Errata

chop. It would be nice to have a chomp index entry.

There are two operators, chomp() and chop(). Unfortunately, the index in
Learning Perl doesn't have an entry for chomp() and further combines chomp()
references with those of chop(), which is an incorrect assumption.

I also noted that the die() operator is listed in the index as having a first
reference on page 23, when the actual first reference appears on page 19. It's
also curious that the book lists these elements as functions, for example,
"die function" whereas the index lists the element as the "die() operator."

chop() is supposed to be chomp()

piece of code is repeated):

Shouldn't "$words" read "%words"? Compare with line 8 on page 19.

"The while loop reads lines from the wordslist file (via the WORDSLIST
file-handle) one line at a time. Each line is stored in the $name variable."

THIS IS NOT QUITE RIGHT.... as the storing is alternated first to $name,
then $word. See the code above and the explanation after which correctly
states the process.

So you might say "Each line is stored alternately, first to $name variable
via ($name = <WORDSLIST>), then to $word variable via $word = <WORDSLIST>;"

reads:


The while loop reads lines from the wordslist file (via
the WORDSLIST file handle) one line at a time.

should read:

The while statement reads lines from the wordslist file
(via the WORDSLIST file handle) a line at a time.

Note: The whole loop brings in two lines from the wordslist file, assuming it
isn't the end of the file, once in the while statement and another at $word =
<WORDSLIST>.

file in our main directory with this:

while ( defined($filename = glob("*.secret")) )

Then in "The Final Programs," this syntax is changed to

while ($filename = <<*.secret>)

without explanation.

From booktech: It is possible to use file name globbing within angle
brackets, so

while ( defined($filename = glob("*.secret")) )

and

while ($filename = <*.secret>)

are essentially the same thing. Basically the angle bracket method is older.
It won't work with variables.

$foo = '*.secret';
@fnames = <$foo>;

won't work, but

$foo = '*.secret';
@fnames = glob($foo);

will work.

On the very top of page 24 (or on the very bottom of page 23) the line:

while ($name = <WORDSLIST>) {

is missing from the source listing.

It says, "The actual permissions of the files will be the logical AND of 0666
and your process's current umask."

I suppose that should be ".... _bitwise_ AND of 0666 and _the bitwise NOT of_
your process's current umask." At least, that's how creat(2) does it: (mode
& ~umask).

The truncation and lower-casing of the name should be done here instead of in
good_word, so that lastdb won't contain multiple entries for equivalent names.

The code in these three lines (starting "if ($somename eq "randal")") will
never be reached, that has been already excluded in line 6 of this code
(starting on page 28: "if ($name =~ /^randal/i))."

doesn't handle 'hello
' which prints as hello
not hello<CR>

In fact, 'hello\n' results in hello
, which is exactly the same result. So
it is not necessary to "precede the backslash by a backslash" for
single-quoted strings. Removing that sentence would help, but also the
parenthetical statement on pp 34 top is in question, because, as you can see,
preceding a backslash by a backslash within single quotes apparently doesn't
have special meaning, despite what even perldata says.

"array" should read "list" (it's not yet an array).

(1.2 .. 5.2) # Same as (1.2, 2.2, 3.2, 4.2, 5.2)

It's actually the same as (1, 2, 3, 4, 5).

same error as above.

The third to last line reads:


print("The answer is ",$a,"
); # three element literal array

The statement does not compile ("Use of uninitialized value at list line ...");
It should read:

print("The answer is ",@a,"
); # three element literal array

The use of the list constructor operator is described incorrectly.

"This operator creates a list of values starting at the left scaler
value up through the right scaler value, incrementing by one each
time."

Should read

"This operator craetes a list of values starting at the integer
value of the left scaler value up through the integer
value of the right scaler value, incrementing by one each time."

The example of

(1.2 .. 5.2) #same as (1.2, 2.2, 3.2, 4.2, 5.2)

should be

(1.2 .. 5.2) #same as (1, 2, 3, 4, 5,)

I have tried this on PERL ver 4.018 PL 36 under Slackware ver 2.3 and on PERL
ver 5.005.03 on a server at school.

Example says:

@fred[1,2] = (9,10); # change the last two values to 9 and 10

The wording "the last two values" only applies for the special case where
@fred has only three values. You could instead say, without loss of
generality, "change the second and third values to 9 and 10 respectively".

No slice here.

