Re: Fix for Emacs 23.2 compilation issue

[Kenichi Handa]

In article <address@hidden>, Karel Klic <address@hidden> writes:

> Emacs 23.2 compilation fails in Fedora Rawhide:
> /builddir/build/BUILD/emacs-23.2/src/ftfont.c:1924: undefined reference 
> to `OTF_drive_gsub_with_log'

> M17N_FLT_USE_NEW_FEATURE (revno 99777) in ftfont.c has been enabled even 
> with libotf 0.9.9, which does not implement OTF_drive_gsub_with_log.

> I fixed it by adding some brackets:

> diff -up emacs-23.2/src/ftfont.c.m17ncheck emacs-23.2/src/ftfont.c
> --- emacs-23.2/src/ftfont.c.m17ncheck 2010-05-11 11:22:19.977753842 +0200
> +++ emacs-23.2/src/ftfont.c   2010-05-11 11:22:44.545753979 +0200
> @@ -1578,8 +1578,8 @@ ftfont_otf_capability (font)

>   #ifdef HAVE_M17N_FLT

> -#if ((LIBOTF_MAJOR_VERSION > 1) || (LIBOTF_RELEASE_NUMBER >= 10) \
> -     && (M17NLIB_MAJOR_VERSION > 1) || (M17NLIB_MINOR_VERSION >= 6))
> +#if (((LIBOTF_MAJOR_VERSION > 1) || (LIBOTF_RELEASE_NUMBER >= 10))   \
> +     && ((M17NLIB_MAJOR_VERSION > 1) || (M17NLIB_MINOR_VERSION >= 6)))
>   /* We can use the new feature of libotf and m17n-flt to handle the
>      character encoding scheme introduced in Unicode 5.1 and 5.2 for
>      some Agian scripts.  */

Oops, I'm very sorry for this stupid mistake.  Perhaps I
made some mistake (setting of LD_LIBRARY_PATH?) when I
tested the building of Emacs with libotf 0.9.9 before
committing that change.  I've just installed your patch to
emacs-23 branch.

---
Kenichi Handa
address@hidden