Guillaume Le Vaillant <glv@posteo.net> writes:

> Christopher Baines <mail@cbaines.net> skribis:
>
>> If the new-name matches the original name, just use the original package.
>> This avoids situations where there are several packages matching the name and
>> the behaviour will be inconsistent, occasionally picking different packages.
>>
>> Since there are multiple glibc packages currently, I'm seeing this behaviour
>> with cl-posix-mqueue and ecl-cl-posix-mqueue, occasionally they'll use the
>> hurd glibc variant.
>>
>> * guix/build-system/asdf.scm (package-with-build-system): Use the original
>> input packages unless the new-name differs.
>>
>> Change-Id: I08a1f3ad1290689b5497d31950ada4dc0bfa3a3a
>> ---
>>  guix/build-system/asdf.scm | 10 +++++++---
>>  1 file changed, 7 insertions(+), 3 deletions(-)
>>
>> diff --git a/guix/build-system/asdf.scm b/guix/build-system/asdf.scm
>> index 26b5a5008a..ad0fb993f6 100644
>> --- a/guix/build-system/asdf.scm
>> +++ b/guix/build-system/asdf.scm
>> @@ -155,9 +155,13 @@ (define* (package-with-build-system from-build-system to-build-system
>>  
>>    (define (find-input-package pkg)
>>      (let* ((name (package-name pkg))
>> -           (new-name (transform-package-name name))
>> -           (pkgs (find-packages-by-name new-name)))
>> -      (if (null? pkgs) #f (list-ref pkgs 0))))
>> +           (new-name (transform-package-name name)))
>> +      (if (string=? name new-name)
>> +          pkg
>> +          (let ((pkgs (find-packages-by-name new-name)))
>> +            (if (null? pkgs)
>> +                #f
>> +                (list-ref pkgs 0))))))
>>  
>>    (define transform
>>      (mlambda (pkg)
>>
>> base-commit: 90ee330bafc5a95493f9cdae2e32ddf740104ebc
>
> Yes, that seems logical. I have not yet tried recompiling packages with
> this patch, but visually it looks good to me.

Thanks for taking a look, with the minimal testing I've done locally it
seems OK, so I've pushed this as
15615db61b106c6e77cfabe042edb9e77d578c95.

Chris