#!/bin/bash # This script downloads sources and modifies them set -ex NAME="mariadb" VERSION=`rpmspec -q --srpm --qf '%{VERSION}' "$NAME.spec" ` # SOURCES_URL=`spectool -s 0 "$NAME.spec" | cut -d ' ' -f 2 ` SOURCES_URL="https://downloads.mariadb.org/interstitial/mariadb-$VERSION/source/mariadb-$VERSION.tar.gz" OLD_ARCHIVE_NAME="${NAME}-${VERSION}" NEW_ARCHIVE_NAME="${NAME}-${VERSION}-downstream_modified" # Retrieve the archive: rm -rf "${OLD_ARCHIVE_NAME}.tar.gz" "${OLD_ARCHIVE_NAME}/" "${NEW_ARCHIVE_NAME}.tar.gz" "${NEW_ARCHIVE_NAME}/" wget "$SOURCES_URL" # Modify the archive: # 1/ Change both the name of the archive and the name of the base directory inside of the archive # It will be necessary to change the name in the SPECfile in the %prep phase # This will prevent maintainer to rebase to a non-modified sources archive without changing the SPECfile tar -xof "${OLD_ARCHIVE_NAME}.tar.gz" mv "${OLD_ARCHIVE_NAME}" "${NEW_ARCHIVE_NAME}" # 2/ Remove the code licensed under the PerconaFT license # which was not yet reviewed as suitable for Fedora or RHEL. # # License file: # storage/tokudb/PerconaFT/PATENTS # # The whole storage engine, which requires code under this license # has to be removed before uploading sources to Fedora. rm -r "${NEW_ARCHIVE_NAME}/storage/tokudb" # Pack the extracted files back to the archive tar -czf "${NEW_ARCHIVE_NAME}.tar.gz" "${NEW_ARCHIVE_NAME}" # Remove the decompressed original used to create the archive rm -r "${NEW_ARCHIVE_NAME}/"