Re: Oddities in ./configure --with-pgtk's output

[Po Lu]

Kévin Le Gouguec <kevin.legouguec@gmail.com> writes:

> Mmm.  That was slightly more involved than I expected:
>
> - HAVE_PGTK is defined way after HAVE_WEBP is computed,
>
> - the HAVE_PGTK block seems to depend on $GTK_LIBS, so IIUC it should
>   come after AC_SUBST(GTK_LIBS),
>
> - the part which performs AC_SUBST(GTK_LIBS) comes after the HAVE_WEBP
>   block as well.
>
> The attached patch seems to work (I've ./configure'd --with-pgtk, webp
> support was added automatically, and visiting a webp file with the
> resulting Emacs successfully displays the image), but I have no idea how
> idiomatic my tweaks are.

> +    HAVE_PGTK=yes
> +    AC_DEFINE([HAVE_PGTK], 1, [Define to 1 if you have pure Gtk+-3.])

> @@ -2950,8 +2952,6 @@ AC_DEFUN
>  if test "$window_system" = "pgtk"; then
>    PGTK_OBJ="pgtkfns.o pgtkterm.o pgtkselect.o pgtkmenu.o pgtkim.o 
> xsettings.o"
>    PGTK_LIBS="$GTK_LIBS"
> -  HAVE_PGTK=yes
> -  AC_DEFINE([HAVE_PGTK], 1, [Define to 1 if you have pure Gtk+-3.])

Thanks, LGTM.  While this contribution is small enough to not require
copyright assignment, your previous contributions under the 15-line
exemption may cause this patch to exceed its limit of total lines
contributed.  If you haven't signed the copyright paperwork yet, now
would be a great time to get started!

Eli, has Kevin completed copyright assignment?