Comment # 5
on bug 11514
from Pascal Quantin
(In reply to Xiaochuan Sun from comment #4)
> The initial issue of blank info column after filtering can only be found
> for the several new log formats I added. In those new log formats, I used
> global array infoString[] to display info column string instead of usual
> const string for each dissected protocol. Only Qt wireshark has blank info
> column issue, the GTK wireshark works well. The strange thing is that the
> issue is not easily reproduced on my PCs but easily reproduced on other PCs.
> I wish I could push my personal code to git repository but I have modifed
> so many codes, about 29k LOC, this is why I could not pusch my code.
> Anyway, I found a workaround for crash issue and blank info column issue.
>
> The col_add_str(pinfo->cinfo, COL_INFO, infoString) is widely used in my new
> added log format, which should be the reason for the blank info column issue
> after filtering.
>
I already saw such behavior and it was due to some wrong code in a dissector
that was calling col_XXX functions under a if (tree) test.
Qt GUI does not seem to cache column info yet (while GTK UI seems to) and when
filtering recalls the packets dissection with a NULL tree.
As indicated in doc/README.dissector chapter 2.9:
"Note, however, that you must fill in column information, create
conversations, reassemble packets, do calls to "expert" functions,
build any other persistent state needed for dissection, and call
subdissectors regardless of whether "tree" is NULL or not."
So double check your code as most probably this is a bug in your dissector that
Qt GUI is exhibiting. By fixing it, you will probably not need your hack.
You are receiving this mail because:
- You are watching all bug changes.