dev-libs/ffcall: revert patch of support for clang

which is a problem for release 2.4 but fixed in 2.5.

The bug is:

> libtool: compile:  clang -O2 -pipe -g -fPIC -x none -c avcall-x86_64.s -fPIC -DPIC -o .libs/avcall-x86_64.o
> avcall-x86_64.s:591:2: error: changed section flags for .eh_frame, expected: 0x2
>  .section .eh_frame,"aw",@progbits

Which is fixed by change flags "aw" to "a" on all platforms, see commit below:
> https://git.savannah.gnu.org/cgit/libffcall.git/commit/?id=580f0bb144c0d63560c61229291e172e55971437

So the old patch (add "-no-integrated-as" for clang) is reverted:
> https://git.savannah.gnu.org/cgit/libffcall.git/commit/?id=fe6d9f089d8326b077420c10fec8c2174557be30

Fixes: 5b89df93ef
Signed-off-by: Z. Liu <zhixu.liu@gmail.com>
Part-of: https://github.com/gentoo/gentoo/pull/42572
Closes: https://github.com/gentoo/gentoo/pull/42572
Signed-off-by: Sam James <sam@gentoo.org>

dev-libs/ffcall/ffcall-2.5.ebuild

@@ -3,7 +3,7 @@
 
 EAPI=8
 
-inherit autotools dot-a flag-o-matic libtool
+inherit dot-a flag-o-matic libtool
 
 MY_PV="libffcall-${PV}"
 
@@ -19,10 +19,6 @@ LICENSE="GPL-2+"
 SLOT="0"
 KEYWORDS="~alpha amd64 ~arm ~arm64 ~hppa ~mips ppc ppc64 ~riscv ~sparc x86 ~amd64-linux ~x86-linux ~ppc-macos"
 
-PATCHES=(
-	"${FILESDIR}"/ffcall-2.5-add-support-for-clang.patch
-)
-
 src_prepare() {
 	# The build system is a strange mix of autogenerated
 	# files and manual tweaks on top. Uses $CFLAGS / $LDFLAGS randomly.
@@ -38,7 +34,6 @@ src_prepare() {
 	done
 
 	default
-	eautoreconf
 	elibtoolize
 
 	if use kernel_linux ; then

dev-libs/ffcall/files/ffcall-2.5-add-support-for-clang.patch

@@ -1,53 +0,0 @@
-https://git.savannah.gnu.org/cgit/libffcall.git/commit/?id=fe6d9f089d8326b077420c10fec8c2174557be30
-
-From fe6d9f089d8326b077420c10fec8c2174557be30 Mon Sep 17 00:00:00 2001
-From: Bruno Haible <bruno@clisp.org>
-Date: Tue, 20 Aug 2024 13:38:19 +0200
-Subject: [PATCH] Add support for Android.
-
-* m4/cc-gcc.m4 (CL_CC_GCC): If using clang as compiler, add the option
--no-integrated-as to GCC_X_NONE.
-* NEWS: Mention Android as supported.
-
-diff --git a/m4/cc-gcc.m4 b/m4/cc-gcc.m4
-index 1e12d8e..97c2dac 100644
---- a/m4/cc-gcc.m4
-+++ b/m4/cc-gcc.m4
-@@ -1,5 +1,5 @@
- dnl -*- Autoconf -*-
--dnl Copyright (C) 1993-2017 Free Software Foundation, Inc.
-+dnl Copyright (C) 1993-2024 Free Software Foundation, Inc.
- dnl This file is free software, distributed under the terms of the GNU
- dnl General Public License as published by the Free Software Foundation;
- dnl either version 2 of the License, or (at your option) any later version.
-@@ -27,6 +27,27 @@ AC_DEFUN([CL_CC_GCC],
-   if test $cl_cv_prog_cc_gcc = yes; then
-     CC_GCC=true
-     GCC_X_NONE='-x none'
-+
-+    dnl On arm, there are two slightly different syntaxes for assembly language:
-+    dnl The "divided" syntax (standard since ca. 2000) and the "unified" syntax
-+    dnl (supported by GNU binutils since 2009).  See
-+    dnl https://sourceware.org/binutils/docs/as/ARM_002dInstruction_002dSet.html
-+    dnl While GNU as and thus GCC supports both, clang by default uses an
-+    dnl "integrated" assembler that supports only the "unified" syntax, leading
-+    dnl to errors "error: invalid instruction" on instructions such as stmeqia.
-+    dnl We therefore need to tell clang to use the external assembler.
-+    AC_CACHE_CHECK([whether using clang], [cl_cv_prog_cc_clang],
-+      [AC_EGREP_CPP([yes],
-+         [#ifdef __clang__
-+          yes
-+          #endif
-+         ],
-+         [cl_cv_prog_cc_clang=yes],
-+         [cl_cv_prog_cc_clang=no])
-+      ])
-+    if test $cl_cv_prog_cc_clang = yes; then
-+      GCC_X_NONE="$GCC_X_NONE -no-integrated-as"
-+    fi
-   else
-     CC_GCC=false
-     GCC_X_NONE=''
--- 
-2.45.2
-