bug#13915: 24.3.50; `font-lock-syntactic-keywords' - really obsolete?

bug#13915: 24.3.50; `font-lock-syntactic-keywords' - really obsolete?

[Drew Adams]

NEWS says that the variable is obsolete, and font-lock.el says that this
is so since 24.1:
 
(make-obsolete-variable 'font-lock-syntactic-keywords
                        'syntax-propertize-function "24.1")
 
But font-lock.el uses it, with no apparent attempt to do without it.
 
If the very file that declares it obsolete cannot be updated to reflect
that, why tell users they should consider it obsolete?
 
The description in NEWS does not make clear, to me at least, how use of
the variable should be replaced.  If I see how that gets done in
font-lock.el then perhaps I will understand better.  If not, it sounds
like not practicing what you preach.  Or perhaps making the change is
not so simple?
 
There are only a few lines in font-lock.el that involve
`font-lock-syntactic-keywords'.  If these were not changed when it was
declared obsolete, and still have not been, 2 or 3 releases later, is
that a sign that something is not quite right?
 
Seems like those few lines would have been fixed immediately.  Now,
presumably some of them are there for backward compatibility.  But I
don't get the impression that they all are, or even that most of them
are.  Seems like the switchover to the new was simply not finished.
 
It would be helpful too if the NEWS description provided more guidance
wrt replacing use of the obsolete variable with use of the new.
 
 
 
In GNU Emacs 24.3.50.1 (i386-mingw-nt5.1.2600)
 of 2013-03-04 on ODIEONE
Bzr revision: 111935 yamaoka@jpl.org-20130304102733-4qy111z41qwoh2as
Windowing system distributor `Microsoft Corp.', version 5.1.2600
Configured using:
 `configure --with-gcc (4.7) --no-opt --enable-checking --cflags
 -IC:/Devel/emacs/build/include --ldflags -LC:/Devel/emacs/build/lib'
 

bug#13915: 24.3.50; `font-lock-syntactic-keywords' - really obsolete?

[Stefan Monnier]

> If the very file that declares it obsolete cannot be updated to reflect
> that, why tell users they should consider it obsolete?
 
Hi Drew!  How's life on your side of the world?
Oh, and by the way, what would you say if we removed support for
font-lock-syntactic-keywords without first marking it obsolete?

Hmm... thought so,


        Stefan

bug#13915: 24.3.50; `font-lock-syntactic-keywords' - really obsolete?

[Drew Adams]

> > If the very file that declares it obsolete cannot be 
> > updated to reflect that, why tell users they should
> > consider it obsolete?
>  
> Hi Drew!  How's life on your side of the world?
> Oh, and by the way, what would you say if we removed support for
> font-lock-syntactic-keywords without first marking it obsolete?
> 
> Hmm... thought so,

No one suggested that you should remove support for it before marking it
obsolete!  You are right to suppose that I would never have suggested such a
thing.  So why bring it up as a straw man?

But perhaps it says something that that was your only reaction to the bug
report, though that was not at all what the report was about.  What do you
think?

1. NEWS does not make clear how to change code that currently uses
`font-lock-syntactic-keywords'.

2. `font-lock.el' does not make that clear either, as it has itself apparently
not been so converted.

I made clear that some of the occurrences in `font-lock.el' obviously provide
ongoing support for it.  But only a few of them, would be my guess.  Am I wrong
about that?

3. You closed the bug.  `font-lock-syntactic-keywords' was marked obsolete in
24.1.  24.3 has now been released.

4. But the 24.3 NEWS is no clearer about this.  Both the 24.3 Elisp manual and
the 24.3  font-lock.el Commentary and doc strings still speak mainly of
`font-lock-syntactic-keywords' (`syntax-propertize-function' is barely present
in font-lock.el).  As before.

Users are not helped any more than before wrt converting code that uses
`font-lock-syntactic-keywords'.

That's what this bug report is about.

bug#13915: 24.3.50; `font-lock-syntactic-keywords' - really obsolete?

[Stefan Monnier]

> No one suggested that you should remove support for it before marking it
> obsolete!  You are right to suppose that I would never have suggested such a
> thing.  So why bring it up as a straw man?

Because the "uses" you seem to see in font-lock.el are just the code
that *implements* font-lock-syntactic-keywords.  None of it *uses* it.
So removing those "uses" would remove support for it.


        Stefan

bug#13915: 24.3.50; `font-lock-syntactic-keywords' - really obsolete?

[Drew Adams]

> > No one suggested that you should remove support for it before
> > marking it obsolete!  You are right to suppose that I would
> > never have suggested such a thing.  So why bring it up as a
> > straw man?
> 
> Because the "uses" you seem to see in font-lock.el are just the code
> that *implements* font-lock-syntactic-keywords.  None of it *uses* it.
> So removing those "uses" would remove support for it.

I would humbly suggest that most of the occurrences (I did not refer to the
occurrences as "*uses*", so your quotes are inaccurate here) are vestigial, in
doc strings.  They do not *implement* `font-lock-syntactic-keywords'.

And I was clear that any occurrences that remain to *implement* the old should
of course remain.

Since a preferred replacement has already been implemented, that (the new, not
the old) is what should be mentioned, above all, in the doc strings.  That is
not the case.  That is the point about `font-lock.el' occurrences.

But the *important* point is what the bug report is really about:

>> 4. But the 24.3 NEWS is no clearer about this.  Both the 24.3 
>> Elisp manual and the 24.3  font-lock.el Commentary and doc strings
>> still speak mainly of `font-lock-syntactic-keywords'
>> (`syntax-propertize-function' is barely present in font-lock.el).
>>
>> As before.
>> 
>> Users are not helped any more than before wrt converting code 
>> that uses `font-lock-syntactic-keywords'.
>>
>> That's what this bug report is about.

If you update the NEWS & the manual & the doc strings, and leave the code as is
to support both old and new, then the bug will be fixed.

The NEWS, in particular, should tell users how to replace the old with the new.
It is no help wrt that, so far.  It is clearly not as simple as substituting one
(a function) for the other (a variable).

It's about helping users.  If I did not find this unclear myself I would not
have taken the time to file the bug (and reply to your (so far) sarcastic and
unhelpful replies).

bug#13915: 24.3.50; `font-lock-syntactic-keywords' - really obsolete?

[Stefan Monnier]

> But the *important* point is what the bug report is really about:

Then please make another bug report where you focus on what the bug
is about.  It's taking too much time to figure that out in too many of
your reports.


        Stefan

bug#13915: 24.3.50; `font-lock-syntactic-keywords' - really obsolete?

[Drew Adams]

> > But the *important* point is what the bug report is really about:
> 
> Then please make another bug report where you focus on what the bug
> is about.  It's taking too much time to figure that out in too many of
> your reports.

This bug report focuses on it quite clearly, from beginning to end.

If it is really taking time for you to figure it out, that is apparently because
you are belaboring it and mocking.  You have already made clear that you are not
interested in fixing it, and you have closed it.

If you want to play games and create another bug about it, go right ahead.

If you change your mind and decide to fix it, that will be helpful and
constructive.  Start with the NEWS, giving users some useful, operational info
about replacing the old with the new.