[9front] [PATCH] Kill, kill, slay: banish sed, long live awk!

[9front] [PATCH] Kill, kill, slay: banish sed, long live awk!

[Amavect]

[-- Attachment #1: Type: text/plain, Size: 162 bytes --]

I'd like to claim 9front's best commit message for the most mundane patch.
Fixes regexp being parsed in kill. A bit more readable.
See attached.

Thanks,
Amavect

[-- Attachment #2.1: Type: text/plain, Size: 379 bytes --]

from postmaster@9front:
The following attachment had content that we can't
prove to be harmless.  To avoid possible automatic
execution, we changed the content headers.
The original header was:

	Content-Type: text/x-patch; charset="US-ASCII"; name="kill.patch"
	Content-Disposition: attachment; filename="kill.patch"
	Content-Transfer-Encoding: base64
	Content-ID: <f_ljkrol6l0>

[-- Attachment #2.2: kill.patch.suspect --]
[-- Type: application/octet-stream, Size: 1148 bytes --]

From: Amavect <amavect@gmail.com>
Date: Sun, 02 Jul 2023 01:36:23 +0000
Subject: [PATCH] Kill, kill, slay: banish sed, long live awk!


Regexp shall parse no more.
---
diff 47d3af71a33e242099a6a6130dcbe7ab4de9a895 f00e4a39c5d1ae11f2aefbfe23e70eae2d67e426
--- a/rc/bin/Kill
+++ b/rc/bin/Kill
@@ -1,5 +1,5 @@
 #!/bin/rc
 rfork e
 for(i){
-	ps | sed -n '/ '^$i^'$/s%^[^ ]* *([^ ]*).*%chmod 666 /proc/\1/ctl;@{echo kill>/proc/\1/ctl}%p'
+	ps | awk -v 'prog='$i '$7 == prog {print "chmod 666 /proc/" $2 "/ctl;@{echo kill>/proc/" $2 "/ctl} # " $7}'
 }
--- a/rc/bin/kill
+++ b/rc/bin/kill
@@ -1,7 +1,5 @@
 #!/bin/rc
 rfork e
 for(i){
-	ps | sed -n '/^'$user' .* '$i'$/s%[^ ]*  *%~>/proc/%
-	s%  *.* (.*)%/note} # \1%
-	s%~%@{echo kill%p'
+	ps | awk -v 'user='$user -v 'prog='$i '$1 == user && $7 == prog {print "@{echo kill>/proc/" $2 "/note} # " $7}'
 }
--- a/rc/bin/slay
+++ b/rc/bin/slay
@@ -1,7 +1,5 @@
 #!/bin/rc
 rfork e
 for(i){
-	ps | sed -n '/^'$user' .* '$i'$/s%[^ ]*  *%~>/proc/%
-	s%  *.* (.*)%/ctl} # \1%
-	s%~%@{echo kill%p'
+	ps | awk -v 'user='$user -v 'prog='$i '$1 == user && $7 == prog {print "@{echo kill>/proc/" $2 "/ctl} # " $7}'
 }

Re: [9front] [PATCH] Kill, kill, slay: banish sed, long live awk!

[sirjofri]

Hello,

out of curiosity, why s/sed/awk? What's wrong about sed?

sirjofri

Re: [9front] [PATCH] Kill, kill, slay: banish sed, long live awk!

[Amavect]

On Sun, Jul 2, 2023 at 2:46 AM sirjofri <sirjofri+ml-9front@sirjofri.de> wrote:
>
> Hello,
>
> out of curiosity, why s/sed/awk? What's wrong about sed?
>
> sirjofri

The command line arguments are passed to sed through a sed expression,
so the program name is parsed as a sed expression and not as a string.
This affects program names with a dot (6.out).
In practice, this almost never results in name conflicts,
so the real reason is that I find the sed version hard to read.

For fun, I rewrote the sed version, which makes the command generation obvious:
ps | sed -n '/^'$user' .* '$i'$/s%[^ ]* *([0-9]+) .* (.*)$%@{echo
kill>/proc/\1/note} # \2%p'
Compare to the original:
ps | sed -n '/^'$user' .* '$i'$/s%[^ ]*  *%~>/proc/%s%  *.*
(.*)%/note} # \1%s%~%@{echo kill%p'
Compare to awk:
ps | awk -v 'user='$user -v 'prog='$i '$1 == user && $7 == prog {print
"@{echo kill>/proc/" $2 "/note} # " $7}'

Thanks,
Amavect

Re: [9front] [PATCH] Kill, kill, slay: banish sed, long live awk!

[Stuart Morrow]

Can awk even build on (actual)386? (no floating point)

Re: [9front] [PATCH] Kill, kill, slay: banish sed, long live awk!

[Steve Simon]

i think you are correct.

i have a distant memory that awk was one of the few bits of plan9 that did not work on a 386 without a 387.

-Steve

> On 2 Jul 2023, at 8:39 pm, Stuart Morrow <morrow.stuart@gmail.com> wrote:
> 
> Can awk even build on (actual)386? (no floating point)

Re: [9front] [PATCH] Kill, kill, slay: banish sed, long live awk!

[Stuart Morrow]

9front has awk in bootfs so being cautious of awk in utilities is pretty moot.

Re: [9front] [PATCH] Kill, kill, slay: banish sed, long live awk!

[mkf9]

Steve Simon wrote:
> i have a distant memory that awk was one of the few bits of plan9 that did not work on a 386 without a 387.

did it ever built/booted in literal 386?
wow.

Re: [9front] Kill, kill, slay: banish sed, long live awk: Now historic plan9

[Steve Simon]

i certainly ran plan9 on 386s in the beginning, i cannot be sure if you could rebuild the system on a 386 but i think so.

the 386 linker used its own allocator (maybe still does) which is fast but greedy, i remember tweaking it to used malloc so i could link a kernel on a small RAM system, though i have forgotten how small.

the biggest pain in those days was finding a supported VGA card.

-Steve


> On 1 Aug 2023, at 6:38 am, mkf9 <mkf9@riseup.net> wrote:
> 
> Steve Simon wrote:
>> i have a distant memory that awk was one of the few bits of plan9 that did not work on a 386 without a 387.
> 
> did it ever built/booted in literal 386?
> wow.

Re: [9front] Kill, kill, slay: banish sed, long live awk: Now historic plan9

[Roberto E. Vargas Caballero]

Hi,

On Tue, Aug 01, 2023 at 11:10:18PM +0100, Steve Simon wrote:
> 
> i certainly ran plan9 on 386s in the beginning, i cannot be sure if you could rebuild the system on a 386 but i think so.

I run 9front in a soekris that is a 386. I rebuilt the system a few times and I can
ensure that it works.

Regards,