Wireshark-dev: Re: [Wireshark-dev] need some more information on bug 8454
From: "Subh. Singh" <[email protected]>
Date: Thu, 2 May 2013 00:14:26 +0530
Hi Jeff,

I am debugging Wireshark using glibtool  ( using OS X 10.8 ) .

I tried to print the log using following ways :
+ printf(" my method is called ");
+ g_print("my method is called ");
+ g_warning("my method is called ");

but none of them worked for me.

Sorry for previous blank link. 
Please see following log : http://pastebin.com/Gasw5WS7

--Thanks and Regards
Subhash Kumar Singh

On Wed, May 1, 2013 at 8:17 PM, Jeff Morriss <[email protected]> wrote:
On 04/21/13 04:34, Subh. Singh wrote:
Hi Jeff,

I started looking on bug 8454

On running wireshark on glibtool I found that some object files are missing.
Please see thislog <http://hastebin.com/roniheduju.scala>.

For some reason that link doesn't work for me: I just get a blank page with what appears to be a prompt.

Anyway, one thing to keep in mind when debugging a libtool-generated program that hasn't been installed yet: you have to get the libtool stuff out of the way.  To do that with 'gdb' I usually do:

% libtool --mode=execute gdb ./wireshark

  when I tried to add the break point for method get_int_value then I
got following message

(gdb) b get_int_value
Function "get_int_value" not defined.
Make breakpoint pending on future shared library load? (y or [n]) y
Breakpoint 1 (get_int_value) pending.

So please tell me how to debug the get_int_value.

Does the above not work (does it not break when it calls that function)?  That may be because you didn't run gdb from libtool (as per above). Normally pending breakpoints like that work for me.

If that doesn't work, well, you can always resort to putting printf()'s (or g_warning()'s) in that function.

I looked on the code of get_int_value.
Can you please explain me how to approach to fix the bug 8454 same time
please give me some more information on bug 8454.

There's a link to the bug where the problem was originally discovered in the bug; reading that may lend some information.

I think the main point is that the current corner-case behavior is inconsistent and should be made consistent.  I don't have any opinion on what they should do; maybe the bug report has some opinion (so asking the question on the bug might help).