Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.misc > #14799 > unrolled thread

Critique My Script-Fu?

Back to article view | Back to comp.os.linux.misc

Contents

  Critique My Script-Fu? Tom Hardy <rhardy702@gmail.com> - 2015-05-18 16:14 -0500
    Re: Critique My Script-Fu? "Anonymous Remailer (austria)" <mixmaster@remailer.privacy.at> - 2015-05-19 13:45 +0200
      Re: Critique My Script-Fu? Tom Hardy <rhardy702@gmail.com> - 2015-05-19 19:53 -0500
        Re: Critique My Script-Fu? "Anonymous Remailer (austria)" <mixmaster@remailer.privacy.at> - 2015-05-20 15:30 +0200
        Re: Critique My Script-Fu? Dan Espen <despen@verizon.net> - 2015-05-20 09:59 -0400
          Re: Critique My Script-Fu? Tom Hardy <rhardy702@gmail.com> - 2015-05-21 11:36 -0500
            Re: Critique My Script-Fu? Dan Espen <despen@verizon.net> - 2015-05-21 13:31 -0400

#14799 — Critique My Script-Fu?

A couple of years ago, I described the screen visibility problems of 
driving around, day and night, with a laptop mounted in my car, and 
described my /usr/local/bin/.invert-display.sh which I can trigger 
with a panel button.  It uses xcalib to perform operations on the 
existing LUT entries and is independent of desktop environments.

I got comments, but no critique of the script, so it couldn't have 
been too bad.  

Well, I recently used perl to add graphing capability in a decidedly 
substandard manner.  I really tried to use awk, but couldn't swing 
it, so this is my introduction to perl.  The script is reproduced at 
the bottom.  It also requires the xcalib and plotutils packages.

xcalib -printramps -alter produces output like the following:

...
<approx 760 lines like the following>
...
Warning - red gamma table not monotonic
Warning - green gamma table not monotonic
Warning - blue gamma table not monotonic
Warning - red gamma table not monotonic
Warning - green gamma table not monotonic
Warning - blue gamma table not monotonic
ffff ffff ffff
ff4b ff4b ff4b
fe97 fe97 fe97
fde2 fde2 fde2
fd2d fd2d fd2d
fc79 fc79 fc79
...
<250 more lines, the last one being 0 0 0>
...

The warnings are produced when the display is inverted, as shown 
above, but xcalib otherwise performs normally.

I first thought to use awk to convert hex to the decimal needed by 
plotutils' 'graph'. It has a '--non-decimal-data' option that seemed 
promising, but I couldn't figure it out.

So, perl.  I am really overloading the hex function, first because 
it apparently evaluates strings like 'Warning' to 0, which I've 
excluded from the output, so the warnings are gone.  Second, it 
apparently only reads the first record off a string, which is 
sufficient for me.  And third, I am excluding 0 from output.  There 
should actually be one, either at the beginning or the end.

There are likely other blunders or issues worthy of comment.  For 
instance, couldn't 'test' be replaced by a shell built-in?  Could 
the files '~/.altered-display' and '~/inverted-display' be replaced 
by a single file that's never deleted but simply moved to a new 
name, or have configuration contents that change?  Would there be 
any advantage in doing so?

Can I use awk in place of perl?  Should I do the whole thing in 
perl?  What about formatting or overall program structure?

Comments welcome.

#!/bin/bash

if test -f ~/.inverted-display; then

   if test -f ~/.altered-display; then

      xcalib -clear -alter;
      rm -f ~/.altered-display ~/.inverted-display;

   else

      xcalib -co 60 -alter;  #contrast
      touch ~/.altered-display;

   fi;

else

   if test -f ~/.altered-display; then

      xcalib -clear -alter;
      xcalib -gc 0.7 -invert -alter;  #gamma correction
      mv ~/.altered-display ~/.inverted-display;

   else

      xcalib -gc 1.5 -alter;
      touch ~/.altered-display;

   fi;

fi;

if test Z$1 = 'Zgraph'; then

xcalib -p -a | perl -e 'while(<>) {my $a = hex $_; print $a, "\n" \
if $a ne 0;}' | graph -a -T X -x 0 254 -y 0 65535 -m-1 -S 1;
xcalib -l -a;

fi;

-- 
Tom Hardy <rhardy702@gmail.com>

[toc] | [next] | [standalone]

#14800

Tom Hardy <rhardy...@gmail.com> [TH]:
TH> Can I use awk in place of perl?

