MirOS Project - Discussion list

Luc GMail | 3 Jul 2009 16:14

Picon

Emulate Ctrl-Delete in mksh

Hi. I really like most of the default key bindings in mksh. I don't
need to set them manually like in other ports of the Korn shell. Even
the Delete key works fine: it deletes one character to the right of
the insertion point. So does Ctrl-d.

The difference is that Ctrl-d will close the session (exit) if no
characters are left to be deleted. The Delete key won't do that. But I
like the Delete key, I think that Ctrl-d is annoying to type. Is there
any way I can add that Ctrl-d feature to the Ctrl-Delete key
combination? I can map Delete or Ctrl-letter, but I cannot map
Ctrl-Delete. I can't find the right string that says "Ctrl-Delete".
Everything I try makes mksh complain that the string "is too long".
Any ideas?

TIA

--

-- 
Luc

Thorsten Glaser | 3 Jul 2009 18:58

Picon

Re: Emulate Ctrl-Delete in mksh

Luc GMail dixit:

>I can't find the right string that says "Ctrl-Delete".

It's terminal dependent. A GNU screen inside wscons does, for example,
not differentiate between these two at all. I think Ctrl-X only exists
for ' <at> ' <= X <= '_' at all.

>Everything I try makes mksh complain that the string "is too long".

Your terminal may have one of the strings like ^[[1;2D or something.
These aren't supported.

>Any ideas?

$ bind '^X3~=eot-or-delete'

This maps the ^D function to the Del key.

HTH, HAND
//mirabilos
--

-- 
“It is inappropriate to require that a time represented as
 seconds since the Epoch precisely represent the number of
 seconds between the referenced time and the Epoch.”
	-- IEEE Std 1003.1b-1993 (POSIX) Section B.2.2.2

Luc GMail | 3 Jul 2009 19:44

Picon

Re: Emulate Ctrl-Delete in mksh

On 7/3/09, Thorsten Glaser wrote:

> Your terminal may have one of the strings like ^[[1;2D or something.
> These aren't supported.

Yes, Ctrl-v tells me that Delete equals '^[[3~' and Ctrl-Delete equals
'^[[3;5~'.

> $ bind '^X3~=eot-or-delete'
> This maps the ^D function to the Del key.

Yeah, I had the same result with bind -m '^[[3~'=^d. But that's not
what I wanted, I wanted Ctrl-Delete.  

Well, thanks for your attention anyway.

Best regards,

--

-- 
Luc

Thorsten Glaser | 3 Jul 2009 20:01

Picon

Re: Emulate Ctrl-Delete in mksh

Luc GMail dixit:

>Yes, Ctrl-v tells me that Delete equals '^[[3~' and Ctrl-Delete equals
>'^[[3;5~'.

I use cat(1) for that... anyway, this is a long string. mksh currently
(it _is_ on the rewrite-in-the-future list, but don't hold your breath
for it) uses a simple table of prefix (^X ^[ and ^[[), character and a
tilde (optional).

>Yeah, I had the same result with bind -m '^[[3~'=^d. But that's not
>what I wanted, I wanted Ctrl-Delete.  

There's an alternative, from a Launchpad bug about Ctrl-CurLeft and
Ctrl-CurRight not working: it was suggested to tell the *terminal
emulator* to map these keys to Esc+b or Esc+f, respectively. So if
you can change your terminal emulator (probably via xresources) to
send ^D for Ctrl-Delete, you're set.

bye,
//mirabilos
--

-- 
“It is inappropriate to require that a time represented as
 seconds since the Epoch precisely represent the number of
 seconds between the referenced time and the Epoch.”
	-- IEEE Std 1003.1b-1993 (POSIX) Section B.2.2.2

Thorsten Glaser | 4 Jul 2009 22:21

Picon

Re: CVS: herc.mirbsd.org: src

Dixi quod:

>Commit ID:	1004A4E4AD759B5482B

>For total 8-bit transparency, there MUST NOT be any multibyte sequence
>matching (in CESU-8 encoding) U+EF80..U+EFFF in the multibyte stream
>converted as if it were CESU-8. Instead, treat these sequences, ranging
>from EE.BE.80..BF to EE.BF.80..BF, as raw octets and convert them to 3
>(three) instead of 1 (one) wide character (OPTU-16). To make it simple,
>there is no Unicode/PUA character whose OPTU-8 encoding begins with EEh
>followed by BEh or BFh.

This is not needed for mksh, because mksh does not use OPTU-16 inter-
nally. It does support using the woctet notation, but only for single
character conversions in base-1 digits.

