Install check?

[Nicolai Waniek]

Hi,

I agree with you and as soon as I have some spare time (and a win32 machine
under my fingers) and if you guys wish it, i'll write an inno setup script for D
so walter could package a Win32 installer. Inno Setup is for free, but i'll have
to read the licence so there are no licence infringements.

This could - of course - then be adopted by the tango team :)

regards and give me some time, possibly I have some this evening







Georg Wrede wrote:
> It seems to be usual to "download D" and then not have the thing work
> off the bat. (d.D.learn has some excellent examples, and many of them by
> able people, some even with a long C++ background.)
> 
> With two compilers, all the platforms (Windows, Mac, various Linuxes)
> and at least two lucrative base libraries (Tango and Phobos) -- and
> especially since the newcomer can't simply choose like "compiler, OS,
> lib, root/luser" etc, and simply have a self-installing complete binary
> automatically sent to you (hint, hint), I suggest the following:
> 
> Just like hello.d we should have a checkinstall.d that you should try to
> build. It would have static asserts (or whatever it takes) that inform
> the user about what's wrong. Probably one should first run a
> checkinstall.{sh|bat} too to inform of the most basic problems.
> 
> Such should be standard issue with all distributions.
> 
> Today, many end up tweaking paths, library paths, choosing install
> directories, contents and placements of the "ini" file, and whatnot. Of
> course such could ultimately be avoided by adequate and prominent
> documentation and targeted binary installs. But the checkinstall stuff
> is more robust and could in principle be bullet proof. As in, either it
> passes, or it discovers what's wrong. Oh, and how to remedy that.
> 
> After all, there are only so many places where the train can derail, but
> figuring out these is unnecessary hard for precisely the people who do
> stumble on them.