Closed Bug 90976 Opened 24 years ago Closed 24 years ago

Completely empty HTTP response does not lead to any user-visible error message

Categories

(Core :: Networking: HTTP, defect)

defect
Not set
normal

Tracking

()

RESOLVED DUPLICATE of bug 92224

People

(Reporter: elaineking, Assigned: darin.moz)

References

()

Details

From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:0.9.2+) Gecko/20010716 Netscape6/6.1b1 BuildID: 2001071604 When trying to surf through the accuweather.com site, neither the tab, nor the icon for "travel by land" render a page. The browser searches endlessly. IE declares "page cannot be displayed," 4.7 says document contained no data. Reproducible: Always Steps to Reproduce: 1.open 6.1 2.go to accuweather.com 3.click on travel tap in top row 4. click either the "land" tap, or the "travel by land" tab 5. browser will not find page
Connecting to the page directly: ~% telnet www.accuweather.com 80 Trying 207.242.93.24... Connected to www.accuweather.com. Escape character is '^]'. GET /adcbin/land_travel?nav=travel HTTP/1.0 Host: www.accuweather.com Connection closed by foreign host. ~% Over to networking:HTTP. We should be detecting this as an empty page.
Assignee: bclary → neeti
Status: UNCONFIRMED → NEW
Component: Evangelism → Networking: HTTP
Ever confirmed: true
OS: Windows 98 → All
QA Contact: zach → benc
Hardware: PC → All
Summary: within site page "travel by land" never renders → Completely empty HTTP response does not lead to any user-visible error message
We have some code which basically says that if we get disconnected without a response, then we assume that the server timed out the keep-alive connection after we sent the request, but before the server recieved it. I guess we should not be doing that if this is the first transaction for a connection, maybe?
Probably depends on: bug 19073. It might help if we returned errors when a URL doesn't actually return a file as the payload.
Depends on: 19073
darin, thoughts?
Assignee: neeti → darin
Oh, and this is bad because we don't do exponential backoff....
*** This bug has been marked as a duplicate of 92224 ***
Status: NEW → RESOLVED
Closed: 24 years ago
Resolution: --- → DUPLICATE
You need to log in before you can comment on or make changes to this bug.