Temporary file and directory creator for node.js

Temporary directory management for Rust

JUnit Jupiter `@TempDir` extension based on the in-memory file system Jimfs

Flysystem adapter for a self-destroying temporary directory

RAII-wrappers for temporary files and directories for modern C++

A Nim library to create and manage temporary directories. Inspired by the Rust library of the same name: https://github.com/rust-lang-nursery/tempdir

Create a temporary directory

An implementation of Python's tempfile.tempdir algorithm in C

No description available

Simple Elixir Library for creating self-cleaning tmp directories

Simple creation and automatic deletion of temporary directories in an easy to use crystal shard

Adds support for a predictable and repeatable temporary directory

Tempdir like Tempfile class.

ttempdir detects temporary directories not using t.TempDir

tempdirs that stick around when tests fail

Tempfile like behaviour for directories

No description available

Provides a temporary directory object, that will kill remove the whole directory on destruction or shutdown.

TempDir is a lightweight C++17 library designed to provide an easy-to-use solution for managing temporary directories and files.

Utility that lets you write to a temporary directory, that automatically cleans up after itself.

Build temporary directories and files with YAML definition.

Easy temporary directories associated with packages

Capture changes performed by system() calls in a tempdir

A Memoryfilesystem based @TempDir provider for JUnit

Inflate any archive to a temporary directory and work in it.

A thin wrapper around libc's mkstemps and mkdtemp

A group of intelligent DI services that provide disk paths to important locations such as tempDir, wwwDir, vendorDir, logDir and other system constants for your applications. Dependency management can be easily solved from one place.

