Unify on encoding

Unify on encoding

[Stefan Monnier]

Is there any objection to the patch below ?

It turns on unify-8859-on-encoding-mode by default (in loadup.el) which has
the effect of allowing a wider choice of coding-systems when saving a file.
It should not have any other effect (i.e. if you choose one of the
coding-systems which would be available even without unification-on-encoding,
the resulting output is exactly the same).

With it, you can load a latin-9 file and save it as latin-1 or utf-8.

A second request (but separate) is whether there is any objection to removing
the ability to turn off unify-8859-on-encoding-mode.


	Stefan


Index: loadup.el
===================================================================
RCS file: /cvsroot/emacs/emacs/lisp/loadup.el,v
retrieving revision 1.118
diff -u -u -b -r1.118 loadup.el
--- loadup.el	20 Jan 2002 15:35:09 -0000	1.118
+++ loadup.el	17 Mar 2002 22:26:04 -0000
@@ -112,6 +121,10 @@
 (load "language/misc-lang")
 (load "language/utf-8-lang")
 (load "language/georgian")
+
+(load "international/ucs-tables")
+(unify-8859-on-encoding-mode 1)
+
 (update-coding-systems-internal)
 
 (load "indent")


_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Eli Zaretskii]

On Sun, 17 Mar 2002, Stefan Monnier wrote:

> A second request (but separate) is whether there is any objection to removing
> the ability to turn off unify-8859-on-encoding-mode.

FWIW, I object in principle to removing user options, especially for 
brand-new features we never tested in a public release.  I think we 
should revisit this specific issue after Emacs 21.3 is released and
used for a while, based on user reports.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Stefan Monnier]

> On Sun, 17 Mar 2002, Stefan Monnier wrote:
> > A second request (but separate) is whether there is any objection to removing
> > the ability to turn off unify-8859-on-encoding-mode.
> FWIW, I object in principle to removing user options, especially for
> brand-new features we never tested in a public release.  I think we
> should revisit this specific issue after Emacs 21.3 is released and
> used for a while, based on user reports.

Since this user option has not been introduced in any release yet,
I fail to see why this would be considered as "removing user option".

Instead I think it's a mistake to introduce this new user option since
there's no reason to ever turn off the feature.


	Stefan


_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Eli Zaretskii]

On Mon, 18 Mar 2002, Stefan Monnier wrote:

> Since this user option has not been introduced in any release yet,
> I fail to see why this would be considered as "removing user option".

I'm guessing that quite a few people are using the CVS version nowadays, 
so the option was, in a certain sense, ``released''.  But you are right 
that it's less grave than if it were present in a released version.

> Instead I think it's a mistake to introduce this new user option since
> there's no reason to ever turn off the feature.

I think we don't know yet if there is a reason to turn that off.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Richard Stallman]

    With it, you can load a latin-9 file and save it as latin-1 or utf-8.

It sounds good to me.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Richard Stallman]

    I'm guessing that quite a few people are using the CVS version nowadays, 
    so the option was, in a certain sense, ``released''.

As a matter of principle, this is NOT a reason to keep a feature which
we decide is unnecessary.  People who use the CVS version do so at
their own risk.  We must firmly reject the idea that the CVS version
exists for other than our development purposes.

    > Instead I think it's a mistake to introduce this new user option since
    > there's no reason to ever turn off the feature.

    I think we don't know yet if there is a reason to turn that off.

Let's take it out and see if anyone complains.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Eli Zaretskii]

On Tue, 19 Mar 2002, Richard Stallman wrote:

>     I'm guessing that quite a few people are using the CVS version nowadays, 
>     so the option was, in a certain sense, ``released''.
> 
> As a matter of principle, this is NOT a reason to keep a feature which
> we decide is unnecessary.  People who use the CVS version do so at
> their own risk.

