Debian Bug report logs - #376760
w3m: don't ignore HOME and END on standard keyboards

Reply or subscribe to this bug.

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Dan Jacobson <jidanni@jidanni.org>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: w3m: don't ignore HOME and END on standard keyboards

Date: Wed, 05 Jul 2006 02:45:39 +0800

Package: w3m
Version: 0.5.1-4
Severity: wishlist

Lynx knows about the HOME and END keys on standard keyboards, but not w3m.

Message #10 received at 376760@bugs.debian.org (full text, mbox, reply):

From: Fumitoshi UKAI <ukai@debian.or.jp>

To: Dan Jacobson <jidanni@jidanni.org>

Cc: 376760@bugs.debian.org

Subject: Re: Bug#376760: w3m: don't ignore HOME and END on standard keyboards

Date: Thu, 06 Jul 2006 15:02:53 +0900

At Wed, 05 Jul 2006 02:45:39 +0800,
Dan Jacobson wrote:

> Lynx knows about the HOME and END keys on standard keyboards, but not w3m.

It works for me.
What terminal are you using?

Regards,
Fumitoshi UKAI

Message #15 received at 376760@bugs.debian.org (full text, mbox, reply):

From: Dan Jacobson <jidanni@jidanni.org>

To: 376760@bugs.debian.org

Subject: w3m vs. rxvt HOME and END keys

Date: Tue, 18 Jul 2006 23:52:53 +0800

In rxvt HOME and END do nothing in w3m. I xterm they work fine.
$ infocmp -d rxvt xterm|...
	kEND: '\E[8$', '\E[1;2F'.
	kHOM: '\E[7$', '\E[1;2H'.

Message #20 received at 376760@bugs.debian.org (full text, mbox, reply):

From: Thomas Dickey <dickey@radix.net>

To: Dan Jacobson <jidanni@jidanni.org>, 376760@bugs.debian.org

Subject: Re: Bug#376760: w3m vs. rxvt HOME and END keys

Date: Wed, 19 Jul 2006 06:51:44 -0400

[Message part 1 (text/plain, inline)]

On Wed, Jul 19, 2006 at 02:10:07AM +0200, Dan Jacobson wrote:
> In rxvt HOME and END do nothing in w3m. I xterm they work fine.
> $ infocmp -d rxvt xterm|...
> 	kEND: '\E[8$', '\E[1;2F'.
> 	kHOM: '\E[7$', '\E[1;2H'.

Those are the shifted home/end keys that you're citing.
Names of shifted function-keys are mixed case.

w3m may/may not check for those.

Note also that w3m is a termcap application, so it would be looking for
the corresponding termcap names as shown in the terminfo(5) manpage.

Very few termcap applications look for the shifted keys.
If you're not actually using the shifted keys, perhaps the problem lies
in a different area.

-- 
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net

[Message part 2 (application/pgp-signature, inline)]

Message #25 received at 376760@bugs.debian.org (full text, mbox, reply):

From: Dan Jacobson <jidanni@jidanni.org>

To: Thomas Dickey <dickey@radix.net>

Cc: 376760@bugs.debian.org

Subject: Re: Bug#376760: w3m vs. rxvt HOME and END keys

Date: Mon, 31 Jul 2006 01:19:58 +0800

On my Thinkpad, in xterm, pressing home and end give:
$ cat -v
^[[H
^[[F
$ infocmp |sed 2!d
xterm|X11 terminal emulator,
$ infocmp -1|grep '\\E\[[FH]'
        clear=\E[H\E[2J,
        home=\E[H,
$ infocmp -1 -C|grep '\\E\[[FH]'
        :cl=\E[H\E[2J:\
        :ho=\E[H:\

Well for some reason lynx and w3m know what end sends even though it
isn't listed above. Home OK too.

In rxvt,
$ cat -v
^[[7~
^[[8~
$ infocmp|sed 2!d
rxvt-unicode|rxvt-unicode terminal (X Window System),
$ infocmp -1 |grep '\\E\[[78]~'
	kend=\E[8~,
	khome=\E[7~,
$ infocmp -1 -C|grep '\\E\[[78]~'
warning: rxvt-unicode entry is 1134 bytes long
	:kh=\E[7~:\

Here we have different names, and termcap lesser entries, and all I
know is lynx still knows what to do but w3m doesn't respond to home
and end.

The Thinkpad r50e does not have a keypad, and I didn't hit shift, not
tinker in ~/.Xresources.

