Re: [AD] Regarding backwards compatibility of 4.1.x and 4.x where x >= 2

[Chris <kcat@xxxxxxxxxx>]

Grzegorz Adam Hankiewicz wrote:
> At the end of Appendix E you
> can find the text:
>
>  "...but based on my experience with as recent a version as 4.1.12,
>  which is not compatible with 4.0.3... I am only officially
>  supporting 4.0.3..."
>
> Which kind of contradicts the proclaimed intention, unless the non
> compatibility issue is with regards to some obscure features.

Did he happen to mention what the problem was? We can't even hope to fix 
it if he doesn't say.

> AFAICS "deprecated" tries to bring
> attention to the developer that a symbol is going to disappear, but
> if the developer only has to add a new include to the source code, I
> would prefer if those warnings forecasting the apocalypse went away.

AFAIK, "deprecated" means no longer supported. However, I have argued 
for removing those warnings in the past because the deprecation warning 
GCC generates /will/ cause compilation to fail if the programmer uses 
-Werror (which is good practice durring development, IMO). And saying 
that the programmer doesn't have to switch to the new API (immediately) 
is contradictory to that.

- Kitty Cat