History API heads up

Alex Kiernan alexk at demon.net
Mon Jul 9 09:06:08 UTC 2001


Russ Allbery <rra at stanford.edu> writes:

> Alex Kiernan <alexk at demon.net> writes:
> 
> > I'm going to start landing this in the next few days, whilst the changes
> > look widespread, much of it is just consolidating the history access
> > stuff to go through the library.
> 
> > Theres a few wrinkles need tidying:
> 
> > - dbz.c, dbz.h should probably move into the libhist/hisv6
> 
> Yup, definitely.  Can we call the directory history rather than libhist?
> (I hope that doesn't break anything!)  It should be clearer.
> 

Oops, just committed that bit :-( I could go on a quick renaming in
the tree, before anyone else gets to check it out?

> > - tagged hash history needs hauling out into a separate implementation
> 
> One could almost use a C preprocessor to do that from the existing dbz.c.
> 

Good point.

> > - `#ifdef CANT_DO_THIS' in expire.c needs resolving (the -i and -x
> > options)
> 
> That may need some work with the interfaces down the road to make sure the
> hooks are available to do this sort of thing.
> 

I suspect we can probably do this with magic options into the various
underlying layers.

> > AFAIK its fully backward compatible - I've been swapping between 2.3.2
> > and 2.4-current regularly without rebuilding stuff.
> 
> Cool!
> 
> Feel free to land stuff as-is as long as it all builds so that we have it
> in CVS and can start cleaning up the remaining wrinkles.  The only thing
> that would be cool to change before committing it all is the directory
> name (history instead of libhist) if I was reading your note above
> correctly.
> 

Hopefully I'm committing in an order whereby if you check out at just
about any point you'll get something which builds.

-- 
Alex Kiernan, Principal Engineer, Development, Thus PLC


More information about the inn-workers mailing list