Bug 33689 - WebSocket: Missing Request-URI, when no tralling slash in host
Summary: WebSocket: Missing Request-URI, when no tralling slash in host
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: 528+ (Nightly build)
Hardware: Mac (Intel) OS X 10.6
: P2 Normal
Assignee: Nobody
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-01-14 14:20 PST by y8
Modified: 2010-01-20 18:41 PST (History)
2 users (show)

See Also:

Attachments
Patch (4.19 KB, patch)
2010-01-15 00:28 PST, Fumitoshi Ukai 		no flags Details | Formatted Diff | Diff
Patch (14.63 KB, patch)
2010-01-20 01:14 PST, Fumitoshi Ukai 		ap: review+
Details | Formatted Diff | Diff
Show Obsolete (1) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description y8 2010-01-14 14:20:14 PST 
If host provided to WebSocket constructor does'n have tralling slash, for example "ws://domain:port" or "ws://domain", WebSocket send GET request without Request-URI.

var ws = new WebSocket("ws://domain") 

Produces: 

GET HTTP/1.1
Upgrade: WebSocket
Connection: Upgrade
Host: localhost
Origin: http://localhost
Comment 1 Fumitoshi Ukai 2010-01-14 19:00:17 PST 
Confirmed it on Safari nightly build.
(no problem with Chromium, though)
Comment 2 Fumitoshi Ukai 2010-01-15 00:28:24 PST 
Created attachment 46648 [details]
Patch
Comment 3 Alexey Proskuryakov 2010-01-15 08:47:22 PST 
Comment on attachment 46648 [details]
Patch

Code changes look good, but I think I'll ask to rework the test. It has absolutely no reason to be script-tests-based - it doesn't have even a single shouldBe anywhere! A bigger issue is that test output doesn't really mention pass/fail criteria, and that it relies on assertions that may be gone at the same time the bug is reintroduced.

A more reliable way to test this would be for the server to report invalid requests in some manner (possibly by setting a state that can be retrieved later, or maybe by intercepting these and forwarding to a specific handler whose response we could check).
Comment 4 Fumitoshi Ukai 2010-01-20 01:14:59 PST 
Created attachment 46989 [details]
Patch
Comment 5 Alexey Proskuryakov 2010-01-20 08:36:59 PST 
Comment on attachment 46989 [details]
Patch

> +    print STDERR "args: @args\n";

Did you mean to leave this in?

r=me
Comment 6 Fumitoshi Ukai 2010-01-20 18:41:36 PST 
Committed r53592: <http://trac.webkit.org/changeset/53592>