[bop-devel] Patch to Business-OnlinePayment 3.00_03

[Ivan Kohler]

On Mon, Aug 28, 2006 at 11:29:15PM -0400, Phil Lobbes wrote:
> Hi Ivan,
> 
> Any hope of getting this diff/patch applied to BOP 3.x?

Sure.  Especially if you're willing to do it yourself as you offer 
below!

> I'll generate a similar patch to 2.x if you're willing to apply it.

Let's just release 3.00_04 and then 3.00 instead.  There's nothing in v3 
which should break v2 payment modules and there's already a handful of 
B:OP:HTTPS-based modules that require v3.

> In fact I could add/enhance the test cases for 2.x/3.x if you are 
> interested (you can take a look at my tests for things I have done for 
> my new PayPal backend:

Please feel free to do all the adding and enhancing of test cases you 
want.  :)

> PS. I'd be happy to apply these myself if you're willing to give me some
> access.

I've sent you private email to get this setup.  It isn't necessary to 
mail both bop-devel@ and myself; I'm subscribed.  :)

> Here's the diff against Business-OnlinePayment 3.00_03:

All your patches look good to me.

> * get rid of possible use an undefined value as a HASH reference
> * stop redefining subs if they already exist
>   - NOTE: it might be nice to do away with this completely and use a
>     notion of $self->param('name' => 'value') and $self->delete('name')
>     instead of creating subs on the fly.  This would be a little cleaner
>     design for v3 I would think.  It would get rid of some (in theory)
>     unnecesary subs and remove some uses of eval which might be a little
>     better for performance (probably minor but haven't benchmarked to
>     know for sure) and perhaps a bit more friendly to people that use
>     mod_perl too.

I don't agree we should do away with this completely - I think we should 
think *really* long and hard about API changes that affect B:OP users / 
app writers in a backwards-incompatible way.

If you felt it was important, perhaps we could get rid of the string 
eval's without breaking the API or processor modules by using an 
AUTOLOAD sub?

Why is this unfriendly to mod_perl users?  I use B:OP with mod_perl all 
the time, and this was never a problem for me.

My most most annoying peeve with mod_perl usage is when Crypt::SSLeay or 
Net::SSLeay loads up a different OpenSSL library verison that the one 
compiled into the web server and the whole thing segfaults without a 
trace of what happened.  :)

-- 
_ivan