From xemacs-m  Thu Mar 27 23:48:50 1997
Received: from altair.xemacs.org (steve@xemacs.miranova.com [206.190.83.19])
	by xemacs.org (8.8.5/8.8.5) with ESMTP id XAA01928
	for <xemacs-beta@xemacs.org>; Thu, 27 Mar 1997 23:48:49 -0600 (CST)
Received: (from steve@localhost)
	by altair.xemacs.org (8.8.5/8.8.5) id WAA00431;
	Thu, 27 Mar 1997 22:01:31 -0800
Mail-Copies-To: never
To: xemacs-beta@xemacs.org
Subject: XEmacs 20.1 and XEmacs 20.2, recent gripes
X-Url: http://www.miranova.com/%7Esteve/
X-Face: #!T9!#9s-3o8)*uHlX{Ug[xW7E7Wr!*L46-OxqMu\xz23v|R9q}lH?cRS{rCNe^'[`^sr5"
 f8*@r4ipO6Jl!:Ccq<xoV[Qz2u8<8-+Vwf2gzJ44lf_/y9OaQ`@#Q65{U4/TC)i2`~/M&QI$X>p:9I
 OSS'2{-)-4wBnVeg0S\O4Al@)uC[pD|+
X-Attribution: sb
From: Steven L Baur <steve@miranova.com>
Mime-Version: 1.0 (generated by tm-edit 7.106)
Content-Type: text/plain; charset=US-ASCII
Date: 27 Mar 1997 22:01:25 -0800
Message-ID: <m22090a72i.fsf@altair.xemacs.org>
Lines: 91
X-Mailer: Gnus v5.4.37/XEmacs 20.1(beta11)

[Followups to xemacs-beta-discuss, please]

Some recent gripes have been aired, and what's very sad is that most
of them have either been addressed, or work is underway to address
them.

1.  XEmacs 19.15 needs more time for bugfixes.
There has been an effort little short of miraculous the last couple
months in stomping out bugs.  In some cases, bugs dating back to 19.12 
and earlier have been fixed.  Special thanks to David Moore, Darrell
Kindred, and Kyle Jones as the major contributors in the grimiest
code.

We can put off the release indefinitely for some reason or another
because in a program with a million source lines of code[1] there's
always going to be a problem somewhere.  XEmacs 19.15 is almost
obsolete and pretty decent quality as it stands now.  It's time.

I have every expectation that XEmacs 20.1 will have better stability
than an XEmacs 19.12 or 19.13.  *Everyone* who's submitted a fix
should pat themselves on the back, but that's my opinion.

2.  Posting major package bug reports to xemacs-beta.
I think this is a mistake for one of several reasons.  The only
reason that mailing lists like this exist are for the convenience
of the developers & maintainers.[2]  When it ceases to be convenient
for them and they stop following it, the value begins to decline
dramatically. I would point as an example of a once-useful forum
the Linux Kernel mailing list that Linus Torvalds now avoids like
the plague (and it shows :-().

Bug reports need to go somewhere more permanent than a mailing list
where messages get scattered to the wind.  Bug reports are much more
important than that.  There needs to be a central database, and
coordination of submission of bug reports to the actual maintainers
of the code in question.  Users and testers also need the ability to
look at the database and see what's been reported already, and who has 
been working on the problem or what the solution is (like the HPUX
strcat fiasco, like the Linux h_errno fiasco, etc.).

Fortunately a system that appears to be able to support all this is in
place and awaiting testing and debugging.  This is Cygnus GNATS.  I
have a brand new fresh, never-been-used bug database, and some front
end emacs lisp code for reporting bugs sitting in the latest XEmacs
20.1 betas (all the GNATS manuals are in the standard info
directories).  There is also a very neat Web interface to allow
querying & updating the database.  I believe this is the same system
used by the Apache httpd server people.

Once this is trouble shot and declared working, we will 
release XEmacs 20.1.  Bug reports will go to the developer who needs
to see it, be tracked by a central database for all to watch the
progress of, and everyone will be happy, right?

3.  We need to keep the distributions split.  Now between 19.15 and
    20.x, later between 20.x.experimental and 20.x.stable.
No.  I'm not sure my health will permit it[3], unless we have a
younger volunteer to take my place.

I guarantee this will be revisited in 20.2 where the major development 
effort will be to split the XEmacs monolithic distribution into
core+packages.  I still don't make any promises, but trimming things
down by 50% to 75% will make a lot of difference in my evaluation.

As it stands now a 1 million source line of code project split into
two development paths is too big for one person to maintain for long.

4.  The volume on xemacs-beta is excessive.
Please bear with me until 19.15 is dumped.  The 20.1 source tree is
almost ready for automated submission of build reports (once automated
they can be mechanically filtered if necessary), and is almost ready to
remove much of the repeated bug report traffic.  What will remain on
xemacs-beta will be discussion of bugs, the future, & other boring
technical stuff[4].

[1]  O.K.  I exaggerate, the source line count in 20.1-b10 is only
995,697.

[2]  In a good situation one can submit a bug report and get a
response within hours.  We're in danger of losing completely that
capability.  What is convenient for the developer's and maintainers is 
highly advantageous for the tester & user.

[3]  I'd kind of like to have a life again, and maybe some time to
hack together footnote-minor-mode so I don't have to do these by
hand. :-)

[4]  All of my most favorite material, actually.
-- 
steve@miranova.com baur
Unsolicited commercial e-mail will be billed at $250/message.

