[TUHS] shell PS1/PS2 which survive cut-paste

[TUHS] shell PS1/PS2 which survive cut-paste

[George Michaelson]

I tried a long time ago to set PS1 and PS2 to character sequences
which would permit cut-paste to work. I failed, but I didn't try very
hard.

The choice of "# " and "> " interests me. Because roots prompt of
"hash" has a side effect of making a cut-paste over it a comment in
most shells.

But.. it predates being *able* to cut-paste so it has to be a
side-effect, not a design choice.

"> " is more dangerous in cut-paste. Which again feels like a
side-effect both by anachronistic time effects, and intent: nobody
would be that suicidally mad to make cut-paste invoke shell
pre-forming the command sequence which neccessarily zeros out the >
targets before executing the cmd...

-G

[TUHS] shell PS1/PS2 which survive cut-paste

[Kurt H Maier]

On Wed, Dec 06, 2017 at 11:22:52AM +1000, George Michaelson wrote:
> I tried a long time ago to set PS1 and PS2 to character sequences
> which would permit cut-paste to work. I failed, but I didn't try very
> hard.

I've used :; and ::; to good effect. Since : is a no-op, it just inserts
noise into the history.

khm

[TUHS] shell PS1/PS2 which survive cut-paste

[Lyndon Nerenberg]

> On Dec 5, 2017, at 5:28 PM, Kurt H Maier <khm at sciops.net> wrote:
> 
> I've used :; and ::; to good effect. Since : is a no-op, it just inserts
> noise into the history.

This fragment lives in my ${HOME}/.env on pretty much every machine I log in to.  If you 'export ENV=${HOME}/.env' in .profile, then configure sudo/doas to pass ENV through unmolested, this will follow you as you switch uids.

 _hostname=$(hostname | sed -e 's/\.hushmail\(inc\)*\.com$//')
 awd=$(which awd 2>/dev/null)
 if [ "${awd}" == "" ] ; then
        awd=':'
 fi
 setprompt () { PS1=": $(id -un)@$_hostname:$(pwd); " ; }
 cd () { command cd "$@" && setprompt && ${awd} ; }
 setprompt

It works with pretty much every Bourne/Korn shell variant I've stumbled across.  Adjust the _hostname bit to match the domains of the systems you regularly log in to, and drop the awd bits if you don't use plan9port.

--lyndon

[TUHS] shell PS1/PS2 which survive cut-paste

[Steve Nickolas]

On Wed, 6 Dec 2017, George Michaelson wrote:

> I tried a long time ago to set PS1 and PS2 to character sequences
> which would permit cut-paste to work. I failed, but I didn't try very
> hard.
>
> The choice of "# " and "> " interests me. Because roots prompt of
> "hash" has a side effect of making a cut-paste over it a comment in
> most shells.
>
> But.. it predates being *able* to cut-paste so it has to be a
> side-effect, not a design choice.
>
> "> " is more dangerous in cut-paste. Which again feels like a
> side-effect both by anachronistic time effects, and intent: nobody
> would be that suicidally mad to make cut-paste invoke shell
> pre-forming the command sequence which neccessarily zeros out the >
> targets before executing the cmd...
>
> -G
>

I think this is why rc used a ";" ?

Maybe use "; " for PS1 and "  " for PS2?

-uso.

[TUHS] shell PS1/PS2 which survive cut-paste

[Grant Taylor]

On 12/05/2017 06:22 PM, George Michaelson wrote:
> I tried a long time ago to set PS1 and PS2 to character sequences 
> which would permit cut-paste to work. I failed, but I didn't try very 
> hard.

What do you mean by "permit cut (and) paste to work"?  Do you want to be 
able to cut (triple click the entire line in Xterm) and then paste it 
(middle click a new line) and have it execute?  Or do you mean that you 
want it to not execute as a safety measure?

Or are you wanting to go one step further and have something that 
toggles input interpretation off until a character sequence is provided, 
possibly by the prompt?  Ideally so that you would not accidentally copy 
& paste swaths of screen output and have it executed.



