshot

I tried a few v7 stable builds and it just fails silently.

I tried v9 dp11 and it script fails with this info: button "Button": execution error at line 4 (snapshot: no screen), char 1

so if it is working for BN on Mac v9 dp11 I suspect that they tried to fix the silent fail at some point and fixed for Mac but not for Windows.

Maybe a few more tests can determine exactly what version it happened.

UPDATE: after a little more testing I can see that my theory is probably wrong. I can see that when it failed in previous versions it fails at that line in the script and doesn't continue. I think the only difference in v9 is that it can now report the error in the script editor and express what the execution error is.

[windowsLC]

I've determined the bug surfaced at least on windows after 6.6.5 stable. Bug exists in 6.7 stable. I have filed a bug report.

[bogs]

Nice work testing, I didn't have time to return to it myself today.