[GRASS5] GRASS for Windows

[Glynn Clements]

Radim Blazek wrote:

> > > There is no user interface and not all modules compile/run,
> > > the work is in progress. Vector modules from CVS do not compile
> > > but with few changes I was able to compile and run also vectors but
> > > without attributes (missing fork()).
> >
> > This would be a good point to make a concerted effort to decide upon a
> > common interface for spawning child processes.
> 
> For dbmi drivers I have done that already last Thursday using _spawnl.
> Do you se any problems with?

It's insufficiently flexible for the general case, e.g. it doesn't
support redirection.

mon.start uses spawnl() for starting the driver, and it works OK
there.

> > I have done some work on this, in lib/gis/spawn.c, but it is currently
> > untested, non-portable, and nothing uses it.
> >
> > It would be useful if someone with access to a Windows development
> > environment could attempt to implement G_spawn[_ex] on Windows and let
> > me know which bits are problematic (signal handling?).
> >
> > Let me know if you need any information about what those functions are
> > supposed to do.
> 
> I have seen G_spawn but because it is not implemented for windows I just
> started looking for something else.
> 
> _spawnl is described here:
> http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vclib/html/_crt__spawnl.2c_._wspawnl.asp
> Would it be difficult to implement G_spawn using _spawnl?

CreateProcess would probably be a better choice, as it provides more
functionality.

-- 
Glynn Clements <glynn at gclements.plus.com>