Jakub Zawadzki wrote:
On Tue, Mar 29, 2011 at 11:07:14AM -0400, Jeff Morriss wrote:
Of course, the fact that they don't work does imply we might want to
comment them out for the time being--unless someone wants to get them
working.
Ok, I've fixed composite tests #0 and #2 :) Patch attached.
Other tests are imho broken, e.g. composite #1
composite is created by two tvb_smalls which are created by:
tvb_small[i] = tvb_new_real_data(small[i], 16, 17);
where small[i] is some random data -- lets assume 01234567890ABCDEF
so total reported_len is 34, and length is 32.
Data:
0 1 2 3 4 5 6 7 8 9 A B C D E F ? (packet truncated by 1 byte)
0 1 2 3 4 5 6 7 8 9 A B C D E F ? (packet truncated by 1 byte)
but expected data is set to: 01234567890ABCDEF01234567890ABCDEF,
and it doesn't care about truncated data.
For the record, your patch was checked in with rev 36440.
You raise an interesting point: from the point of view of a TVB that
/might/ be the appropriate behavior (though even then it's
questionable), but it's certainly not correct if it were being used for
reassembly.