Re: [Chicken-hackers] state of affairs on win32

[Brandon J. Van Every]

felix winkelmann wrote:
> Hi!
>
> I currently don't know what to do about the 
> chicken+cmake-on-windows-VS problem
> (the hang during compilation). Since I'd like to push for a new 
> release soon, I
> inquire hereby if it would be acceptable to leave this problem as it is?

Bear in mind the problem is specific to VS 2005 Express.  Although I'd 
like to see the problem get fixed, it's acceptable to me to stick it in 
the bugtracker, and let it rot until someone wishes to take it on.  I 
had hoped that one of the newbies running VS 2005 Express would be 
interested in conjuring a debug version of Chicken and instrumenting it, 
so that we get a better idea where the error is.  i.e. Those who really 
want it, could help to get it fixed.  I'd like to think that's the open 
source way, but of course the other OSS way is, "this tool doesn't work 
very well, and I'm not bothering with it again until it does."  We lose 
a certain percentage of newbies with VS 2005 Express being broken.  
That's a concern to me, but not enough of a concern to drop what I'm 
doing right now and deal with it.

I could do it myself, but signature gathering season has begun and I now 
lack time.  It is 3:15 pm as I write this, I haven't done a lick of 
paying work today, and that's because I stayed up late last night 
finishing the "reuse the PCRE object files during builds" functionality 
in CMake.  Plus my boss just called and I told him I took the morning 
off.  That's my perogative, but it doesn't sound good at the beginning 
of a run, and I can't keep doing that.  So, I'm not interested in 
creating new debugging / maintenance projects for myself right now.  I 
finished the PCRE one after a lapse of 2 weeks, just so it would be done 
before I (likely) hit the road and lose my computer access, and so it 
isn't left unfinished and forgotten.

I think the best interim answer for chagrined newbies is "MinGW works 
fine.  Use that."

> Of course, the basic problem is that we don't have a Windows maintainer
> (i.e. one who has all the development tools and is able to debug the 
> .c code).
>
> Is there one? Please show yourself!

I feel I am a primary Windows tester, particularly with regards to MinGW 
as that's my primary development environment.  MSVC is a distant second 
for me.  I don't develop with it, I just make sure the build is 
working.  I mainly test with Visual Studio .NET 2003 and things work 
fine there.  I note also the value of John Cowan's Cygwin testing.  I 
think the testing of Windows is covered well enough, short of an 
automated nightly build-and-test suite.  But I don't think either of us 
would describe ourselves as Windows "maintainers," in the sense of 
fixing broken code.  I did some of that to get the build working, once 
upon a time, but I wouldn't describe my knowledge of Chicken as very 
deep.  I still haven't coded anything with it, and I'd much rather 
become known as an OpenGL maintainer than a Windows maintainer.

> I can not do Windows maintenance. This means that Windows support will
> be lacking until someone picks this up. This is bad. Brandon will not 
> like this.
> Chicken can and should run on Windows, even with M$'s broken tools.

Do remember that it does, on everything but VS 2005 Express.  The 
unified build system is doing a lot of the job.  I of course have no 
problem being known as the Windows / CMake *build* maintainer, as that's 
exactly what I am.  Generally it's the Linux side of things that I don't 
want to get deeply involved with.  Especially since there are oodles of 
people available to do Linux stuff.  I don't feel that manpower is an 
issue on Linux, only people's will / discipline to do what needs to be 
done.  On Windows there's a notable shortage of developers, so I'm more 
likely to do things myself, out of a sense of "no one else will."


Cheers,
Brandon Van Every