I didn't have a GitBook account and appears I'm unable to make one with the "legacy" version of the site when clicking on this link:

Download ebook for offline reading - link

401
Unauthorized
New accounts should be created on the new version of GitBook. Head to https://www.gitbook.com

edit: able to download but wasn't able to preview or attempt to login on the legacy URL

google told us there is some 'error 005' in epub file .I can't detect this in error.So I could not publish this book in google book.

In gnu_grep.md (around line ) and gnu_sed.md (around line 1402) there is an example of equalivent word and space character classes. Unfortunately these are valid-ish markdown, and so do not get rendered correctly. They should be escaped as code in both places.

This:

| Character classes | Description |
| ------------- | ----------- |
| \w | Same as [0-9a-zA-Z_] or [[:alnum:]_] |
| \W | Same as [^0-9a-zA-Z_] or [^[:alnum:]_] |
| \s | Same as [[:space:]] |
| \S | Same as [^[:space:]] |

Should be this:

| Character classes | Description |
| ------------- | ----------- |
| `\w` | Same as `[0-9a-zA-Z_]` or `[[:alnum:]_]` |
| `\W` | Same as `[^0-9a-zA-Z_]` or `[^[:alnum:]_]` |
| `\s` | Same as `[[:space:]]` |
| `\S` | Same as `[^[:space:]]` |

The ticks around the slash classes aren't necessary for formatting, but I think are good style (elsewhere in the document these are generally formatted as code).

For consistency and clarity, it might also make sense to apply this pattern to the table a couple paragraphs above this. I don't see any markdown issues there, but I think it would generally be clearer to list the actual regex patterns as code.

In section Filtering, subsection Fixed string matching, the command $ awk 'index($0,"a+b")' eqns.txt should return all lines of the input file, since a+b appears in each line, right?

What did you have in mind for this? I could maybe throw something together for the xmllib tools (xmllint, xsltproc) and/or xmlgawk (gawk with a SAX-like XML parser). Xmllib tools also have a —html option to parse XHTML (maybe regular HTML as well, but haven’t tried).

I’ve also wrote awk scripts to convert various patterned text to XML, if that sounds useful.

$ cat -E marks_2015.txt
cat: illegal option -- E
usage: cat [-benstuv] [file ...]

First, I'm sorry for submitting a PR (#11) without reading the README more closely. Just wanted to contribute what I hope will be some useful knowledge to the project.

IMHO, less is a largely underutilized tool that can greatly improve upon, if not replace, the functionality of several other commands--namely cat & tail--as you'll see in my PR. Also less is a tool of convenience because how many times have you opened a file with it, then exited less to open it in your editor? Yet you can jump straight into your editor from within less! Lastly, you can view multiple files with less in a manner--and keystrokes--akin to changing buffers in vim.

Decoding your rot13 email address at the bottom of page 2 does not work in solaris, unless GNU utils are installed. The string1 string2 patterns for tr in solaris are literal.

rot13 code/decode that works with all versions of "tr" and handles both upper and lower case.
the alias is wrapped in this text box but belongs all on the same line.
alias rot13="tr 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' 'nopqrstuvwxyzabcdefghijklmNOPQRSTUVWXYZABCDEFGHIJKLM'"

Now you can simply pipe text to rot13.
echo '[email protected]' | rot13

perl_the_swiss_knife.md: dyas -> days
gnu_awk.md: intial -> initial
gnu_sed.md: whereever -> wherever
perl_the_swiss_knife.md: wtih -> with
Found with: https://github.com/ss18/grep-typos

Submitting my solution for question 1 of ex13_regex_character_class_part2 results in the below:

Mismatch for question 1:
Expected output is:
a[2]
foo_bar
appx_pi
greeting
food[4]
b[0][1]

But it is the exact same output I'm getting. I've even tried with the solution from .ref_solutions, without any luck.

Using GNU bash, version 4.4.23(1)-release (x86_64-pc-msys)

XML: xmllint --format
JSON: cat $file | python -mjson.tool

That kind of thing might be useful for this collection.

In gnu_sed.md around line 182, there should be an additional cat command that results in the output being shown.

This:

# original file gets preserved in 'greeting.txt.bkp'
Hi there
Have a nice day

Should be this instead:

# original file gets preserved in 'greeting.txt.bkp'
$ cat greeting.txt.bkp
Hi there
Have a nice day

I contribute to https://github.com/uutils/coreutils

Any way I can help? I would like to get more regression tests comparing GNU and rust coreutil resource usage.

Adding fixed-width data would be interesting: https://www.gnu.org/software/gawk/manual/html_node/Fixed-width-data.html#Fixed-width-data

$NF points to last field ### This is not true. It contains the last field "value". NF (without the $ sign) indirectly points to the last field, by holding the total Number of Fields found in $0 (the current record).

NF is built-in variable and can be used in expressions ### to reference the last, or number of, fields in the current record being examined/processed. More simply, NF is the total Number of Fields found in the current record.

$(NF-1) points to second last field and so on ### Again, this would be the "contents" of the second to last field in the current record. NF-1 (without the $ sign) represents the field Number of the second to last field.

I think it would be a good idea to mention the tee command, probably somewhere in the "Cat, Less, Tail and Head" chapter