From: John Snow <jsnow@redhat.com>
To: qemu-devel@nongnu.org
Cc: "Warner Losh" <imp@bsdimp.com>, "Beraldo Leal" <bleal@redhat.com>,
"John Snow" <jsnow@redhat.com>, "Kyle Evans" <kevans@freebsd.org>,
"Paolo Bonzini" <pbonzini@redhat.com>,
"Thomas Huth" <thuth@redhat.com>,
"Daniel Berrange" <berrange@redhat.com>,
"Reinoud Zandijk" <reinoud@netbsd.org>,
"Wainer dos Santos Moschetta" <wainersm@redhat.com>,
"Cleber Rosa" <crosa@redhat.com>,
"Ryo ONODERA" <ryoon@netbsd.org>,
"Philippe Mathieu-Daudé" <philmd@linaro.org>,
"Ani Sinha" <ani@anisinha.ca>,
"Michael S. Tsirkin" <mst@redhat.com>,
"Alex Bennée" <alex.bennee@linaro.org>
Subject: [RFC PATCH v3 07/20] mkvenv: add nested venv workaround
Date: Mon, 24 Apr 2023 16:02:35 -0400 [thread overview]
Message-ID: <20230424200248.1183394-8-jsnow@redhat.com> (raw)
In-Reply-To: <20230424200248.1183394-1-jsnow@redhat.com>
Python virtual environments do not typically nest; they may inherit from
the top-level system packages or not at all.
For our purposes, it would be convenient to emulate "nested" virtual
environments to allow callers of the configure script to install
specific versions of python utilities in order to test build system
features, utility version compatibility, etc.
While it is possible to install packages into the system environment
(say, by using the --user flag), it's nicer to install test packages
into a totally isolated environment instead.
As detailed in https://www.qemu.org/2023/03/24/python/, Emulate a nested
venv environment by using .pth files installed into the site-packages
folder that points to the parent environment when appropriate.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: John Snow <jsnow@redhat.com>
---
python/scripts/mkvenv.py | 72 ++++++++++++++++++++++++++++++++++++++--
1 file changed, 69 insertions(+), 3 deletions(-)
diff --git a/python/scripts/mkvenv.py b/python/scripts/mkvenv.py
index 445f4eb092..45d1b772e5 100644
--- a/python/scripts/mkvenv.py
+++ b/python/scripts/mkvenv.py
@@ -51,9 +51,11 @@
import os
from pathlib import Path
import re
+import site
import stat
import subprocess
import sys
+import sysconfig
import traceback
from types import SimpleNamespace
from typing import (
@@ -74,6 +76,11 @@
logger = logging.getLogger("mkvenv")
+def inside_a_venv() -> bool:
+ """Returns True if it is executed inside of a virtual environment."""
+ return sys.prefix != sys.base_prefix
+
+
class Ouch(RuntimeError):
"""An Exception class we can't confuse with a builtin."""
@@ -82,9 +89,15 @@ class QemuEnvBuilder(venv.EnvBuilder):
"""
An extension of venv.EnvBuilder for building QEMU's configure-time venv.
- The only functional change is that it adds the ability to regenerate
- console_script shims for packages available via system_site
- packages.
+ The primary differences are:
+
+ (1) It adds the ability to regenerate console_script shims for
+ packages available via system_site_packages for any packages
+ specified by the 'script_packages' argument
+
+ (2) It emulates a "nested" virtual environment when invoked from
+ inside of an existing virtual environment by including packages from
+ the parent.
Parameters for base class init:
- system_site_packages: bool = False
@@ -99,11 +112,51 @@ class QemuEnvBuilder(venv.EnvBuilder):
def __init__(self, *args: Any, **kwargs: Any) -> None:
logger.debug("QemuEnvBuilder.__init__(...)")
self.script_packages = kwargs.pop("script_packages", ())
+
+ # For nested venv emulation:
+ self.use_parent_packages = False
+ if inside_a_venv():
+ # Include parent packages only if we're in a venv and
+ # system_site_packages was True.
+ self.use_parent_packages = kwargs.pop(
+ "system_site_packages", False
+ )
+ # Include system_site_packages only when the parent,
+ # The venv we are currently in, also does so.
+ kwargs["system_site_packages"] = sys.base_prefix in site.PREFIXES
+
super().__init__(*args, **kwargs)
# Make the context available post-creation:
self._context: Optional[SimpleNamespace] = None
+ def compute_venv_libpath(self, context: SimpleNamespace) -> str:
+ """
+ Compatibility wrapper for context.lib_path for Python < 3.12
+ """
+ # Python 3.12+, not strictly necessary because it's documented
+ # to be the same as 3.10 code below:
+ if sys.version_info >= (3, 12):
+ return context.lib_path
+
+ # Python 3.10+
+ if "venv" in sysconfig.get_scheme_names():
+ return sysconfig.get_path(
+ "purelib", scheme="venv", vars={"base": context.env_dir}
+ )
+
+ # For Python <= 3.9 we need to hardcode this. Fortunately the
+ # code below was the same in Python 3.6-3.10, so there is only
+ # one case.
+ if sys.platform == "win32":
+ return os.path.join(context.env_dir, "Lib", "site-packages")
+ return os.path.join(
+ context.env_dir,
+ "lib",
+ "python%d.%d" % sys.version_info[:2],
+ "site-packages",
+ )
+
def ensure_directories(self, env_dir: DirType) -> SimpleNamespace:
logger.debug("ensure_directories(env_dir=%s)", env_dir)
self._context = super().ensure_directories(env_dir)
@@ -124,6 +177,19 @@ def post_post_setup(self, context: SimpleNamespace) -> None:
"""
The final, final hook. Enter the venv and run commands inside of it.
