Errata

Dear O'Reilly team,

the filename containing the sources to the book (page xvii - xix) is specified as

progs.tar.gz

in the folder /published/oreilly/nutshell/sedawk_2

However, on the server there is no such file, but the following file

progs.tar.Z

By using the commands in the book (either ftp, ftpmail, etc.), the file cannot be download, if not using the correct name progs.tar.Z

With best regards,
Marco

on page 24 the following code omits a critical line
----------------------------------------------------
#!/bin/sh
awk -F, '{
print $4 ", " $0
}' $* |
sort |
awk -F, '
$1 == LastState {
print " " $2
}
$1 != LastState {
LastState = $1
print $1
}'
--------------------------------------------------------
on page 25 the correct piece of code is shown

$1 == LastState {
print " " $2
}
$1 != LastState {
LastState = $1
print $1
print " " $2
}'

Errata for the errata at

http://oreilly.com/catalog/errata.csp?isbn=9781565922259

$ grep " ["[{(]*book[]})"?!.,;:'s]* " bookwords

should be

$ grep " [\"[{(]*book[]})\"?\!.,;:'s]* " bookwords

Code example should be:

/^\.Ah/{
s/\.Ah */\
@A HEAD = /
s/"//g
s/$/\
/
}


The period must be escaped, i.e. \.Ah, else the regex will match any expression with any single character preceeding the Ah, e.g. ZAh, etc.

Ooops.. In addition to the escaped \. regexp, also need to eliminate the 3rd line, i.e.

/^\.Ah/{
s/\.Ah */\
\
@A HEAD = /
s/"//g
s/$/\
/
}


Should be:

/^\.Ah/{
s/\.Ah */\
@A HEAD = /
s/"//g
s/$/\
/
}


Else you get one too many additional blank lines in the substitution.

The example, as "printed" in both ePub and Book is correct:

/^\.XX /s/substitution/substitute/

Errata is incorrect, as in:

/^.XX /s/substitution/substitute/

The period must be escaped in order to match a literal period, else otherwise will be a regex match for any single character.

The Month, day, year Regex `[A-Z][a-z]\{3,9\} [0-9]\{1,2\}, [0-9]\{4\}` Will not match the month of 'May' due to the {3,9} modifier. It should read `{2,9} `

I just wanted to report a few errors in the examples from the book "sed &
awk", second edition by Dale Dougherty and Arnold Robbins.

I am referring to the examples in:

ftp://ftp.ora.com/published/oreilly/sed2/progs.tar.gz

The first error is that the name of the directory is listed in the book as:

/published/oreilly/nutshell/sedawk2/
---------------------------


Secondly, in the directory called ch06, the example entitled "phrase" is
correct in the book, but is missing a final line in the example source. The
last line should be "done". (I numbered the lines to refer to them):

01 #! /bin/sh
02 # phrase -- search for words across lines
03 # $1 = search string; remaining args = filenames
04 search=$1
05 shift
06 for file
07 do
08 sed '
09 /'"$search"'/b
10 N
11 h
12 s/.*\n//
13 /'"$search"'/b
14 g
15 s/ *\n/ /
16 /'"$search"'/{
17 g
18 b
19 }
20 g
21 D' $file
22 done <------------------------------ this is missing in the example

Finally, the "N" in line 10 should be "$!N" or the last line will always
print.

I like the book very much, and point out these items only to make it better...

HTH and Best regards,


Glenn
--------
Glenn R Williams

ePub version reads:

echo inner1 inner2 inner1 is executed. Its output (the words inner1 inner2 inner3) is placed into the next command to be executed.

There is no "inner3". Should be "inner1", as in:

echo inner1 inner2 inner1 is executed. Its output (the words inner1 inner2 inner1) is placed into the next command to be executed.

This is printed correctly in the dead tree version, but not the ePub.

This code breaks on repeated capital letters. "aPPle" becomes "aaale". The "for" loop repeats twice, once for each "P". The first time gsub replaces both "P"s with "p"s. The second time "x" has advanced. CAP = "p" and index() tries to find "p" in the string "upper" which fails and returns "0" which is stored in CHAR. The substr() function in my version of awk (GNU AWK 4.0.1) treats any starting index less than 1 as 1 so gsub ends up substituting all "p"s with "a"s.