Cannot capture frames from Edge/Chrome because API is None

[noahgamboa]

Hello! I'm having the following issue:

Windows 10 Version 21H2 (OS Build 19044.1706) RenderDoc v 1.20 Chrome 102.0.5005.63

Steps to repro:

1. Open command prompt
2. run this:

C:\Windows\System\cmd.exe /c "SET RENDERDOC_HOOK_EGL=0 && START "" ^"C:\Program Files\Google\Chrome\Application\chrome.exe^" --disable-gpu-sandbox --gpu-startup-dialogue"

3. Observe chrome launch correctly, but no GPU debugging info at the top left corner. Collect the PID from the dialogue
4. in render doc if I inject into the PID
5. Observe that the connection in render doc is established but API is None
6. navigate to Google Earth or Google Maps and set the 3D view as expected
7. Go back to render doc
8. Observe that "Capture Frame(s) immediately" is blurred out and API is None

Tried a bunch of different ways to go with different versions of stuff but still nothing. Any hints? I'm using an Nvidia Titan Xp as my gpu not sure if that matters though...

[eliemichel]

You should swap steps 4 and 3: Injection must occur between the apparition of the pop-up window and the validation (that is then starting Chrome/Edge). If you don't see the popup it means that you did not fully close Edge/Chrome prior to starting this process.

[noahgamboa]

Ah gotcha. ya so currently I:

1. open chrome, dialogue appears, main chrome window is all white
2. Get the pid from the popup
3. close the popup
4. inject into the pid in renderdoc

But you're saying I should:

1. open chrome, dialogue appears, main chrome window is all white
2. get the pid from the popup
3. inject into the pid in renderdoc
4. close the popup