Re: Hash seed breaks 5.8.1 binary API; fix suggested

[Steve Hay <steve.hay@uk.radan.com>]

Chip Salzenberg wrote:

> According to Graham Barr:
>  
>
> > Below is a count of downloads [...]
> >    
>
> Excellent.  Now the question remaining is: How many different 5.8.0
> APIs are there?
>
> In other words, what are the config options that create binary
> incompatibility, and which therefore must be tested separately?
>
> (I am *not* counting system-driven API changes like gcc options.)
>
> The only ones I know of are the threading options.
>  
The win32/Makefile contains a PERL_MALLOC option (equivalent to 
-Dusemymalloc) which is commented as affecting binary compatibility.

There is also a USE_LARGE_FILES option; I don't know if that affects 
compatibility or not?

- Steve