Document the reality that some dumps were faked for purpose of testing
corner cases.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
---
 tests/qemucapabilitiesdata/README.rst | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/tests/qemucapabilitiesdata/README.rst b/tests/qemucapabilitiesdata/README.rst
index 90e240f800..c5ddc6544d 100644
--- a/tests/qemucapabilitiesdata/README.rst
+++ b/tests/qemucapabilitiesdata/README.rst
@@ -116,3 +116,15 @@ The tool validates and updates the numbering of the entries in the QMP dump in
 case something was modified.

 The tool also allows programatic modification of the ``.replies`` file.
+
+Fake test data dumps for certain architectures
+==============================================
+
+For some architectures it was impossible or impractical to fetch real capability
+dumps. To ensure coverate of certain cases the dumps were collected from
+corresponding binaries running on a different architecture.
+
+The following files are fake:
+
+ - ``caps_7.2.0_ppc``
+   Captured on x86_64 host running fedora.
-- 
2.43.0
_______________________________________________
Devel mailing list -- devel@lists.libvirt.org
To unsubscribe send an email to devel-leave@lists.libvirt.org

On Tue, Feb 27, 2024 at 05:36:04PM +0100, Peter Krempa wrote:
> +Fake test data dumps for certain architectures
> +==============================================
> +
> +For some architectures it was impossible or impractical to fetch real capability
> +dumps. To ensure coverate of certain cases the dumps were collected from
> +corresponding binaries running on a different architecture.
> +
> +The following files are fake:
> +
> + - ``caps_7.2.0_ppc``
> +   Captured on x86_64 host running fedora.

This list is guaranteed to go stale over time, and it's already
missing some entries right now (caps_5.1.0_sparc, caps_*_riscv64).

I would use more generic, future-proof language instead: something
along the lines of

  Capabilities dumps for the following architectures are usually
  produced on real hardware:

  - x86_64
  - aarch64
  - ppc64
  - s390x

  In most other cases, x86_64 will be used as the host architecture.

If there's a simple way to check whether a specific capability dump
has been produced on real hardware (presence of KVM support?) that
could be mentioned here as well for completeness' sake.

-- 
Andrea Bolognani / Red Hat / Virtualization
_______________________________________________
Devel mailing list -- devel@lists.libvirt.org
To unsubscribe send an email to devel-leave@lists.libvirt.org