From: Carlos <angus@...>
Date: 2004-12-27T13:33:42+09:00
Subject: Re: HTML and CSS validation

Bil Kleb wrote:

> http://www.htmlhelp.com/tools/validator/upload.html
>  http://validator.w3.org/file-upload.html
>
> because I'm reluctant to install W3's source due
> to it's large dependency chain and I'm too stupid
> to figure out how to automate the interaction
> with these file upload interfaces.
>
> From looking at their form source and reading a bit
> about file upload forms, I gather that I have to send
> a form with enctype="multipart/form-data", but it
> is unclear just how to do this.


Your request should be something like this (not tested):

# let's say that what you write to stdout goes to the server
# ('validator-page' is the path that appears in the 'action'
# attribute of the <form>
separator = "----------------------86428764287643287642"
print "POST /validator-page HTTP/1.1\r\n"
# you first send a header for your request
print "Host: www.example.org\r\n" # the server, probably www.htmlhelp.com
print "User-Agent: RubyValidator/1.0 :)\r\n"
print "Content-Type: multipart/form-data; boundary=", separator, "\r\n"
print "\r\n"
# and then goes every form field, prefixed by separator and a header
print separator, "\r\n"
print "Content-Disposition: form-data; name=\"fieldname\"\r\n" # the 
name in the form field
print "Content-Type: text/plain; charset=UTF-8\r\n" # for almost every 
field type
print "Content-Transfer-Encoding: 8 bit\r\n"
print "\r\n"
print "These are the contents of the text field.\r\n"
# for a file upload, the headers change slightly
print separator, "\r\n"
print "Content-Disposition: form-data; name=\"fieldname\"" # the name in 
the form field
print                   "; filename=\"file-name.html\"\r\n"  # ..and in 
this case the filename
print "Content-Type: application/octet-stream\r\n" # the content type 
also changes...
# ...but I guess you can put "text/html; charset=..." as well
print "Content-Transfer-Encoding: 8 bit\r\n" # or whatever
print "\r\n"
print your_file_in_a_string
print "\r\n"
# you finalize your request with the separator plus two dashes
print separator, "--\r\n"

A tip: create a simple webrick servlet that only outputs its request. 
Then copy any form you want to automate to your local machine, and 
change its "action" attribute to point to your servlet. Then you'll know 
exactly what you need to send.

Good luck.