I'd kind of like to rewrite mksh to use OPTU-16 internally though, as
the "cheap shot at WTF-8" I took at first helps a lot, but only real-
ly in the short term. (And make the parser recursive while there. And
fix all bugs in the command line editing code. Hell yeah.)

//mirabilos
--

-- 
“mysql is about as much database as ms access” – “MSSQL at least descends
from a database” “it's a rebranded SyBase” “MySQL however was born from a
flatfile and went downhill from there” – “at least jetDB doesn’t claim to
be a database”	-- myself, Tonnerre and psychoschlumpf in #nosec

Thorsten Glaser | 6 Jul 2009 17:10

Picon

pdksh/mksh 'stop () {' handling (was Re: CVS: herc.mirbsd.org: src)

Dixi quod…

>Commit ID:	1004A5212AD490123A5

>be more robust against things like Debian #535970
>reverts and rewrites the code from cid 10047C1EBA57E4F4AF0
>
>XXX find out if this is done right

Hi Robert,

I originally used the last_terminal_was_bracket patch from you in
mksh as well, after having encountered Debian initscripts doing
│stop() {
but now there are some which do
│stop () {
so I tried a different approach. As the author of the original
pdksh Debian-specific patch, do you think this is safe?

Index: src/bin/mksh/lex.c
diff -up src/bin/mksh/lex.c:1.88 src/bin/mksh/lex.c:1.89
--- src/bin/mksh/lex.c:1.88	Thu Jun 11 12:42:19 2009
+++ src/bin/mksh/lex.c	Mon Jul  6 15:06:23 2009
 <at>  <at>  -159,7 +159,6  <at>  <at>  yylex(int cf)
 	char *wp;		/* output word pointer */
 	char *sp, *dp;
 	int c2;
-	bool last_terminal_was_bracket;

  Again:

(Continue reading)

Thorsten Glaser | 7 Jul 2009 21:09

Picon

Re: CVS: herc.mirbsd.org: www

Dixi quod...

>Commit ID:	1004A53955F75A543E8

forgot: agreed bsiegert <at> 

//mirabilos
--

-- 
“It is inappropriate to require that a time represented as
 seconds since the Epoch precisely represent the number of
 seconds between the referenced time and the Epoch.”
	-- IEEE Std 1003.1b-1993 (POSIX) Section B.2.2.2

Thorsten Glaser | 7 Jul 2009 21:19

Picon

Re: CVS: herc.mirbsd.org: X11

Dixi quod...

>Commit ID:	1004A4F3D3510A8D45A

>maybe this will fix the EeePC’s [...]

Still don't get any output on it with the i810 driver. I think
it's an EeePC 701 or something, the colleagues say one of the
first models.

//mirabilos
--

-- 
“It is inappropriate to require that a time represented as
 seconds since the Epoch precisely represent the number of
 seconds between the referenced time and the Epoch.”
	-- IEEE Std 1003.1b-1993 (POSIX) Section B.2.2.2

Thorsten Glaser | 7 Jul 2009 22:14

Picon

Re: "$ <at> " vs. nounset

Hi Chet,

I found this via a Debian bug; you wrote:
|The interpretation process went back and forth, and we finally decided
|that set -u will apply to references to every variable, parameter, and
|special parameter *except* $ <at>  and $*.  That's going to be in the next
|revision of the standard.

I would like to get a confirmation that the next POSIX shell standard
will mandate this, because I maintain http://mirbsd.de/mksh which aims
to be at least pretty POSIX compatible (while not losing Korn Shell hi-
story though).

Hi David,

what will AT&T ksh93 do in this regard? Follow POSIX?

Thanks in advance to both of you for an answer.

Sincerely,
//mirabilos
--

-- 
I believe no one can invent an algorithm. One just happens to hit upon it
when God enlightens him. Or only God invents algorithms, we merely copy them.
If you don't believe in God, just consider God as Nature if you won't deny
existence.		-- Coywolf Qi Hunt

Thorsten Glaser | 8 Jul 2009 19:35

Picon

Re: "$ <at> " vs. nounset

Chet Ramey dixit:

>This was the working group consensus.  It will take time to get into
>the official standard:

But it will eventually make it? Okay.

