Re: why is imp_xxh stored in an SV?

On Wed, Apr 18, 2012 at 11:55:45AM +0100, Dave Mitchell wrote:
> 
> To keep things as simple as possible, I've left mg_obj pointing to
> the SV containing the struct, but added using mg_ptr (with mg_len == 0)
> that caches a direct pointer to the structure, skipping the need to
> indirect via the SV and PVX(sv).
> 
> I've also updated dbih_getcom() to directly handle the common code path
> when retrieving the handle, rather than doing a dTHX and punting to
> dbih_getcom2().
> 
> This gives a very slight performance boost, and doesn't have any binary or
> backwards compatibility issues (hopefully).

Looks good. I've uploadded DBI-1.618_901.tar.gz for cpantesters to chew on.

> Note that this and the just-submitted DBIS patch represent the end of
> my work on improving performance on DBI, (not counting any remedial work
> that may be required).

Many thanks again for your contributions to the DBI, Dave.
They are very much appreciated.

Tim.

• why is imp_xxh stored in an SV? by Dave Mitchell
• Re: why is imp_xxh stored in an SV? by Tim Bunce
• Re: why is imp_xxh stored in an SV? by Dave Mitchell
• Re: why is imp_xxh stored in an SV? by Dave Mitchell
• Re: why is imp_xxh stored in an SV? by Tim Bunce
• Re: why is imp_xxh stored in an SV? by Martin J. Evans
• Re: why is imp_xxh stored in an SV? by Tim Bunce
• Re: why is imp_xxh stored in an SV? by H.Merijn Brand