Create .merlin file for a project with all your ocamlfind packages and .opam sources in there

merlin-init.sh

#!/bin/sh

if test -f .merlin; then

    echo ".merlin already exists, bailing out ..." >&2
    exit 1

else

    # You could add your default EXT's and such to this list:
    cat >.merlin <<EOF
S .
B _build
EOF

    # Add PKG's:
    ocamlfind list \
	| awk '{ print "PKG "$1 }' >>.merlin

    # See https://github.com/the-lambda-church/merlin/wiki/Letting-merlin-locate-go-to-stuff-in-.opam
    find ~/.opam -name '*.cmt' -print0 \
	| xargs -0 -I{} dirname '{}' \
	| sort -u \
	| awk '{ print "S "$0"\nB "$0 }' >> .merlin

fi

Very useful – thanks!

Loading all packages is a bit of an overkill and makes my editor take very long to load. I'm using it as follows:

merlin-init.sh | grep endian >> .merlin

And changed the script to print to stdout:

#!/bin/sh

# Add PKG's:
ocamlfind list \
    | awk '{ print "PKG "$1 }'

# See https://github.com/the-lambda-church/merlin/wiki/Letting-merlin-locate-go-to-stuff-in-.opam
find ~/.opam -name '*.cmt' -print0 \
    | xargs -0 -I{} dirname '{}' \
    | sort -u \
    | awk '{ print "S "$0"\nB "$0 }'

Super useful. Thanks to both of you!

Here is a slightly updated version that works with opam switches.

#!/bin/sh

eval $(opam config env)

# Add PKG's:
ocamlfind list \
    | awk '{ print "PKG "$1 }'

# See https://github.com/the-lambda-church/merlin/wiki/Letting-merlin-locate-go-to-stuff-in-.opam
find "$OPAM_SWITCH_PREFIX" -name '*.cmt' -print0 \
    | xargs -0 -I{} dirname '{}' \
    | sort -u \
    | awk '{ print "S "$0"\nB "$0 }'