X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f
X-Recipient: djgpp AT delorie DOT com
Date: Tue, 06 Nov 2007 01:27:24 -0500
From: Ethan Rosenberg <ethros AT earthlink DOT net>
Subject: Re: C programming errors,  and some other things  Resend
In-reply-to: 
 <OFDF7D3DAD DOT FD0D9F79-ON87257363 DOT 005733DB-87257363 DOT 0057C9A2 AT seagate DOT com>
To: Gordon DOT Schumacher AT seagate DOT com, DJGPP List <djgpp AT delorie DOT com>
Message-id: <0JR20077ZKI0JVC0@mta3.srv.hcvlny.cv.net>
MIME-version: 1.0
X-Mailer: QUALCOMM Windows Eudora Version 7.1.0.9
Content-type: text/plain; charset=us-ascii; format=flowed
Content-transfer-encoding: 7BIT
References: <200709270405 DOT l8R452ET008178 AT delorie DOT com>
 <OFDF7D3DAD DOT FD0D9F79-ON87257363 DOT 005733DB-87257363 DOT 0057C9A2 AT seagate DOT com>
Reply-To: djgpp AT delorie DOT com

At 10:58 AM 9/27/2007, you wrote:
>Ethan Rosenberg <ethros AT earthlink DOT net> wrote on Wed, 26 Sep 2007 00:45:00
>-0400:
>
># 3] However... previously when I compiled with RHIDE, the screen would
># show the compile commands; eg, gcc -g -o bla.c  bla.o [I am probably
># not accurate with what I just wrote, but you have the idea.]  Now all
># it says it "compiling".
>#
># 3] If I set a breakpoint, and wish to run till breakpoint [F4], RHIDE
># crashes with a diagonal display on the screen showing some of the C
># code, and some messages "code not created for ....".  I think it is
># being compiled w/o debugging code and is being optimized.  Is there
># any compiler flag that will force the system not to optimize?
>#
># 4]  If I step thru the program,  [F7 or Shift F7], the user screen is
># never displayed.  When the program needs input, there is no way to do it.
>#
># FYI -  I was using ver 5 of RHIDE [rhid15ab.zip] and was experiencing
># the above problems.  I switched to an earlier version, [rh1501b.zip]
># and the problem persists.
>
>My first suggestion - to you and others that have been requesting
>help with vaguely similar issues - would be, give up on RHIDE.  It's
>not been supported (for DJGPP, at any rate) for years now - which
>means that there are a host of problems with using it with the modern
>toolchains (and often with modern machines as well).  For instance,
>under GCC 3.x or greater, you cannot put breakpoints in C++
>constructors or destructors; they will simply be ignored.  You often
>cannot inspect member variables.  Breakpoints often crash or give
>unpredictable results... etc., etc.  (And yes, that's even with
>optimization disabled.)
>
>My suggestion would be to use "gdb -tui" instead; the latest GDB
>download supports the TUI mode.  This still requires that you learn
>the GDB command-line, but it's not as bad as one might think.  And
>as you get more used to it, you'll discover that you can do things
>with it that you could never do with RHIDE.
>http://sourceware.org/gdb/documentation
>
>It'd be nice to get RHIDE/RHGDB fixed, or get another frontend...
>but until then I don't think it's worthwhile to try and fight with
>it.
++++
Gordon -

Much thanks.

I am investigating using gdb-tui.  Until I become proficient in gdb, 
I am using RHIDE with the following work-around:

When I debug, I first force the compilation with debug by using 
Shift-F8.  That will bring me to the first executable statement.  If 
I then wish to go to a breakpoint, or step, RHIDE is stable.

I can also see the user and compile screen.

The previous version of RHIDE seems to be more stable.

Much thanks to all.

Ethan