#!/usr/bin/env bash #/ Usage: ghe-backup-repositories #/ Take an online, incremental snapshot of all Git repository data. #/ #/ Note: This command typically isn't called directly. It's invoked by #/ ghe-backup. set -e # This command is designed to allow for transferring active Git repository data # from a GitHub instance to a backup site in a way that ensures data is # captured in a consistent state even when being written to. # # - All Git GC operations are disabled on the GitHub instance for the duration of # the backup. This removes the possibly of objects or packs being removed # while the backup is in progress. # # - In progress Git GC operations are given a cooldown window to complete. The # script will sleep for up to 60 seconds waiting for GC operations to finish. # # - Git repository data is transferred in a specific order: auxiliary files, # packed refs, loose refs, reflogs, and finally objects and pack files in that # order. This ensures that all referenced objects are captured. # # - Git GC operations are re-enabled on the GitHub instance. # # The script uses multiple runs of rsync to transfer repository files. Each run # includes a list of filter rules that ensure only specific types of files are # transferred. # # See the "FILTER RULES" and "INCLUDE/EXCLUDE PATTERN RULES" sections of the # rsync(1) manual for more information: # <http://rsync.samba.org/ftp/rsync/rsync.html> # Bring in the backup configuration # shellcheck source=share/github-backup-utils/ghe-backup-config . "$( dirname "${BASH_SOURCE[0]}" )/ghe-backup-config" bm_start "$(basename $0)" # Set up remote host and root backup snapshot directory based on config host="$GHE_HOSTNAME" backup_dir="$GHE_SNAPSHOT_DIR/repositories" # Location of last good backup for rsync --link-dest backup_current="$GHE_DATA_DIR/current/repositories" # Verify rsync is available. if ! rsync --version 1>/dev/null 2>&1; then echo "Error: rsync not found." 1>&2 exit 1 fi # Perform a host-check and establish GHE_REMOTE_XXX variables. ghe_remote_version_required "$host" # Split host:port into parts port=$(ssh_port_part "$GHE_HOSTNAME") host=$(ssh_host_part "$GHE_HOSTNAME") # Add user / -l option user="${host%@*}" [ "$user" = "$host" ] && user="admin" hostnames=$host ssh_config_file_opt= tempdir=$(mktemp -d -t backup-utils-backup-XXXXXX) remote_tempdir=$(ghe-ssh "$GHE_HOSTNAME" -- mktemp -d -t backup-utils-backup-XXXXXX) routes_list=$tempdir/routes_list remote_routes_list=$remote_tempdir/remote_routes_list opts="$GHE_EXTRA_SSH_OPTS" # git server hostnames under cluster if [ "$GHE_BACKUP_STRATEGY" = "cluster" ]; then ssh_config_file="$tempdir/ssh_config" ssh_config_file_opt="-F $ssh_config_file" opts="$opts -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -o PasswordAuthentication=no" hostnames=$(ghe-cluster-nodes "$GHE_HOSTNAME" "git-server") ghe-ssh-config "$GHE_HOSTNAME" "$hostnames" > "$ssh_config_file" fi # Make sure root backup dir exists if this is the first run mkdir -p "$backup_dir" # Removes the remote sync-in-progress file on exit, re-enabling GC operations # on the remote instance. cleanup() { for pid in $(jobs -p); do kill -KILL $pid > /dev/null 2>&1 || true done # Enable remote GC operations for hostname in $hostnames; do ghe-gc-enable $ssh_config_file_opt $hostname:$port || true done ghe-ssh "$GHE_HOSTNAME" -- rm -rf $remote_tempdir rm -rf $tempdir } trap 'cleanup' EXIT trap 'exit $?' INT # ^C always terminate # Disable remote GC operations for hostname in $hostnames; do ghe-gc-disable $ssh_config_file_opt $hostname:$port done # If we have a previous increment, avoid transferring existing files via rsync's # --link-dest support. This also decreases physical space usage considerably. if [ -d "$backup_current" ]; then link_dest="--link-dest=../../current/repositories" fi # Calculate sync routes. This will store the healthy repo paths for each node # # This gets a repo path and stores the path in the $node.sync file # a/nw/a8/3f/02/100000855 dgit-node1 >> dgit-node1.sync # a/nw/a8/bc/8d/100000880 dgit-node3 >> dgit-node3.sync # a/nw/a5/06/81/100000659 dgit-node2 >> dgit-node2.sync # ... # One route per line. # # NOTE: The route generation is performed on the appliance as it is considerably # more performant than performing over an SSH pipe. # bm_start "$(basename $0) - Generating routes" echo "github-env ./bin/dgit-cluster-backup-routes > $remote_routes_list" | ghe-ssh "$GHE_HOSTNAME" -- /bin/bash ghe-ssh "$GHE_HOSTNAME" -- cat $remote_routes_list | ghe_debug bm_end "$(basename $0) - Generating routes" bm_start "$(basename $0) - Fetching routes" ghe-ssh "$GHE_HOSTNAME" -- gzip -c $remote_routes_list | gzip -d > $routes_list cat $routes_list | ghe_debug bm_end "$(basename $0) - Fetching routes" bm_start "$(basename $0) - Processing routes" if [ "$GHE_BACKUP_STRATEGY" != "cluster" ]; then server=$host fi cat $routes_list | awk -v tempdir="$tempdir" -v server="$server" '{ for(i=2;i<=NF;i++){ server != "" ? host=server : host=$i; print $1 > (tempdir"/"host".rsync") }}' ghe_debug "\n$(find "$tempdir" -maxdepth 1 -name '*.rsync')" bm_end "$(basename $0) - Processing routes" if [ -z "$(find "$tempdir" -maxdepth 1 -name '*.rsync')" ]; then echo "Warning: no routes found, skipping repositories backup ..." exit 0 fi # Transfer repository data from a GitHub instance to the current snapshot # directory, using a previous snapshot to avoid transferring files that have # already been transferred. A set of rsync filter rules are provided on stdin # for each invocation. rsync_repository_data () { port=$(ssh_port_part "$1") host=$(ssh_host_part "$1") #check if we are syncing from a given