xemacs vs emacs

[David Kastrup]

"Stephen J. Turnbull" <stephen at xemacs.org> writes:

> David Kastrup writes:
>
>  > Oh puleeze!  The _only_ party that is _objecting_ to AUCTeX trying
>  > to support XEmacs are the XEmacs core team.  They don't accept our
>  > package in distribution.
>
> That last is true,

Fine.  Remember that.

> but it is precisely the same policy that we face with respect to
> Debian (and I assume every other distribution).

But they don't call you "objecting in principle" to Debian if you have
no developers willing to provide and maintain the complete Debian
support.

> They will not distribute our tarballs; they insist that they must
> build the SUMOs themselves and wrap them up in .debs themselves,
> according to their procedures.  There's nothing unusual or
> unreasonable about this, and I do not understand why you think it
> should be any other way.

Oh, that's all fine.  Except that Debian does not constantly abuse and
sabotage and blame upstream developers who have no interest of their own
in Debian.

>  > They demand that we have to scrap our own package build and
>  > directory structure and adapt the XEmacs style of installing
>  > packages
>
> No one has ever made such a suggestion, let alone demanded it.  If
> there is no support for the XEmacs package system, though, it's not an
> XEmacs package.  The point of the XEmacs package system is not merely
> producing tarballs of a specified format.

It is producing _broken_ tarballs of specified format apparently.

> The point of our package system is building each package in the
> environment in which it will be used, and verifying that it is happy
> to build and run in that environment, that other packages don't cause
> any problems for it, and that it doesn't cause any problems for other
> packages.

Unfortunately, that is bunk since the XEmacs package repositories are
not platform-specific.  Which means that the packages have to be
designed to build and run _independently_ of the build system.

> We automatically do a *full* build of the whole tree for *every*
> change committed to the packages.[1] Why should AUCTeX be an exception
> to that practice?

Apparently because somebody would have to work on it.  And you have
repeatedly stated that you don't see sufficient user interest in a
working XEmacs package of AUCTeX to warrant working on it.

>  > (never mind that this is much less versatile)
>
> But we don't care about versatility; we care that it work in the
> context of our package system.  This is the whole point, the only
> point.  We don't tell you how to run your project, and we don't demand
> effort from you.  We *offer* CVS access and *request* that you do as
> much of the integration work as you will, because that is the cheapest
> way for the community to get an updated AUCTeX as an XEmacs package.

No, it isn't.  XEmacs packaging of AUCTeX requires expert work, as
witnessed by the events of the last years.  The "integration work" is
not possible for an outsider to do with the existing documentation.  Not
in a reasonable timeframe.  I can't waste my sparse developer time on
fixing the broken package build system of XEmacs (which can't even be
_used_ without CVS access: there are no tarballs or anything).

> This obviously means extra work all around, with most of the burden
> falling on you if you will accept it.  Nobody says you are under any
> obligation, though.

I have neither time nor inclination to accept it.  I respect the people
who decide to waste their time fighting XEmacs policies and
documentation and integration and code.  And I'll support their
endeavours.  But I am not doing the work myself.

If nobody else is interested in a working XEmacs, it will certainly not
be me who leaves last and switches off the lights.  Though it sometimes
feels like I am pretty close already.

>  > What I get in return is ostracization of the package we produce and
>  > tirades on the XEmacs list what a bad person I am for objecting to
>  > supporting XEmacs and similar crap.
>
> *chuckle* David, we have our differences, but that doesn't make you a
> bad person.  Still, we'd all get along a lot better if you'd admit
> that we just have goals that differ,

And if you'd accept that you are (and nobody else is) responsible for
reaching your goals.

> and that getting everybody satisfied is going to take more effort than
> what it takes to satisfy *you* that *you* have done everything *you*
> think is necessary.

But the effort is not mine to make.

-- 
David Kastrup