Shader technique nil pin

Think I saw that one around, but just was thinking making it a nice report.

Looks like with quite a few shaders we can end up with nil techniques(specially when moving around renderers). All compiled properly, just a red nil technique enum.

It looks like to happen a bit random (never the same one getting red).
It’s definitely not related to gfx memory (had it with like 5-6 texturefx, which is nowhere near 1gb on my graphics card). Moving renderer across screen make it easier to happen tho (but even without still does).

Quite hard to reproduce as well, but a few texture fx generally does the job.

Quite a pretty annoying bug (quite a show killer sometimes really), so would be good to have that one sorted.

Hi people,
here the same,
if you change the render window between the monitors in dual mode, sometimes it happens, mostly with some render passes and a video camera connected.

Saving it with the new configuration and restart uses to solve it… (sometimes).

+1 !
may i suggest from some bad experiences with 23 and 24:
memory card will always shows up it s okay, but in fact this bug appears when too much things loaded inside of it.
26.1 is less doing it, keeping it only on moiving renderers.
Uncs shaders are avoiding it. never had this bug with them, only from shaders home made or contributions from community.

yes for video connected: video stream seems to be responsible of something

+1
always happens to me if i move the renderer from the primary to the second screen. also i get some kinda errors every frame in the tty renderer and the EX9 renderer turns black. but i never saved it. i’ll dump it later. however if i save the patch with the renderer on the second screen exit vvvv and reopen the patch, it works as it should.

Happens with any shader really, including unc’s ones.

Moving renderers across screen it not really a solution, well kinda works but not all the time, also happens in single screen (also multi screen is more common have to admit). And moving something like 50 renderers on a semi decent texture post is not really great way (specially as it doesnt always solve it).

Also can have moved renderer, but then as soon as you add another fx it might happen again, so you need restart.

Just to notify this still happens in b27, had simple patch with a few post with switch (around 5-10 TextureFX).

the same with b27.sometimes even without moving across screens
also strange: if i have many instances of the same fx, it could be 3 of them OK and one red with nil

any news re: random red nil technique enum?
i have beta31 DX9 with same problem, even with fix screen conf + starting /dda 1
any tips how to fix it “on the fly” without restart?

Yeah still see this sometimes. May not be a good solution, but I’ve found just opening the code editor for the shader in question seems to make it work again.

Yep, when i open code editor this error disappearing, but let’s consider that this is not a solution ).

Same problem here, but I disabled enabled the shader to have it back.

open and close code editor helps. thanks.

I just bought a new GTX 780 and Im currently having a problem like this with the shader going nil.
But before i even noticed the shader issue i was struggling with my projector connected by hdmi as second monitor dropping out while having renderer in fullscreen and my main monitor started flashing until i have “ctrl alt del and task manager” the screen stops flashing and projector is back up. this could happen every 3-20min.

This also occurred with win7 alone, so googled around and found its a Microsoft bug with not being able to Disable “Monitor OFF Detection in windows”.
But back to vvvv:) when i create a new shader like the one thats nil or Disable/Enable it, the projector drops out and main screen flashing.

Have found out that DVI port is more stable but with pin 16 pulled out (pin 16=Hot plug) and it also is less nils in the vvvv shaders

Any ideas for what could be stabilizing from vvvv´s side on this issue?