4

When I access certain HTTP URLs, such as http://www.ssa.gov/framework/images/icons/png/ in Firefox (running on Linux), I surprisingly get a "Problem loading page" browser error, containing the "File not found" message that is used for local files:

Firefox

The text doesn't seem to make sense, since there is no file involved.

At first I thought it might just be a Firefox bug, but the same thing happens in Chromium:

Chromium

When I tell Wget to download it, it reports 500 Server Error. But it can't be that browsers display "File not found" on receiving a 500 error, because it doesn't happen when I access https://httpbin.org/status/500.

What causes this error message, and what does it mean?

Mechanical snail
  • 7,963
  • 5
  • 48
  • 67

4 Answers4

16

I think I've found the reason for this.

When a web server encounters an error, it normally serves a document (usually an HTML document describing the error) to the browser, indicating the error condition using the HTTP status code.

According to this bug report, Firefox originally always displayed the returned document; normally this is what you want. However, a user found an issue with a misconfigured AOL server: when requesting a nonexistent EXE file, the server would serve the 404 page, but with an incorrect Content-Type. That caused Firefox to offer to download the HTML document with a .exe extension, which was confusing since there was no indication that any error occurred. They changed the behavior with a simple hack (not warranting the effort of writing a new error message page, since it's an uncommon case, instead reusing the "not found" page, which makes sense in the specific example given by the bug's reporter).

From the bug report that @m4573r found, it sounds like the current behavior when Firefox receives a response with an HTTP code signaling an error, and the response's Content-Type is something other than HTML, then Firefox displays a "File not found" error page.

The vast majority of web servers are configured to serve an HTML document on error, which is why you don't normally see this. But in this corner case, the error message doesn't make sense.

wget -d http://www.ssa.gov/framework/images/icons/png/ confirms what's going on here:

---response begin---
HTTP/1.1 500 Server Error
Server: Generic Web Server 1.0
Date: Thu, 20 Feb 2014 23:29:57 GMT
Cache-control: public
Content-type: magnus-internal/directory
Transfer-encoding: chunked

---response end---

It's serving the error page with the bogus Content-type of magnus-internal/directory, triggering the behavior in Firefox.

Evidently Google thought that this behavior made sense and implemented it similarly in Chromium.

Community
  • 1
Mechanical snail
  • 7,963
  • 5
  • 48
  • 67
1

It's been reported as a bug in... 2008 :/ Seems that it's on a low priority list.

m4573r
  • 5,701
1

Since you said in ur question that there is no file involved so I think that telling you a little about how web works may clear your doubt.

When you send a request to a server, you are indeed requesting a file stored on the server or a file generated by the server if the request is asking for dynamic content.

Url is divided into 3 parts:

1) Http:// << the protocol used

2) www.whatever.com << the dns used to identify the server(comp from where u are asking for content)

3) /music/song.mp3 << the location of file on the server

So this error message just states that the file that you have requested is not found. I have taken example of a music file but even a webpage with .html extension can be requested and same error might be generated.

This error just says that the resource ie webpage, music file, video or whichever type of file you have requested could not be found on the server.

Ankit Goel
  • 111
-1

I'd just like to point out that what you were looking for was only PART of the path for the URL. It's missing both a file name and file type. Essentially, it was telling you that it could not find the non-existent 'non-file' you were looking for. The reason this 'bug' hasn't been fixed is because this isn't a bug.

After all, if you go to a complete stranger's house and ask where their ______ is, they won't be able to help you find it until there's something filling that blank part of the question, and really, that's what you were asking the browser to do.

BTW, if you take out the period at the end of that partial URL, it should give you a list of files that are in the site's png folder. Some people consider doing that to be hacking, though. Personally, I don't and instead leave occasional 'easter eggs', like extra downloadable content, for users that understand how websites are structured better than others. I've always liked rewarding people for having that extra bit of knowledge that others don't.

saywha
  • 1