Re: [PATCH] eglibc: fix as and ld check in libc dir

[Giuseppe CONDORELLI <giuseppe.condorelli@st.com>]

[-- Attachment #1: Type: text/plain, Size: 1789 bytes --]

 

 

From: Martin Jansa [mailto:martin.jansa@gmail.com] 
Sent: venerdì 18 gennaio 2013 09:55
To: Giuseppe CONDORELLI
Cc: Giuseppe Condorelli; Patches and discussions about the oe-core layer
Subject: Re: [OE-core] [PATCH] eglibc: fix as and ld check in libc dir

 

 

On Fri, Jan 18, 2013 at 9:32 AM, Giuseppe CONDORELLI <giuseppe.condorelli@st.com> wrote:

Yes Martin,

but I discovered the issue building (erroneously because eglibc was actually
not required at all) an external-toolchain a little bit old (gcc and
binutils) against the master oe-core repository,
which contains eglic-2.16 as you know. And libc/configure comes with the
issue I highlighted.
So, please, think to my patch as a nice to have and not as a mandatory one.
I don't know if other people will fall in my same condition.

Thanks so much for your support.
Giuseppe


And are you using eglibc-2.16 built with binutils-2.19 in runtime on some targets?

 

In fact I wrote it was my fault, indeed eglibc has not to be built at all given my external toolchain fully provides all the infrastructure.

However this mistake highlighted the possible issue

 

My point is that if upstream developers decided to support binutils-2.20+ only, they probably had some reason to do it and failing configure check is better way to show you're building unsupported combination then something failing badly in runtime on target.

 

http://sourceware.org/git/?p=glibc.git;a=commit;f=configure.in;h=bec039bcefe6a0494a249e2b78e112a0ab60893b

This commit also doesn't say why..

 

Ok, no problem. The patch can remain unapplied if you think it could be useless. At any rate it will remain as document if other people will fall in that for any reason.

 

Cheers,

Cheers,


[-- Attachment #2: Type: text/html, Size: 5466 bytes --]