bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

[Drew Adams]

emacs -Q

C-x b *Messages*

C-h k 7

Tells you that 7 is bound to `digit-argument'.

Why?  It should NOT be.  It is useful to be able to write text in
*Messages*, e.g., to set off or group sets of debug messages.  There is
no need for digit N to be a short cut for C-u N in buffer *Messages*.

C-x C-q ; Make the buffer writable

Unfortunately, Emacs has recently screwed *Messages* so that most keys
are not self-inserting.  This was misguided.  Be that as it may...

C-q H C-q e C-q l C-q p C-q !

Painful, but OK so far.  Now try this:

C-q 7

Nothing.  Nada.

C-q 7 C-q 7 C-q 7.  You see ^G inserted each time (except the first)
that you hit C-q (not when you hit 7), because 7 acts as a prefix arg
for the following ^Q.  Wunderbar.  A feature, no doubt.

C-q 7 7 7.  Now Emacs is hosed - in Neverneverland.  ESC ESC ESC no
good; C-g no good.  No mouse, no keys, nada.  And no CPU use.  Just
hosed.

The only recourse is to kill the Emacs task from the Windows Task
Manager.

OK, you might not see this freezing behavior each time.  Sometimes C-g
does quit OK, as does C-b, muose-1 etc.  Hit C-q and then hold 7 down
for a few seconds, then C-g.  If Emacs is not frozen solid and C-g works
then you get helpful feedback such as this:

Quit
user-error: A-C-H-M-S-s- is not a valid character

Other times, Emacs freezes solid.  Regardless of whether it freezes each
time, if you hit C-q and then 7 one or more times there is no feedback -
already a no-no (bug).

C-q should simply quote the 7, as it should quote other characters you
type (including ^U, BTW).

Trying to "improve" *Messages* for Emacs 24 was misguided.  It is far
less useful now than it was before.  Is there even one change that was
made that represents progress - something useful?  Please consider
reverting it to what it was since Day One.  Barring that requested
return to sanity, please at least fix the C-q behavior.  Thx.


In GNU Emacs 24.4.50.1 (i686-pc-mingw32)
 of 2014-06-28 on ODIEONE
Bzr revision: 117431 rgm@gnu.org-20140628015517-eku6hj8mpgcvfnso
Windowing system distributor `Microsoft Corp.', version 6.1.7601
Configured using:
 `configure --prefix=/c/Devel/emacs/snapshot/trunk
 --enable-checking=yes,glyphs 'CFLAGS=-O0 -g3'
 LDFLAGS=-Lc:/Devel/emacs/lib 'CPPFLAGS=-DGC_MCHECK=1
 -Ic:/Devel/emacs/include''

bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

[Drew Adams]

> C-q should simply quote the 7, as it should quote other characters you
> type (including ^U, BTW).
        ^^^^^^^^^^^^^^^^^
(I meant that it already does correctly quote ^U.)

bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

[Glenn Morris]

Part of this is you trying
http://lists.gnu.org/archive/html/emacs-devel/2013-09/msg00284.html

again. This is about the third time now. Give it a rest.

For the rest, read the help of quoted-insert.

If you want to type text, maybe try text-mode?

If you have a clear, reproducible recipe to freeze Emacs starting from
emacs -Q, I suggest opening a new report that just gives that recipe,
minus all the invective.

bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

[Drew Adams]

> If you have a clear, reproducible recipe to freeze Emacs starting from
> emacs -Q

I provided such a recipe.  Did you try it?

bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

[Eli Zaretskii]

> Date: Wed, 9 Jul 2014 19:35:05 -0700 (PDT)
> From: Drew Adams <drew.adams@oracle.com>
> 
> C-q 7
> 
> Nothing.  Nada.
> 
> C-q 7 C-q 7 C-q 7.  You see ^G inserted each time (except the first)
> that you hit C-q (not when you hit 7), because 7 acts as a prefix arg
> for the following ^Q.  Wunderbar.  A feature, no doubt.

As Glenn points out, this is normal (and very old) behavior of C-q,
which is well documented.

> C-q 7 7 7.  Now Emacs is hosed - in Neverneverland.  ESC ESC ESC no
> good; C-g no good.  No mouse, no keys, nada.  And no CPU use.  Just
> hosed.

Not here, it isn't.  After "C-q 7 7 7", any key gets something
inserted, and there's no hang.  None.

> Regardless of whether it freezes each time, if you hit C-q and then
> 7 one or more times there is no feedback - already a no-no (bug).

That's indeed a bug, worthy of a separate bug report.

But everything else you describe isn't, or is not reproducible here.

> C-q should simply quote the 7, as it should quote other characters you
> type (including ^U, BTW).

