On Tue, Dec 8, 2015 at 8:07 AM, <villasc...@gmail.com> wrote: > I am struggling on some general concepts. My past experience with > server-side code is mostly limited to PHP and websites. I have some file > called "whatever.php", the browser accesses it, and PHP parses it and returns > HTML, JSON, etc. Every now and then, I need to run some background PHP > script, or have some PHP script initiated by a CRON job, or have some PHP > script initiated by the command line running in an endless loop, and while it > works, feel other languages might be more appropriate. > > So, my interest in Python... > > I've read up on Python, and while some aspects seem similar to PHP, some > don't. I have learned how to create Python script, have run it from the > command line, and have even accessed it with Apache by placing > http://example.com/myscript.py in the browser. I am used to seeing .php > extensions, but never .py extentions, and even visited multiple sites which I > knew were written in Python, but never saw the browser expose the .py > extensions. I am obviously missing something. > > Why don't I see the .py extension in my browser? > > Is Python event driven like PHP, or is it somehow different? > > How should I view Python differently than PHP?
Hi there! Welcome to the list, and thank you for trying to understand the differences, instead of simply expecting Python to behave like the thing you already know. To demonstrate what I'm talking about, I'm going to use a tiny Python-based web site that I knocked together a little while ago. The source code is on GitHub, and the site itself is public, so you can see both: https://github.com/Rosuav/MinstrelHall http://minstrelhall.com/ The most important file is mh.py - click on that to see the code. In PHP-based web sites, the structure of your source code exactly matches the structure of your web site. The file system *is* your web site; any time the web server (say, Apache) comes across something that has the .php extension, it goes and runs it as code. In Python-based web sites like this one, the layout of the source tree follows a structure that keeps logically-similar things together, and the structure of the web site is defined by code. As you read through mh.py, you'll see a number of lines saying @app.route("...") - those are creating URLs (sloppy terminology, but close enough). There's one that says @app.route("/"), which handles http://minstrelhall.com/, and another just underneath it that says @app.route("/campaign/<int:id>"), which handles http://minstrelhall.com/campaign/6 (and any other integer ID that might be passed in). One advantage of this kind of setup is that your URLs don't depend on your back end. I could replace all this with a Ruby on Rails site, and nobody would notice the difference. I could put together something using Node.js to replace the Ruby site, and continue to provide content at http://minstrelhall.com/campaign/6 - again, nobody would care. Another *massive* advantage is a complete separation of code and data. You can't see it in Minstrel Hall, but here's a web site that has a static-files collection: https://github.com/Rosuav/Flask1 http://rosuav.com/1/ (Note that this Flask site isn't running the whole domain, but only one subdirectory. What Python sees as @app.route("/foo") is actually http://rosuav.com/1/foo.) Any files in the static/ directory are made available by the framework - but they are, by definition, static files. Any file that I put there will be delivered as-is. Suppose I create a world-writable directory called "uploads", and then have a script that accepts a form submission with an attachment, saving the attachment into the uploads directory. (This, by the way, is not an engineered example; it's a real-world example from another web site that I host.) If the uploads directory is inside static, anything saved to there will be viewable as http://rosuav.com/1/uploads/some_file_name, but - and this is the important part - no matter what its file extension, *it will not be run*. You can upload a .py file, and it'll be visible as Python source code. You can upload a .cgi script, and it'll be visible as plain text. You can upload a .php file, and nothing will run it. This is quite tricky to do with PHP, and requires some careful management of Apache config files (or presumably the equivalent in nginx or other servers). With Python+Flask (or any other code-based server setup), it's automatic. This safety makes a *huge* difference to the compromisability of your web site; you can accept uploads without worrying about remote users running code and creating reverse shells and stuff. So that's a quick potted summary of why the URLs don't reflect the language used. Python is event-driven, but instead of defining events at the file level, the way PHP does, they're defined at the function level. Of course, if you *want* to put ".py" on the end of all your URLs, Python won't stop you... :) ChrisA -- https://mail.python.org/mailman/listinfo/python-list
