diff mbox

FIXUP: [PATCH 01/10] gnu: cross: Use CROSS_*_INCLUDE_PATH for system headers.

Message ID 87vb2hcahb.fsf_-_@drakenvlieg.flower
State New
Headers show

Commit Message

Jan Nieuwenhuizen May 14, 2016, 8:25 a.m. UTC
Jan Nieuwenhuizen writes:

Fix for

    ./pre-inst-env guix build --target=mips64el-linux-gnuabi64 hello

attached.  Adds srfi-26 to glibc for cut.

As Andy suggested I will be factoring-out the for-each (setenv) CROSS_*
stanzas but I'm still experimenting with that, with cross-libtool and
the whole lilypond stack.

Greetings,
Jan

Comments

Ludovic Courtès May 14, 2016, 8:05 p.m. UTC | #1
Jan Nieuwenhuizen <janneke@gnu.org> skribis:

> From 0361374c4469b1e9b573668a6020b19e7abfbfe2 Mon Sep 17 00:00:00 2001
> From: Jan Nieuwenhuizen <janneke@gnu.org>
> Date: Sat, 14 May 2016 10:13:21 +0200
> Subject: [PATCH] gnu: cross-base: Add srfi-26 for glibc.  Fixes glibc-based
>  cross builds.
>
> * gnu/packages/cross-base.scm (glibc): Add (srfi srfi-26) to (arguments

Thanks for your quick response, applied with an adjusted commit log.

I concur with Mark: this is a very sensitive area that takes time to
test and adjust, which is the main reason why it takes me so much time
to review and such ;-), so we want to make sure we don’t break it.

In the future, please make sure to test individual patches with the
known-good *-linux-gnu targets.

Thanks!

Ludo’.
Jan Nieuwenhuizen May 14, 2016, 8:50 p.m. UTC | #2
Ludovic Courtès writes:

> Thanks for your quick response, applied with an adjusted commit log.

Thanks.

> I concur with Mark: this is a very sensitive area that takes time to
> test and adjust, which is the main reason why it takes me so much time
> to review and such ;-), so we want to make sure we don’t break it.

Sure, I agree.  These two real silly problems I introduced only makes me
appreciate more that it takes some time to get the patches in ;-)

> In the future, please make sure to test individual patches with the
> known-good *-linux-gnu targets.

Yes.  I wrongly assumed (though that I knew and saw happen) that
cross-gcc and cross-libc were used in the bootstrap process that I was
thus testing them...I've seen a lot of my build host compilers and libc
getting built and rebuilt.  Now I know that's not the case and need to
test for that when touching this area.  That's still no excuse for
not testing the cross-base.scm patches individually :-/

I would still like to explore getting rid of the patched gcc and all
the environment variables and use a --sysroot instead; but I would
need some help/inspiration to get that experiment going.

Greetings, Jan
Ludovic Courtès May 16, 2016, 9:02 p.m. UTC | #3
Jan Nieuwenhuizen <janneke@gnu.org> skribis:

> I would still like to explore getting rid of the patched gcc and all
> the environment variables and use a --sysroot instead; but I would
> need some help/inspiration to get that experiment going.

I’m not sure this is doable or desirable.  AIUI --sysroot is pretty much
biased towards FHS.

My 2¢,
Ludo’.
diff mbox

Patch

From 0361374c4469b1e9b573668a6020b19e7abfbfe2 Mon Sep 17 00:00:00 2001
From: Jan Nieuwenhuizen <janneke@gnu.org>
Date: Sat, 14 May 2016 10:13:21 +0200
Subject: [PATCH] gnu: cross-base: Add srfi-26 for glibc.  Fixes glibc-based
 cross builds.

* gnu/packages/cross-base.scm (glibc): Add (srfi srfi-26) to (arguments
---
 gnu/packages/cross-base.scm | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/gnu/packages/cross-base.scm b/gnu/packages/cross-base.scm
index 73ac76a..7ed4ae4 100644
--- a/gnu/packages/cross-base.scm
+++ b/gnu/packages/cross-base.scm
@@ -323,6 +323,11 @@  XBINUTILS and the cross tool chain."
            ;; itself.
            #:implicit-cross-inputs? #f
 
+           ;; We need cut from srfi-26
+           #:modules ((guix build gnu-build-system)
+                      (guix build utils)
+                      (srfi srfi-26))
+
            ,@(package-arguments glibc))
        ((#:configure-flags flags)
         `(cons ,(string-append "--host=" target)
-- 
2.7.3