No, numeric characters after C-q behave specially.

bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

[Eli Zaretskii]

> Date: Thu, 10 Jul 2014 06:55:21 -0700 (PDT)
> From: Drew Adams <drew.adams@oracle.com>
> 
> > If you have a clear, reproducible recipe to freeze Emacs starting from
> > emacs -Q
> 
> I provided such a recipe.  Did you try it?

I did.  It didn't hang for me.

bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

[Eli Zaretskii]

> Date: Thu, 10 Jul 2014 17:57:49 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: 17983@debbugs.gnu.org
> 
> > Regardless of whether it freezes each time, if you hit C-q and then
> > 7 one or more times there is no feedback - already a no-no (bug).
> 
> That's indeed a bug, worthy of a separate bug report.

I've just reported it.

bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

[Drew Adams]

> > C-q 7 C-q 7 C-q 7.  You see ^G inserted each time (except the first)
> > that you hit C-q (not when you hit 7), because 7 acts as a prefix arg
> > for the following ^Q.  Wunderbar.  A feature, no doubt.
> 
> As Glenn points out, this is normal (and very old) behavior of C-q,
> which is well documented.

Yes, for inserting a char via an octal sequence.  And yes, any number
of octal digits can be used.  I had forgotten that this single 7 fell
into that use case.

So the behavior is normal: each C-q after the first ends the octal
sequence, so it is just 7, which is the code for ^G.  Thanks for the
reminder that this is what I was seeing here.

It would be good, IMO, if there were some echo-area feedback when
you hit an octal digit after C-q.  Please consider that enhancement
request.

> > C-q 7 7 7.  Now Emacs is hosed - in Neverneverland.  ESC ESC ESC no
> > good; C-g no good.  No mouse, no keys, nada.  And no CPU use.  Just
> > hosed.
> 
> Not here, it isn't.  After "C-q 7 7 7", any key gets something
> inserted, and there's no hang.  None.

As I said before, sometimes it happens and sometimes it does not.
I was not able to determine just when or why.  Each time, I started
with emacs -Q, and the result was sometimes a hang and sometimes not.
I reproduced the hang several times, including several times in a
row (always from emacs -Q), but the same recipe sometimes did not
result in a hang.

> > Regardless of whether it freezes each time, if you hit C-q and then
> > 7 one or more times there is no feedback - already a no-no (bug).
> 
> That's indeed a bug, worthy of a separate bug report.

See above.  There should be no need for a separate report.  This was
part of this report, from the beginning.  If you want to create a
separate report for it or rename the subject to better reflect this,
feel free to do so.

I feel that I've reported the problem and asked that it be fixed.
If you find that part of what I reported is not a bug, and because
of that you want to ignore other parts and not fix them, that's
your prerogative.

> But everything else you describe isn't, or is not reproducible here.

See above.

> > C-q should simply quote the 7, as it should quote other characters you
> > type (including ^U, BTW).
> 
> No, numeric characters after C-q behave specially.

(Only octal digits, to be specific - or whatever is specified by
`read-quoted-char-radix'.)

Yes, thanks for that reminder.  I did not realize that that's what
I was seeing here (even though I have long used C-q to input an octal
sequence.)

bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

[Drew Adams]

> > > Regardless of whether it freezes each time, if you hit C-q and then
> > > 7 one or more times there is no feedback - already a no-no (bug).
> >
> > That's indeed a bug, worthy of a separate bug report.
> 
> I've just reported it.

Very good.  Thx.

bug#17983: 24.4.50; digit keys in *Messages* unhelpful, hang Emacs, etc.

[Eli Zaretskii]

> Date: Thu, 10 Jul 2014 09:25:50 -0700 (PDT)
> From: Drew Adams <drew.adams@oracle.com>
> Cc: 17983@debbugs.gnu.org
> 
> It would be good, IMO, if there were some echo-area feedback when
> you hit an octal digit after C-q.  Please consider that enhancement
> request.

It's not an enhancement: all Emacs versions before 24.4 did echo the
input.  This is a regression that should be fixed.

> > > C-q 7 7 7.  Now Emacs is hosed - in Neverneverland.  ESC ESC ESC no
> > > good; C-g no good.  No mouse, no keys, nada.  And no CPU use.  Just
> > > hosed.
> > 
> > Not here, it isn't.  After "C-q 7 7 7", any key gets something
> > inserted, and there's no hang.  None.
> 
> As I said before, sometimes it happens and sometimes it does not.

I tried many times, and couldn't get it to hang.

> > > Regardless of whether it freezes each time, if you hit C-q and then
> > > 7 one or more times there is no feedback - already a no-no (bug).
> > 
> > That's indeed a bug, worthy of a separate bug report.
> 
> See above.  There should be no need for a separate report.  This was
> part of this report, from the beginning.  If you want to create a
> separate report for it or rename the subject to better reflect this,
> feel free to do so.

There are too many issues in this report, so I filed a new one for the
lack of echo.