---
title: Eboard 04  The Bash shell - a refresher
number: 04
section: eboards
held: 2017-02-16
---
CSC 282.01, Class 04:  The Bash shell - a refresher
===================================================

_Overview_

* Preliminaries
    * Notes and news
    * Upcoming work
    * Questions
* Refresher: Some basic tasks
* What makes a good solution?
* Sample solutions 
* Bash

### News / Etc.

* Sorry for not getting homework out last week.   It was more chaotic than
  I thought it would be.  (It also sounded like some of you needed a break.)

### Upcoming Work

*Forthcoming*

### Good things to do

* Today's CS Extra on the 4-1 program
* Swim meet, all weekend
* Final home basketball games of the year
* Art house event Saturday at 1pm
* Concert Saturday at 2pm with "Sound painting"
* Slavic Coffee House Saturday at 5:30 pm in Bucksbaum Rotunda
* Tuesday's CS Table on Net Neutrality
* Today, 11am, pay Noteworthy to serenade people
* Monday, bone marrow drive

### Questions

Refresher: Some basic tasts

* Given a DOS-formatted text file (lines end with `\r\n` rather than just
  `\n`), convert it to a standard text file.
* Given a standard text file, convert all uppercase letters to lowercase.
* Given a standard text file, remove all blank spaces at the end of lines.
* Make a list of all misspelled words in a text file.
* Given a CSV file in which each line has the form  
  `LastName,FirstName,Assignment,NumericGrade`
  find the the five highest grades on homework 2.
* Given an HTML file, find the URLs of all images.  In case you don't
  know HTML, those will typically look like  
  `<img ... src="*URL*" ...>`
  + The `img` can have any capitalization (`img`,
    `IMG`, `Img`,  `iMg`, etc.)
  + There can be other text between the `img` and the
    `src`.  (That text cannot include a greater than sign.)
  + You may find it easier to start this problem by assuming that there's only one image tag on a line.


What makes a good solution?

* It works correctly
* As little work for the programmer as possible
* Efficient
* Clear
* Modifiable
* Elegant / clever / not garbage
* You can learn from it

Some sample solutions

Given a DOS-formatted text file (lines end with `\r\n` rather than just
`\n`), convert it to a standard text file.

* Backstory: Typewriters
* Useful tool: `od`, octal dump, look at the contents of files
* Solution one: Write a C program
     * Twenty minutes to an hour for student
     * Five minutes Sam (actually, 1:30 for a mediocre version)
* Solution two: In vi, use `:1,$s/\r\n/\n/g`
    * :1,$ means every line
    * s mean substitute
    * / is the separator
    * \r\n is the pattern
    * \n is the replacement
    * g is "everywhere"
    * Didn't work
    * Didn't work with `sed`.  Why not?
    * Debugging: Long time
* Solution three: R (using a scripting language)
    * Load the file into a string
    * Did a search and replace
    * Saved
    * About five minutes
* Solution four: `tr` - translates or deletes characters
    * tr -d \r
    * One minute

Given a standard text file, convert all uppercase letters to lowercase.

* Solution one: C program
    * Five minutes to fifteen minutes
* Solution two: Python (or other scripting language)
    * Two minutes
* Solution three: tr
    * `tr [A-Z] [a-z] < oldfile > newfile`
    * `tr [:upper:] [:lower:] < oldfile > newfile`
    * Why upper and lower?  More readable.
    * `:upper:` and `:lower:` support i18n. 

Given a standard text file, remove all blank spaces at the end of lines.

* Solution one: C program
    * Whoops!  Buffer length!  (Someone will always find a way to overflow your buffer.)
    * Thirty minutes
* Solution two: vim plugin
    * Install whitespace plugin
    * Save file
    * Five minutes
* Solution three: emacs
    * M-x white-cleanup
    * Five seconds
* Solution four: Scripting language
    * Replace ' \n' with '\n'
    * Three minutes
* Solution five: vi search and replace
    * `:1,$s/ *$//`
    * Ten seconds
* Solution six: sed search and replace
    * sed is "stream editor", it's a non-interactive editor.  You give
      it editing commands and it modifies things

Given a CSV file in which each line has the form  
`LastName,FirstName,Assignment,NumericGrade`
find the the five highest grades on homework 2.

* Solution one: Load it into Excel, etc.
* Solution two: R or scripting language
    * Read CSV file
    * Subset by things that have column 3 labeled HW2
    * Sort
    * Thirty seconds
* Solution three: C program
* Solution four: `cat file | grep ',HW2,' | sort -t, -n -k3 | cut -f4 | head -5 `
Given an HTML file, find the URLs of all images.  In case you don't