[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Public WebGL] Uniform and varying packing in the WebGL spec



On Tue, Jul 16, 2013 at 8:06 AM, Olli Etuaho <oetuaho@nvidia.com> wrote:
> I think that would be okay, yes. Or move the last paragraph there, "The WebGL API further requires that if the packing algorithm fails either for the uniform variables of a shader or for the varying variables of a program, compilation or linking must fail. "

Your suggestion sounds cleaner.
http://www.khronos.org/registry/webgl/specs/latest/ has been updated
with this change.

-Ken


> By the way, I'll be away on vacation for a couple of weeks starting this Thursday, but hopefully I'll have time to look into the remaining issues in the variable packing tests after I get back to work in August.
> ________________________________________
> From: Kenneth Russell [kbr@google.com]
> Sent: Monday, July 15, 2013 10:46 PM
> To: Olli Etuaho
> Cc: public_webgl@khronos.org
> Subject: Re: [Public WebGL] Uniform and varying packing in the WebGL spec
>
> Sorry for the long delay replying.
>
> I agree that the sentence is ambiguous. The intent is that only
> shaders and programs which pass this conservative variable packing are
> allowed to execute. Would it be clearer if it were simply changed to:
>
>   "The WebGL API imposes further requirements."
>
> and then let the subsequent bullet points and paragraph provide the details?
>
> The difficulty in specifying precise errors for calls like
> compileShader is that it's legal for implementations to defer shader
> compilation until program link time.
>
> -Ken
>
>
>
> On Thu, Jul 4, 2013 at 9:52 AM, Olli Etuaho <oetuaho@nvidia.com> wrote:
>>
>> Hi all,
>>
>> This sentence in the WebGL specification section "6.24 Packing Restrictions for Uniforms and Varyings" confuses me:
>>
>> "The WebGL API further requires that this packing algorithm succeed for all shaders and programs submitted to the API."
>>
>> Based on the rest of the section, failing/succeeding means whether you can fit the data into the available space. So, to me, this seems to say that the algorithm can pack any kind of data at all into limited space, which is obviously false. If this is referring to that there are extra requirements posed for the shaders by the WebGL API compared to GL ES API (which allows implementation-specific tricks in varying packing, for example), then that is redundant information, since a paragraph further down clearly states when compilation or linking must fail.
>>
>> The wording was changed in February, but I don't think I get the point that the previous version is trying to convey either, and it seems to me that this needs to be looked into again.
>>
>> -Olli
>> -----------------------------------------------------------
>> You are currently subscribed to public_webgl@khronos.org.
>> To unsubscribe, send an email to majordomo@khronos.org with
>> the following command in the body of your email:
>> unsubscribe public_webgl
>> -----------------------------------------------------------
>>

-----------------------------------------------------------
You are currently subscribed to public_webgl@khronos.org.
To unsubscribe, send an email to majordomo@khronos.org with
the following command in the body of your email:
unsubscribe public_webgl
-----------------------------------------------------------