Re: next release of Lout is approaching (mingw vs cygwin)

[KHMan]

Valeriy E. Ushakov wrote:
> On Sat, Aug 16, 2008 at 18:04:01 +0800, KHMan wrote:
> > If anyone has experience using Lout on Cygwin, let us know if you
> > want and I will put the material up on a new wiki page later.
>
> Last time I tried (which, admittedly, was *years* ago), lout's use of
> ftell confused the hell out of cygwin unless you was careful with CLRF
> options, but IIRC, Jeff changed db files to be read in binary mode, so
> that probably no longer matters.

Roger that, will take a look at that and your older posting below.

> I'd prefer a "native" win32 port compiled with whatever is the
> "native" version of gcc for windows, lcc-win32, or with the free
> variety of Visual Studio (Express, or whatever microsoft's fancy
> marketing name for it is).  I don't see a need to tie lout for windows
> to cygwin.

Of course, "native" is desirable, but only in a more perfect 
world, like the Linux kernel where lots of developer resources is 
available and Linus et al. can even afford to be choosy or 
insulting to the large pool of Linux kernel developers.

But from my position, the our problem is this: Small FLOSS 
projects like Lout have very little manpower resources. I make my 
decisions based on very limited availability of manpower for small 
projects like Lout. Transitioning to Cygwin is manpower-efficient.

And you won't catch me using VS... it's at least a 460MB download 
plus the SDK, and I still don't want to touch it (I already have 
it installed, but urggghh...) I'd rather use MinGW because using 
VS will be a significant commitment. One wonders why there are 
still many people using the old Visual C++ after so long... In a 
nutshell, I'd much more rather stick to fat-free Unix-style tools 
with simple and clean config needs.

Lout is primarily a source release only, and Lout is currently 
more or less in maintenance mode. I would rather use my time for 
better purposes than maintain a legacy binary release. Someone who 
releases a binary release has to answer for it (or not, but that 
wouldn't be nice) and I don't see myself maintaining MinGW 
binaries forever.

So unless someone seriously wants to take over the MinGW port, 
transitioning to Cygwin is best for everyone; I no longer need to 
make those binary releases, and users should be able to compile 
and use a Cygwin-based Lout with a minimum of effort and fuss -- 
their toolchain won't be in any jeopardy even if a meteor wipes 
out South East Asia.

In the end, talk is just talk, and in order to get things done, 
people must step up. The primary constraint here is scarcity of 
developer resources. Also, what's best in the long term when faced 
with this scarcity. This means it may not be possible to give 
prospective users the best possible outcomes. While we fail to do 
the best for users, the second best solution may be one that is 
the most sustainable in the long term.

> Google has found me this old mail of mine, that still pretty
> accurately reflects my thoughts on the matter (ignore the part about
> ftell, that should be fixed now, I think)
>
> http://lists.planix.com/pipermail/lout-users/2000q1/001726.html

Thanks, will read up on this.

--
Cheers,
Kein-Hong Man (esq.)
Kuala Lumpur, Malaysia