All of lore.kernel.org
 help / color / mirror / Atom feed
From: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
To: buildroot@busybox.net
Subject: [Buildroot] [PATCH] bash: fix linking for static builds with uClibc toolchains
Date: Wed, 22 Oct 2014 10:17:08 +0100	[thread overview]
Message-ID: <54477614.1040209@imgtec.com> (raw)
In-Reply-To: <5446C703.7070501@mind.be>

Dear Arnout Vandecappelle,

On 10/21/2014 09:50 PM, Arnout Vandecappelle wrote:
> On 21/10/14 15:21, Vicente Olivert Riera wrote:
>> ...and also use configure options instead of environment variables.
>>
>> bash fails to link for static builds with uClibc toolchains due to
>> getenv redefinitions. This is caused because bash is unable to check if
>> getenv is already defined when cross-compiling, so it defaults to 'yes':
>>
>> configure:14438: WARNING: cannot check getenv redefinition if cross
>> compiling -- defaulting to yes
>>
>> We can avoid this redefinition by adding bash_cv_getenv_redef=no to the
>> configure options.
>>
>> At the same time, we use configure options instead of environment
>> variables because there is no need to pass those options to the
>> configure script as environment variables. Doing it in this way we
>> avoid the risk of environment variables leaking into places we don't
>> expect.
>
>   That change from _ENV to _OPTS should definitely be a separate patch.
> Especially because I don't see the reason to do it. The environment variables
> will leak into exactly one place: the configure script. And that script will
> just export all assignments passed on the command line into its environment. So
> the effect is exactly the same.
>
>   If you do make this change, then it should probably be done for all packages,
> which means it should probably be done in pkg-autotools.mk.
>
>   So you have a concrete example of where the _ENV causes trouble?

No, I haven't. I just prefer using configure options rather than 
environment variables. Anyway..., no problem, I will send another patch 
without that change.

>>
>> Related:
>> 	http://lists.gnu.org/archive/html/bug-bash/2012-03/msg00052.html
>>
>> Fixes:
>> 	http://autobuild.buildroot.net/results/a20/a2007e6dbcfe53e7cd837ae642869ee26376826a/
>>
>> Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
>> ---
>>   package/bash/bash.mk |    5 ++++-
>>   1 files changed, 4 insertions(+), 1 deletions(-)
>>
>> diff --git a/package/bash/bash.mk b/package/bash/bash.mk
>> index 34a3a73..9a13f69 100644
>> --- a/package/bash/bash.mk
>> +++ b/package/bash/bash.mk
>> @@ -13,7 +13,7 @@ BASH_CONF_OPTS = --with-installed-readline
>>   BASH_LICENSE = GPLv3+
>>   BASH_LICENSE_FILES = COPYING
>>
>> -BASH_CONF_ENV += \
>> +BASH_CONF_OPTS += \
>>   	ac_cv_rl_prefix="$(STAGING_DIR)" \
>>   	ac_cv_rl_version="$(READLINE_VERSION)" \
>>   	bash_cv_job_control_missing=present \
>> @@ -28,6 +28,9 @@ BASH_MAKE = $(MAKE1)
>>   # The static build needs some trickery
>>   ifeq ($(BR2_PREFER_STATIC_LIB),y)
>>   BASH_CONF_OPTS += --enable-static-link --without-bash-malloc
>> +ifeq ($(BR2_TOOLCHAIN_USES_UCLIBC),y)
>> +BASH_CONF_OPTS += bash_cv_getenv_redef=no
>> +endif
>
>   It's probably better to also set it to yes explicitly in the else branch.

No problem.

>   Also, a short comment would be good.

No problem.

>   Also, does musl allow redefining getenv?

Well, I did a test with a static build using the musl toolchain and it 
worked fine. So, yes, it seems that musl is ok with the getenv redefinition.

>
>
>   Regards,
>   Arnout
>
>>   endif
>>
>>   # Make /bin/sh -> bash (no other shell, better than busybox shells)
>>
>
>

Best regards,

-- 
Vicente Olivert Riera
Graduate Software Engineer, MIPS Processor IP
Imagination Technologies Limited
t: +44 (0)113 2429814
www.imgtec.com

  reply	other threads:[~2014-10-22  9:17 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-10-21 13:21 [Buildroot] [PATCH] bash: fix linking for static builds with uClibc toolchains Vicente Olivert Riera
2014-10-21 20:50 ` Arnout Vandecappelle
2014-10-22  9:17   ` Vicente Olivert Riera [this message]
2014-10-22  9:35     ` Arnout Vandecappelle
2014-10-22  9:41       ` Vicente Olivert Riera

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=54477614.1040209@imgtec.com \
    --to=vincent.riera@imgtec.com \
    --cc=buildroot@busybox.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.