Re: [Gretl-devel] Feature request for plotting: GUI accesss to xtics and ytics
On Sun, 25 Oct 2015, Riccardo (Jack) Lucchetti wrote:
The following piece of conversation somehow dropped off-list, but
IMO is of general interest, so I'm re-posting it.
On Sun, 25 Oct 2015, Logan Kelly wrote:
> I saw an older post about building under Windows, but there was a
> need for someone to maintain the system. If this is something of
> value, I'm willing to help. However, I haven't had any problems
> with the snapshot not being updated quickly enough. So maybe the
> marginal benefits are limited.
I do think the marginal benefit is limited, but in case anyone is
interested in building gretl from git on MS Windows, here are a few
thoughts.
First, should you use MSVC or a GNU-like alternative? And if the
latter, which alternative? My recommendation is to use mingw. I
successfully built gretl using mingw on Windows a few years back, so
I know it can be done. Cygwin might be OK too, but if you want a
build that can be distributed to other Windows users you'd have to
be careful not to build in any cygwin-specific dependencies. MSVC is
unknown territory for me. In principle it should be possible to
build gretl using MS proprietary tools but I can't offer any help on
that.
In the gretl source tree there's a "win32" subdirectory that
contains various files relevant to a Windows build using mingw.
These files are set up such that you can choose between "cross"
(which I use, compiling for Windows on Linux) and "native". See in
particular "Makefile" and "mkwindist".
On the gretl server at WFU there's a directory "winbuild" --
http://ricardo.ecn.wfu.edu/pub/gretl/winbuild/ -- with some material
that was potentially helpful about 3 years ago ;-) If anyone is
seriously interested in trying a gretl build on Windows I could
update that material. Otherwise, doing so is not a good use of my
time.
Allin