Hi Dave,

Thanks for your feedback.

> Do you have copyright assignment paperwork on file?
> https://gcc.gnu.org/contribute.html#legal

My paperwork is done.

> The autotools are not my strongest suit.

> In a previous life I was a Windows developer, but I think it's been
> about 18 years since I've done any coding on Windows, so I'm going to
> have to trust your Windows expertise.

I guess the best solution would be to use libtool and Automake, but
that require major changes in the GCC build system. I came up with a
solution that uses an if statement to choose the way to build the
dynamic library.

> > It is not necessary to use --enable-host-shared in Windows (I tested
> it),
> > but I
> > don't know the proper way to disable that check.

I used case statement to check the value of $(target).

> What's the issue with fchmod on Windows?  Does it not exist, or does it
> not work?
> Please add a comment explaining why.

fchmod does not exist in Windows.

> Please can you fix the indentation here to stay within 80 columns.
> It looks like there's very similar (identical?) code to this below in
> several places.
> Can it be put in a subroutine to avoid repetition?

I created a new file "jit-w32.{h,c}" to store the definitions of
win_mkdtemp() and print_last_error().

> Is there a cast to HMODULE everywhere that m_dso_handle is used?
> If so, does it make the code cleaner if the field becomes an HMODULE on
> Windows?

I added a new type "result::handle_t" that abstracts over this
difference.

> This seems like a very awkward way to get a temporary directory, but
> perhaps it's the best that can be done on Windows?

We could copy an implementation of mkdtemp () for Windows or write
something using a random number generator. I did it this way because
it was the fastest.

> Sorry if this seems excessively like nitpicking - thanks again for the
> patch.

No problem.

Here is a new version.

Nico.