Re: [PATCH 03/12] populate_sdk_base: Update extraction script for multilibs

[Mark Hatle <mark.hatle@windriver.com>]

On 10/1/12 10:17 AM, Laurentiu Palcu wrote:
>
>
> On 10/01/2012 05:55 PM, Mark Hatle wrote:
>> On 10/1/12 8:17 AM, Richard Purdie wrote:
>>> On Sat, 2012-09-29 at 19:19 -0500, Mark Hatle wrote:
>>>> When multilibs are enabled, there will be more then one environment
>>>> file created.  We need to be sure to process each environment file.
>>>> The next function can simply use the last environment file processed
>>>> to get the magic value(s) that it requires.
>>>>
>>>> Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
>>>> ---
>>>>    meta/classes/populate_sdk_base.bbclass |    5 +++--
>>>>    1 files changed, 3 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/meta/classes/populate_sdk_base.bbclass b/meta/classes/populate_sdk_base.bbclass
>>>> index 6eb6726..1bc1438 100644
>>>> --- a/meta/classes/populate_sdk_base.bbclass
>>>> +++ b/meta/classes/populate_sdk_base.bbclass
>>>> @@ -158,8 +158,9 @@ echo "done"
>>>>
>>>>    printf "Setting it up..."
>>>>    # fix environment paths
>>>> -env_setup_script=$(find $target_sdk_dir/ -name "environment-setup-${REAL_MULTIMACH_TARGET_SYS}")
> The reason to replace globbing with ${REAL_MULTIMACH_TARGET_SYS} was to
> match only the file we're interested in and avoid matching other files
> whose names start with environment-setup. However, since all the
> environment-setup directories reside in $target_sdk_dir/ we would
> probably be safer, and faster, by doing a simple ls and avoid searching
> all the tree.
>
> env_setup_script=$(ls $target_sdk_dir/environment-setup-*)

That seems reasonable to me, and a lot faster then the find.

>>>> -sed -e "s:$DEFAULT_INSTALL_DIR:$target_sdk_dir:g" -i $env_setup_script
>>>> +for env_setup_script in `find $target_sdk_dir/ -name "environment-setup-*"` ; do
>>>> +  sed -e "s:$DEFAULT_INSTALL_DIR:$target_sdk_dir:g" -i $env_setup_script
>>>> +done
> We could probably do this without the for loop since sed accepts
> multiple input files.

The end result of the env_setups_script though, needs to only contain -one- 
environment script (doesn't matter which one).  That's why I used the loop.

>>>>
>>>>    # fix dynamic loader paths in all ELF SDK binaries
>>>>    native_sysroot=$(cat $env_setup_script |grep OECORE_NATIVE_SYSROOT|cut -d'=' -f2|tr -d '"')

The code snippet above is the reason why it can only contain the one item.. but 
it happens that OECORE_NATIVE_SYSROOT is the same in all environment files (and 
has to be...)  So it's safe to just use the last processed item.

A for loop shouldn't be any slower then calling the set w/ multilib inputs and 
then later adjusting the variable to only be one item.  (We're talking at a 
maximum 3 environment scripts.. thus why it shouldn't be an issue.)

>>>
>>> This is on course to conflict with
>>>
>>> http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=1b6019086c4242c550b4e0551c7b5d206a0d52e1
>>>
>>> Can you please talk with Laurentiu and come up with a solution that
>>> works for everyone.
>>
>> I think this is a better fix for the problem.  The other change limits the
>> environment file to a single file.  However, there is nothing in the function --
>> other then the use of 'env_setuo_script' -- that wants or needs a single file
>> loaded.
>>
>> So it's better to find htem all and just iterate over them all.  Also the order
>> of processing, and the last item processed does not matter.  The later chunks of
>> functionality just look for a static value that is supposed to be the same in
>> all of the environment files.
>>
>> --Mark
>>
>>> Cheers,
>>>
>>> Richard
>>>
>>