bug#58567: Some grafts use a different input derivation than computed by --no-grafts

[zimoun <zimon.toutoune@gmail.com>]

Hi Marius,

On lun., 17 oct. 2022 at 13:36, Marius Bakke <marius@gnu.org> wrote:

>> 1: <https://yhetil.org/guix/874jy87gcl.fsf@gmail.com>
>
> Not sure if it's the same problem, but it is weird.

About mesa, the graft machinery produces two different items with the
same content.  It is because multi-outputs; as pointed in [1].


> The bug here is that Guix treats grafted origins as different, even
> though their outputs are known in advance.

About python-patch-ng, I miss the origin of the difference, e.g.,

--8<---------------cut here---------------start------------->8---
diff -r --no-dereference /gnu/store/a58yf1jbryyffzs4i8zp8ywns1b5hrvs-python-patch-ng-1.17.4/lib/python3.9/site-packages/patch_ng.py /gnu/store/d6nhxbayyal1nximx048bvk6zx4phcap-python-patch-ng-1.17.4/lib/python3.9/site-packages/patch_ng.py
1c1
< #!/gnu/store/9qz2zckx1mlcg8lijl7rb4fyxygv32ml-python-wrapper-3.9.9/bin/python
---
> #!/gnu/store/slsh0qjv5j68xda2bb6h8gsxwyi1j25a-python-wrapper-3.9.9/bin/python
Binary files /gnu/store/a58yf1jbryyffzs4i8zp8ywns1b5hrvs-python-patch-ng-1.17.4/lib/python3.9/site-packages/__pycache__/patch_ng.cpython-39.pyc and /gnu/store/d6nhxbayyal1nximx048bvk6zx4phcap-python-patch-ng-1.17.4/lib/python3.9/site-packages/__pycache__/patch_ng.cpython-39.pyc differ
--8<---------------cut here---------------end--------------->8---

and then, this python-wrapper reads,

--8<---------------cut here---------------start------------->8---
$ guix gc --derivers /gnu/store/9qz2zckx1mlcg8lijl7rb4fyxygv32ml-python-wrapper-3.9.9
/gnu/store/xm26mvbldnqa081mbnnlcikn30xxvzrg-python-wrapper-3.9.9.drv

$ guix gc --derivers /gnu/store/slsh0qjv5j68xda2bb6h8gsxwyi1j25a-python-wrapper-3.9.9
/gnu/store/28b2j5m498bry3x33by2y7h8ms5fsxmk-python-wrapper-3.9.9.drv
/gnu/store/d78g4awha9cplcxmz7ssxdd1jgn55iym-python-wrapper-3.9.9.drv
/gnu/store/l2x2bh5l37cjiifv9qws9700vb0h583j-python-wrapper-3.9.9.drv
--8<---------------cut here---------------end--------------->8---


About telegram-desktop, it is probably another origin of difference.
See [2].

2: <https://issues.guix.gnu.org/msgid/8735bmn2t9.fsf@gmail.com>


> Probably the grafting machinery should ignore fixed-output derivations
> somehow?

Well, I do not know what could be fixed about all these mysteries. :-)


Cheers,
simon