https://github.com/telent/meddle

Replay Firefox's Tamper Data XML sessions in Ruby
https://github.com/telent/meddle

Last synced: 13 days ago
JSON representation

Replay Firefox's Tamper Data XML sessions in Ruby

• Host: GitHub
• URL: https://github.com/telent/meddle
• Owner: telent
• Created: 2010-08-07T16:34:25.000Z (over 14 years ago)
• Default Branch: master
• Last Pushed: 2010-08-19T18:29:19.000Z (over 14 years ago)
• Last Synced: 2024-10-12T00:29:55.398Z (about 1 month ago)
• Language: Ruby
• Homepage:
• Size: 129 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README

Awesome Lists containing this project

README

        
= Meddle

A less resource-heavy way to do realistic regression tests (and

eventually load tests) than controlling an actual web browser 

a la watir.  

* Interact with your web site using Firefox.  

* Capture the requests sent with Tamper Data, and export as XML

* Replay them from the command line

 - with realistic timing

 - with SSL support

 - with a 'rewrite' step that lets you programmatically change the

   request data before sending it (e.g. to switch hostnames from

   production to test, or vice versa)

 - using the single-threaded low-overhead goodness of EventMachine

 - using "approximately real" rules for number of parallel requests

   per hostname (default 2), persistent connections, and timeout values

== Usage

See example.rb

== BUGS/ISSUES/TODO

0) It's pre-pre-pre alpha in a big way.  I put it on github in the

expectation that it will *someday* be generally useful, not as an

advertisement that it is currently.  Because it's not.

1) See the TODO file for more rapidly changing detail

== Other Tools

1) what about watir?

Tried it, didn't like it ;-)

More seriously, it doesn't do the same job.  Meddle replays real

human-generated sessions: watir simulates them.  Meddle runs from the

command line and doesn't need a web browser to be installed or

running.  Eventually, if I can make it run fast enough (or if it's run

against sites slow enough ;-), it will be useful for load testing.

2) What about Mechanize?

Again, it doesn't do the same job.  Mechanize will do you for

functional testing ("did I get this page when I clicked that button or

followed the link?") but the headers it sends aren't the headers your

browser sent, and it doesn't download images, and so on and so forth.

3) What about WWW::TamperData (on CPAN)?

That's actually pretty similar to what this does.  But it's in Perl

and I wanted one in Ruby.

Also, as of v 0.08 it sends out all the requests at once and pays no

heed to the timings in the XML.  This is fine for a single shot but

probably not so hot for simulating actual user load