🚀 go-pugleaf

Forwarded-For: jkh@meepmeep.pcs.com (Jordan K. Hubbard)

I recently composed a long and lengthy reply to this question, only
to nuke it by mistake (gotta stop using emacsclient; it's got some really
unhealthy quirks). Rather than type in the original 4 page reply again,
I will try to do it in a nutshell.

It was asked whether or not someone was working on xgdb, whether I was,
and whether or not work was continuing. The answer to both questions
is "Yes."

The second question that then arises is "Why in the hell is it taking so
long?"

The answer to that requires a quick synopsis of xgdb's history.

About a year ago, I did a preliminary release that used the Athena
widgets and a weird, but fairly powerful, regexp based definition file
to parse the incoming data from gdb. The basic concept worked, but it
quickly became apparent that to do a really decent widget based interface
"right" was going to be a lot more work than I had envisioned. The tech-
nology behing gdb itself has been proven and refined through the years,
the X Toolkit and Athena widgets have not. At the time, everything was
extremely rough. I managed to make the whole thing work fairly well,
nonetheless, but used some non-standard widgets and toolkit calls that
subsequently broke under R4. After R4 had been released for awhile,
I gazed upon the now non-functioning bit of software and pondered awhile.
Other projects came from The Powers That Be as well, which used up much
of my time. Fortunately, these projects all involved toolkit programming,
the efforts on which have proved to be somewhat synergistic with the
xgdb effort. I have written several powerful tools that promise to make
the new xgdb system substantially more powerful that it was before,
both in terms of functionality and future extensibility.

The next release of xgdb will be in 3 stages:

The termCanvas and xterm+ widgets, comprising the new "xterm+" terminal
emulator system (hoped to replace the current xterm), also done over the
last year. xgdb will use this widget for its main I/O window, allowing
all the current symbol-completion/history features of gdb's to be preserved.
Many other cute features, like displaying gdb's output and the user's
input in different fonts, will also be supported through the auspices of
the xterm+ widget. If the user is also using the xterm+ based terminal
emulator as their "xterm", they will get seamless operation with xgdb.
This will be released in 30 days or less.

The AWL widget programming language (allowing high level, widget set indep-
endent programming) I've written over the last year to provide more rapid
prototyping and development of X toolkit based applications. This was
alpha released about 4 months ago, but has since undergone major
improvements. All screens for xgdb and parts of the inspector are written
in AWL. The release date on this is a bit foggier, but I hope to make rapid
progress on it once the xterm+ stuff is released. Say 60-90 days, ballpark.

A beta test release of xgdb itself, without the Inspector (the facility
for examining complex data and types in a nifty saber-like fashion) initially
implemented. All the features of the original xgdb (mine, not xgdb.c) will
be supported, though a common "style guide" approach will be used this time
to make all views orthogonal. Full on-line help will be supported with
the xinfo widget (already released). Release date also foggy, but probably
co-incident with the AWL release date. The usage of AWL has made development
an order of magnitude faster, so I have high hopes. I would really welcome
user feedback as to what they'd like the Inspector to look like, since that's
still pretty much open-ended.

Any and all questions should be directed to me: jkh@meepmeep.pcs.com.

Thanks.

					Jordan
--
			PCS Computer Systeme GmbH, Munich, West Germany
	UUCP:		pyramid!pcsbst!jkh jkh@meepmeep.pcs.com
	EUNET:		unido!pcsbst!jkh
	ARPA:		jkh@violet.berkeley.edu or hubbard@decwrl.dec.com