Comment # 3
on bug 9101
from Kaithar
(In reply to comment #2)
> The packet has an additional, (malformed?), "00a8" header line immediately
> after the X_D_INTSPEC line. This does not look like a header line to
> Wireshark, so is being included in the body text, but this is throwing off
> the length calculations.
>
> I have been able to find no reference to these header types in any spec, so
> as far as I can tell Wireshark is doing the right thing. These headers are
> making the packet malformed.
Ya, that's not a header as such... it's the chunked encoding length spec.
There's a better explanation available at [1] and [2] but basically for chunked
encode you specify the length of the chunk, in hex, on a line followed by \r\n.
The transmission is terminated by a 0 length chunk. Form is: 0\r\n\r\n
I'm reasonably confident my implementation of it is is correct, wget, curl and
chrome all process it correctly, but if I've an error there I'm happy to
correct it and retest.
[1] http://en.wikipedia.org/wiki/Chunked_encoding
[2] http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.6.1
You are receiving this mail because:
- You are watching all bug changes.
