Re: [exim] More integer annoyances in 4.65

On 03/01/2007 10:20, Philip Hazel wrote:
>> I agree the change for empty strings is [not] very good. OTOH, using
>> an empty string in integer comparisons is definitely a config flaw.
>
> That is the dilemma. It seems that it is a common practice.

I suspect even very-close-to-default configs and some of the examples in
the spec rely on this but I haven't combed through to make sure.

Perhaps the solution is for future Exim 4.x to throw a warning but
continue to operate as if the empty string == 0. Then people's systems
won't break on upgrade, but they'll know they have a possible config
flaw. And we can have a wiki page to document this previously
undocumented feature :-)

Cheers,

John.