Example says:

@y = sort(@y); # @y gets 1,16,2,32,4,64,8

This should read:

@y = sort(@y); # @y gets (1,16,2,32,4,64,8)

Similarly, the line two lines up should read:

# @x gets ("large","medium","small")

if (keys(%somehash)){ # if keys() not zero:
...; #array is non empty

It is better to write as

if (keys(%somehash)){ # if keys() not zero:
#array is non empty
...;

Since "#array is non empty" is also a part of comment of "if
(keys(%somehash))", but not "...". It will be consistent with the style of the
book.

%league{keys %score} = values %score

should be

@league{keys %score} = values %score

"or put it as the first character within the list".

So if we want find a line containing any "]" or an "a", we can use /[]a]/, but
it does not work.

If we put a backslash, then it is OK. It also works for [. (At first I think [
and ] are treated differently in perl.)

Now reads:

"In this case , making that match lazy (with a trailing to "?")
will actually simplify the work that Perl has to perform, so you
may want to consider that".

In fact I think ? will do the forward tracking. Therefore

$_ = "a xxx ce xxxxci ,,,cd"
/a.*?ci.*d/

will also match. It is not true that "the a.*?c now matches the fesest
characters between the a and c."

The difference between ? and no ? is not only in speed, but some other place
as well, see the following example

$_= "a xxx c xxxxxxx c xxx d";";
s/c.*d/KK="a xxx c xxxxxxx KK"
s/c.*?/KK="a xxx KK"

text: /+/ matches "x+y" but not "++" or " + ". This is as the same as
-,*,/,=,++, Notice that + and * have different meaning if there is no
before. I guess before the symbol changes b to B and vice versa. I suggest
to say that "Notice the usage of /,+,*, if we want to search these symbols
themselves, we have to put backslash before them, However it will change the
meaning of  and B."

Add  and B.

The answer suggested by the authors to this exercise uses the && operator,
which is not introduced in the text until p. 106.

"@result;" should be "return @result;"

The last underscore in "@_[1..$#_]" may be incorrect.

"From." should be "From: ", i.e., that in line -7.

should start "File is executable or directory is searchable"

"to" should be "in".

print "$name
"; # prints ., .., passwd, group, and so on

should read

print "$name
"; # prints ., .., group, passwd, and so on

because the list has been sorted.

Swap "to" and "from"; link example, line 2: "fred to bigdumbguy" should be
"bigdumbguy to fred"; last paragraph, line 2: swap "to" and "from".

"neighbor" should be "barney"

Should "for user" be "for owner"? Also, here the definiton of "owner" and
"group" appears (in section Modifying Ownership) after they are used (in
section Modifying Permissions), it seems not suitable.

Actually, the time value of UNIX does not overflow in the year 2000, but on
Tuesday, January 19, 2038 at 03:14:08 UTC. (Source: David A. Curry: UNIX
Systems Programming for SVR 4, O'Reilly & Associates 1996, p. 171.)

system("date >right_now") || die "cannot create right_now";

rather than:

system("date >right_now") && die "cannot create right_now";
^^

Should "angle brackets" be "single quotes"?

"shell" should be "shell command".

Paragraph 4, sentence 2: "A signal is a one-bit message..."

If the signals are numbered with small numbers like 15 or 31, as noted later
in the same paragraph, perhaps this should read "A signal is a one-byte
message..."

Or perhaps I'm being far too literal.

"huge_array" should be "huge_list".

In page 68, hash function, usually we use double quotes for keys, here we use
single quotes. It is better to say something in page 68, like "we can also
use single quotes".

The 3rd paragraph states that substr will default the starting position to 0
if the starting position argument is a negative number larger than the length
of the string. That's not true; substr behaves as if it created a temporary
string and prepended null characters whose number equals the "excess" length.
Examples:

print(substr("abc",-5,1)."
"); # prints nothing
print(substr("abc",-5,2)."
"); # prints nothing
print(substr("abc",-5,3)."
"); # prints a

The 2nd paragraph makes reference to an "ending position", but there is no
ending position in the syntax for substr.

"def became abc and ghi becomes a signle d".

It is right. But it will give readers the impression that tr is a command for
string replacement (def became abc) and in fact, tr is a command for character
replacement.

