Devel::LeakTrace

[Nigel Rantor]

Andy Armstrong wrote:
> On 1 Mar 2006, at 13:06, Jacqui Caren wrote:
> 
>> I would have though so - feed the changes back to the author but if  
>> they do not accept them (they have the right to refuse) then you  have to
>> either fork or keep a patch & apply as required.
> 
> 
> OK.
> 
>> We usually feed changes back to the author having created test  cases for
>> before and after, making sure the change is minimum reqd and has no
>> side effects. It can take a while to do this and somethimes takes  over a
>> month to get a positive response, so patience is reqd :-)
> 
> 
> The problem with keeping the change to the minimum required in this  
> case is that apart from a bit of boiler plate the module just  
> implements a fairly simple algorithm - and it's the algorithm that  
> doesn't scale. So most of it will have to change.
> 
>> We currently have four such modules that have elicited negative of  no 
>> response from the author. We keep patches for all of them.
> 
> 
> Uh huh. Well I'll just keep working on it and if I get something  
> decisively better I'll see what happens, thanks :)

Well, look, why not just wait until Rich has some bandwidth to get back 
to you. He may say "take over number one", in which case you have just 
volunteered yourself. Otherwise he'll accept patches in which case 
everything is groovy and you don't have to pick up a maint hat for a module.

Failing that if you think the amount of code that you would retain is so 
minimal then just release your code under a different name... 
Devel::TraceLeak perhaps? ;-)

   n