your patch does not compile.
it broke the structure print_text_ops in print.c
On Tue, 1 Feb 2005 21:54:53 -0800, Chris Eagle <cseagle@xxxxxxxxxxxx> wrote:
> First, does anyone read these emails? I have posted before on this topic
> and never seen any replies. From browsing the archives, it is clear that I
> am not the only one that believes a problem exists in the "Follow TCP
> Stream" display
>
> Bottom line on this email is that the attached patch fixes several problems
> with the Follow TCP Stream dialog.
>
> Now, why are these changes needed:
>
> 1. There is currently no way to save the raw binary data of a stream to a
> file. This is because the current implementation converts all non-printable
> ASCII characters to '.' before writing to the GTK window or a file.
> Carriage returns are also added. This renders the saved data useless for
> binary analysis. The attached patch adds a "Raw" radio button that provides
> the same text display as the ASCII button, but allows for raw binary save
> when the "Save As" button is clicked.
>
> 2. The "Hex Dump" and "C Arrays" options are useless for all but the
> smallest of streams. When I tried to view a "Hex Dump" of a 1.4Mb stream,
> it took 18 minutes to generate, I gave up on the "C Arrays" display when it
> had failed to produce any output after 2 hours. This on a fairly beefy
> laptop. While ethereal was attempting to generate these two displays, it
> was utterly useless and appeared to the casual observer to be completely
> locked up. No hour glass icon or anything. The attached patch improves
> performance so that each of the two displays above takes about 8-10 seconds
> to generate.
>
> Point number 1 has been raised in the past. I have no idea why it has yet
> to be fixed. People have even offered patches in the past. A "Save As"
> feature that does not faithfully save the stream is essentially useless,
> raising the question "why have it at all." Likewise, the excessive times
> required to generate hex dumps and c arrays make those "features" useless as
> well. If no one has any intention of incorporaing fixes for the problem, or
> at least acknowledging that a problem exists, then the functionality should
> be removed altogether.
>
> Regards,
>
> Chris
>
>
> _______________________________________________
> Ethereal-dev mailing list
> Ethereal-dev@xxxxxxxxxxxx
> http://www.ethereal.com/mailman/listinfo/ethereal-dev
>
>
>
>
