Open Bug 328980 Opened 19 years ago Updated 2 years ago

Wrong TCPIP behaviour when loading bmp (special case)

Categories

(Core :: Graphics: ImageLib, defect)

1.8 Branch
defect

Tracking

()

UNCONFIRMED

People

(Reporter: manfred.schmid, Unassigned)

Details

Attachments

(1 file)

User-Agent:       Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322)
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv:1.8.0.1) Gecko/20060111 Firefox/1.5.0.1

When requesting a bmp file with Mozilla Firefox from a webserver, then the action fails in this special case:
The webserver answers to the GET with the Bitmap header in the 1st packet and then bitmap data in the 2nd packet.
The action only works when the header is followed by at least 1 Byte of the picture payload.

Reproducible: Always

Steps to Reproduce:
1.Set up a webserver that sends bmp header and bmp payload in 2 different packets.
2.Create a simple webpage that loads this picture.
3.Fails 98% of the time




The behaviour was logged with Ethereal.
PC --> Webserver         GET
Webserver --> PC         ACK
Webserver --> PC         HTTP/1.1 OK 200 + 54 byte bmp header
Webserver --> PC         bmp payload
PC --> Webserver         RST
Connection is closed: Sometimes Firefox shows the errormessage: bmp corrupted
Attached file ethereal log
This if the Ethereal log of Mozillas behaviour. RST is send, which should not occur.
After combining header and data Mozilla worked correct.
Assignee: general → pavlov
Component: General → ImageLib
Product: Mozilla Application Suite → Core
QA Contact: general
Version: unspecified → 1.8 Branch
Difficult to reproduce without a specific webserver. Is there a more easier testcase to use?
Assignee: pavlov → nobody
QA Contact: imagelib
as far as imagelib is concerned images are just streams. So if this bug is valid, it's a necko bug. reassigning.
Component: ImageLib → Networking
QA Contact: imagelib → networking
Bobby, that's not really true. Imagelib sees the data in chunks. I find it more likely that the BMP decoder mishandles the case that the header end falls on a chunk boundary than that necko is screwing this up.
Component: Networking → ImageLib
QA Contact: networking → imagelib
Status: UNCONFIRMED → RESOLVED
Closed: 13 years ago
Resolution: --- → FIXED
Status: RESOLVED → UNCONFIRMED
Resolution: FIXED → ---
Sorry, accidentally closed this bug.
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: