Hi

On Thu, Jul 03, 2025 at 10:56:39AM +0200, Nicolas George wrote:
> Michael Niedermayer (HE12025-07-02):
[...]
> To be clear, you suggest, IIUC:
> 
> - The developer registers the error code.
> - The developer writes the explanation for the error when registering
>   the error code.
> - The developer uses the error code when returning the error.
> - The user gets an error code.
> - The user looks up the error code to get the explanation for the error.

no, just:

return av_adv_err_new(AVERROR_INVALIDDATA, "Garbled foobar data", "Foo triangle quantum decoder"
                      __FILE__, __LINE__, NULL, "Whaetver you like %s", favorite_food);

teh return type is int64_t here

this also cannot fail as it allocates nothing
it also needs no context but would use a mutex or thread local storage

the message length would be bound by a maximum, set at compile time
and the number of messages that this keeps track of also would be
bound by a maximum, set at compile time


> 
> What I propose:
> 
> - The developer writes the explanation for the error when returning the
>   error.
> - The user gets the explanation for the error directly.
> 
> Or, in terms of code:
> 
> return AVERROR_MESSAGE(ctx, "Garbled foobar data at offset ${offset}",
>     "offset", ctx->offset, NULL);

I am not sure if passing a context around is going to find the volunteers
to implement and maintain. Also it has a performance impact for small and
lightweight functions.

thx

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

For a strong democracy, genuine criticism is necessary, allegations benefit
noone, they just cause unnecessary conflicts. - Narendra Modi