This is a misunderstanding.  The current mode of operation of released 
Emacs versions is that there's no unification on encoding.  The CVS code 
added an option to turn on that unification.  Stefan's question was about 
whether turning it on by default (to which I didn't object) should be 
accompanied by removing the possibility to turn it off.

I only objected to the latter part, because doing so would change an 
important characteristic of the default operation without leaving an
escape route for users who don't want the new default.  I don't think 
I've ever seen defaults changed in Emacs without some option to go back.

In other words, the option that was recommended to be removed ``exists''
in the released version--in the current default operation mode.

> Let's take it out and see if anyone complains.

What's the rush?  An option that is turned off cannot possibly do any 
harm.  Bitter experience with Mule should have taught us that we 
shouldn't make changes without a good reason.  Isn't it better to get 
some experience from having unification turned on by default?

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Jason Rumney]

Eli Zaretskii <eliz@is.elta.co.il> writes:

> I only objected to the latter part, because doing so would change an 
> important characteristic of the default operation without leaving an
> escape route for users who don't want the new default.

The escape route would be to select an iso2022 or emacs-mule coding
system from the expanded list offered.  Or do those get unified as well?


-- 
Jason Rumney


_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Stefan Monnier]

> > I only objected to the latter part, because doing so would change an 
> > important characteristic of the default operation without leaving an
> > escape route for users who don't want the new default.
> 
> The escape route would be to select an iso2022 or emacs-mule coding
> system from the expanded list offered.  Or do those get unified as well?

I believe you're confusing the current issue (unify-on-encoding)
which has no known negative side-effect with the problem discussed
a few weeks/months back that occurs due to unify-on-decoding.


	Stefan


_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Kenichi Handa]

Eli Zaretskii <eliz@is.elta.co.il> writes:
> On Sun, 17 Mar 2002, Stefan Monnier wrote:

>>  A second request (but separate) is whether there is any objection to removing
>>  the ability to turn off unify-8859-on-encoding-mode.

> FWIW, I object in principle to removing user options, especially for 
> brand-new features we never tested in a public release.  I think we 
> should revisit this specific issue after Emacs 21.3 is released and
> used for a while, based on user reports.

I agree on turning on unify-8859-on-encoding-mode by default.

As for removing the ability of turning off
unify-8859-on-encoding-mode, I'm not sure.

If we remove it, that means all Emacs codes can assume
unify-8859-on-encoding-mode unconditionally.  If that
decreases some cost of maintaining some code, I agree with
removing it.

If keeping it just means that keeping a few variables and
functions, we don't have to remove it now.  We can remove it
after confirming that there's no case that
unify-8859-on-encoding-mode does something bad through the
pretest period of 21.3.

---
Ken'ichi HANDA
handa@etl.go.jp


_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Eli Zaretskii]

On 19 Mar 2002, Jason Rumney wrote:

> The escape route would be to select an iso2022 or emacs-mule coding
> system from the expanded list offered.  Or do those get unified as well?

I don't know; I didn't yet have time to try this mode.  Perhaps Stefan 
can answer that.

In general, iso-2022 tags each charset with a special escape sequence, so 
at least in principle, you could have, say, latin-2 characters retagged 
as latin-1 there.  But I don't know if that actually happens.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Unify on encoding

[Stefan Monnier]

> 
> On 19 Mar 2002, Jason Rumney wrote:
> 
> > The escape route would be to select an iso2022 or emacs-mule coding
> > system from the expanded list offered.  Or do those get unified as well?
> 
> I don't know; I didn't yet have time to try this mode.  Perhaps Stefan 
> can answer that.
> 
> In general, iso-2022 tags each charset with a special escape sequence, so
> at least in principle, you could have, say, latin-2 characters retagged
> as latin-1 there.  But I don't know if that actually happens.

No, as I said, any char that can be encoded by a coding-system will be
encoded just as it was before.  The only difference is now some coding
systems can encode more chars than before.


	Stefan


_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel