What do you wish Fiddler could do that it can't... or that you can't figure out how to do?

In this question's comment, EricLaw (the author of Fiddler) wrote:

Fiddler has lots of interesting features, but not all of them are super wel开发者_如何转开发l-documented. A related question would be: "What do you wish Fiddler could do that it can't... or that you can't figure out how to do?" – EricLaw -MSFT- Nov 2 at 2:54

Following the lead - what do you want from Fiddler that it doesn't have now (or you don't know whether it has)?

I'd like it to be able to format and pretty-print XML and JSON request/response bodies, e.g. so a raw:

<SomeElement><Nested><MoreNested>X</SomeElement></Nested></MoreNested>

Could be displayed as:

<SomeElement>
  <Nested>
    <MoreNested>X</SomeElement>
  </Nested>
</MoreNested>

Would be really useful when looking at our API calls.

I know it can do the XML tree view, but I'm more comfortable looking at raw markup so I can see exactly what's going on. I'd just like to look at the raw markup in a nicely formatted and coloured way!

The Inspectors tab has a WebForms button, which is very nice for checking x-www-form-urlencoded POST data, but as soon as the form is multipart/form-data (e.g. forms with file uploads), this button can't display it. Or can it?

I'm not sure this is really a programming question... But, one feature I'd like is to be able to configure a standard set of screens on the right side. I always use raw mode, for example, and have to reset it every time I start the app.

Another is that I would love to have it work in a mode where it snoops on the TCP conversation, rather than acting as a proxy, along the lines of tools such as IBM Page Detailer. The problem is that browsers interact with proxies differently than they do when they're talking directly to hosts.

I hope I can select which process to watch.

not only browser or not-browser option.

I'd really like to increase the latency between the request header of an HTTP POST and the request body. I can see how you can increase the latency of the entire request as a whole, but not the individual packets.