See bug 655661 comment 5.
"A parsing error is often seen, when the server delivers content not conforming to the specified mime-type." The word 'often' implies we trust the quality of our parsers. I'd suggest to also add some helpful options for the impatient user: "Also, the mime-type as sent by the server (text/xml) might be inaccurate. The following options are for explorational use: * View the raw source code [text/plan] * View as normal web contents [text/html] * Our best guess (based on suffix, ...) is 'text/dtd' [Try this]" Where the last option is only available, if firefox was able to come up with a good guess. Employ the logic we have for file:// URLs?
> The word 'often' implies we trust the quality of our parsers. XML parsers are not rocket science. Ours also doesn't get changed much, so things that used to work don't break. We've had something like 2 bugs in the last 10 years in the well-formedness detection in our XML parser. > * View as normal web contents [text/html] Offering this option almost certainly creates a security hole, in general.
You need to log in before you can comment on or make changes to this bug.