Mark H Weaver writes:
> GCC 4.7.4 fails to build since April 2017:
>
> https://hydra.gnu.org/job/gnu/master/gcc-4.7.4.x86_64-linux
> https://hydra.gnu.org/job/gnu/master/gcc-4.7.4.i686-linux
The attached patch should fix this. I tested it on an inherited package
like this below, while attempting to create diverse double compilation
(define-public repro-gcc-4.7
(package
(inherit gcc-4.7-$ORIGIN)
(source
(origin
(inherit (package-source gcc-4.7-$ORIGIN))
(patches (append ((compose origin-patches package-source) gcc-4.7)
(search-patches
"gcc-5-reproducibility-drop-profile.patch"
"gcc-4-cfns-fix-mismatch-in-gnu_inline-attributes.patch"
"gcc-4-build-path-prefix-map.patch")))))
(name "repro-gcc")
(version "4.7.4")
(arguments
(substitute-keyword-arguments (package-arguments gcc-4.7-$ORIGIN)
((#:phases original-phases)
`(modify-phases ,original-phases
(add-before 'configure 'build-prefix-path
(lambda* (#:key inputs #:allow-other-keys)
(setenv "BUILD_PATH_PREFIX_MAP"
(string-append "gcc" "-" ,version "=" (getcwd)))
(format (current-error-port)
"BUILD_PATH_PREFIX_MAP=~s\n"
(getenv "BUILD_PATH_PREFIX_MAP"))))))))))
Greetings,
janneke
>From b2fb0adc3e0de7194493a0c5f1f9bbdbcd0a4087 Mon Sep 17 00:00:00 2001
From: Jan Nieuwenhuizen <jann...@gnu.org>
Date: Mon, 6 Nov 2017 22:50:05 +0100
Subject: [PATCH] gnu: gcc-4.7: Resurrect building with gcc-5.4.0.
* gnu/packages/patches/gcc-4-cfns-fix-mismatch-in-gnu_inline-attributes.patch:
New file.
* gnu/local.mk (dist_patch_DATA): Add it.
* gnu/packages/gcc.scm (gcc-4.7): Use it.
---
gnu/local.mk | 1 +
gnu/packages/gcc.scm | 3 +
...fns-fix-mismatch-in-gnu_inline-attributes.patch | 65 ++++++++++++++++++++++
3 files changed, 69 insertions(+)
create mode 100644 gnu/packages/patches/gcc-4-cfns-fix-mismatch-in-gnu_inline-attributes.patch
diff --git a/gnu/local.mk b/gnu/local.mk
index 5dfcf497b..76aef903a 100644
--- a/gnu/local.mk
+++ b/gnu/local.mk
@@ -637,6 +637,7 @@ dist_patch_DATA = \
%D%/packages/patches/gcc-cross-environment-variables.patch \
%D%/packages/patches/gcc-libvtv-runpath.patch \
%D%/packages/patches/gcc-strmov-store-file-names.patch \
+ %D%/packages/patches/gcc-4-cfns-fix-mismatch-in-gnu_inline-attributes.patch \
%D%/packages/patches/gcc-4.6-gnu-inline.patch \
%D%/packages/patches/gcc-4.9.3-mingw-gthr-default.patch \
%D%/packages/patches/gcc-5.0-libvtv-runpath.patch \
diff --git a/gnu/packages/gcc.scm b/gnu/packages/gcc.scm
index 7870d4513..2991cbd0b 100644
--- a/gnu/packages/gcc.scm
+++ b/gnu/packages/gcc.scm
@@ -136,6 +136,9 @@ where the OS part is overloaded to denote a specific ABI---into GCC
(method url-fetch)
(uri (string-append "mirror://gnu/gcc/gcc-"
version "/gcc-" version ".tar.bz2"))
+ (patches
+ (search-patches
+ "gcc-4-cfns-fix-mismatch-in-gnu_inline-attributes.patch"))
(sha256
(base32
"10k2k71kxgay283ylbbhhs51cl55zn2q38vj5pk4k950qdnirrlj"))))
diff --git a/gnu/packages/patches/gcc-4-cfns-fix-mismatch-in-gnu_inline-attributes.patch b/gnu/packages/patches/gcc-4-cfns-fix-mismatch-in-gnu_inline-attributes.patch
new file mode 100644
index 000000000..861cd4857
--- /dev/null
+++ b/gnu/packages/patches/gcc-4-cfns-fix-mismatch-in-gnu_inline-attributes.patch
@@ -0,0 +1,65 @@
+Taken from https://gcc.gnu.org/cgi-bin/get-raw-msg?listname=gcc-patches&date=2016-01&msgid=1451802493-17406-1-git-send-email-vapier%40gentoo.org
+
+Since the 3.0.3 release of gperf (made in May 2007), the generated func
+has had the gnu_inline attribute applied to it. The gcc source however
+has not been updated to include that which has lead to a mismatch.
+
+In practice, this hasn't been an issue for two reasons:
+(1) Before gcc-5, the default standard was (gnu) C89, and gcc does not
+warn or throw an error in this mode.
+(2) Starting with gcc-4.8, the compiler driver used to build gcc was
+changed to C++, and g++ does not warn or throw an error in this mode.
+
+This error does show up though when using gcc-5 to build gcc-4.7 or
+older as then the default is (gnu) C11 and the C compiler driver is
+used. That failure looks like:
+In file included from .../gcc-4.7.4/gcc/cp/except.c:990:0:
+cfns.gperf: At top level:
+cfns.gperf:101:1: error: 'gnu_inline' attribute present on 'libc_name_p'
+cfns.gperf:26:14: error: but not here
+
+Whether the compiler should always emit this error regardless of the
+active standard or compiler driver is debatable (I think it should be
+consistent -- either always do it or never do it).
+
+2015-08-06 Mike Frysinger <vap...@gentoo.org>
+
+ * cfns.gperf [__GNUC__, __GNUC_STDC_INLINE__]: Apply the
+ __gnu_inline__ attribute.
+ * cfns.h: Regenerated.
+---
+ gcc/cp/cfns.gperf | 3 +++
+ gcc/cp/cfns.h | 3 +++
+ 2 files changed, 6 insertions(+)
+
+diff --git a/gcc/cp/cfns.gperf b/gcc/cp/cfns.gperf
+index 68acd3d..953262f 100644
+--- a/gcc/cp/cfns.gperf
++++ b/gcc/cp/cfns.gperf
+@@ -22,6 +22,9 @@ __inline
+ static unsigned int hash (const char *, unsigned int);
+ #ifdef __GNUC__
+ __inline
++#ifdef __GNUC_STDC_INLINE__
++__attribute__ ((__gnu_inline__))
++#endif
+ #endif
+ const char * libc_name_p (const char *, unsigned int);
+ %}
+diff --git a/gcc/cp/cfns.h b/gcc/cp/cfns.h
+index 1c6665d..6d00c0e 100644
+--- a/gcc/cp/cfns.h
++++ b/gcc/cp/cfns.h
+@@ -53,6 +53,9 @@ __inline
+ static unsigned int hash (const char *, unsigned int);
+ #ifdef __GNUC__
+ __inline
++#ifdef __GNUC_STDC_INLINE__
++__attribute__ ((__gnu_inline__))
++#endif
+ #endif
+ const char * libc_name_p (const char *, unsigned int);
+ /* maximum key range = 391, duplicates = 0 */
+--
+2.6.2
+
--
Jan Nieuwenhuizen <jann...@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar® http://AvatarAcademy.com
--
Jan Nieuwenhuizen <jann...@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar® http://AvatarAcademy.com
