bash read multi-line string with Process Substitution

[georgalis]

I would like to read a multi-line string, with proper exit code. Here Document, and Here String methods are not optimal, for performance, Process Substitution is desired, eg

IFS= read -d "" lines < <(printf "one \ntwo \nthree \n") && echo success || echo error

"works" however it returns an error status.

The following reads the string as expected, with an exit status 0, GNU bash, version 5.2.15(1)-release

eot="$(awk 'BEGIN{printf "%c",4}')"
IFS= read -d "$eot" lines < <(printf "one \ntwo \nthree \n${eot}")
echo "$lines"
one
two
three

Two issues here, 1) the awk statement is the only way I know to reference the ascii EOT character (od could probably work too). Is there any builtin method to generate characters from their numeric equivalent? 2) Shouldn't read have an option to gracefully complete the read, on end-of-transmission, or end of file (ie, the first example)? Typically it would be inconvenient (and with performance impact) to inject this ${eot} with my Process Substitution. I am inclined to simply ignore the exit status. Is there a better option?

[andry81]

@georgalis

https://linuxcommand.org/lc3_man_pages/readh.html

 -d delim continue until the first character of DELIM is read, rather
      than newline

Exit Status: The return code is zero, unless end-of-file is encountered, read times out (in which case it's greater than 128), a variable assignment error occurs, or an invalid file descriptor is supplied as the argument to -u.

The return code is not zero because the end-of-file is reached. The end-of-line is reached first because the delimiter is not found first.

This is by design. Why to bother?