Typical, install instructions are several times more effort in Windows (not to mention Cygwin and having to restart your computer to play nice with VS). Is it not possible to get this working using MSYS and Mingw32?
Windows is pretty much always going to be a second class citizen once you take a half-step outside the mainstream. I would just try to "get used to it" (i.e. run Linux/Mac).
I was hoping this one would course-correct itself... (Reading comprehension). "Windows is pretty much always going to be a second class citizen once you take a half-step outside the mainstream."
He said... Windows is a pain in the ass if you do anything off the really-well-beaten path (like Haskell). And even then, yes, C/C++ can be a pain in Windows (again, see the linked project).
I do. I still lament for others'. There've been more than one occasion where I want to collaborate with my peers that refuse to make the leap and it's always a nightmare.
These days, it seems like POSIX has become just another VM environment, like Smalltalk, Erlang, the JVM or the CLR. Mingw/Cygwin doesn't run *nix programs "on" Windows, really; it just provides them with a really half-assed VM implementation. Why not instead just go all the way and run, say, Ubuntu Server in a headless VMware appliance? If nothing else, it gives your spare cores something to do :)
From my experience of doing software development professionally in several flavors of UNIX and Windows systems, I would say Linux is no better as there is seldom a way that works flawlessly across distributions.
I can't remember the last time I struggled building something on a random distribution that worked on another. What are you doing that you run into incompatibilities across modern Linux distributions?
It's certainly better than: "use VS" or "suffer through any other sort of Windows dev" in my opinion.
I'm constantly running into incompatibilities with my own code, all of them due to Debian. The one that's still causing me trouble is being able to invoke a pyuic4 that has access to both pykde and pyqwt; as far as I can tell that's impossible on Debian.
The third is the only that's tangentially related to the development pains we're discussing here. Packaging is a "fun" issue, but it's apples and oranges with multiple compatibility layers to get to a common denominator to be able to ... build a DCPU toolchain for example.
>Regarding your comment, there are many other ways to develop software for Windows besides VS.
Uh huh, and as seen even... right here... it's normally a (relatively) major pain in the ass to orchestrate.
... I can't tell if you don't understand what the issue is here (from this post and your others), or if you really think that forcing complex workarounds into one-off IDEs is a good way to solve this problem.
I am playing a bit devil's advocate, probably with not so much success.
The thing is, that each operating system has it own way of doing things, and many people in open source tend to think that everywhere else should be like GNU/Linux.
I used to think like that, until I got the opportunity to work with several different types of operating systems, all UNIX flavours, Windows, OS/400, Symbian, VMS and see that there are many ways to do certain things.
Here is an example how someone else did a Windows based development environment for DCPU,
