Sorry, wrong version.

Created attachment 51819 [details] [review]
possbile cleanups

The child_watch_helper_thread thread change looks fine (check to make sure
that GCC doesn't warn about the removal of the trailing return NULL.)

Maybe you can explain the rest of the patch? It doesn't make immediate sense
to me. (Or look like it would even compile.)

icc complained that child_watch_source was set but never used in
g_child_watch_check() so I removed those two lines. While there I noticed that
g_child_watch_prepare() had the same issue and after removing those lines from
there it ended up just setting timeout = -1 and returning. So I looked for uses
of that function in that file since it was static and there were none. So if the
function was unused and static I figured it also didn't need to be ported to
Win32. It does compile here with icc, trying gcc now...it compiles but with the
following warning:

gmain.c:286: warning: initialization from incompatible pointer type
gmain.c:287: warning: initialization from incompatible pointer type

Looking at the code it's clear that removing it is probably the wrong thing to
do, but is icc right in complaining about the variables being set and never used?

You are right that the variables can be removed, but I think you missed
the fact the the return statement does the actual work:

  return check_for_child_exited (source);

The function is used as a function pointer:

GSourceFuncs g_child_watch_funcs =
{
  g_child_watch_prepare,
  g_child_watch_check,
  g_child_watch_dispatch,
  NULL
};

This gives a list of functions for the main looop to use at various
stages. If you remove the first line, then you have check in the prepare
slot and dispatch in the check slot, and NULL in the dispatch slot,
and things wont' work at all well :-)

Indeed I did ;-)

So, after removing the fluff we have two matching functions with different names
only one of them takes a gint parameter that it never uses for anything...

Created attachment 51863 [details] [review]
updated patch

The *timeout = -1 (note the '*') is not extraneous and cannot be removed.

Created attachment 51864 [details] [review]
updated patch

Hopefully this time it's correct then :-)

Yes, looks correct now. Please commit to head only.

Commited.