Marco Antoniotti <[EMAIL PROTECTED]> writes:

>> Hi,
>> 
>> I'm working on a new emacs interface to cmucl debugger that (IMHO) is
>> more user-friendly than the listener and not so annoying as the window
>> debugger.
>
> Interesting.  How will it interplay with ILISP?
>

It's being done _in_ the ILISP framework.

I started to use ILISP + CMUCL last week.  Previously, I was using
fi-eli + allegro but due to the heap limitations of the recent allegro
compiler my programs don't compile.

Unfortunately, my transition hasn't been as smooth as I was expecting.
I'm almost sure there are a few bugs in CMUCL that prevent my code to
run properly.  While trying to solve those bugs, I searched for but
couldn't find a good CMUCL debugger interface.  The line interface is
unapropriate and the graphical debugger is not much of an improvement.

As a result, I decided to build my own, following the ideas of my
beloved Explorer Lisp Machine.  The screen is devided into areas
(emacs windows, actually), each one show different aspects of the
debug session.  There's one window for the backtrace, another one to
show local variable information, another one for the listener, another
one for the code, highlighting the source code in the same way hemlock
is supposed to do (I use very similar functions) but using emacs
instead.  It even includes a kind of follow mode where movement on the
backtrace automatically updates all the other windows.  It's already
running and I like it!  But I'm missing the critical information I
asked for.

I'm planning to also develop an inspector that will be also integrated
in the debugger.

By the way, I guess there's a bug in ILISP regarding the
lisp-send-region function.  This function is using buffer-substring
instead of buffer-substring-no-properties.

Best regards,

Ant�nio Leit�o.

Reply via email to