Re: luacheck configuration for pandoc Lua filters?

[Albert Krewinkel <albert+pandoc-9EawChwDxG8hFhg+JK9F0w@public.gmane.org>]

Norman Ramsey <fellswalker-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> writes:

> luacheck works with one module at a time, so all I need to do is figure out
> the _names_ of the callable objects.  I could poke at metatables or could
> assume conservatively that all objects are callable.  What I don't know how
> to do is find callable names that _aren't_ materialized in the `pandoc`
> table but are available only via an `__index` metamethod.

We don't do any metatable magic on the pandoc module itself, so
iterating over the module table should give us all objects. As BPJ
points out, some functions are actually callable tables, so we have to
check for those. Also, the three objects 'Block', 'Inline', and
'MetaValue' are callable, but always lead to errors when called.
Therefore, the following code should print all the names we are after:

    local dont_call = pandoc.List{'Block', 'Inline', 'MetaValue'}

    for k, v in pairs(pandoc) do
      if not dont_call:includes(k) and
        (type(v) == 'function' or
         getmetatable(v) and getmetatable(v).__call) then
        print(k)
      end
    end

The refactored Lua subsystem that I'm working on will do away with these
silly inconsistencies -- just testing if an object has type "function"
will be enough.


> On Tuesday, May 11, 2021 at 4:21:47 AM UTC-4 BP wrote:
>
>> Den tis 11 maj 2021 01:16Norman Ramsey <fells...-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> skrev:
>>
>>> I've recently discovered `luacheck`, a fantastic static-analysis tool for
>>> Lua programs.  I'd like to use it with my Pandoc filters, but it needs a
>>> configuration to tell it what functions in the global `pandoc` table are OK
>>> to call.  If anybody happens to have made such a configuration, please let
>>> us know.
>>
>>
>> I guess the fact that most of the constructors in the pandoc library are
>> actually callable objects rather than functions may be a problem one way or
>> another.


--
Albert Krewinkel
GPG: 8eed e3e2 e8c5 6f18 81fe  e836 388d c0b2 1f63 1124