>"Actually, $? and $! are the only special parameters other than $*
>and $ <at>  that can be used when unset if -u is on

Ah. I wonder what all the other special parameters do:

* $0, $1..$9, ${10}... (probably not special)
vs.
* $# (probably always set => uncritial)
* $$ (same)
* $- (in ksh: set -<letter> current value, same I think)

The next major release of mksh will thusly change the behaviour to
match future POSIX, unless there are major concerns, for example from
the AT&T ksh side.

Maybe it would have been better to agree to disagree (i.e. make the
behaviour wrt. $* and $ <at>  vendor-defined)?

bye,
//mirabilos
--

-- 
Sometimes they [people] care too much: pretty printers [and syntax highligh-
ting, d.A.] mechanically produce pretty output that accentuates irrelevant

(Continue reading)

Group

gmane.os.miros.general.

Project Web Page

MirOS Project - Discussion list

Search Archive

Page

1 | 2 | 3 | 4

Articles in period: 37

July 2009

Language

Change language

Options

Current view: Threads only / Showing only 20 lines / Not hiding cited text.
Change to All messages, whole messages, or hide cited text.

Post a message
NNTP Newsgroup
Classic Gmane web interface
XML

XML

RSS Feed
List Information

About Gmane

Recent entries

Warning: could not send message for past 4 hours (24 May 2013)
Warning: could not send message for past 4 hours (22 May 2013)
Warning: could not send message for past 4 hours (22 May 2013)
[ANN] The MirOS Project type foundry presents: FixedMisc [MirOS] 20130 (19 May 2013)
Testmessage (11 May 2013)
Funding and Feature Request Management For Open Source (12 Apr 2013)
pkgsrc-2013Q1 MirBSD 10uB0/i386 2013-03-29 20:11 (9 Apr 2013)
pkgsrc (24 Mar 2013)
Die Quittung #879677 ist bezahlt (6 Mar 2013)
[FYI] Name dispute filed (4 Mar 2013)
kernel panic by long Unix domain socket name. (27 Feb 2013)
mksh R42 results on Cygwin (15 Feb 2013)
Notification of new releases (15 Feb 2013)
mksh does not compile with gcc47 under FreeBSD (1 Feb 2013)
EOVERFLOW and ERANGE (31 Dec 2012)
sshd in the installer? (28 Dec 2012)
[code review] support for _SC_GET{GR,PW}_MAX (28 Dec 2012)
Adding ECANCELED in errno.h? (17 Dec 2012)
Fwd: Call for Presentations: BSD developer's room at FOSDEM 2013 (19 Nov 2012)
bug in ports/mailnews/pine when forwarding multipart MIME mails (13 Nov 2012)
mksh patch (16 Oct 2012)

Archives

9	May 2013
30	April 2013
27	March 2013
17	February 2013
10	January 2013
35	December 2012
45	November 2012
32	October 2012
24	September 2012
31	August 2012
29	July 2012
11	June 2012
13	May 2012
13	April 2012
27	March 2012
26	February 2012
39	January 2012
33	December 2011
40	November 2011
8	October 2011
25	August 2011
7	July 2011
33	June 2011
21	May 2011
24	April 2011
8	March 2011
22	February 2011
22	January 2011
24	December 2010
15	November 2010
40	October 2010
25	September 2010
14	August 2010
18	July 2010
22	June 2010
11	May 2010
3	April 2010
10	March 2010
11	February 2010
33	January 2010
31	December 2009
40	November 2009
18	October 2009
25	September 2009
38	August 2009
37	July 2009
31	June 2009
29	May 2009
17	April 2009
12	March 2009
34	February 2009
18	January 2009
30	December 2008
46	November 2008
37	October 2008
34	September 2008
38	August 2008
50	July 2008
33	June 2008
32	May 2008
53	April 2008
68	March 2008
18	February 2008
1	January 2008
24	October 2007
20	September 2007
43	August 2007
31	July 2007
48	June 2007
76	May 2007
19	April 2007
14	March 2007
38	February 2007
39	January 2007
39	December 2006
48	November 2006
34	October 2006
87	September 2006
98	August 2006
44	July 2006
32	June 2006
37	May 2006
35	April 2006
33	March 2006
49	February 2006
65	January 2006
92	December 2005
118	November 2005
40	October 2005
64	September 2005
35	August 2005
78	July 2005
42	June 2005
68	May 2005
55	April 2005
85	March 2005
69	February 2005
75	January 2005
76	December 2004
45	November 2004
66	October 2004
59	September 2004
110	August 2004
81	July 2004
10	June 2004
23	May 2004

Design

Zawodny | Right Menu | Left Menu

Your Own Design

Paste the URL of your CSS below. Download CSS template