[SOLVED] How to read multi-line properties from text files in bash with grep

Issue

I want to read properties from a build file using grep 3.1 in bash (gitbash, actually). Most properties in the file have single line value, some have no value while others like REFLIBS contains multiline value.

REFLIBS=module.o testModule1.o libutil.a \
                               libdebug.a \
                               libxfor.a
SOURCEFILES=client.c
FLAGS=
HEADERS=linkedlist.h
#OLD_REFLIBS  ignore lines with bash-style comments
#HEADERS=utils.h

Problem

My grep command: grep -i -m 1 -E '^REFLIBS=' test.txt

The command captures only up to the first \, but my goal is to concatenate all lines into a single line like this:

REFLIBS_CAPTURE=$(grep -i -m 1 -E '^REFLIBS=' test.txt)
echo $REFLIBS_CAPTURE
(desired output)
> module.o testModule1.o libutil.a libdebug.a libxfor.a

Before posting I searched again and found this post suggesting to alter default IFS input delimiter from '\n' to something else, in my case \. Looks interesting but not sure if this is the right solution. https://stackoverflow.com/a/44112055/791406

Solution

You can use sed to delete each backlash-newline sequence, and then match only a single line with grep:

sed -z 's/\\\n *//g' < test.txt | grep -i -m 1 -E '^REFLIBS='

which outputs

REFLIBS=module.o testModule1.o libutil.a libdebug.a libxfor.a

Note that GNU sed is required for the -z option, which treats lines as being NUL-separated instead of newline separated. The means that you effectively read the entire input file into memory as a single "line" for sed to operate on.

Answered By - Brian61354270

Answer Checked By - Timothy Miller (WPSolving Admin)