RE: Are there Windows users of cons?

[Greg Spencer]

I figured I had better chime in here, since I seem to have started some
portion of this snowball.

We use Cons exclusively on Windows here, using my extensions, and little
else.  The reason I haven't updated the extensions to run with the
current release of Cons is simply that I haven't had time, and the older
versions of Cons work fine for us, so it's been hard to justify spending
my work hours on something that already works for us (especially when we
have plenty of other more pressing projects).  For this I apologize,
since I really do want Cons to move forward.

That said, I'd really like to get some time to work on this and "do it
right".  The correct way to integrate into MSDEV, for instance, is not
to use my "hack" of the DSPGenerator.pl (and a hack is exactly what it
is), but rather to write a plugin for MSDEV that can communicate with
cons to build it's environment and make changes to the Conscripts (If I
add a file in MSDEV, for instance, it should be able to add it to the
Conscript).  The problem there is that MSDEV plugins are barely
documented, and don't really seem to include all the functionality
needed.  Maybe that will change with the .NET versions, but I doubt it.

The way Windows (and other platforms) is integrated into Cons is not
really all that first rate either.  It would be much more powerful if
all the compiler/linker/etc specific code were separated from the build
infrastructure so that it would be easy to write perl modules that
supported various platforms under the same build mechanisms.  Cons
should be able to build Java applets, C# controls, or windows help files
just as easily as it builds C++ object files.  This is probably the one
place where "Make" is better than Cons.

So, it's not just windows support that is lacking.  This doesn't mean
that Cons is useless, or that the maintainers have been lax.  It just
means that there's still a lot of work to do, by people who don't really
have time to do it in the first place, since they have other jobs where
people pay them to do other things.

In the interest of speed, if anyone would like to take over stewardship
of my extensions, I'd be happy to help them with any information they
needed.  Otherwise, I'll try and get some time to port them to the
current version of Cons.  Of course, there are some changes that need to
occur to Cons itself before this can work (like the ":S" and "!" changes
already mentioned).

[Steven, if you don't like the "!" and "#" notations because they limit
your namespace, then why not just add an escape mechanism to the
expansion routines?  Prefacing a character with, say, "\" would place
that character into the string literally, for instance, and "\\" would
indicate a backslash.  Maybe another character would be a better choice
(because of the windows backslash path separator), but you know the
drill.]

I agree with Steven that we really need a way to steward each of the
changes through a process, so that we can control the changes going in,
but also so we can make sure that the process moves along, and changes
that everyone wants will make it into the releases in a timely fashion.
Some way to affect the *design* of cons would be good too -- we need to
look at controlling the direction of development, as well as the current
status of development.  One step at a time, though, I suppose.

Them's my $0.02.

                                -Greg.