393002 - "Server not found" message for view-source window doesn't display server name

Status: VERIFIED FIXED

(Core :: DOM: Navigation, defect)

Description

[Niek Kouwenberg]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.8.1.6) Gecko/20070725 Firefox/2.0.0.6
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.8.1.6) Gecko/20070725 Firefox/2.0.0.6

When I try to access a server that doens't exist, I get a message saying "Firefox can't find the server at www.somedomainthatdoesntexists.com.". If I try to view the source of this page, I get the same message, but without the server name.
This also occurs if I try to access the source in the mail browser window, like visiting "view-source:http://www.somedomainthatdoesntexists.com"

Reproducible: Always

Steps to Reproduce:
1. Visit a server that doesn't exists
2. Try to view the source

or

1. Visit the server with the "view-source:" prefix.
Actual Results:  
The following message is displayed: "Firefox can't find the server at ."

Expected Results:  
The following message should be displayed, as is so on a normal request: "Firefox can't find the server at www.somedomainthatdoesntexists.com."

Comment 2

[Dão Gottwald [:dao]]

This is not a duplicate of bug 273968. Bug 273968 could very well be invalid, but this one should be fixed regardless. That is, the host should not be empty.

Comment 3

[Boris Zbarsky [:bzbarsky]]

Issue here is probably that the host of a view-source URI is empty...  Maybe whatever code is involved should be looking at the innermost URI?

Comment 4

[Boris Zbarsky [:bzbarsky]]

biesi, thoughts on comment 3?

Comment 5

[Christian :Biesinger (don't email me, ping me on IRC)]

or perhaps view-source URIs should return the host of their inner URI

Comment 6

[Boris Zbarsky [:bzbarsky]]

They're an nsSimpleNestedURI.  I suppose we could subclass that and just override GetHost....

Comment 7

[Christian :Biesinger (don't email me, ping me on IRC)]

Wouldn't it make sense for SimpleNestedURI in general to forward GetHost to the inner URI?

Comment 8

[Boris Zbarsky [:bzbarsky]]

I guess we could do that.  But then jar: URIs would have the same issue, etc.

We could specify that a nested URI must forward certain nsIURI things to the underlying URI.  But which ones?

It seems that any code that cares about which uri is "really" involved (e.g. the security code) should look at the inner URI.

Comment 9

[Christian :Biesinger (don't email me, ping me on IRC)]

ok, you're probably right

the code in question is this one, I think:
http://lxr.mozilla.org/seamonkey/source/docshell/base/nsDocShell.cpp#2896

(other places in that function too)

Comment 10

[Anant Narayanan [:anant]]

Attachment: patch v1: fixes mentioned problem

Comment 11

[Anant Narayanan [:anant]]

The patch uses GetInnermostURI because that would work better in the case of nested URIs although GetInnerURI should work in most cases.

Comment 12

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 292022 [details] [diff] [review]
patch v1: fixes mentioned problem

I think you want to use NS_GetInnermostURI here instead of largely duplicating it.

Comment 13

[Anant Narayanan [:anant]]

Attachment: patch v2: address issue raised by super-review

Comment 14

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 292031 [details] [diff] [review]
patch v2: address issue raised by super-review

You don't need the QI.  Just call the function and use the URI it returns.

Comment 15

[Anant Narayanan [:anant]]

Attachment: patch v3: remove unnecessary QueryInterface

Sorry, should have taken a look at the implementation of NS_GetInnermostURI first.

Comment 16

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 292088 [details] [diff] [review]
patch v3: remove unnecessary QueryInterface

Excellent!

Comment 17

[Anant Narayanan [:anant]]

Comment on attachment 292088 [details] [diff] [review]
patch v3: remove unnecessary QueryInterface

Request approval for 1.9

Comment 18

[Damon Sicore (:damons)]

Comment on attachment 292088 [details] [diff] [review]
patch v3: remove unnecessary QueryInterface

a=endgame drivers

Comment 19

[Reed Loden [:reed]]

Checking in docshell/base/nsDocShell.cpp;
/cvsroot/mozilla/docshell/base/nsDocShell.cpp,v  <--  nsDocShell.cpp
new revision: 1.876; previous revision: 1.875
done