Right way to cope with user error in make test?

[Alistair Leslie-Hughes]

"Jeremy White" <jwhite at codeweavers.com> wrote in message 
news:482CF239.7000207 at codeweavers.com...
> So...turns out that in this flood of new reporting, that one of the errors
> only happened to me, and it further turns out to be entirely user error;
> I didn't have libxslt.
>
> So, the obvious first solution is for me to actually read my configure
> results and deal with it.
>
> But I think I serve nicely as an example of the sort of incompetent user
> for whom it would still be nice to have make test work cleanly.
>
> I didn't see any obvious standard way of coping with this situation.
> Did I miss it?  I imagined that maybe we'd skip these cases, but I didn't
> see evidence of that.  I could also imagine a facility whereby we note
> that the configure was not clean, and then refuse to run make test
> (or at least refuse to run the full winetest battery).  Should we make
> libxslt non optional (or at least require an explicit --without-libxslt
> in order to build without it)?
Hi Jeremy,

This could be a good option.  libxslt should properly be non-optional
since msxml3 relys on it.