#!/usr/bin/env bash

set -euo pipefail

: "${CARGO:=cargo}"

# A list of the licenses that we currently allow in our code.
#
# If a package supports multiple licenses (using OR), then we are okay
# if it supports _any_ of these licenses.
#
# We don't currently do a good job of understanding AND, so
# interesting license combinations that involve AND may need to be given
# in quotes.
RECOGNIZED_LICENSES=(
    Apache-2.0
    BSD-2-Clause
    BSD-3-Clause
    CC0-1.0
    ISC
    MIT
    Unicode-DFS-2016
    Unlicense
    Zlib
    "(MIT OR Apache-2.0) AND Unicode-DFS-2016"
)

# List of packages that don't list a license.
NO_LICENSE=(
    # The license for "ring" is something like "ISC AND openssl AND
    # ssleay AND MIT"; the openssl license is not up-to-date with
    # modern openssl.  It includes an advertising clause. :P
    #
    # See https://gitlab.torproject.org/tpo/core/arti/-/issues/493 for
    # our related ticket.
    ring
    # License appears to be ISC.
    webpki
    rustls-webpki
)

# List of packages which we allow to use the MPL-2.0 license.
#
# We need to check these individually because, if the party says
# "MPL-2.0" without actually including the text of exhibit A from the
# MPL, it is not unambiguous that they have applied MPL-2.0 to their
# code.
#
# (See https://gitlab.torproject.org/tpo/core/arti/-/issues/845)
MPL_20_OK=(
    # This one is pending: See https://github.com/soc/option-ext/pull/4
    option-ext
    generational-arena
    dynasm
    dynasmrt
)

# List of packages allowed to use the LGPL-3.0-only license.
#
# We aren't including LGPL code in the general Arti dependency tree, this is
# just meant to be a limited whitelist which allows some of our own crates
# we are developing under the LGPL.
LGPL_30_ONLY_OK=(
    equix
    hashx
)

containsElement () {
  local e match="$1"
  shift
  for e; do
      [[ "$e" == "$match" ]] && return 0;
  done
  return 1
}

if ! $CARGO license --help >/dev/null; then
    echo "cargo-license is not installed!"
    echo
    echo "For reasonable results, run:"
    echo "    cargo install cargo-license"
    exit 2
fi

cd "$(dirname "$0")/.."

# The caller might reasonably have set CARGO to something containing spaces.
# So collect the output before we set IFS.
output=$($CARGO license --all-features -t)

problems=0
IFS=$'\n'
for line in $output; do
    package=$(echo "$line" | cut -f1)
    licenses=$(echo "$line" | cut -f5)

    # skip the first line.
    if test "$package" = "name" && test "$licenses" = "license"; then
	continue;
    fi
    if test -z "$licenses"; then
	if ! containsElement "$package" "${NO_LICENSE[@]}"; then
	    echo "$package has no license"
	    problems=1
	fi
	continue
    fi

    if test "$licenses" = "MPL-2.0"; then
	if ! containsElement "$package" "${MPL_20_OK[@]}"; then
	    echo "$package uses MPL-2.0 but has not been allow-listed."
	    problems=1
	fi
	continue
    fi

    if test "$licenses" = "LGPL-3.0-only"; then
	if ! containsElement "$package" "${LGPL_30_ONLY_OK[@]}"; then
	    echo "$package uses LGPL-3.0-only but has not been allow-listed."
	    problems=1
	fi
	continue
    fi

    found_ok=0
    if containsElement "$licenses" "${RECOGNIZED_LICENSES[@]}"; then
	found_ok=1
    else
        # TODO: By Splitting on "OR" without parsing, this can give bogus
        # elements in the output if the license is something like "(A OR
        # B) AND C".  Fortunately the parenthesis will save us from false
        # negatives here, but in the end we should probably switch to a
        # real parser.
        for lic in ${licenses// OR /$'\n'}; do
    	    if containsElement "$lic" "${RECOGNIZED_LICENSES[@]}"; then
		found_ok=1
    		break
	    fi
	done
    fi
    if test $found_ok = "0"; then
	echo "$package does not advertise any supported license!"
	echo "   ($package: $licenses)"
	problems=1
    fi
done

if test "$problems" = 1; then
    echo "You can suppress the above warnings by editing $0..."
    echo "but only do so if we are actually okay with all the licenses!"
fi

exit "$problems"