-- 
Grant. . . .
unix || die

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3982 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20171205/87e5c062/attachment.bin>

[TUHS] shell PS1/PS2 which survive cut-paste

[arnold]

George Michaelson <ggm at algebras.org> wrote:

> The choice of "# " and "> " interests me. Because roots prompt of
> "hash" has a side effect of making a cut-paste over it a comment in
> most shells.

"#" as the root prompt predates # as the comment in the Bourne shell,
not to mention predating copy/paste entirely. (My understanding is that
the do-nothing command, ":" was used for comments. Talk about minimalist!)

Same point for ">", since copy/paste didn't exist in the late 70s when
Bourne was doing the shell, it wasn't an issue.

Arnold

[TUHS] shell PS1/PS2 which survive cut-paste

[George Michaelson]

So.. arnold.. you never used the ! command in ex or vi? because it
will treat a range command as the bounding set of lines to feed into
the command.. if you specify sh or /path/to/shell then its
functionally a cut-paste-exec sequence isn't it?

emacs has the same functionality (in some syntactic suger)

I used fmt all the time in vi, to get <72 for mail in MH. which led me
to find other things you can pipe the editor lines through...

-G

[TUHS] shell PS1/PS2 which survive cut-paste

[Ian Zimmerman]

On 2017-12-06 16:01, George Michaelson wrote:

> I used fmt all the time in vi, to get <72 for mail in MH. which led me
> to find other things you can pipe the editor lines through...

Exactly!  Just put a trivial script in front of the /bin/sh in the pipe
to strip the undesirable stuff like prompts.  I do this for browsing
URLs in mail bodies which may be split across multiple lines or
otherwise mangled.

With emacs, you don't even need an extra process, because the demangling
can be programmed in lisp.

-- 
Please don't Cc: me privately on mailing lists and Usenet,
if you also post the followup to the list or newsgroup.
To reply privately _only_ on Usenet, fetch the TXT record for the domain.

[TUHS] shell PS1/PS2 which survive cut-paste

[Dave Horsfall]

On Tue, 5 Dec 2017, arnold at skeeve.com wrote:

> "#" as the root prompt predates # as the comment in the Bourne shell, 
> not to mention predating copy/paste entirely. (My understanding is that 
> the do-nothing command, ":" was used for comments. Talk about 
> minimalist!)

And as a label for the dreaded "goto" command; it was a while before I 
realised that ":" in and of itself did bugger-all...

-- 
Dave Horsfall DTM (VK2KFU)  "Those who don't understand security will suffer."

[TUHS] shell PS1/PS2 which survive cut-paste

[arnold]

Dave Horsfall <dave at horsfall.org> wrote:

> On Tue, 5 Dec 2017, arnold at skeeve.com wrote:
>
> > "#" as the root prompt predates # as the comment in the Bourne shell, 
> > not to mention predating copy/paste entirely. (My understanding is that 
> > the do-nothing command, ":" was used for comments. Talk about 
> > minimalist!)
>
> And as a label for the dreaded "goto" command; it was a while before I 
> realised that ":" in and of itself did bugger-all...
>
> -- 
> Dave Horsfall DTM (VK2KFU)  "Those who don't understand security will suffer."

"goto" was in the V6 shell, I think (next to no experience with the V6
shell, and it was a LONG time ago anyway).  But there never was a goto
in the V7 shell.

[TUHS] shell PS1/PS2 which survive cut-paste

[Dave Horsfall]

On Wed, 6 Dec 2017, arnold at skeeve.com wrote:

> "goto" was in the V6 shell, I think (next to no experience with the V6 
> shell, and it was a LONG time ago anyway).  But there never was a goto 
> in the V7 shell.

Never got to use the V7 shell; we were stuck with 11/40s[*] :-(  It was 
certainly in the V6 shell, hence by bafflement at why labels were of the 
form ": blah" instead of the sensible "blah: "...

