chicken-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Chicken-users] Win32 maintainer needed


From: Bob McIsaac
Subject: Re: [Chicken-users] Win32 maintainer needed
Date: Mon, 09 Jan 2006 22:55:12 -0500
User-agent: Mozilla Thunderbird 1.0.6 (X11/20050716)

Hi:

Brandon has been advocating the use of Cmake, I guess to minimize
cross-platform build complexity.  I took a quick look at the Cmake site
and agree that the idea has merit.  It appears to use XML to describe
the pattern rules used by various compilers. Correct me if I am mistaken.

My main goal is to use Scheme on Linux and win32 without any fuss.  I
don't have a Watcom agenda.. it just appears to be best maintained of
the free win32 compilers.  And my options are limited because I have win98.

regards,
-Bob-

PS: has anyone tried http://visual-mingw.sourceforge.net/ ?
It makes installing mingw much easier.

Brandon J. Van Every wrote:

I actually posted my questions to you on the Chicken list. But since you have replied privately, I'll reply privately. If that was an accident, and you meant to discuss things publically, we can regroup there.


Bob McIsaac wrote:

Brandon wrote:

1) do you care about building stuff on Windows? Is this actually important to you, or are you one of those Unix-only types?

I use XP at work, linux at home, sometimes booting win98 to research ideas
I can bring to work.  For Scheme,  I need to use the FFI on windows to
access hardware such as serial ports. But I have trouble compiling Scheme
on win98 where I do preliminary stuff.


Good to know you're actually motivated by a desire for working Windows solutions. I sure hate hearing from Unix guys who don't care and just want Windows to go away.

2) what do you propose as a "separate Windows build system," considering that Windows has many compilers, all requiring different build systems? Are you proposing anything other than "well I was just going to support my own favorite compiler and leave other people hanging" ?

Cygwin is ponderous and slow, Mingw is complicated, the free VC++ won't
install on win98.  I believe Watcom would work for everyone.


All these compilers have different licensing, toolchain, and support issues associated with them. For instance, I have been driven primarily by "what's not GPLed and works with Eclipse," and that happens to be only MinGW. Watcom has a decidedly marginal user community. I don't see your claim as in any way realistic. People don't need just a Scheme-only solution, where the backend compiler doesn't matter. They need to do a lot of FFI stuff, and library linkage stuff, and IDE support stuff, and other-people-will-maintain-this-crap stuff. Then the other toolchains matter a great deal. The C universe of a Scheme-to-C compiler is relevant.


3) do you actually know anything about CMake?

I just visited the web site. It looks impressive.



Let me impress you further. We have working CMake builds for Chicken, thanks in no small part to the direct assistance of the pricipal CMake author. You can't beat that kind of business relationship. The only reason you don't see the build in the tarball is because Xmas was hectic, New Year's is still hectic, and I need to take the proper steps of documenting it up and so forth. You either subscribed to the Chicken list very recently, or have been ignoring my posts, or don't read archives before suggesting new courses of action. I've been posting about this CMake stuff for a good 3 months now.


Looks like watcom could be
included in the Templates where MSC and bsc32 now exist.


Sure, if you do the work. They seem to be amenable to that sort of thing. For instance, they added a MinGW-specific target the other day, for no particular reason that I could discern. I guess they thought that would make their lives easier. What would you gain from making this effort? Well, if you want more people to use Watcom, then people already using CMake wouldn't have to do anything special to give Watcom a whirl. They'd just retarget for Watcom, and if they like it, they'll keep it. Of course, my agenda is to get more people using CMake in the 1st place. The mission is to supplant GNU Autoconf, and I see SCons as the principal competition we have to address.


4) are you doing, or going to do, any real work to resolve cross-platform build issues?



I created a watcom makefile for Gambit.. there wasn't much interest there. The one for Chicken is much more difficult given the different libraries and objects that are linked together in various ways. It's about 90% complete.
I need to understand the chicken architecture in order to complete it.

I might be able to contribute to Cmake when I know more about it.


If you're interested in the latter, I can assist you with that. If only politically, but that does matter. I'm part of CMake-Promote, after all. It would help your cause if you could point out some Watcom showcase projects, though. Like, apps and libraries that effectively demonstrate why people should like / care about Watcom. Most people think it's a dead compiler and a dead company. What proves otherwise?

If you aren't into CMake, you and I are going to find ourselves on opposite sides of the build issue. I'm probably going to end up being the Windows build-and-test guy. I'll tell you frankly, I won't build and test anything on Windows that isn't CMake. So if you're really determined to do it "the Watcom way," you'll be doing it all yourself. Failing to leverage other people, often leads to substandard releases, which in turn leads to user apathy about a particular build. Unless of course you really really have lotsa extra time on your hands, to stay on top of things and make other people happy with your Watcom stuff.

Cheers,
Brandon Van Every
"We live in a world of very bright people building
crappy software with total shit for tools and process."
                               - Ed McKenzie










reply via email to

[Prev in Thread] Current Thread [Next in Thread]