You sure can. The following might be a good starting point:

awk '
/^[0-9a-f \r]+$/{                       # select interesting lines
  print "*** Input line:" $0;           # debugging output
  n=split($0,a," ");
  for (i=1; i <= n; i++){               # for all fields
    /* get their decimal value and use it as needed */     
    printf strtonum("0x" a[i]) (i != n ? " " : "\n");
  };
}
'

[toc] | [prev] | [next] | [standalone]

#14812

Anonymous Remailer (austria) wrote:

> 
> Tom Hardy <rhardy...@gmail.com> [TH]:
> TH> Can I use awk in place of perl?
> 
> You sure can. The following might be a good starting point:

<code>

Hm, well, one advantage of the perl code is that it is short.

I will be studying this.  Awk, regex, shell programming are all 
areas where I could use a lot of practice.  Your code gives me 
something to chew on.

-- 
Tom Hardy <rhardy702@gmail.com>

[toc] | [prev] | [next] | [standalone]

#14814

Tom Hardy <rhardy...@gmail.com> [TH]:
TH> one advantage of the perl code is that it is short.

It is (at the cost of increased resources used by the perl
executable, which might be totally acceptable in your situation).

My sample code sacrifices some brevity for the sake of readability.

[toc] | [prev] | [next] | [standalone]

#14816

Tom Hardy <rhardy702@gmail.com> writes:

> Anonymous Remailer (austria) wrote:
>
>> 
>> Tom Hardy <rhardy...@gmail.com> [TH]:
>> TH> Can I use awk in place of perl?
>> 
>> You sure can. The following might be a good starting point:
>
> <code>
>
> Hm, well, one advantage of the perl code is that it is short.

Another advantage is that you can do the whole thing in Perl.
No need to mix Perl and sh.

> I will be studying this.  Awk, regex, shell programming are all 
> areas where I could use a lot of practice.  Your code gives me 
> something to chew on.

Years ago, I'd start a project in sh, and often get to point where
I'd drug in sed, awk, cut, head.  IMO, a complete mess.
Just start out in Perl.

-- 
Dan Espen

[toc] | [prev] | [next] | [standalone]

#14826

Dan Espen wrote:

[...]

> Just start out in Perl.

Trying to think what that would mean, given the simplicity of the 
script; a few tests, touch, mv, xcalib, etc. in the first part, and 
the second part a single line with perl sandwiched in the middle of 
a pipe.  (It could be a separate script as it's not necessary to the 
everyday use of .invert-display.sh.

Ultimately, perl whold be your shell.  Is that in any way feasible?

-- 
Tom Hardy <rhardy702@gmail.com>

[toc] | [prev] | [next] | [standalone]

#14827

Tom Hardy <rhardy702@gmail.com> writes:

> Dan Espen wrote:
>
> [...]
>
>> Just start out in Perl.
>
> Trying to think what that would mean, given the simplicity of the 
> script; a few tests, touch, mv, xcalib, etc. in the first part, and 
> the second part a single line with perl sandwiched in the middle of 
> a pipe.  (It could be a separate script as it's not necessary to the 
> everyday use of .invert-display.sh.

That means, when you think you have a simple shell script to write,
give some thought to what that shell might grow into.

I found a number of times that what started out simple grew steadily
more complex.  In your case, part of your shell turned into perl.
Rather than having mixtures of sh, perl, awk, and a bunch of other
stuff, perl can do the whole job.

Surprisingly, I didn't find any Perl modules for xcalib, but
when there is a perl module, (like for web access or html parsing),
perl becomes even more of a win.

> Ultimately, perl whold be your shell.  Is that in any way feasible?

No.

-- 
Dan Espen

[toc] | [prev] | [standalone]

Back to top | Article view | comp.os.linux.misc

csiph-web