Glad that you were able to figure it out! And yeah, that can be a
non-obvious requirement of operating in a worker. I'll see if there's any
sensible places for us to surface that to the rest of the community.

--Brandon

On Wed, Aug 2, 2023 at 11:26 PM wenhan chong <wenhan....@gmail.com> wrote:

> Hi Brandon,
>
> Thanks for looking at this. I compared you sample code against the code in
> my app and saw that the render loop on my worker thread was causing the
> issue. It seems that having a render loop on the worker prevents control
> from returning to the ui thread, thus suppressing any errors thrown by
> webgpu device. I was able to use the uncapturederror event listener after
> removing the loop.
>
> Thanks,
> Wen Han
>
> On Tuesday, 1 August 2023 at 02:50:15 UTC+8 Brandon Jones wrote:
>
>> Hi Wenhan!
>>
>> I [wrote a quick test](https://codepen.io/toji/pen/ZEmVypz) to try this
>> out and I'm able to see the uncapturederror events in the worker and
>> forward them on to the main thread for logging.
>>
>> Is there a reduced version of the code that's producing the issue that
>> you could share? Because otherwise it looks like it's working in at least
>> this trivial case.
>>
>> --Brandon
>>
>> On Sun, Jul 30, 2023 at 11:54 AM wenhan chong <wenha...@gmail.com> wrote:
>>
>>> Hi All,
>>>
>>> I am trying to use WebGPU uncapturederror listener to detect unexpected
>>> errors that happen during render time. I notice that errors are suppressed
>>> and not captured when I am rendering to an offscreen canvas on a worker. I
>>> tested this by forcing a simple GPUValidationError in the shaders. On an
>>> app that renders directly on the main thread, I am able to catch the error.
>>> However, on an app that renders to an offscreen canvas on a worker thread
>>> before transferring to main, the error is not captured. There is also no
>>> console output indicating any errors without the use of the uncapturederror
>>> listener. The canvas simply doesn't render. Is this a known issue and is
>>> there a workaround?
>>>
>>> Thanks,
>>> Wen Han
>>>
>>> --
>>> You received this message because you are subscribed to the Google
>>> Groups "blink-dev" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to blink-dev+...@chromium.org.
>>> To view this discussion on the web visit
>>> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/2ff4f166-182e-46bb-bedd-9145dd2b675dn%40chromium.org
>>> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/2ff4f166-182e-46bb-bedd-9145dd2b675dn%40chromium.org?utm_medium=email&utm_source=footer>
>>> .
>>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to blink-dev+unsubscr...@chromium.org.
To view this discussion on the web visit 
https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAEGwwi04tWvpcQEZM9SJri6j3r_YFtXGU2YQ0Gczc7BqE3GeTw%40mail.gmail.com.

Reply via email to