Bash脚本未生成预期HTML输出，寻求修改方案

Let's walk through the key issues in your script and how to fix them to get the expected output:

1. Wrong Username Extraction #

Your original cut -d" " -f4 was grabbing the timestamp field instead of the username. We'll use awk to correctly pull the username, and map invalid users to <UNKNOWN>:

grep "Failed password for" "$logfile" | awk '{
    # Check if this is an invalid user line
    if ($11 == "invalid") {
        print "<UNKNOWN>"
    } else {
        # Grab the valid username from field 11
        print $11
    }
}' | sort | uniq -c | sort -k1,1nr -k2,2 > templog

(Note: Adjust field numbers if your log format differs—this assumes the username is in field 11 after "Failed password for".)

2. Stop Editing the Original Log File #

Remove the sed -i 's/invalid/<UNKNOWN>/g' $logfile line. Modifying log files in place is risky and unnecessary now that we handle invalid users during extraction.

3. Correct HTML Tag Rendering #

Your script was writing escaped tags like <html>, which makes browsers display tags as text instead of rendering them. Write raw HTML tags instead:

echo "<html>" > flog.html
echo "<body><h1>Failed Login Attempts Report as of `date`</h1>" >> flog.html
# ... later ...
echo "</body>" >> flog.html
echo "</html>" >> flog.html

4. Properly Escape User Content #

Escape < and > in usernames to avoid breaking HTML (e.g., <UNKNOWN> becomes <UNKNOWN>):

sed -i 's/</\&lt;/g;s/>/\&gt;/g' templog

5. Add Commas to Large Counts #

Format numbers with commas for readability using awk:

awk '{printf "%'\''d %s\n", $1, $2}' templog > temp && mv temp templog

6. Fix Closing Body Tag #

Change echo "/body" to echo "</body>" to produce valid HTML.

7. Clean Up Temporary Files #

Add rm templog at the end to delete the temporary file after generating the report.

Full Corrected Script #

#!/bin/sh
if [ $# -eq 0 ]
then
    echo "Usage : flog LOGFILE"
    exit 1
fi
logfile="$1"

# Extract usernames: map invalid users to <UNKNOWN>
grep "Failed password for" "$logfile" | awk '{
    if ($11 == "invalid") {
        print "<UNKNOWN>"
    } else {
        print $11
    }
}' | sort | uniq -c | sort -k1,1nr -k2,2 > templog

# Add commas to count numbers
awk '{printf "%'\''d %s\n", $1, $2}' templog > temp && mv temp templog

# Escape HTML special characters in usernames
sed -i 's/</\&lt;/g;s/>/\&gt;/g' templog

# Prepend line breaks to each entry
sed -i 's/^/<br \/> /g' templog

# Generate valid HTML file
echo "<html>" > flog.html
echo "<body><h1>Failed Login Attempts Report as of `date`</h1>" >> flog.html
cat templog >> flog.html
echo "</body>" >> flog.html
echo "</html>" >> flog.html

# Show the generated HTML
cat flog.html

# Clean up temporary file
rm templog

Expected Output (HTML Source) #

This script will generate a valid HTML file that renders correctly in a browser. The source will look like:

<html>
<body><h1>Failed Login Attempts Report as of Thu Dec 25 07:00:00 MST 2014</h1>
<br /> 1,325 &lt;UNKNOWN&gt;
<br /> 505 root
<br /> 17 adm
<br /> 16 ftp
<br /> 10 apache
<br /> 10 mail
<br /> 2 backuppc
<br /> 2 games
<br /> 2 news
<br /> 2 squid
<br /> 1 bin
<br /> 1 nobody
<br /> 1 operator
<br /> 1 smmsp
<br /> 1 sshd</body>
</html>

When opened in a browser, this will display a heading followed by each login attempt count on a new line, with <UNKNOWN> for invalid users.

内容的提问来源于stack exchange，提问作者ShreddedSaber