file list if [[ "$2" == *".rsync" ]]; then files_list="$2" shift shift ghe-rsync -avr \ -e "ssh -q $opts -p $port $ssh_config_file_opt -l $user" \ $link_dest "$@" \ --rsync-path='sudo -u git rsync' \ --include-from=- --exclude=\* \ --files-from="$files_list" \ --ignore-missing-args \ "$host:$GHE_REMOTE_DATA_USER_DIR/repositories/" \ "$backup_dir" 1>&3 2>&3 else shift ghe-rsync -avr \ -e "ssh -q $opts -p $port $ssh_config_file_opt -l $user" \ $link_dest "$@" \ --rsync-path='sudo -u git rsync' \ --include-from=- --exclude=\* \ --ignore-missing-args \ "$host:$GHE_REMOTE_DATA_USER_DIR/repositories/" \ "$backup_dir" 1>&3 2>&3 fi } sync_data (){ # Sync all auxiliary repository data. This includes files and directories like # HEAD, audit_log, config, description, info/, etc. No refs or object data # should be transferred here. echo 1>&3 echo "* Transferring auxiliary files ..." 1>&3 rsync_repository_data $1:122 $2 -z <<RULES - /__*__/ - /info/ + /*/ + /*/*.git - /*/*.git/objects - /*/*.git/refs - /*/*.git/packed-refs - /*/*.git/logs + /*/*.git/** + /*/??/ + /*/??/??/ + /*/??/??/??/ + /*/??/??/??/gist/ + /*/??/??/??/gist/*.git - /*/??/??/??/gist/*.git/objects - /*/??/??/??/gist/*.git/refs - /*/??/??/??/gist/*.git/packed-refs - /*/??/??/??/gist/*.git/logs + /*/??/??/??/gist/*.git/** + /*/nw/??/??/??/ + /*/nw/??/??/??/*/ + /*/nw/??/??/??/*/*.git - /*/nw/??/??/??/*/*.git/objects - /*/nw/??/??/??/*/*.git/refs - /*/nw/??/??/??/*/*.git/packed-refs - /*/nw/??/??/??/*/*.git/logs + /*/nw/??/??/??/*/*.git/** RULES # Sync packed refs files. This is performed before sync'ing loose refs since # loose refs trump packed-refs information. echo 1>&3 echo "* Transferring packed-refs files ..." 1>&3 rsync_repository_data $1:122 $2 -z <<RULES - /__*__/ - /info/ + /*/ + /*/*.git + /*/*.git/packed-refs + /*/??/ + /*/??/??/ + /*/??/??/??/ + /*/??/??/??/gist/ + /*/??/??/??/gist/*.git + /*/??/??/??/gist/*.git/packed-refs + /*/nw/??/ + /*/nw/??/??/ + /*/nw/??/??/??/ + /*/nw/??/??/??/*/ + /*/nw/??/??/??/*/*.git + /*/nw/??/??/??/*/*.git/packed-refs RULES # Sync loose refs and reflogs. This must be performed before object data is # transferred to ensure that all referenced objects are included. echo 1>&3 echo "* Transferring refs and reflogs ..." 1>&3 rsync_repository_data $1:122 $2 -z <<RULES - /__*__/ - /info/ + /*/ + /*/*.git + /*/*.git/refs + /*/*.git/refs/** + /*/*.git/logs + /*/*.git/logs/** + /*/??/ + /*/??/??/ + /*/??/??/??/ + /*/??/??/??/gist/ + /*/??/??/??/gist/*.git + /*/??/??/??/gist/*.git/refs + /*/??/??/??/gist/*.git/refs/** + /*/??/??/??/gist/*.git/logs + /*/??/??/??/gist/*.git/logs/** + /*/nw/??/ + /*/nw/??/??/ + /*/nw/??/??/??/ + /*/nw/??/??/??/*/ + /*/nw/??/??/??/*/*.git + /*/nw/??/??/??/*/*.git/refs + /*/nw/??/??/??/*/*.git/refs/** + /*/nw/??/??/??/*/*.git/logs + /*/nw/??/??/??/*/*.git/logs/** RULES # Sync git objects and pack files. Compression is disabled during this phase # since these files are already well compressed. echo 1>&3 echo "* Transferring objects and packs ..." 1>&3 rsync_repository_data $1:122 $2 -H <<RULES - /__*__/ - /info/ + /*/ + /*/*.git + /*/*.git/objects - /*/*.git/objects/**/tmp_* + /*/*.git/objects/** + /*/??/ + /*/??/??/ + /*/??/??/??/ + /*/??/??/??/gist/ + /*/??/??/??/gist/*.git + /*/??/??/??/gist/*.git/objects - /*/??/??/??/gist/*.git/objects/**/tmp_* + /*/??/??/??/gist/*.git/objects/** + /*/nw/??/ + /*/nw/??/??/ + /*/nw/??/??/??/ + /*/nw/??/??/??/*/ + /*/nw/??/??/??/*/*.git + /*/nw/??/??/??/*/*.git/objects - /*/nw/??/??/??/*/*.git/objects/**/tmp_* + /*/nw/??/??/??/*/*.git/objects/** RULES echo 1>&3 } # rsync all the repositories bm_start "$(basename $0) - Repo sync" for file_list in $tempdir/*.rsync; do hostname=$(basename $file_list .rsync) repo_num=$(cat $file_list | wc -l) ghe_verbose "* Transferring $repo_num repositories from $hostname" sync_data $hostname $file_list & done for pid in $(jobs -p); do wait $pid done bm_end "$(basename $0) - Repo sync" # Since there are no routes for special data directories, we need to do this # serially for all hostnames. Good candidate for future optimizations. bm_start "$(basename $0) - Special Data Directories Sync" for h in $hostnames; do # Sync __special__ data directories, including the __alambic_assets__, # __hookshot__, and __purgatory__ directories. The __nodeload_archives__, # __gitmon__, and __render__ directories are excludes since they act only as # caches. # # Under v2.x and greater, only the special __purgatory__ directory remains under # /data/repositories. All other special user data directories have been moved under # the /data/user directory. echo 1>&3 echo "* Transferring special data directories from $h..." 1>&3 rsync_repository_data $h:122 -z <<RULES - /__nodeload_archives__/ - /__gitmon__/ - /__render__/ + /__*__/ + /__*__/** + /info/ - /info/lost+found/ + /info/* RULES echo 1>&3 done bm_end "$(basename $0) - Special Data Directories Sync" if [ -z "$GHE_SKIP_ROUTE_VERIFICATION" ]; then bm_start "$(basename $0) - Verifying Routes" cat $tempdir/*.rsync | uniq | sort | uniq > $tempdir/source_routes (cd $backup_dir/ && find * -mindepth 5 -maxdepth 6 -type d -name \*.git | fix_paths_for_ghe_version | uniq | sort | uniq) > $tempdir/destination_routes git --no-pager diff --unified=0 --no-prefix -- $tempdir/source_routes $tempdir/destination_routes || echo "Warning: One or more repository networks and/or gists were not found on the source appliance. Please contact GitHub Enterprise Support for assistance." bm_end "$(basename $0) - Verifying Routes" fi bm_end "$(basename $0)"

No description available

:open_file_folder: :snake: temp.tempdir(), temp.tempfile() functions