Hi, IMO, I do not like split a repo into several repos... Reasons: 1. For users, they have to download the repos one by one. (git submodule can solve the issue more or less, but git submodule is not very easy to use) 2. For new developers, it brings new steps for them to load all the sub projects into one project, and debug them among subprojects.. 3. the issue management and Pull request will be decentralized, which is also not a good idea..
BTW, 400,000 lines of codes are changes, not the alive codes, right? Maybe we can consider how to prune the the repo history and branches to reduce the repo size. Best, ----------------------------------- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Dawei Liu <[email protected]> 于2021年5月21日周五 下午6:37写道： > Hi all, > > > After rapid development, IOTDB code has reached 400,000 + lines > roughly[1], > which is a good news, but there are also some problems, > which will make our compilation process extremely slow. > > > Therefore, I propose to split the existing project once. > If you have any idea, please reply to me. > My preliminary idea is : > > > * IoTDB-Connector: session, jdbc, hive-connector, spark-connector, > spark-tsfile, flink-connector, flink-tsfile, Hadoop > > > * IoTDB-client-go > > > * IoTDB-Client: client-cpp, client-py, IoTDB-client-go(sub) > > > * IoTDB-Example > > > * IoTDB-Visual: grafana, zeppelin > > > * IoTDB: antlr, cli, cluster, server, server-roc, thrift, thrift-cluster, > thrift-sync, tsfile, testcontainer, docker > > > > > > > > > > > [1] git log --pretty=tformat: --numstat | awk '{ add += $1; subs += $2; > loc += $1 - $2 } END { printf "added lines: %s, removed lines: %s, total > lines: %s\n", add, subs, loc }' - > >
