Historical errata for Text Processing with Ruby

all source codes with echo commands and \
doesn’t work in bash
I think printf should be used instead of echo.

Please, if possible, include a chapter on how to extract data from PDF files.
In linux and windows there are several utilities with dozens of options and integration with Ruby would be very interesting.

“… each_line will result in a noticeably more solution.”

This phrase is missing a word.

“Allows you to hone in on” - should be HOME in on

There should not be a trailing forward slash after
www.example.com

(first line of output in the first example under the heading “Running the Script”)

Spelling mistake in p.26 paragraph 2 “… searchs the filesystem…” : “searchs” should be “searches”

When I run the first example on page 32 I don’t get precisely the same output.

Running zsh I get a trailing ‘%’ character after ‘bar’, indicating the text ended without a newline character, and running bash the command prompt starts immediately after the output (the ‘bar’ in this case).

My suggestion is that for this example and similar ones you might consider adding a newline character to the end of the input i.e. ‘foo\
bar\
’ in this case. This way the output on both bash and zsh would match the book (I think).

Misspelling of matches

This sentence seems awkward and considering you have not mentioned a Log file before this line, it seems a bit off.

“One example of this sort of log file that you might have encountered before is from the popular web server Apache.”

Maybe put an inline reference to an URL where one can get this data?

“The National Oceanic and Atmospheric Administration releases data ”

In the subsection “String-specific encodings” there’s an extra ‘the’ which should be removed.

The sentence reads as follows:

The result is a string encoded as ASCII that has had “the” any characters that are invalid ASCII stripped from it.

That “the” can be removed.

Duplicated verb:

If we know that the string we have is is in a specific format,

on section Detecting Encodings, a method is referred by the wrong name. The paragraph in question reads like so:

If we want to both detect the encoding and also change Ruby’s internal encoding of the string - […] we can do so with the detect_encoding method.

however that method is called detect_encoding! as can be seen on the follow-up example.

The one-liner for finding all home directories of all users can be slightly simplified. It currently reads as:

$ ruby -F: -ane ‘puts File.expand_path(“~” + $F[0]) unless /^#/’ /etc/passwd

However, the home directory is listed in passwd itself at the 5th position (6th if you’re counting starting at 1). That means the one-liner:

$ ruby -F: -ane ‘puts $F[5] unless /^#/’ /etc/passwd

is slightly shorter and equivalent, at least for systems which provide /etc/passwd with all fields. I did notice that the example User Database doesn’t show home directories, though.

the code from section “The Logic” can be slightly improved if we use join, instead of iterating over each url. Like below:

#!/usr/bin/env ruby

$stdin.each_line do |line|
urls = line.scan(${https?://\\S+})
puts urls.join(“\
”)
end

Wondering if I’m missing something as to why it couldn’t be done like above.

$ cat foo | bar

can be written as:

$ bar < foo

If you want to pipe the contents of a file, you have two options. The first is to
use cat. The second is to have the shell pass the file as input, something we
can achieve with >: (should this symbol be pointing the other way or vice versa to match the example below?)

user@computer:~ $ cat Documents/birthday-list.txt | grep Ruby
Text Processing with Ruby by Rob Miller

user@computer:~ $ grep Ruby < Documents/birthday-list.txt
Text Processing with Ruby by Rob Miller

When having process ids with different digit numbers (ie 3003 and 12755), then using the cut with the space delimiter will not display the smaller number so by using the following:
ps ax | grep ruby | cut -d’ ’ -f1

we will get the following output:
(blank line here as process id is four digits)
12755
12770
28883

when using a different command like awk:
ps ax | grep ruby | awk ‘{print $1}’
we will get:

3003
12755
12770
28883

In Chapter 3 on the bottom page 7 (Streaming Lines with the -n Switch) of the free excerpt the text after the code snippet " $ printf “foo\
bar\
” | ruby -ne ‘puts $_.upcase’ " is not properly set due to a markup problem.

There seems to be some HTML mark up left at the bottom of the page (below the FOO/BAR)

$ printf “foo\
bar\
” | ruby -ne ‘puts $_.upcase’
FOO
BAR

There’s more to $_ than this, though.

In chapter 5 (delimited data) there are a couple of places where you refer to $; as a record separator and others as a field separator - also there are a few places where you say “record separator” when I think you mean “field separator”

Second line of “Deleting Elements” section has “… is to to allow …”.

One to too many

In paragraph 2, ‘detect_encoding’ should be ‘detect_encoding!’ (i.e. with a ‘!’ suffix).

In nlp/simple-search.rb: the return value of ‘articles.select’ should be assigned to the variable ‘matches’, not ‘articles’.

In erb/purchase-ledger-decorators.erb: ‘Currency’ doesn’t match the class name defined in purchase-ledger-decorators.rb, which is ‘Dollars’.

[In the source code download, there’s an additional instance of this in the computation of ‘Total’, and there are two spurious ‘<%’ lines.]

The code example for LeagueTable#initialize is incorrect, as url as passed in as a method argument, but the first line uses url ||= 'default' without ever considering the method argument. Either the default value for the url should be a method argument default, and the code should use url = url, or the first line of the method should be @url = url || “default”.

The text and code are out of sync and in conflict.

- The text is talking about specifying the assignment token, but the code snippet starts with defining newline tokens, using “alternation” which isn’t discussed until pg 146… “In the delimiter definition, we’re introduced to another new concept: alternation…” The text on pg. 138 talks about the snippet as if this line doesn’t exist.

- The method `.absent?` is introduced without talking about it until page 146 in a different code snippet, which explicitly states “The .absent? method is something we haven’t seen before.”

- Further down the text refers to code that doesn’t exist:
“… the config file is made up of a name, then the assignment operator, then a value. We then allow for whitespace at the end of the line, followed by a newline character.”
The whitespace token is not accounted for in the code.
```
rule(:item) { name >> assignment >> value.as(:value) >> newline }
```
The reader is left trying to understand how the newline token is included between `value` and `newline` (it’s not). `whitespace`, was previously defined but only included in the definition for either side of `assignment`.

The book incorrectly specifies that: “The first column, containing the date of the week in
question, is always twelve characters long.” - taken from the data found in: wksst8110.for.

There first column containing the date is actually just 10 characters long. This is also evident from the code sample “noaa-first- row-simple.rb”.

The line “file.read(10)” only reads 10 bytes - not 12 as suggested.

Third paragraph, last sentence reads: “At times like these, it would nice if we could introduce Ruby …”

I think it was meant to say, “At times like these, it would be nice …”

The return value is not given for the command to return the longest line from ` a\
bb`

I’m thinking it should be `bb`, but when I ran the command, it returned `a`.

My suggestion is to include the return value on the next print of the book, or append some `chomp`s on `$_` so that the expected `bb` is returned

In the “noaa-all-rows.rb”, to skip to the next line, you must file.read(2) instead of 1.

Seeking through the file. Must read 64 characters to skip a whole line.

In “noaa-seek.rb” / “get_week” block. to skip lines must multiply by 64 instead of 63.