This is a Standard Fetcher for use with a full TCP/IP stack (e.g. FreeNet).
<P>
It should support the following...
<ul>
<li>HTTP fetches directly.
<LI>FTP, GOPHER and WAIS fetches via a <a href="#proxy">proxy</a>.
<LI>POSTing of Forms.
<LI>Authentication. (usernames and passwords for access to protected areas)
<LI>The Extended URL protocol.
<LI>Decoding of HTML and TEXT as they are received. (This requires messages that at the moment are not part of the <a href="protocol">Standard Fetchers protocol</a>).
<LI>Should work with any other WebBrowser, although I have only tested it with Webster.
<LI>Requests jpeg images rather than gif images if the server has them. These are usually smaller than the equivalent gif.
</ul>
The current version does however have some limitations ...
<ul>
<LI>It is not guaranteed with anyother TCP/IP stack except FreeNet as this is the only one that has been tested.
<LI>It's fairly simple at the moment, only supporting HTTP fetches, POSTing, Authentication, Fetching everything else via a <A HREF="#proxy">proxy</A>.
<li>The current version loads any form data for the POST request into memory. So if you want to send large amounts of data on a form I suggest you ensure there is lots of memory available.
<LI>POSTing of forms has not been tested. It should work though as it's a simple extension of the code that sends the mime headers.
</ul>
<h2>Setting up</h2>
At this version all setting up must be done by editing a text file called <a href="../fetchers/!freeweb/config">config</a> and reloading the fetcher (No problem as Webster can be told to load any fetcher as needed)
<P>
The values in the file are as follows.
<dl>
<dt>SUPPORTED
<dd>A | (vertical line) separated list of the methods supported (in one way or another) by this fetcher.
<dt>QUITWITHBROWSER
<dd>A flag with a value of either 0 or 1.
<dl>
<dt>0<dd>means the fetcher will not quit when the Browser does.
<dt>1<dd>means it will.
</dl>
<DT>ICONBAR
<dd>A flag with a value of either 0 or 1.
<dl>
<dt>0
<dd>Don't Create an iconbar icon, You should also set QUITWITHBROWSER to 1.
<dt>1
<dd>Create an iconbar.
</dl>
<dt>CONNECTIONS
<dd>Maximum number of connections supported at one time. Other fetch requests will be queued and acted upon when possible.
<dt>HTTP, GOPHER, FTP and WAIS
<dd>More flags, this time with three possible values, 0, 1, and 2.
<dl>
<dt>0<dd>means the fetcher will ignore this method.
<dt>1<dd>that this fetcher can fetch it directly.
<dt>2<dd>that this fetcher knows a proxy that can fetch it. (More on proxies <A HREF="#proxy">below</A>)
</dl>
</dl>
<H2>Using the Fetcher</h2>
To actually use this Fetcher with Webster, you need to do the following.
<ol>
<LI>Open Webster's Choices window from Webster's iconbar menu.
<LI>Swich to the Fetchers subsection of the Choices windows
<LI>Select the <em>HTTP</em> method from the menu.
<LI>Select the <em>Call Standard Fetcher (Extended URL)</em> option for this method.
<LI>Click on the <kbd>Save</kbd> button.
</ol>
You can also get this fetcher loaded automatically as follows...
<ol>
<LI>Open Webster's Choices window from Webster's iconbar menu.
<LI>Swich to the Fetchers subsection of the Choices windows
<LI>Select the <em>HTTP</em> method from the menu.
<LI>Select the <em>On Fails, Call Command Delow & ReTry</em> button.
<LI>In the Text field below put the following text<BR><samp><Webster$Fetch>.!freeWeb</samp>
<LI>Click on the <kbd>Save</kbd> button.
</ol>
<H2><a name="proxy">Proxies</a></H2>
A proxy is usually an http server (or daemon) that accepts complete URLs rather than simple pathnames. The daemon then fetches what was required using these complete URLs and returns it the the requester. When you are using a proxy your machine never needs to connect to any other machine, it's the proxy that does the connecting (at least for the protocols the proxy copes with).
<P>
Using a proxy may be faster that going direct as many Proxies also Cache pages and files (This is often their main purpose). Proxies may also allow FTP, WAIS and Gopher connections without requireing a specific fetcher. This is because the data returned from proxies for directories and gopher menus for example, is already in HTML format.
<P>
If your Internet provider has a proxy for any protocol (even http), then you should use it, as it should make your fetches faster and it also simplifies the job your machine has to do.
<HR>
PS Clicking the iconbar icon with Adjust while holding Shift lists the connections that are open.
<hr>
<address>Andrew Pullan</a>, Sun 11th August 1996</address>
