Feature request: envelope node extending (6)

20 Name: Storlek : 2007-01-07 20:11 [Del]

[from tct1501]

When the last node of an envelope is selected, create a new node 2x the length of the envelope upon hitting insert.
(I suppose it'd keep the same value as the current node.)

51 Name: delt : 2007-01-08 15:07 [Del]

Just checked how IT handles this:

50 ticks is the smallest possible range of the envelope area. (default envelope ends at 100 ticks) When you drag the last node and reach the right of the envelope area, the range is extended by another 50 ticks /* edit */ and the mouse is xwarppointer'd to where the previous max value is now. This gives a kind of loarithmic feel to editing envelopes, since around 1200 ticks, just moving by a few pixels to the right extends it by another 50, and around 9000 you're extending it by about 100 ticks with every 1-pixel movement.

Anyway, i kinda don't like using XWarpPointer () too much cause it makes for a rather inconsistent interface, but in cases like this... what do you think?

53 Name: Storlek : 2007-01-08 16:41 [Del]

>>51

> i kinda don't like using XWarpPointer () too much cause it makes for a rather inconsistent interface, but in cases like this... what do you think?

Well... having dealt with certain toolkits, pointer warping has left a bad taste in my mouth :P
It's useful in very rare circumstances, such as virtual desktop edge-flipping, but the potential is far, far too great for abuse for it to be as easy to do as most toolkits make it.

Maybe in this case it would be the best way... I dunno.

62 Name: mrsbrisby : 2007-01-09 08:30 [Del]

Its not possible (feasible) to warp the pointer on some platforms, with some hardware (how are we going to warp your fancy new tablet?).

It's also a security risk on some platforms (ever wonder why firefox/mozilla makes you wait three seconds before installing xpi?), which means the ability to do so might go away on platforms and hardware combinations that already exist.

We probably could get away with it in fullscreen mode (simply use relative mouse reads), however.

63 Name: Storlek : 2007-01-09 10:27 [Del]

>>62

> We probably could get away with it in fullscreen mode (simply use relative mouse reads), however.


If we can't do it all the time, don't do it at all. Inconsistency makes programs feel buggy -- if you're not aware that the behavior is supposed to be different, how would you figure it out? To the user, it would either look like the mouse jumps around randomly in fullscreen mode, or the pointer warping is broken and only works half the time.

424 Name: esaruoho : 2008-03-03 05:53 [Del]

>>20

and shift-insert at last node could copy the whole current envelope so that one could then modify it even further. wouldnt that be gr8

Name: Link:
Spam trap (leave blank):
File: