c++/2169: cant step into more than 1 function call on a single line

Daniel Jacobowitz drow@false.org
Fri May 18 20:28:00 GMT 2007


The following reply was made to PR c++/2169; it has been noted by GNATS.

From: Daniel Jacobowitz <drow@false.org>
To: intangir@gmail.com, Thomas McGuire <thomas.mcguire@gmx.net>
Cc: gdb-gnats@sources.redhat.com
Subject: Re: c++/2169: cant step into more than 1 function call on a single
	line
Date: Fri, 18 May 2007 16:23:39 -0400

 Sorry I didn't see this report earlier.
 
 On Fri, Sep 08, 2006 at 09:23:28PM -0000, intangir@gmail.com wrote:
 > >Description:
 > when debugging C++, if there is more than one call (or overloaded operators) on a single line, when you step in, it steps in the first call, if its not the one you wanted to step into you have to step into every single line in that function till you get back out to the original line you want to step into.. 
 > 
 > if you step over then when it returns to the original scope it steps over the rest of the line that you needed to step into.. and you miss your chance to step into it ;(
 
 Use "finish":
 
 (gdb) start
 Breakpoint 1 at 0x40056b: file z.cc, line 13.
 Starting program: /space/codesourcery/mips64-build/z
 main () at z.cc:13
 13        functioniwanttostepinto(givea(), giveb()); // *
 (gdb) s
 giveb () at z.cc:2
 2       int giveb() { return 2; }
 (gdb) fin
 Run till exit from #0  giveb () at z.cc:2
 0x0000000000400570 in main () at z.cc:13
 13        functioniwanttostepinto(givea(), giveb()); // *
 Value returned is $1 = 2
 (gdb) s
 givea () at z.cc:1
 1       int givea() { return 1; }
 (gdb) fin
 Run till exit from #0  givea () at z.cc:1
 0x0000000000400577 in main () at z.cc:13
 13        functioniwanttostepinto(givea(), giveb()); // *
 Value returned is $2 = 1
 (gdb) s
 functioniwanttostepinto (a=1, b=2) at z.cc:6
 6         int c = a + b; // try and debug to THIS line. it is a pain
 
 The behavior that "next" steps over other calls on the same line is
 unfortunate in your case, but deliberate - it makes some other cases
 clearer.
 
 -- 
 Daniel Jacobowitz
 CodeSourcery



More information about the Gdb-prs mailing list