Bash/Sed - multiline sed operation printing lines out of order - bash

Question

I am having some trouble with using sed to edit a log file. I have built it into a function which is supposed to replace the text between two search strings with the output from another function. It is almost working correctly, but is printing the lines to the log file out of order. For the life of me I can't figure out why, and most adjustments I have made while trying to fix it have actually had less desirable results.
My sed function:
log_edit(){
"$3" > temp.txt
sed -i -n "/$1/{
:loop
n
/$2/!b loop
x
r temp.txt
G
s/$2/\n\n&/
}
p" "$FILE"
rm temp.txt
}
I am using the "=== text ===" dividers as my start and stop strings to pass along to the function, and using the same functions that built the log in the first place to fill the temporary text file.
The problem is occurring somewhere near/related to the 'G' command. Rather than appending the hold pattern line to the end of the string, it appears to be attaching it to the beginning of the string.
Original log sample/Desired output:
=== Metech ITAMS Log ===
Metech Recycling
ITAMS Hardware Report
Date: Thu Mar 2 08:01:38 PST 2017
Tech: SP
=== Manufacturer Information ===
# dmidecode 2.12
...
Unfortunately, the output I am getting looks like this:
=== Manufacturer Information ===
=== Metech ITAMS Log ===
Metech Recycling
ITAMS Hardware Report
Date: Fri Mar 3 09:39:02 PST 2017
Tech: SS
# dmidecode 2.12
...
Would someone be able to help me understand what I'm doing wrong, or propose a fix? This is my first question ever to SO, if more information is necessary I am happy to provide it. Thanks in advance.
Edit #1: As requested a snippet of the code that calls the function:
2)
printf "\n"
text_prompt "Please enter Tech initials: "
set_tech_id
text_prompt "Please enter Traveler ID: "
set_travel_id
mv "$FILE" "$TRAVEL_ID $TECH_INITIALS"
FILE="$TRAVEL_ID $TECH_INITIALS"
log_edit "=== Metech ITAMS Log ===" \
"=== Manufacturer Information ===" "print_header"
unset TECH_INITIALS
unset TRAVEL_ID
;;
This is part of a menu function, and it would be overkill to include the whole thing, just be aware that there will be several calls to log_edit with different start/stop strings (though all follow the === === pattern), but usually calling different functions to fill the temp.txt with.
Edit 2: For added clarity, I thought I should add the function being called with $3:
print_header(){ #Prints log header.
print_div "Metech ITAMS Log"
printf "Metech Recycling\nITAMS Hardware Report\nDate: $(date)\nTech: %s\n" \
"$TECH_INITIALS"
}
and print_header calls print_div:
print_div(){ #Prints a divider. Required parameter: $1=Text for divider.
printf "\n=== %s ===\n\n" "$1"
}
Edit 3: For question clarity, my issue is that the $2 string is being written to the log before the contents of temp.txt, rather then after.
Final Edit: A solution was found. I thought I would post the working code below just in case it's helpful to others. A big portion of my problem was a misunderstanding with how sed uses the 'r' command. There's another part to this solution that came from the accepted answer that I still don't understand, and that is the substitute commands that add backslashes, this was key to making it work. I don't know why it works, but it does.
log_edit() { #Works!!
"$3" > temp.txt
sed -i -n '/^'"$1"'$/ {
:loop
n
/^'"$2"'$/!b loop
i\
'"$(sed 's/\\/\\&/g;s/$/\\/' -- "temp.txt")"'
#Blank line terminates i command.
}
p' "$FILE"
rm temp.txt
}

Answer 1

The r command copies out the file before the next read, not when it is evaluated, and does not modify pattern-space. However the file can be inserted into the script as part of an i command:
log_edit() {
sed -n '/^'"$1"'$/ {
p
:loop
n
/^'"$2"'$/!bloop
i\
'"$("$3" | sed 's/^[[:space:]]/\\&/;s/\\/\\&/g;s/$/\\/')"'
# The blank line above is part of the `i' command,
# and appends a newline to the inserted text.
}
p' "$FILE" > "$FILE.mod" && mv -f -- "$FILE.mod" "$FILE"
}
The command-substitution "$("$3" | sed '...')" filters the output
of $3 for use with sed's i command. The i command prints
a series of lines will all but the last ending with a \.
$ echo three | sed 'i\
> one\
> two
> '
one
two
three

Answer 2

Looks like just a few things out of order there. Try this:
log_edit(){
"$3" > text.tmp
sed -i -n "/$1/{
r text.tmp
:loop
N
/$2/!b loop
s/.*\n/\n\n/g
}
p
" "$FILE"
rm text.tmp
}
print_header(){ #Prints log header.
print_div "Metech ITAMS Log"
printf "Metech Recycling\nITAMS Hardware Report\nDate: $(date)\nTech:%s" "$TECH_INITIALS"
}
print_div(){ #Prints a divider. Required parameter: $1=Text for divider.
printf "\n=== %s ===\n\n" "$1"
}
log_edit "=== Metech ITAMS Log ===" "=== Manufacturer Information ===" "print_header"

Answer 3

Try the csplit program, which can divide a file up into sections according to a pattern:
csplit $3 "/\($1\|$2\)/" "{*}"
This means take file $3, and break it into files xxNN (where NN starts at 00 and goes up) according to sections demarcated by an unlimited number ({*}) of patterns $1 OR $2 (two alternate patterns, separated by \| and grouped by escaped parentheses). The demarcation lines will remain in the output. You can then write ancillary code to delete the files you don't want. You can also change the name of the output filename and pattern.
# cat foo
a
b
#
c
d
%
e
f
#
g
h
%
i
j
# csplit foo '/\(#\|%\)/' '{*}'
4
6
6
6
6
# more xx0*
::::::::::::::
xx00
::::::::::::::
a
b
::::::::::::::
xx01
::::::::::::::
#
c
d
::::::::::::::
xx02
::::::::::::::
%
e
f
::::::::::::::
xx03
::::::::::::::
#
g
h
::::::::::::::
xx04
::::::::::::::
%
i
j
Note: You'll need to tweak it if your demarcation lines can repeat/occur out of order. This is very simple; the breaks occur at any point that one pattern OR the other is seen, regardless of order.