[oe] external-toolchain-csl-1.0-r6 fails

[Koen Kooi]

Op 23 nov 2010, om 16:16 heeft Tom Rini het volgende geschreven:

> On 11/23/2010 07:57 AM, Víctor M. Jáquez L. wrote:
>> On Tue, Nov 23, 2010 at 07:40:18AM -0700, Tom Rini wrote:
>>> On 11/23/2010 04:37 AM, Víctor M. Jáquez L. wrote:
>>>> I'm trying to build minimal-image with CSL toolchain, but I got this error:
>>>> 
>>>> NOTE: package external-toolchain-csl-1.0-r6: task do_install: Started
>>>> ERROR: TaskFailed event exception, aborting
>>>> ERROR: Build of /media/DISK/vjaquez-angstrom-setup-scripts/sources/openembedded/recipes/meta/external-toolchain-csl.bb do_install failed
>>>> ERROR: Task 187 (/media/DISK/vjaquez-angstrom-setup-scripts/sources/openembedded/recipes/meta/external-toolchain-csl.bb, do_install) failed with 256
>>>> ERROR: '/media/DISK/vjaquez-angstrom-setup-scripts/sources/openembedded/recipes/meta/external-toolchain-csl.bb' failed
>>>> ERROR: Function do_install failed
>>>> NOTE: Task failed: ('function do_install failed', '/media/DISK/vjaquez-angstrom-setup-scripts/build/tmp-angstrom_2008_1/work/armv7a-angstrom-linux-gnueabi/external-toolchain-csl-1.0-r6/temp/log.do_install.22227')
>>>> ERROR: Logfile of failure stored in: /media/DISK/vjaquez-angstrom-setup-scripts/build/tmp-angstrom_2008_1/work/armv7a-angstrom-linux-gnueabi/external-toolchain-csl-1.0-r6/temp/log.do_install.22227
>>>> Log data follows:
>>>> | cp: cannot stat `/media/DISK/sgpp/arm-2010q1/arm-angstrom-linux-gnueabi/libc/lib/*': No such file or directory
>>>> | ERROR: Function do_install failed
>>>> NOTE: package external-toolchain-csl-1.0-r6: task do_install: Failed
>>>> ERROR: TaskFailed event exception, aborting
>>>> 
>>>> For some reason the variable ${TARGET_SYS} is evaluated as
>>>> arm-angstrom-linux-gnueabi instead of arm-none-linux-gnueabi which is the
>>>> correct value.
>>> 
>>> Please post your local.conf settings as well as what version of
>>> openembedded (and what branch) you are using.
>> 
>> I'm using the openembedded's branch org.openembedded.dev (HEAD)
> [snip]
>> 
>> If I apply this patch[1] the problem seems to be solved.
>> 
>> 1. http://www.mail-archive.com/openembedded-devel@lists.openembedded.org/msg11415.html
> 
> I thought the patch had been applied.  Koen, do you still object after it was explained that currently angstrom is always going to win setting that variable unless it's made a ?= ?  Thanks!

I still object to that patch, since a different solution has been applied already. And I still don't see why people want to use known broken toolchains...