Re: [PROPOSAL 5.4] alternate way of format/pack/unpack (in the string library)

[Tom N Harris <telliamed@...>]

I wouldn't say minilanguages are "liked". Just that C++ iostream is even 
less liked. And some of the concerns seemed frivolous to me. Why is it a 
problem that format strings can't validate argument types ahead of time 
when Lua is a dynamic language?

On 5/26/2017 5:14 PM, François Perrad wrote:
> An implementation based on a minilanguage doesn't allow user extension,
> for example, you cannot add a "%Q" option to `string.format` in pure Lua.

It isn't because format strings are a minilanguange that makes it 
unextendable. The current implementation delegates interpretation of 
format strings to the C runtime. A hypothetical Lua library could parse 
the format strings itself while being extensible.

That said, an attempt to make the formatting language more Lua-like I 
think should take cues from LPEG. Use simple composable objects to 
define the format. A back-of-the-hand example is:

   xy_format = "x = " .. dec() .. " y = " .. dec()
   print(xy_format.format(10, 20))

   pi_format = "pi = " .. fixed(nil, 4)
   print(pi_format.format(math.pi))

   dec2 = pad(dec(), 2, "0")
   dec4 = pad(dec(), 4, "0")
   date_format = dec2()/"day".."/"..dec2()/"month".."/"..dec4()/"year"
   print(date_format.format(os.date"*t"))

--
tom (telliamed@whoopdedo.org)