On Wed, Aug 22, 2012 at 12:32:59PM +0200, Dominic Tarr wrote:
> I agree with Tim Caswell. Lets face it, that exists exists is embarassing.
> removing it from the docs (replace with a link to stat) but leaving it
> in the code (with a big comment) is the right compromise. once
> everyone has forgotten about it, it will be easier to remove it.
>
> I understand isaac's position. he doesn't want to remove exists,
> because although it's a breaking change, and although we LOVE breaking
> changes in node, what we really love about breaking changes is when a
> breaking change is a breaking improvement.
>
> but this is only a cosmetic improvement. there will be no associated
> performance improvement...
Consistency is a hobgoblin and all that. I'm glad that node doesn't waste time
on the sort of refactoring that plagued open source a few years back; endlessly
refactoring for the sake of a taxonomy that is closer to the Platonic form.
I really appreciate the way decisions to alter node are made.
However, that `exists` has a different signature from everything else has always
felt like a burr. I'm surprised it lives on. I'm surprised there is a function
that's been left in node that whenever anyone uses it, they are told not to,
because checking for existence is an anti-pattern, and `stat` is better.
That is a waste of community energy: having a bright shinny function that draws
newbies like moths, only to have the community zap them when they use it.
It it a test?
--
Alan Gutierrez -
http://twitter.com/bigeasy