How to pass string literal containing newlines to grep from bash script - bash

Question

I am trying to pass the "strings" from a file as input to grep using the -F (fixed string) parameter.
From grep the man page, the expected format is newline-separated:
-F, --fixed-strings
Interpret PATTERN as a list of fixed strings (instead of regular expressions), separated by newlines, any of which is to be matched.
How can this be done in bash? I have:
#!/bin/bash
INFILE=$1
DIR=$2
# Create a newline-separated string array
STRINGS="";
while read -r string; do
STRINGS+=$'\n'$string;
done < <(strings $INFILE);
cd $DIR
for file in *; do
grep -Frn \"$STRINGS\" .
done;
But grep reports error at run-time regarding input formatting. Grep is interpreting the passed string arguments as parameters -- hence the need to pass them as one large string literal.
Debugging bash with -xand passing the first parameter (INFILE) as the script itself gives:
+ grep -Frn '"' '#!/bin/bash' 'INFILE=$1' 'DIR=$2' [...]

Answer 1

Try the following:
#!/bin/bash
inFile=$1
dir=$2
# Read all lines output by `string` into a single variable using
# a command substitution, $(...).
# Note that the trailing newlines is trimmed, but grep still recognizes
# the last line.
strings="$(strings "$inFile")"
cd "$dir"
for file in *; do
grep -Frn "$strings" .
done
string outputs each string found in the target file on its own line, so you can use its output as-is, via a command substitution ($(...)).
On a side note: strings is used to extract strings from binary files, and strings are only included if they're at least 4 ASCII(!) characters long and are followed by a newline or NUL.
Note that while the POSIX spec for strings does mandate locale-awareness with respect to character interpretation, both GNU strings and BSD/macOS strings recognize 7-bit ASCII characters only.
If, by contrast, your search strings come from a text file from which you want to strip empty and blank lines, use strings="$(awk 'NF>0' "$inFile")"
Double-quote your variable references and command substitutions to ensure that their values are used as-is.
Do not use \" unless you want to pass a literal " char. to the target command - as opposed to an unquoted one that has syntactical meaning to the shell.
In your particular case, \"$STRINGS\" breaks down as follows:
An unquoted reference to variable $STRINGS - because the enclosing " are \-escaped and therefore literals.
The resulting string - "<value-of-$STRINGS>" - due to $STRINGS being unquoted, is then subject to word-splitting
(and globbing), i.e., split into multiple arguments by whitespace. As a result, because grep expects the search term(s) as a single argument, the command breaks.
Do not use all-uppercase shell variable names in order to avoid conflicts with environment variables and special shell variables.