Re: [MinGW-dvlpr] Shell evironment for MinGW

[Earnie B. <ear...@ya...>]

"Steve D. Perkins" wrote:
> 
>     Just wanted to forward this snippet from Joerg Bruehe, from a dialogue
> in the general [MinGW-users] mailing list pertaining to website updates:
> 
> > If this is true, I propose the team contacts the owner of
> >    http://www.weihenstephan.de/~syring/win32/UnxUtils.html
> > and includes a link to that location,
> > because I have good experiences with the binaries ("zsh", "tar"
> > as well as several others) from there.
> 
>     I didn't want to bring this up on that list... but a month or two ago
> there was a brief flurry of discussion out here about expanding the scope of
> MinGW to include a shell environment (basically for the "MinGW environment"
> to become Cygwin without the bloat and licensing issues).  Where did that
> discussion leave off, and has there been any progress towards the ideas that
> were talked about?
> 

I'm still working on it.  I lost a lot with a NT kernel loss and the
only means of restoration at this company is a new image. :(  I've
modified the Cygwin winsup/cygwin source and removed the registry
dependency for the mount points.  I've removed the /cygdrive reference
in favor of /a, /c, /d, etc.  I've made changes to Cygwin's gcc to allow
for -mmsys, -mcygwin and -mmingw so that I can bootstrap my
environment.  I'm finding necessary other changes to other libraries
such as newlib and w32api to support the __MSYS__ and __INSIDE_MSYS__
flags.

>     One of the reasons I've been so eager to improve the website and
> documentation for MinGW is that I would like to make it easier for other
> library and application vendors to design their products such that they will
> build using MinGW.  I get endlessly frustrated with apps and libraries that
> support building with Cygwin, but have bugs/issues with MinGW because the
> environment was too much of a sketchy and fluid for developers to target.
> I'm hoping that more robust documentation will help promote a "standard
> MinGW" that developers can target... the new single-file distribution format
> is another step in that direction, and the inclusion of a shell and toolset
> would seal the deal.
> 

Yep, me too.  ZSH is reported to work and I've used it but there are
other issues that reside.  The reported link used to reside on our pages
and was removed because the owner refused to distribute source.  This
has been rectified and the above link actually has a SourceForge base.

>     I told Joerg via the [MinGW-users] list that I wouldn't mind including
> the above link on the MinGW website... but after considering it more
> carefully, I'm having second thoughts.  I don't want a half-dozen various
> shell environments to be seen as "officially endorsed" by MinGW... as this
> would run counter to the goal of a stable environment vendors can target.
> If the discussion of a standard MinGW shell environment has been vetoed or
> put on hold indefinitely, I'd rather position the MinGW-with-Cygwin combo as
> the standard target to be used by vendors whose products require a shell
> environment in which to build.
> 

Since MSYS isn't yet ready, consider adding it now and then removing it
later in favor of MSYS.

Earnie.

_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com