Re: Lua 5.1 (final,rc) now available

[Mike Pall <mikelu-0601@...>]

Hi,

Luiz Henrique de Figueiredo wrote:
> I'm not sure what to do. We can just include those incantations suggested
> by the community and hope that they work; but we cannot test them all.
> What is worse, "solaris" or "bsd" or "macosx" may not mean the same thing
> to different persons...

I guess it's difficult to draw a line. Judging from the mailing
list traffic, the Lua user community consists mainly of
developers. Lacking a well organized collection of easily
installed libraries, Lua is not (yet) attractive enough as a
general scripting replacement for 'the masses' (like what Perl or
Python is used for). (*)

Developers are obviously skilled enough to fix the Makefile and
change luaconf.h. They may have to support weird configurations
anyway (cross-compiling, embedded targets, game consoles).

But a good first impression still counts. Evaluating a language
means toying around with it. But if it doesn't compile on the
first try, it gets dropped. Developers are busy persons. Why fix
something before I even know whether I could use it at all?

This means the decision on which targets to support out of the
box depends on the common _development systems_ and not
necessarily on the huge variety of _target systems_ that Lua
works for.

Getting facts on significant market shares of developer systems
is hard. My personal guess is: Windows, Linux, BSD, Mac OS X and
Solaris (not necessarily in this order). Anything else has a low
enough market share or is different enough that the 'posix' or
'generic' targets should suffice.

Asko Kauppi wrote:
> And please, only readline-enabled. Working with Lua without history
> is a chore, and it's better people install the readline, or modify
> Makefile, than think Lua sucks because of the command line.

Yes, but only if readline is well supported on the target. This
is probably the case for 99% of Linux installations. It might be
the case for modern FreeBSD and NetBSD installations (but I guess
not for OpenBSD). I'm not so sure every developer has already
upgraded to Mac OS X 10.4 (let alone users).

OTOH adding readline/no-readline targets is 'cheap' because it
doesn't require much more testing.

BTW: Seems BSD wants "-ledit -ltermcap". Is the "-lreadline" for
Mac OS X 10.4 verified? Is this a symlink?

Bye,
     Mike

(*) But it has the potential. Making it work out of the box on
many common operating systems is a step in the right direction.