Python、Django、Django-cms升级顺序及指定版本升级工具咨询

Hey there! Let's walk through your upgrade plan step by step—this is a common jump from Python 2.7-era stacks, and following the right order will save you a lot of frustration. Here's what you need to know:

The golden rule here is transition incrementally, and move from "low-level" to "high-level" components once each layer is stable. For your specific stack, we can't jump straight from Python 2.7 to 3.7 with your current Django/Django CMS versions—we need intermediate steps to avoid breaking everything at once.

Your target is Python 3.7 → Django 2.2 → Django CMS 3.6. Here's the safest path:

1. First, Prepare Your Safety Net #

• Backup everything: Tag your code repo, export a full database dump, and make sure you can roll back to your current working state at any time.
• Set up an isolated test environment: Use virtualenv or venv to create a copy of your production stack—never test upgrades directly on live code.

2. Transition to Python 3-compatible base versions (still on Python 2.7) #

Directly jumping from Python 2.7 to 3.7 is risky. First, get your project running on the latest Python 2.7-compatible versions of Django and Django CMS that also support Python 3:

• Upgrade Django to 1.11.x (LTS): This is the last Django version that supports Python 2.7, and it also works with Python 3.5-3.7. Install it with pip install django==1.11.29 (grab the latest 1.11 stable release). Run python manage.py check and python manage.py migrate, fixing any deprecation warnings or errors that pop up.
• Upgrade Django CMS to 3.4.x latest: Django CMS 3.4 supports Django 1.8-1.11. Install the newest 3.4 release (e.g., pip install django-cms==3.4.6), run checks and migrations, and confirm your project still runs smoothly on Python 2.7.

3. Upgrade Python to 3.7 #

Now that your stack is Python 3-ready:

• Create a new Python 3.7 virtual environment: python3.7 -m venv venv37, activate it, and reinstall your dependencies. Note: Some Python 2-only packages (like MySQL-python) need replacement—swap them with Python 3 alternatives (e.g., mysqlclient).
• Use 2to3 to auto-fix Python 2 → 3 syntax: Run 2to3 -w your_project_root to handle common issues like print statements, unicode/str differences, and xrange → range changes.
• Manual cleanup: Fix remaining issues like iteritems() → items(), except Exception, e → except Exception as e, and encoding-related bugs. Run your project and test suite to catch any edge cases.

4. Upgrade Django to 2.2 (LTS) #

Django 2.2 is fully compatible with Python 3.7 and is a long-term support release:

• Review Django's upgrade docs for 2.0-2.2 to spot deprecated APIs: Key changes include replacing url() with path()/re_path(), mandatory on_delete parameters for foreign keys, and moved modules like django.core.urlresolvers → django.urls.
• Install Django 2.2: pip install django==2.2.28 (latest stable 2.2 release).
• Run python manage.py check --deploy and python manage.py migrate, fixing all deprecations. Tools like django-upgrade can automate many of these fixes for you.

5. Upgrade Django CMS to 3.6 #

Django CMS 3.6 supports Django 2.0-2.2 and Python 3.5-3.7, matching your target stack:

• Install Django CMS 3.6: pip install django-cms==3.6.0 (or the latest 3.6.x release).
• Run migrations: python manage.py migrate to handle any database schema changes from Django CMS.
• Test custom plugins and templates: Some older plugin patterns or template tags might break in 3.6. Test every page and component to fix frontend/backend compatibility issues.

Static analysis and automation tools will save you hours of manual work:

• 2to3: Python's official tool for converting Python 2 code to Python 3 syntax—handles most routine changes automatically.
• django-upgrade: A Django-specific tool that fixes deprecated API calls (like replacing old url() patterns, adding on_delete parameters) across your codebase.
• pyupgrade: Upgrades Python syntax to modern standards (e.g., f-strings instead of str.format(), pathlib over os.path).
• pylint-django: A Pylint plugin that checks for Django-specific code issues, including version compatibility warnings, to catch potential problems early.

内容的提问来源于stack exchange，提问作者Shankar Thiyagaraajan