RFC/WIP - The issue being worked around here is the fact that if we
install our local python packages to the pyvenv environment, we need
additional libraries available to actually "build" the package to
install it; our alpine linux environment doesn't have the necessary
packages and so testing fails at this step unless we allow it to use the
existing packages already installed to perform the build.

I'm not sure why it's just Alpine, or if there's a smarter way to solve
this, or why this issue doesn't impact our existing methods of
installing the local packages.

For me to investigate.

Signed-off-by: John Snow <jsnow@redhat.com>
---
 python/scripts/mkvenv.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/python/scripts/mkvenv.py b/python/scripts/mkvenv.py
index e38292b63a6..9cf7a6e2b2a 100644
--- a/python/scripts/mkvenv.py
+++ b/python/scripts/mkvenv.py
@@ -677,6 +677,7 @@ def pip_install(
         "-m",
         "pip",
         "install",
+        "--no-build-isolation",
         "--disable-pip-version-check",
         "-v" if loud else "-q",
     ]
-- 
2.51.1