[Pdx-pm] short job

Chris Dawson cdawson at webiphany.com
Fri Apr 23 19:09:53 CDT 2004

Hi there,

I have need for someone who can write a small program for me.  I am not 
sure how to agree to payment, so I will just describe the job first.  I 
am happy to pay by the hour, or I would be happy if someone wanted to 
propose a complete solution for a fixed price.  I would like to see some 
work samples, of course.

Here is what I want:

I need a test script which:
-- receives a hostname or IP from the command line
-- read in XML files
-- each XML file specifes a request to a webserver
-- the XML contains a URL and a method, GET or POST should be fine for now
-- the XML will provide a mechanism to POST variables, and can also POST 
files, by referencing a path on the same machine.
-- verifies that output from the URL matches a regexp once the request 
has processed
-- can chain output from one request and use as input to the next 
request by a regexp (1)
-- the XML will provide a way to run pre-processing and post-processing 
scripts (probably perl scripts) before/after each URL access (2)
-- the output of each pre-process script will be passed along with the 
output from accessing the URL, to the post-processing script (3)
-- each stage will be fully logged
-- might want to pay for writing some XML files as well

Here is an example of the command line:


Here is a simple XML:

<variable name="foo">Hello</variable>
<variable type="file" name="foo_file">/home/cdawson/test.txt</variable>
<ouput_match><!-- This is an example of #1 -->
<regexp>abc(.{3})(.{17})xyz</regexp><!-- This might need to be CDATA?  -->
<variable name="foo"/>
<variable name="bar"/>
<pre_process path="./scripts/second-pre.pl" variable="foo"/>
<variable type="interpolated" name="foo"/>
<post_process path="./scripts/second.pl" variable="bar" /><!-- This is 
an example of #2 -->

To explain the XML file:  First, the test_server.pl accesses the URL  It posts to the URL the variable 
"foo" with value "Hello"  It also posts the file named 
"/home/cdawson/test.txt"  The output of accessing this URL is sent to 
the "output_match" filter.  Let's pretend that the output is 
abdefg...xyz (all 26 letters).  The first grouping match would be set to 
the variable $foo.  The second would be set to $bar.  So, $foo would 
equal "def" and bar would be "ghi...w" (18 letters).  This is completely 
arbitrary for this example, of course.  Then, after accessing this URL, 
the test script runs the "./scripts/second-pre.pl" script, with the 
environment variable "TEST_VAR_foo" set to "def".  I think this is a 
simple way to pass variables which the second-pre.pl script can 
process.  Then, the script accesses the URL 
""  Notice that it used the $foo 
variable and interpolated the value of it in the URL as a GET request.  
Then, the output of the preprocessing script, the output of the request 
itself are both passed as environment variables (using the keys 
"TEST_preprocess_output" and "TEST_request_output"), in addition to the 
variable bar passed as "TEST_VAR_bar" with the "ghi...w" value. 

There are probably some holes in what I am suggesting, so I am open to 
reworking the XML, changing requirements etc.

My design requirements:
-- the test_server.pl script should run under -w and strict
-- everything should reside inside a .pm, so test_server.pl is basically 
a wrapper.  I think this makes it easier to unit test the 
-- use variables named "test_variable" not "testVariable" or 
"strTestVariable"  (absolutely don't want hungarian notation!)
-- perfectly happy to use any perl modules from CPAN.  If you can figure 
out a way to do this with ten lines wrapping around an existing perl 
module, I'd be happy to pay you for this.
-- If someone is more comfortable doing this in python or ruby to have 
the solution done in a different language, I am happy with either of 
these two languages, if there are some lurkers on this list.

Anyone want to tackle this?


More information about the Pdx-pm-list mailing list