Re: Putting version information, description, copright into modules

[Philipp Janda <siffiejoe@...>]

Am 12.02.2016 um 17:09 schröbte Peter Aronoff:
> On Friday, February 12, 2016 at 01:52PM, Roberto Ierusalimschy wrote:
> > > > The manual says "names", not "variables".
> > >
> > > I'm sorry to be dense, but I don't see what you mean. Is a "name" simply
> > > a toplevel variable? Does it mean something more specific here?
> >
> > name == identifier. It can be a global variable name, a field name,
> > a parameter name, a local variable name, etc. All names are names :-)
>
> Ok. Unfortunately both of your responses are too cryptic for me. I don't
> know if you're answering my original question. And if you are answering,
> I don't know what your answer is.
>
> My thinking is that even though my module exports a table with an
> identifier, a name, __VERSION, it's ok because that identifier is
> namespaced within any program that imports (requires) it. (It's a field in
> a table that no Lua-Rio code touches directly.)
>
> Thus, even though the manual appears to say this is bad, there can't be any
> actual conflict in practice.

I'm sure we all agree that a general table should be able to hold any 
string key including ones starting with _ + uppercase letter. However, a 
module table is not a general table, and there already has been Lua-Rio 
code in Lua 5.1 that used the `_M`, `_NAME`, and `_PACKAGE` fields in 
module tables, so the reservation of those names probably does not apply 
to global variables (and the registry) only.

For personal use I'd say it doesn't matter much -- if a new Lua version 
gets released that starts to use `_VERSION`, you just fix your module. 
Chances are that you have to fix *some*thing anyway. But it's probably a 
bad name to make a community standard.

> Am I wrong? Or were you just commenting on my imprecise use of variable
> instead of name?
>
> Thanks, Peter

Philipp