[*]
Although I did shove as much of the V7 kernel into V6 as I could (such as 
XON/XOFF), thereby earning me the sobriquet "Mr Unix 6.5" or something; 
having the AUSAM buffer scheme helped a lot.

-- 
Dave Horsfall DTM (VK2KFU)  "Those who don't understand security will suffer."

[TUHS] shell PS1/PS2 which survive cut-paste

[arnold]

> On Wed, 6 Dec 2017, arnold at skeeve.com wrote:
>
> > "goto" was in the V6 shell, I think (next to no experience with the V6 
> > shell, and it was a LONG time ago anyway).  But there never was a goto 
> > in the V7 shell.

Dave Horsfall <dave at horsfall.org> wrote:
> Never got to use the V7 shell; we were stuck with 11/40s[*] :-(  It was 
> certainly in the V6 shell, hence by bafflement at why labels were of the 
> form ": blah" instead of the sensible "blah: "...

Come now, since then you've *never* used any kind of Bourne shell?
BSD on vax? SunOS? System III or System V on anything else? You've been
stuck in a time warp for 35 years?  :-)

Arnold

[TUHS] shell PS1/PS2 which survive cut-paste

[Dave Horsfall]

On Thu, 7 Dec 2017, arnold at skeeve.com wrote:

>> Never got to use the V7 shell; we were stuck with 11/40s[*] :-( It was 
>> certainly in the V6 shell, hence by bafflement at why labels were of 
>> the form ": blah" instead of the sensible "blah: "...
>
> Come now, since then you've *never* used any kind of Bourne shell? BSD 
> on vax? SunOS? System III or System V on anything else? You've been 
> stuck in a time warp for 35 years?  :-)

Sigh; there's one in every bunch :-)  OK, I never got to use the V7 shell 
on V7, because I never got to use V7 on my machines (and the use of "goto" 
had been thoroughly beaten out of me by then, so I never thought to try 
it).

-- 
Dave Horsfall DTM (VK2KFU)  "Those who don't understand security will suffer."

[TUHS] shell PS1/PS2 which survive cut-paste

[arnold]

Hi.

Dave Horsfall <dave at horsfall.org> wrote:

> On Thu, 7 Dec 2017, arnold at skeeve.com wrote:
>
> >> Never got to use the V7 shell; we were stuck with 11/40s[*] :-( It was 
> >> certainly in the V6 shell, hence by bafflement at why labels were of 
> >> the form ": blah" instead of the sensible "blah: "...
> >
> > Come now, since then you've *never* used any kind of Bourne shell? BSD 
> > on vax? SunOS? System III or System V on anything else? You've been 
> > stuck in a time warp for 35 years?  :-)
>
> Sigh; there's one in every bunch :-)

Sorry. I tend to be too literal-minded sometimes. That can help for
doing software development, but can get in the way at other times.

> OK, I never got to use the V7 shell on V7, because I never got to use
> V7 on my machines (and the use of "goto" had been thoroughly beaten out
> of me by then, so I never thought to try it).

I see. Much better. :-)

Thanks,

Arnold

[TUHS] shell PS1/PS2 which survive cut-paste

[Clem Cole]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 700 bytes --]

On Wed, Dec 6, 2017 at 6:57 PM, Dave Horsfall <dave at horsfall.org> wrote:

>
> Never got to use the V7 shell; we were stuck with 11/40s[*]


​Dave I don't understand this statement.  We ran Seventh Edition on 11/40
class processors more than anything else.  The issue was lack of separate
I/D space and the kernel got large so you ran out of physical memory.   The
later was solved with things like the Able ENABLE, which I admit I put one
 couple of 11/34 and our 11/60 (which are 11/40 class).  But V7 ran fine on
it.

