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

[Philipp Janda <siffiejoe@...>]

Am 17.04.2014 13:30 schröbte steve donovan:
> On Thu, Apr 17, 2014 at 12:57 PM, Philipp Janda <siffiejoe@gmx.net> wrote:
> > 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
>
> Absolutely!
>
> >    [*] almost (e.g. the return values of `os.exit` are incompatible)
>
> That's when a disciplined patcher has to stop.  For Penlight, I got
> around this by providing utils.execute and utils.load which have 5.2
> semantics.

Ah, I meant `os.execute` not `os.exit` (which usually doesn't return at 
all) ...

> Plus, utils.getfenv is always defined, since kind people
> pointed out that checking for existence of getfenv is a common enough
> way to detecting 5.2.  The only global alterations left are that
> table.pack and table.searchpath are added in 5.1.

So maybe we should try to define what "acceptable" monkey-patches are 
(in case legacy code might be involved) ...

*   adding new functions/fields to existing modules?
*   adding additional (optional) parameters to existing functions?
*   adding additional return values?
*   handling arguments that previously would've raised an error?
*   increasing the value range of parameters and/or return values of an 
existing function?
*   ... ?

Philipp