Wireshark · Wireshark-dev: Re: [Wireshark-dev] Patch for bug 310

["Steve Schaeffer" <schasj@xxxxxxxxxxxx>]

Ulf Lamping wrote:

> Steve Schaeffer wrote:
> > Let us review...
> >
> > I submitted this bug report 18 months ago and you said you'd have a look 
> > at it.
> I *had* a look at that problem some time ago, and tried a few hours to fix 
> it - when I remember correct I dropped my test because of a completely 
> wrong approach.

Didn't know that. Perhaps an entry in the bug report was called for?

> > Almost 3 months ago I pointed you to the source of the problem, as I see 
> > it.
> I guess you're right on the cause.

Thank you.

> > I finally have some time to dig into it myself and provide a patch that 
> > *does* fix the issue to my satisfaction, and yes I did test it, again to 
> > my satisfaction, and all you can do is make snide comments?
> Well, I'm sorry if my comments annoyed you - but getting a patch that 
> doesn't work (and it didn't worked after a very simple test with a few 
> packets) is just annoying to me as well.

If the patch didn't work, I would agree with you. However it does work and 
happens to have a side effect you find objectionable. For the way I use 
Wireshark it is much preferable to have the list cleared when applying a 
filter and be able to navigate using the list as it gets created afterward. 
The clearing of the list was not an intentional act and I'm not sure how 
that crept in. Of course it would be better if it did not get cleared.

> Just clicked on a few packets, applied a display filter and all history 
> entries were gone.

I don't consider this a test of the issue in the bug report.

> > If you don't like what I've submitted then why not expand on what I've 
> > done to *your* satisfaction?
> I don't understand you here.

I mean take my fix and find out why the clearing is occuring and remedy that 
without rejecting my needed fix.

> > You might want to take a look at just how it is you encourage input to an 
> > open source project.
> Well, in the long run having a patch that doesn't work won't motivate me 
> to look at patches from other people as well.

Again, if the patch didn't work I'd be sympathetic to this point. You are a 
huge contributor to Wireshark and I've complimented you on that before. I 
don't blame you for being annoyed by crap. I don't think my patch is crap or 
I wouldn't have sent it. I think it could be considered a design change that 
the list gets cleared, but I won't argue the point because it wasn't what I 
intended.

> The question about attaching your patch to the bug report was serious, I 
> don't see a good reason to have a bug report but doing most communication 
> without it. Bugzilla is intended for such purposes, so while your patch 
> might not work for me, some one else (including me or yourself) may have a 
> look at it later to get a final fix.

According to the instructions (http://www.wireshark.org/develop.html): "You 
can send the patch to the wireshark-dev mailing list or attach it to a 
http://bugs.wireshark.org/."; Since I didn't have my password handy (required 
for adding an attachment to a bug report) and I didn't want to create a 2nd 
set of credentials I opted for the quicker route.

I attached the patch to the bug report earlier today.

Look, I know that Wireshark is a huge piece of work and few, if any, people 
understand all of the parts. This is exactly why I was reluctant to mess 
with something I didn't have time to fully understand. However, since it 
appears that I'm the only one who cares about this issue I decided to take a 
chance. Perhaps not the best thing to do, but I was tired of waiting.

Maybe the history list does need a major overhaul, but I'm not going there! 
:-)

> Regards, ULFL

Peace