Re: bin/56134: patch: add feedback to tee(1)

["James K. Lowden" <jklowden%schemamania.org@localhost>]

The following reply was made to PR bin/56134; it has been noted by GNATS.

From: "James K. Lowden" <jklowden%schemamania.org@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: bin/56134: patch: add feedback to tee(1)
Date: Mon, 3 May 2021 17:35:05 -0400

 On Thu, 29 Apr 2021 22:15:02 +0000 (UTC)
 "Jeremy C. Reed" <reed%reedmedia.net@localhost> wrote:
 
 > From: "Jeremy C. Reed" <reed%reedmedia.net@localhost>
 > To: gnats-bugs%netbsd.org@localhost
 > Cc: 
 > Subject: Re: bin/56134: patch: add feedback to tee(1)
 > Date: Thu, 29 Apr 2021 17:12:38 -0500 (CDT)
 > 
 >  > >How-To-Repeat:
 >  > 	$ find /usr/bin | ./tee -v | wc 
 >  > 	682617.19 KB/s (7,689 bytes in 0.000011 seconds)
 >  > 	     486     486    7689
 >  
 >  Not a perfect alternative, but also see
 >  
 >  t1:reed$ find /usr/bin | progress -e cat | wc    
 >    7887        4.81 MiB/s 
 >       498     498    7887
 
 Not an exact replacement, but your suggestion is probably better.  
 
 I implemented your idea as /usr/local/bin/meter: 
 
 	#! /bin/sh 
 	exec /usr/bin/progress -e cat
 
 How-To-Repeat: 
 
 $ find /usr/bin | meter | wc
   7689        4.98 MiB/s 
      486     486    7689
 
 That's pretty hard to beat, actually.  
 
 >  (Or have you seen a "tee" like this?)
 
 I have not.  
 
 Modifying progress(1) looks to be almost as much work as starting from
 scratch, unless cat(1) were made the default argument.  That would be
 only imperceptibly more efficient than the 2-liner above. 
 
 If there's interest in wrapping progress around cat -- one way or the
 other -- I could do that.  Otherwise, thanks for suggesting a solution
 I didn't think of.  
 
 BTW, the above script would be much nicer if this worked: 
 
 	#! /usr/bin/progress -e cat
 
 Is there interest in modifying the kernel to reliably pass, say, 9
 arguments, instead of just 1 on the interpreter line? Or are we dead
 set against?  Asking for a friend.  
 
 --jkl