[gentoo-commits] proj/sci:master commit in: scripts/, .github/workflows/

["Andrew Ammerlaan" <andrewammerlaan@riseup.net>]

commit:     c8511a14fec6323b99300f881502ba49c9878b32
Author:     Andrew Ammerlaan <andrewammerlaan <AT> riseup <DOT> net>
AuthorDate: Wed Feb 17 19:38:52 2021 +0000
Commit:     Andrew Ammerlaan <andrewammerlaan <AT> riseup <DOT> net>
CommitDate: Wed Feb 17 19:38:52 2021 +0000
URL:        https://gitweb.gentoo.org/proj/sci.git/commit/?id=c8511a14

.github/workflows/duplicates.yml: add script to check for duplicates

Signed-off-by: Andrew Ammerlaan <andrewammerlaan <AT> riseup.net>

 .github/workflows/duplicates.yml | 25 +++++++++++++
 scripts/check-duplicates.sh      | 76 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 101 insertions(+)

diff --git a/.github/workflows/duplicates.yml b/.github/workflows/duplicates.yml
new file mode 100644
index 000000000..2db452228
--- /dev/null
+++ b/.github/workflows/duplicates.yml
@@ -0,0 +1,25 @@
+name: duplicates
+
+on:
+  push:
+    branches: [ dev ]
+  pull_request:
+    branches: [ dev ]
+  schedule:
+    - cron: '0 0 * * *'
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v2
+    - name: Setup master gentoo repository
+      run: |
+        sudo mkdir -p /var/db/repos/gentoo /etc/portage /var/cache/distfiles
+        wget -qO - "https://github.com/gentoo-mirror/gentoo/archive/master.tar.gz" | sudo tar xz -C /var/db/repos/gentoo --strip-components=1
+        sudo wget "https://www.gentoo.org/dtd/metadata.dtd" -O /var/cache/distfiles/metadata.dtd
+        sudo wget "https://gitweb.gentoo.org/proj/portage.git/plain/cnf/repos.conf" -O /etc/portage/repos.conf
+        sudo ln -s /var/db/repos/gentoo/profiles/default/linux/amd64/17.1 /etc/portage/make.profile
+    - name: Check for duplicates
+      run: |
+        ./scripts/check-duplicates.sh

diff --git a/scripts/check-duplicates.sh b/scripts/check-duplicates.sh
new file mode 100755
index 000000000..8f6b89af7
--- /dev/null
+++ b/scripts/check-duplicates.sh
@@ -0,0 +1,76 @@
+#!/usr/bin/env bash
+# Maintainer: Andrew Ammerlaan <andrewammerlaan@riseup.net>
+#
+# This checks if packages in ::guru are also in ::gentoo
+#
+# Note that this is not going to be 100% accurate
+#
+#
+
+printf "\nChecking for duplicates....\n"
+
+gentoo_location="/var/db/repos/gentoo"
+guru_location="."
+
+gentoo_packs=$(find ${gentoo_location} -mindepth 2 -maxdepth 2 -printf "%P\n" | sort | grep -Ev "^(.git|.github|metadata|profiles|scripts)/")
+guru_packs=$(find ${guru_location} -mindepth 2 -maxdepth 2 -printf "%P\n" | sort | grep -Ev "^(.git|.github|metadata|profiles|scripts)/")
+
+pack_overrides="" pack_close_match_in_cat="" pack_close_match=""
+for guru_pack in ${guru_packs}; do
+	# separate category and packages
+	guru_pack_cat="${guru_pack%%/*}"
+	guru_pack_name="${guru_pack##*/}"
+
+	# convert all to lowercase
+	guru_pack_name="${guru_pack_name,,}"
+
+	# stip all numbers, dashes, underscores and pluses
+	guru_pack_name="${guru_pack_name/[0-9-_+]}"
+
+	for gentoo_pack in ${gentoo_packs}; do
+		# separate category and packages
+		gentoo_pack_cat="${gentoo_pack%%/*}"
+		gentoo_pack_name="${gentoo_pack##*/}"
+
+		# convert all to lowercase
+		gentoo_pack_name="${gentoo_pack_name,,}"
+
+		# stip all numbers, dashes, underscores and pluses
+		gentoo_pack_name="${gentoo_pack_name/[0-9-_+]}"
+
+		#TODO: check DESCRIPTION, HOMEPAGE and SRC_URI for close matches
+
+		if [[ "${gentoo_pack_name}" == "${guru_pack_name}" ]]; then
+			if [[ "${gentoo_pack_cat}" == "${guru_pack_cat}" ]]; then
+				if [[ "${gentoo_pack}" == "${guru_pack}" ]]; then
+					pack_overrides+="\t${guru_pack}::guru exact match of ${gentoo_pack}::gentoo\n"
+				else
+					pack_close_match_in_cat+="\t${guru_pack}::guru possible duplicate of ${gentoo_pack}::gentoo\n"
+				fi
+			else
+				pack_close_match+="\t${guru_pack}::guru possible duplicate of ${gentoo_pack}::gentoo\n"
+			fi
+		fi
+	done
+done
+
+if [ -n "${pack_close_match}" ]; then
+	printf "\nWARNING: The following packages closely match packages in the main Gentoo repository\n"
+	printf "${pack_close_match}"
+	printf "Please check these manually\n"
+fi
+
+if [ -n "${pack_close_match_in_cat}" ]; then
+	printf "\nWARNING: The following packages closely match packages in the main Gentoo repository in the same category\n"
+	printf "${pack_close_match_in_cat}"
+	printf "Please check these manually\n"
+fi
+
+if [ -n "${pack_overrides}" ]; then
+	printf "\nERROR: The following packages override packages in the main Gentoo repository\n"
+	printf "${pack_overrides}"
+	printf "Please remove these packages\n"
+	# do not exit fatally on ::science
+	# exit 1
+fi
+exit 0