Clem​
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20171207/c6d4cdf8/attachment.html>

[TUHS] shell PS1/PS2 which survive cut-paste

[Dave Horsfall]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1295 bytes --]

On Thu, 7 Dec 2017, Clem Cole wrote:

> > Never got to use the V7 shell; we were stuck with 11/40s[*] 
> 
> Dave I don't understand this statement.  We ran Seventh Edition on 11/40
> class processors more than anything else.  The issue was lack of separate
> I/D space and the kernel got large so you ran out of physical memory.   The
> later was solved with things like the Able ENABLE, which I admit I put one
> couple of 11/34 and our 11/60 (which are 11/40 class).  But V7 ran fine on
> it.

I tried, but no way it would work; the only non-DEC gear we ended up with 
was extra memory (nothing special, but of course not covered by Field 
Circus) and an Ampex[*] RP-03 clone (I think).  For all intents and 
porpoises it was a maxed-out 11/40 (then the /60 later on).

As I mentioned before, the AUSAM buffer scheme certainly helped, and we 
kissed goodbye to deadlocks; we even modified character devices (LP. XY, 
LV etc) to use the buffer pool instead of the character queues, and they 
went damned fast.

I should mention that the /40s were originally specced to run RSX-11D 
(yuck), and when the ACM article came out the rest was history...

[*]
My spell checker suggested "Tampax" for "Ampex"...

-- 
Dave Horsfall DTM (VK2KFU)  "Those who don't understand security will suffer."

[TUHS] shell PS1/PS2 which survive cut-paste

[Derek Fawcus]

On Wed, Dec 06, 2017 at 11:22:52AM +1000, George Michaelson wrote:
> I tried a long time ago to set PS1 and PS2 to character sequences
> which would permit cut-paste to work. I failed, but I didn't try very
> hard.

h=spuggy$; ssh copper
h=copper$;
h=copper$; pwd
/usr/home/derek
h=copper$; h=copper$; pwd
/usr/home/derek
h=copper$;


.pdf

[TUHS] shell PS1/PS2 which survive cut-paste

[Don Hopkins]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 627 bytes --]

I just put a newline at the end of my prompt, so the command has the whole next line to itself and there’s no prompt to get in the way of copying and pasting.

-Don


> On 16 Dec 2017, at 18:42, Derek Fawcus <dfawcus+lists-tuhs at employees.org> wrote:
> 
> On Wed, Dec 06, 2017 at 11:22:52AM +1000, George Michaelson wrote:
>> I tried a long time ago to set PS1 and PS2 to character sequences
>> which would permit cut-paste to work. I failed, but I didn't try very
>> hard.
> 
> h=spuggy$; ssh copper
> h=copper$;
> h=copper$; pwd
> /usr/home/derek
> h=copper$; h=copper$; pwd
> /usr/home/derek
> h=copper$;
> 
> 
> .pdf

[TUHS] shell PS1/PS2 which survive cut-paste

[Paul Ruizendaal]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 988 bytes --]


>> The choice of "# " and "> " interests me. Because roots prompt of
>> "hash" has a side effect of making a cut-paste over it a comment in
>> most shells.
> 
> "#" as the root prompt predates # as the comment in the Bourne shell,
> not to mention predating copy/paste entirely. (My understanding is that
> the do-nothing command, ":" was used for comments. Talk about minimalist!)
> 
> Same point for ">", since copy/paste didn't exist in the late 70s when
> Bourne was doing the shell, it wasn't an issue.

As early as V5 the (thompson) shell prompts were “#” and “%”, and “:” for
a label. As the goto command exists in V4 (there is a man page for it), I
would assume that those characters were used in V4 as well. So it would
seem to go back to 1974.

In the V7 (bourne) shell the default non-root prompt is “$”. Goto is
dropped at this point.

Don’t know when or where “>” was first used on Unix as a prompt character
(on my boxes it still is “$”).

Paul

[TUHS] shell PS1/PS2 which survive cut-paste

[arnold]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 287 bytes --]

Paul Ruizendaal <pnr at planet.nl> wrote:

> Don’t know when or where “>” was first used on Unix as a prompt character
> (on my boxes it still is “$”).

The V7 shell uses it as the secondary prompt, for when a construct
isn't complete:

	$ echo 'abc
	> def'
	abc
	def

Arnold