>>>> "KK" == Karl Kleinpaste
<karl(a)justresearch.com> writes:
KK> "John A. Turner" <turner(a)blueskystudios.com> writes:
> I hope you don't mean that as harshly as it sounds. The
whole point
> of the beta process is to beat on new features and (hopefully) help
> fix them.
> If you're only interested in code that's "right", wait for a
release.
KK> Well, I meant it strongly enough to indicate that I'm annoyed that an
KK> odd definition of "progress" is present, and I think such annoyance is
KK> justified.
I agree.
I wear Karl's hat on the egcs project, and I'm not happy at the large
numbers of bugs I've found pre-testing egcs snapshots, so I know where
his head is at. I wish they would use successful compilation of
XEmacs with C and C++, with optimization and debug, as a criterion in
their test suite.
The control of quality on free software projects is a hard problem,
given that hacking on new features is generally cooler. We'll see if
the commercialization of free software will make a difference. And
personally, I'm trying to make a difference.
KK> I certainly expect new features to have bugs to be shaken out. But
KK> tabs began appearing sometime in the early 21.2.teens, here we are at
KK> .29, and today tabs can't even correlate with buffers present.
KK> No, I don't expect code that's merely "right". This is a beta
list,
KK> after all, and I know that. I'm on it because I like/want/need some
KK> of the features present in 21.2 and I'm pretty good at stressing
KK> certain odd corners of XEmacs to uncover bugs; when things get bad
KK> enough, as they unavoidably do, I just sit tight on N-1 and wait for
KK> the N+1 version in hope of something less unstable. On the other
KK> hand, a concept of "improvement" needs to be considered, not merely
KK> "updated" and "different": I ask, in an honest and earnest
tone,
KK> whether the .29 state of tabs was tested at all, by anybody, for it to
KK> be in this condition.
XEmacs 21.2 is pretty stable for me if you disable gutter tabs (as I do):
;;; No 21.2 tabs, please
(defmacro safe-set-specifier (specifier value)
`(when (and (boundp ',specifier)
(specifierp ,specifier))
(set-specifier ,specifier ,value)))
(safe-set-specifier default-gutter-visible-p nil)
Karl, I suggest you use this as well.
KK> Andy, I don't mean to offend you personally, but one does have to
KK> wonder what experimentation with a fix happens before it lands in CVS.
KK> Is it any surprise that such wonderment exists, when tabs literally
KK> don't work at all?
KK> Regrets to any others I might have also offended.
We should be looking towards a release of XEmacs 21.2 soonish. The
things holding us up:
- We'd like to turn on the portable dumper by default, but it's not
actually quite done yet.
- We'd like to solve any redisplay/gutter code stability problems.
- I'd like to do something about the packaging of the packages. Not
sure quite what, though.
I'm willing to help to do whatever it takes to make these things
happen, with the understanding that Olivier, Andy, Steve, and Michael
all have more expertise on the features they implemented than I do.
Andy, I'd like to see these things for the gutter feature:
- documentation
- add to the automated test suite (might be hard - currently there are
no automated gui tests).
- turn off by default for non-beta releases. Or, we could hold a vote
in xemacs-beta or even comp.emacs.xemacs as to whether this feature
is one they are using. Unfortunately, getting an unbiased result of
such a vote is tough.
Martin