"/etc/passwd line split apart" should be "/array/list line split apart" and
"an array" should be "a list">

$idnum = getpwuid ("daemon");
$login = getpwnam (25);

This should be:

$idnum = getpwnam("daemon");
$login = getpwuid(25);

Which means the sentence "They each return the thing you've asked them to get"
seems incorrect as well. They really return the other thing. Getting the
login name returns the ID, getting the ID returns the login name.

>From perldoc -f getpwuid:

"In scalar context, you get the name, unless the function was a lookup by
name, in which case you get the other thing, whatever it is. (If the entry
doesn't exist you get the undefined value.) For example:

$uid = getpwnam($name);
$name = getpwuid($num);
"

Running the examples seems to bear this out.

"string value" should be "the string value of their home directory."

"either ABCD or DCBA, depending on [endian]"

I don't have a little-endian to test, but I though it should be "CDAB".

"array" should be "hash"; paragraph 4, line 2: "array" should be "hash".

"array" should be "hash"; example 1: ARRAYNAME should be HASHNAME; paragraph
2, line 1: ARRAYNAME should be HASHNAME; example 2, comment: Should be "open
mydatabase into %FRED."

"DBM array" should be "DBM hash".

"after" should be "before".

After END_of_Start, there is a useless my in :

my $favorite = param("flavor:);

It should read :

$favorite = param("flavor");

The added "my" bugs the code.

What's your flavor? <INPUT NAME="favorite" VALUE="mint">

The problem is that the CGI script it should talk to (ice_cream, version 2, p.
187) expects input of the form "...?flavor=mint", not "...?favorite=mint".
In order for these two to interface properly, the input line in the form
should be

What's your flavor? <INPUT NAME="flavor" VALUE="mint">

But actually, you are talking about the previous example: Ice Cream Stand.
Therefore I suggest to change it to:

If instread of popup_menu flavor, we want to use scrolling_list(), we can
replace the corresponding line with the following code.

print p("What flavor: ", scrolling_list(
-NAME => "flavor",
-VALUES => [ qw(mint chocolate cherry vanilla peach) ],
-LABELS => {
mint => "Mighty Mint",
chocolate => "cherished chocolate",
cherry => "Cheery Cherry",
vanilla => "Very Vanila",
peach => "Perfectly Peachy",
},
-SIZE => 3,
-MULTIPLE => 'true',
));

And notice that it is better to change "flavors" to "flavor" in the NAME row.

You use bail here, but bail isn't mentioned or described so far (it is
described much later in the book).

"executable" should be "searchable".

Now reads:


if ($webdoc->is_success) { # found it
print STDOUT "$url: ", $webdoc->title, "
";
} else { # something went wrong
print STDERR "$0: Couldn't fetch $url
";

Should read:

if ($webdoc->is_success) { # found it
print STDOUT "$url: ", $webdoc->title, "
";
} else { # something went wrong
print STDERR "$0: Couldn't fetch $url
";

The printed string should end with a newline instead of a blank.

You need to integerize the random number with int(rand(@b)).

is a bit 1st edition and a bit 2nd edition. We think that the end of the
first para (after the program listing) and the beginning of 2nd para are not
right any longer.

Now reads:


print if
(/^[^aeiou]*a[^eiou]*e[^aiou]*i[^aeou]*o[^aeiu]*u[^aeio]*$ );

Should read:

print if
(/^[^aeiou]*a[^eiou]*e[^aiou]*i[^aeou]*o[^aeiu]*u[^aeio]*$/i );

and in that order. The errata on-line has a correction to permit mixed upper
and lower case, but that string is still incorrect.

The correct string should read:

/^[^aeiou]*a[^aeiou]*e[^aeiou]*i[^aeiou]*o[^aeiou]*u[^aeiou]*$/i

The regular expression string for vowel matching doesn't have a terminating
slash /.

"array element" should be "hash element."

Swap "new" and "old".

The if statement should begin

if ($slash == 0) {
$head = "/";
$tail = substr($_, 1);
} elsif ($slash > 0) {

The first item in the module list should be "integer" with a small "i" -
currently it's listed as Integer, which doesn't exist.

"or" should be ||.

Now reads:


$remote = IO::Socket::INET->new( Proto => "tcp",
PeerAddr => $host,
PeerPort => "http(80)",
);

Should read:

$remote = IO::Socket::INET->new( Proto => "tcp",
PeerAddr => $host,
PeerPort => "http(80)",
);

"or" should be ||.