 |
lua-l archive |
|
|
||
|
On Wed, Mar 2, 2011 at 2:54 PM, Andrew Wilson <agrwagrw@gmail.com> wrote:To elaborate on this point, there are two objections to dropping use
> difficult to peg for a really improved Lua for Windows, one which
> isn't dependent on ms c runtime libraries
of MS toolchains in favour of GCC:
- some important libraries (like LuaInterface or LuaCOM) may not
survive the transition
- people on Windows often work with the MS development environment.
LuaCOM can be built using MinGW, thanks to the work of David M - not
yet sure about LuaInterface.
As for the second point, yes, people do like Visual Studio, but it's
increasingly unlikely that they will be using VS 2005, which is the
LfW standard. Instead, there are common problems when people try to
use these components with newer environments. I feel that including
Lua and some libraries in a Visual Studio project is really not
complicated, and that LfW's job is not to be a binaries provider
anyway (that's the role of LuaBinaries)
LfW is designed to be convenient for people writing Lua programs, not
embedding Lua in Windows programs.
Together with the move to a more sensible toolchain comes the Holy
Grail of 'build the world!'. That is, a person could rebuild all of
LfW from source without having to hunt around. Either LuaRocks or
LuaDist could provide that role.