On my desktop 2.6.16-1-k7, all the above is the same. And there is an
additional numerical keypad at the right, where there are additional
home and end keys, and they send:
^[Ow ^[Oq in rxvt (lynx uses them, w3m doesn't), and again ^[[H ^[[F in xterm.

In /etc/X11/xorg.conf on both machines I do
	Option		"XkbRules"	"xorg"
	Option		"XkbModel"	"pc104"
	Option		"XkbLayout"	"us"
but apparently that is not the problem.

Message #30 received at 376760@bugs.debian.org (full text, mbox, reply):

From: Thomas Dickey <dickey@radix.net>

To: Dan Jacobson <jidanni@jidanni.org>

Cc: 376760@bugs.debian.org

Subject: Re: Bug#376760: w3m vs. rxvt HOME and END keys

Date: Sun, 30 Jul 2006 19:25:47 -0400

[Message part 1 (text/plain, inline)]

On Mon, Jul 31, 2006 at 01:19:58AM +0800, Dan Jacobson wrote:
> On my Thinkpad, in xterm, pressing home and end give:
> $ cat -v
> ^[[H
> ^[[F
> $ infocmp |sed 2!d
> xterm|X11 terminal emulator,
> $ infocmp -1|grep '\\E\[[FH]'
>         clear=\E[H\E[2J,
>         home=\E[H,
> $ infocmp -1 -C|grep '\\E\[[FH]'
>         :cl=\E[H\E[2J:\
>         :ho=\E[H:\

The terminfo's written to assume that the application sends smkx (which
puts the keypads into application mode).  That's standard (sending smkx),
though some terminfo's were designed to make the bash users happy (mainly
the "linux" description).  You can see the effect of changing the modes
easily in xterm using the "VT Options" menu.
 
> Well for some reason lynx and w3m know what end sends even though it
> isn't listed above. Home OK too.
> 
> In rxvt,
> $ cat -v
> ^[[7~
> ^[[8~
> $ infocmp|sed 2!d
> rxvt-unicode|rxvt-unicode terminal (X Window System),
> $ infocmp -1 |grep '\\E\[[78]~'
> 	kend=\E[8~,
> 	khome=\E[7~,
> $ infocmp -1 -C|grep '\\E\[[78]~'
> warning: rxvt-unicode entry is 1134 bytes long
> 	:kh=\E[7~:\
> 
> Here we have different names, and termcap lesser entries, and all I
> know is lynx still knows what to do but w3m doesn't respond to home
> and end.

I seem to recall that rxvt's terminfo is (like "linux" ;-), written for
bash-users, and assumes the terminal is started in application mode.
Checking, I see smkx differs between the two:

        smkx: '\E[?1h\E=', '\E='.

xterm is on the left; you'll note that xterm uses two escape (\E) sequences
while rxvt uses one.  The \E[?1h sets the cursor-keys into application
mode.  I know offhand that (because of the way X defines the keycodes),
the home/end keys are classified as cursor keys - for xterm.  I'm not
sure about rxvt since it bends the rules differently.  Since rxvt isn't
setting that mode explicitly, and because its arrow keys are specified
in application mode, it's possible that some application reset it to
the normal mode.  (That might affect the home/end keys - I'm not sure
w/o testing it).

Oddly a quick check of rxvt's rs1/rs2 strings (used in "reset") seems to
show it resetting the terminal's cursor keys into normal mode.  But running
"reset" and then using cat -v still shows ^[[A (normal mode), which is
consistent.  If I use TERM=xterm, then it would change to ^[OA (application
mode).  But a quick check doesn't show home/end changing.  It's still
^[[7~ and ^[[8~ respectively.

> The Thinkpad r50e does not have a keypad, and I didn't hit shift, not
> tinker in ~/.Xresources.
> 
> On my desktop 2.6.16-1-k7, all the above is the same. And there is an
> additional numerical keypad at the right, where there are additional
> home and end keys, and they send:
> ^[Ow ^[Oq in rxvt (lynx uses them, w3m doesn't), and again ^[[H ^[[F in xterm.

lynx has some hardcoded fallbacks for vt320 escape sequences that might
be catching this case.  In the default keyboard configuration, xterm
sends the keys as labeled, e.g., home/end.
 
> In /etc/X11/xorg.conf on both machines I do
> 	Option		"XkbRules"	"xorg"
> 	Option		"XkbModel"	"pc104"
> 	Option		"XkbLayout"	"us"
> but apparently that is not the problem.

I don't think so either (I've done most of my testing with pc104/us).

-- 
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net

[Message part 2 (application/pgp-signature, inline)]

Send a report that this bug log contains spam.