Re: Monkey Patching is Bad, Unless Really Needed (was Re: A guide to building Lua modules)

[Philipp Janda <siffiejoe@...>]

Am 16.04.2014 23:46 schröbte Hisham:
> On 16 April 2014 14:55, Philipp Janda <siffiejoe@gmx.net> wrote:
> > > There are always better solutions; the mechanisms that allow monkey
> > > patching are there for low-level metaprogramming, not for regular
> > > module/application development.
> >
> > I agree in general, yet we both have contributed to a publicly available
> > module[1] that does _nothing but_ monkey-patching ...
>
> Well, yeah, but one could argue that this is almost the kind of
> low-level plumbing I was referring to. Steve made us a good defense by
> arguing that it has a disciplined interface.

A disciplined interface doesn't help if the affected code isn't aware of 
that interface (which happens if you have old or foreign code). Sure, 
*new* code can handle the extended interface just fine, but old code 
doesn't expect it.
So, IMHO, what makes this case different is that the monkey-patches are 
intended as conveniences for new code only, not to change the behavior 
of old code (`compat52.lua` strictly[*] _extends_ the 5.1 standard 
library). If you make incompatible changes you have to isolate 
old/foreign code somehow (which is why `compat52.strict` is separate and 
optional, and `compat52.mstrict` works on a per-chunk basis) ...

> -- Hisham

Philipp

  [*] almost (e.g. the return values of `os.exit` are incompatible)