emacs-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Semantic: update or remove?


From: Alfred M. Szmidt
Subject: Re: Semantic: update or remove?
Date: Tue, 18 Mar 2025 16:20:36 -0400

   The current maintainers of Emacs are Eli Zaretskii, Stefan Kangas,
   and Andrea Corallo.  At least one of them has explicitly stated
   support for moving Semantic to GNU ELPA.

That is false.  The maintainers of Emacs are, as per GNU records:

  Andrea Corallo, Eli Zaretskii, John Wiegley, Lars Ingebrigtsen,
  Richard Stallman, Stefan Kangas.

The only one stated anything is you.

   > Not to mention the question was technical, i.e. why
   [...]
   > which was entierlly ignored.

   Alfred, you've been around long enough to know that repeating a
   single-word question like "why?" without engaging substantively is
   unlikely to elicit detailed responses.

I asked why, because no reason was given.  As a maintainer, you should
be able to give such a reason, and I was interested in the reaasons so
I asked why.  You're being dishonest at best in trying to envoke "you
should know better" -- assume people are posting in good faith.

   - Semantic is currently unmaintained.  There have been almost no
     non-trivial changes to the lisp/cedet/ directory since 2019, and the
     grammar files are no longer kept up-to-date with changes in the
     underlying languages.  To be generous, this is unlikely to change
     unless someone steps up.

Many parts of emacs get no non-trivial changes, plenty of Emacs works
becasue it is stable.  For example, outline-mode.  Po Lu, by your own
admission, a core contributor, said objected which seems to have been
not good enough.

   - Rebasing Semantic on Tree-sitter or LSP has been suggested, but, as
     Stefan Monnier noted, this would be a major undertaking with uncertain
     payoff.  No one has volunteered to do that work.

Not pertinent to well tested, and working code.   

   - Performance remains a concern.  Improving it would likely require
     significant effort, and we have no volunteers for this either.

Hogwash, I'm using it on a 2 million line code base and it works just
fine.

   - It is difficult to use and extend.  It has a steep learning curve,
     both for users and for potential contributors.  Its original author
     remarked in 2022 (see below) how hard it was to get CEDET to "just
     work", and how often users gave up and turned to simpler alternatives.

So are plenty of other things in Emacs.  

   - Built-in IDE functionality is progressing -- just not via CEDET.
     Tools like Eglot and Tree-sitter are now the focus of Emacs's
     development in this area.  They have broad language support and large
     upstream communities, which is essential given our limited resources.
     (This was a major factor behind our decision to integrate and start
     promoting Tree-sitter support, for example.)

This is more to preference than anything else.  What next? vc-mode to
be deprecated because magit is "progressing more"? outline-mode to be
deprecated because org-mode is getting more commits?

   - Keeping CEDET in-tree risks misleading new users.  They may stumble
     across it, spend time trying to get it working, and come away
     frustrated or confused.  This reflects poorly on Emacs overall.[2]

Yes, somethings take effort to setup (GNUS for example!), and there is
nothing wrong with that.  How are users mislead exactly?  I stumbled
upon GNUS once, I never could set it up .. does that also reflect
badly on Emacs?

   - Moving it to GNU ELPA preserves it for those who want it.  This is not
     deletion.  It acknowledges the reality of the situation, avoids
     promoting Semantic as actively maintained or recommended, and keeps it
     available for those who still find it useful, or indeed anyone who
     might want to revive it.

It breaks everything for those who expect it, and gains nothing to
those who do not want to use CEDET.  Nobody forces you to even take a
look at CEDET.

     He concluded:

         "Overall, I think that is fine though – having many projects
         experimenting with different techniques, and having the best
         solution win is the benefit of free software."

Which reads to me, as keeping CEDET for those who want it, and those
who prefer something else can use something else.

   If you're still unconvinced, I suggest reading this thread in full with
   an open mind.

I did.  Did you?

   > The uproar is really that, future users of Emacs will have to
   > explicitly install something that have been part of Emacs for many
   > decades.  This is a major breakage, which Kangas takes so very lightly
   > that is is absurd that it cannot be taken seriously.

   We are not obligated to continue shipping features that are
   unmaintained, no longer viable, and already superseded by better
   alternatives.  Users having to explicitly install such packages is a
   reasonable and appropriate outcome.

Should outline-mode to be deprecated in favor of org-mode?
outline-mode has not gotten a non-trivial change in decades.

Alan Mackenzie was on point, this is it all over again.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]