Mastering Regular Expressions
Errata
Last updated 1-Oct-2012
Please see my contact information if you'd like to send in an erratum.
Most errata tend to be just inconsequential typos or grammar nits, but there are also some factual mistakes that cause confusion during a close reading. In the list below, the factual mistakes are marked like this.
page xvii
| Printed: | It was published in 2002. |
| Fixed: | The second edition was published in 2002. |
Clarifies some poor wording.
page 13, last line of 3rd paragraph
| Printed: | understand that (first|1st) and |
| Fixed: | understand that (First|1st) and |
It's not wrong how it was presented, but to match better with the prior discussion, it makes sense to capitalize the 'f'.
page 49, first list item
| Printed: | Perl's appear to be of the same general type as egrep's, but has a.... |
| Fixed: | Perl's appears to be of the same general type as egrep's, but Perl's has a.... |
page 63, table, 3rd row
| Printed: | ..., but this not of much practical... |
| Fixed: | ..., but this is not of much practical... |
page 69, last line
| Printed: | ... or perhaps [^ \t\r]*$ to allow for |
| Fixed: | ... or perhaps [^ \t]*$ to allow for |
See the next errata item for the description.
page 70, first line
| Printed: | spaces, tabs, and the carriage return that some systems have before the line-ending newline. |
| Fixed: | spaces and tabs. |
Including \r creates problems on the very systems mentioned (removing the \r required at the end of line).
page 70, second paragraph
| Printed: | [^ \t\r]*$ |
| Fixed: | [^ \t]*$ to allow for |
See prior item.
page 70, mid page
| Printed: | $text =~ s/[^ \t\r]*$/<p>/mg; |
| Fixed: | $text =~ s/[^ \t]*$/<p>/mg; |
See prior item.
page 74, last line
| Printed: | maybe some tech stuff at http://www.slashdot.com! |
| Fixed: | maybe some tech stuff at http://www.slashdot.com! |
The "!" should be included in the underlined text.
page 94, last line of 2nd paragraph
| Printed: | style is use by most... |
| Fixed: | style is used by most... |
page 96, 2nd code snippet
| Printed: | if (! line.matches("\\s*", )) |
| Fixed: | if (! line.matches("\\s*")) |
page 199, mid page
| Printed: | s/\s*(.*?)\s*$/$1/s |
| Fixed: | s/^\s*(.*?)\s*$/$1/s |
page 210, 2nd-to-last paragraph
| Printed: | ...better to use [1235-9] instead... |
| Fixed: | ...better to use [01235-9] instead... |
| or | ...better to use [0-35-9] instead... |
Forgot that zero is a digit. Doh!
page 218, middle of third paragraph
| Printed: | ..., would I do add this? |
| Fixed: | ..., would I add this? |
page 261, third paragraph
| Printed: | ... the distribution of the colon from (?:this|that):) to ... |
| Fixed: | ... the distribution of the colon from (?:this|that): to ... |
page 298, bottom of second paragraph
| Printed: | ... the whole reason we've been looking localization: ... |
| Fixed: | ... the whole reason we've been looking at localization: ... |
page 370, last paragraph
| Printed: | (it renders LF and all the other non-newline... |
| Fixed: | (it renders CR and all the other non-newline... |
Reported by Alan Moore, 11-Oct-2006
page 378, first code snippet
| Printed: | String regex = "(?x) ^(https?):// ([^/:]+) (?:(\\d+))?"; |
| Fixed: | String regex = "(?x) ^(https?):// ([^/:]+) (?: :(\\d+) )?"; |
page 447, last paragraph
| Printed: | ..., '+' is treated as '*?' and vice-versa |
| Fixed: | ..., '+' is treated as '+?' and vice-versa |
Reported by Jan-Pieter Cornet, 14-Jan-2006
page 457, last line
| Printed: | they retained, ... |
| Fixed: | they are retained, ... |
page 459, last line
| Printed: | the variables are note interpolated at the wrong time. |
| Fixed: | the variables are not interpolated at the wrong time. |
page 467, first code snippet
| Printed: | $parts = preg_split('/\r? \n \r? \n/x', $response, 2); |
| Fixed: | $parts = preg_split('/\r? \n \r? \n/xS', $response, 2); |
The parenthetical note just after this snippet refers to the S modifier, so it should have been there.
page 473, comment on the if line in the middle of the page
| Printed: | /* '/' followed by '\' or EOS */ |
| Fixed: | /* '\' followed by '/' or EOS */ |
page 481, mid-page regex
| Printed: | ^((?:<(\w++)[^>]*+(?<!/)>(?1)</\2>|[^<>]++|<\w[^>]*+/>)*+)$</tt> |
| Fixed: | ^((?:<(\w++)[^>]*+(?<!/)>(?1)</\2>|[^<>]++|<\w[^>]*/>)*+)$</tt> |
This is a pretty major error — the use of a possessive greedy quantifier on '[^>]' means that the '/' that follows can never match (since if there were a slash for the '/' to match, the '[^>]' would have consumed it). In an attempt to make the expression a bit more efficient, I made it incorrect, and that's very bad. If PHP had lazy possessive quantifiers, I could use them here, but it doesn't, so the thing to do is to drop the minor efficiency gains from the possessive quantifiers, and use normal greedy quantifiers. That way, it will at least be correct.
Reported by Jan-Pieter Cornet, 14-Jan-2006
page 482, middle paragraph
| Printed: | The third alternative, <\w[^>]*+/>, matches... |
| Fixed: | The third alternative, <\w[^>]*/>, matches... |
Reported by Jan-Pieter Cornet, 14-Jan-2006
page 482, middle paragraph
| Printed: | As before, the use of a possessive quantifier here may be overkill but it certainly doesn't hurt. |
| Fixed: | (that sentence removed) |
Jan-Pieter pretty much summed this one up when he concluded the errata report with “Ouch.”
Reported by Jan-Pieter Cornet, 14-Jan-2006
page 484, top regex
| Printed: | | <\w[^>]*+/> # self-closing tag |
| Fixed: | | <\w[^>]*/>, # self-closing tag |
Same issue as reported by Jan-Pieter Cornet, 14-Jan-2006
page 484, bottom regex
| Printed: | | <\w[^>]*+/> # self-closing tag |
| Fixed: | | <\w[^>]*/>, # self-closing tag |
Same issue as reported by Jan-Pieter Cornet, 14-Jan-2006