Here’s an incomplete list of GNU Emacs releases. See also GNU Emacs Release History, the Emacs Timeline by JamieZawinski, and EmacsHistory for more information.


Jokes about releasing Emacs 21

The rest of this page contains interesting conversations from EmacsMailingLists about releasing Emacs 21.

The standard response is this one:

From: StefanMonnier
Subject: Re: emacs 21
Newsgroups: comp.emacs
Date: 2001-02-06 08:25:20 PST

“Bas” writes:

Does anybody know when it will be released ?

As soon as it’s ready,


An excerpt from Irving Stone’s Agony and the Ecstasy.

During all these months the Pope kept insisting that Michelangelo complete his ceiling quickly, quickly! Then one day [Pope] Julius climbed the ladder unannounced.

‘When will it be finished?’

‘When I have satisfied myself.’

‘Satisfied yourself in what? You have already taken four full years.’

‘In the matter of art, Holy Father.’

‘It is my pleasure that you finish it in a matter of days.’

‘It will be done, Holy Father, when it will be done.’

We’ve had different answers, too, however. It all started with a mail by Thaddeus L. Olczyk in summer 01 [1]:

From: Thaddeus L. Olczyk
Subject: Status of Emacs 21.
Newsgroups: comp.emacs
Date: 2001-07-10 15:29:55 PST

People are once again asking about Emacs 21, and it’s been about four weeks since I asked this question. So once again I ask this question to stall all the posters that will soon come asking.

Will Emacs 21 be released:
Immenently (next 2-3 weeks)?
soon ( 2-3 months)?
not so soon (6 months to a year)?
a long time off ( one year to five years)?
almost never (10 years)?

From: EliZaretskii
Subject: Re: Emacs 21: Two questions
Newsgroups: comp.emacs
Date: 2001-06-14 08:06:39 PST


It’s not that we know the date but don’t want to tell. We really don’t know. As long as there are reports about crashes on popular platforms, we won’t release. And just the other day I saw two crash reports, after at least 2-3 weeks without a single one.

So will Emacs 21 be released:
Immenently (next 2-3 weeks)?


soon ( 2-3 months)?

Most probably.

not so soon (6 months to a year)?


a long time off (one year to five years)?
almost never (10 years)?


From: JuanManuelBarranquero
Subject: Re: Status of Emacs 21.
Newsgroups: comp.emacs
Date: 2001-07-16 03:28:22 PST

“Allan Peda” wrote:

I too am curious, but a “very stable/not very stable” reply would be better than nothing.
 91  2
 92 10
 93 13
 94  6
 95 23
 96 13
 97  2
 98 16
 99 18
100  9
101 22
102  1
103 33
104 66

First column is prerelease number, second one is the number of days between the prerelease and the previous one.

Discounting the one-digit rows as quick corrections of small oversights, it is easy to see that things seems to be converging towards stabilization.

(BTW, I’m not speaking for the Emacs maintainters, which I am not).


From: EliZaretskii
Subject: Re: Status of Emacs 21.
Newsgroups: comp.emacs
Date: 2001-07-17 09:39:47 PST

Thaddeus L. Olczyk wrote:

OK. This would seem to be bad news. It means we have to wait at least 66 days after 104 was released before we can be sure that there won’t be a 105.

EliZaretskii wrote:

Actually, it gets worse: the release criteria require that after each pretest 21.0.N, we need to wait N^Pi days, and if there no crashes or other similarly grave problems reported, only then we can make a release.

Jonadab the Unsightly One wrote:

That’s several millenia! I’m not waiting that long.

And what exactly are you going to do about that, huh? Switch to XEmacs, perhaps??

From: Stefan Monnier
Subject: Re: Status of Emacs 21.
Newsgroups: comp.emacs
Date: 2001-07-17 11:25:30 PST

Jonadab the Unsightly One writes:

That’s several millenia!

But think how good it’ll be,


From: EliZaretskii
Subject: Re: Status of Emacs 21.
Newsgroups: comp.emacs
Date: 2001-07-17 20:18:16 PST

GalenBoyer wrote:

I am perfectly happy with 20.7, and I’m really excited to see what wonders await with 21, but I would be perfectly happy with 20.7 for quite a few years.

those who know me have no need of my name wrote:

you’d better be, since the formula eli gave (presuming ‘Pi’ meant the irrational number 3.1415…) calls for about 3000 years of additional testing.

David Koppelman wrote:

That would be 5943 years (104^pi days), which works out to AD 14 January 7946, according to the Mathematica date functions based on a 13 July release date for Emacs 104, using
      ToDate[FromDate[{2001,7,13,0,0,0}] + Floor[N[104^Pi]] 3600 24])
Emacs’ calendar package couldn’t handle this:
        (+ 5944 (calendar-absolute-from-gregorian (calendar-current-date)))))

I think your calculations are wrong (3000 years is much closer to the accurate result than 5944), and I didn’t have any trouble adding that with calendar (you just have to add days instead of years). But if you think something is wrong with calendar, please make a bug report. E.g., “Emacs’s Calendar is unable to compute the date of its own final release”.

Pretest numerology

  1. First pretest for Emacs 21 released on 2000-11-17 [2]
  2. Emacs 21 released on 2001-10-28 (346 days in pretest)
  3. First pretest for Emacs 22 released on 2006-11-27 [3]
  4. Emacs 22 released on 2007-06-03 (189 days in pretest)
  5. First pretest for Emacs 23 released on 2009-02-01 [4]
  6. Emacs 23 released on 2009-07-29 (179 days in pretest)
  7. First pretest for Emacs 24 released on 2011-09-25 [5]
  8. Emacs 24 released on 2012-06-10 (260 days in pretest)
  9. First pretest for Emacs 25 released on 2016-01-30 [6]
  10. Emacs 25 released on 2016-09-17 (245 days in pretest)
  11. First pretest for Emacs 26 released on 2017-10-11 [7]

CategoryHistory CategoryHumor