Errata

In the example:
egrep 'Victor(ia)*' file

This will actually match something like 'Victoriaia..ia' as opposed to the '(Victor)|(Victoria)' wanted.

The example should actually be
egrep 'Victor(ia)?' file

The "-l. --lines" option for wc user command is "Print line count only."
I use the gnu wc program supplied with openSUSE.
Its help for "-l, --lines" states: "print the newline counts."
If the last line in the file is not terminated with a newline,
the newline count will be one less than the number of lines in the file.

>python3 -c 'print("one line", end="")' > oneline.txt
>wc -l oneline.txt
0 oneline.txt
>cat oneline.txt
one line>
>
>

NB. this description is based on Bash version 3.2.48, whereas the book mentions on p.596 that all references are to Bash version 4. Regardless, these issues may still apply and should be verified.


"s1 =~ s2" entry should specify "(Valid only within [[ ]] construct.)" since the =~ operator is not valid with the `test' or [ ] constructs.

Conversely, the statement "(Valid only within [[ ]] construct.)" should be *removed* from the entries "s1 < s2" and "s1 > s2", since the `<' and `>' operators *are* valid with the `test' and [ ] constructs, though in the latter cases they must be quoted or escaped (eg. s1 \< s2) to prevent being interpreted as redirection operators (this being a potentially disastrous mistake).

Lastly, in the "s1 =~ s2" entry, the instruction "Quote s2 to keep the shell from expanding embedded shell metacharacters" is apparently incorrect:
$ [[ a =~ '[a]' ]]; echo $?
1

$ [[ a =~ [a] ]]; echo $?
0