Test file line length limits in a shell in Linux

test_len.sh

#!/bin/sh

# test_len.sh

# This script is used to test the maximum line length that can be processed
# using read -r and eval.

# testing line lengths divided by 10:
test_lengths='409 410 900 1800'

# select a line generator:
# line_generator=generate_line
# line_generator=generate_line_with_spaces
line_generator=generate_line_with_quotes

if test "$line_generator" = generate_line_with_quotes ; then
    # If set to 1 add quotes around each word in the line to compensate for
    # the fact that eval removes quotes around words.
    eval_add_quotes=1
fi

tmp_dir=$(mktemp -d)

generate_line_part () {
    printf %s "$2"
    for i in $(seq 0 $(( $1 - 1 ))) ; do
        printf %s $(( i % 10 ))
    done
    printf %s "$3"
}

generate_line () {
    for i in $(seq 0 "$1") ; do
        generate_line_part 10 '' ''
    done
}


generate_line_with_spaces () {
    for i in $(seq 1 "$1") ; do
        generate_line_part 9 ' ' ''
    done
}


generate_line_with_quotes () {
    for i in $(seq 1 "$1") ; do
        generate_line_part 7 ' "' '"'
    done
}


generate_lines () {
    for i in "$@" ; do
        "$line_generator" "$i"
        echo
    done
}


eval_print_line () {
    if test "$eval_add_quotes" -eq 1 ; then
        eval "printf ' \"%s\"' $1"
    else
        eval "printf ' %s' $1"
    fi
}


get_file_names () {
    for i in "$@" ; do
        printf '%s/test_len_%s0.txt\n' "$tmp_dir" "$i"
    done
}


generate_files () {
    cat /dev/null > "$tmp_dir"/test_len_combined.txt
    for i in "$@" ; do
        file_name=$(get_file_names "$i")
        ("$line_generator" "$i" ; echo ) > "$file_name"
        cat "$file_name" >> "$tmp_dir"/test_len_combined.txt
    done
}


get_files () {
    for file_name in $(get_file_names "$@") ; do
        cat "$file_name"
    done
}


cleanup () {
    rm -rf "$tmp_dir"
}


printf 'Using %s; example line of 100 characters:\n' "$line_generator"

"$line_generator" 10
printf '\n\n'


trap cleanup EXIT
# shellcheck disable=SC2086  # word splitting is needed for multi-word options
generate_files $test_lengths
# ls -l "$tmp_dir"/


echo "Lengths of generated lines:"

# shellcheck disable=SC2086
# get_files $test_lengths | awk '{print length}'
for file_name in $(get_file_names $test_lengths) ; do
    < "$file_name" awk '{print length}'
done


echo "Lengths of these lines read using 'read -r line':"

# shellcheck disable=SC2086
# get_files $test_lengths |
for file_name in $(get_file_names $test_lengths) ; do
    while read -r line ; do
        printf ' %s' "$line" | awk '{print length}'
    done < "$file_name"
done


echo "Lengths of these lines additionally processed using eval:"

# shellcheck disable=SC2086
# get_files $test_lengths |
for file_name in $(get_file_names $test_lengths) ; do
    while read -r line ; do
        if test "$eval_add_quotes" -eq 1 ; then
            eval "printf ' \"%s\"' $line" | awk '{print length}'
        else
            eval "printf ' %s' $line" | awk '{print length}'
        fi
    done < "$file_name"
done


echo "Lengths of these lines additionally processed using eval inside a function:"

# shellcheck disable=SC2086
# get_files $test_lengths |
for file_name in $(get_file_names $test_lengths) "$tmp_dir"/test_len_combined.txt ; do
    while read -r line ; do
        eval_print_line "$line" | awk '{print length}'
    done < "$file_name"
done