"""
+ if self.use_parent_packages:
+ # We're inside of a venv and we want to include the parent
+ # venv's packages.
+ parent_libpath = sysconfig.get_path("purelib")
+ logger.debug("parent_libpath: %s", parent_libpath)
+
+ our_libpath = self.compute_venv_libpath(context)
+ logger.debug("our_libpath: %s", our_libpath)
+
+ pth_file = os.path.join(our_libpath, "nested.pth")
+ with open(pth_file, "w", encoding="UTF-8") as file:
+ file.write(parent_libpath + os.linesep)
+
args = [
context.env_exe,
__file__,
--
2.39.2
next prev parent reply other threads:[~2023-04-24 20:09 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-04-24 20:02 [RFC PATCH v3 00/20] configure: create a python venv and ensure meson, sphinx John Snow
2023-04-24 20:02 ` [RFC PATCH v3 01/20] python: update pylint configuration John Snow
2023-04-25 16:38 ` Daniel P. Berrangé
2023-04-24 20:02 ` [RFC PATCH v3 02/20] python: add mkvenv.py John Snow
2023-04-24 20:02 ` [RFC PATCH v3 03/20] mkvenv: add console script entry point generation John Snow
2023-04-24 20:02 ` [RFC PATCH v3 04/20] mkvenv: Add better error message for missing pyexpat module John Snow
2023-04-24 20:02 ` [RFC PATCH v3 05/20] mkvenv: generate console entry shims from inside the venv John Snow
2023-04-24 20:02 ` [RFC PATCH v3 06/20] mkvenv: work around broken pip installations on Debian 10 John Snow
2023-04-24 20:02 ` John Snow [this message]
2023-04-24 20:02 ` [RFC PATCH v3 08/20] mkvenv: add ensure subcommand John Snow
2023-04-24 20:02 ` [RFC PATCH v3 09/20] tests/docker: add python3-venv dependency John Snow
2023-04-25 16:42 ` Daniel P. Berrangé
2023-04-24 20:02 ` [RFC PATCH v3 10/20] tests/vm: Configure netbsd to use Python 3.10 John Snow
2023-04-25 16:43 ` Daniel P. Berrangé
2023-04-24 20:02 ` [RFC PATCH v3 11/20] tests/vm: add py310-expat to NetBSD John Snow
2023-04-25 16:45 ` Daniel P. Berrangé
2023-04-25 16:57 ` John Snow
2023-04-24 20:02 ` [RFC PATCH v3 12/20] scripts/make-release: download meson==0.61.5 .whl John Snow
2023-04-24 20:02 ` [RFC PATCH v3 13/20] configure: create a python venv unconditionally John Snow
2023-04-24 20:02 ` [RFC PATCH v3 14/20] configure: use 'mkvenv ensure meson' to bootstrap meson John Snow
2023-04-24 20:35 ` Warner Losh
2023-04-24 20:41 ` John Snow
2023-04-24 21:20 ` Warner Losh
2023-04-24 20:02 ` [RFC PATCH v3 15/20] configure: add --enable-pypi and --disable-pypi John Snow
2023-04-24 20:02 ` [RFC PATCH v3 16/20] tests: Use configure-provided pyvenv for tests John Snow
2023-04-24 20:02 ` [RFC PATCH v3 17/20] configure: move --enable-docs and --disable-docs back to configure John Snow
2023-04-24 20:02 ` [RFC PATCH v3 18/20] mkvenv: add diagnose() method for ensure() failures John Snow
2023-04-24 20:02 ` [RFC PATCH v3 19/20] configure: use --diagnose option with meson ensure John Snow
2023-04-24 20:02 ` [RFC PATCH v3 20/20] configure: bootstrap sphinx with mkvenv John Snow
2023-04-25 17:17 ` [RFC PATCH v3 00/20] configure: create a python venv and ensure meson, sphinx Daniel P. Berrangé
2023-04-25 17:22 ` John Snow
2023-04-25 17:34 ` John Snow
2023-04-25 18:10 ` Daniel P. Berrangé
2023-04-25 18:58 ` John Snow
2023-04-26 8:21 ` Daniel P. Berrangé
2023-04-26 8:35 ` Paolo Bonzini
2023-04-25 18:03 ` Daniel P. Berrangé
2023-04-26 8:05 ` Paolo Bonzini
2023-04-26 8:49 ` Paolo Bonzini
2023-04-26 16:16 ` John Snow
2023-04-26 19:10 ` Paolo Bonzini
2023-04-26 8:53 ` Daniel P. Berrangé
2023-04-26 9:08 ` Paolo Bonzini
2023-04-26 16:32 ` John Snow
2023-04-26 19:23 ` Paolo Bonzini
2023-05-01 19:20 ` John Snow
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230424200248.1183394-8-jsnow@redhat.com \
--to=jsnow@redhat.com \
--cc=alex.bennee@linaro.org \
--cc=ani@anisinha.ca \
--cc=berrange@redhat.com \
--cc=bleal@redhat.com \
--cc=crosa@redhat.com \
--cc=imp@bsdimp.com \
--cc=kevans@freebsd.org \
--cc=mst@redhat.com \
--cc=pbonzini@redhat.com \
--cc=philmd@linaro.org \
--cc=qemu-devel@nongnu.org \
--cc=reinoud@netbsd.org \
--cc=ryoon@netbsd.org \
--cc=thuth@redhat.com \
--cc=wainersm@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.