7fc3b3ce76531995f0978618fbb5c5e1fccf31cf braney Sun Sep 20 06:58:05 2026 -0700 hgcentralregress: a central database that tests may write to Every hgcentral a developer can reach is shared. hgcentraltest holds thousands of rows of other people's sandbox sessions, and the production centrals are not reachable from hgwdev at all, so a test that needs to create a session, a login or an api key has had nowhere to put it. Every test written for #38391 so far only reads, and that is the limit this lifts. makeHgCentralRegress.sh creates the database and its tables, copying each schema from the central the caller's hg.conf already names, so a column added to namedSessionDb reaches it on the next run and there is no second copy of the schema in the tree to forget. It is idempotent and meant to run before a test. ONE GRANT IS STILL MISSING and no test uses the database yet. The account the browser uses for the central has global SELECT, INSERT, CREATE and DROP, but UPDATE and DELETE only on databases it has been granted them on individually. So a test can create rows here and cannot take them down: the delete comes back 1142. The script says what to ask for. Measured rather than assumed, and one guess along the way was wrong: the hgcentraltest prefix carries no wildcard grant, hgcentraltestregress behaves the same as any other new name. hubSpaceKeysTester is written against it and is deliberately NOT in the test target, so nothing here goes red while it waits. It covers the rules an api key lives by -- one key per user, a new key revokes the old, an adopted key from a peer mirror replaces both, a revoked key names nobody -- and the signature a peer checks before accepting a sync. refs #38323. refs #38391 diff --git src/hg/lib/tests/makeHgCentralRegress.sh src/hg/lib/tests/makeHgCentralRegress.sh new file mode 100755 index 00000000000..666b4ea0a98 --- /dev/null +++ src/hg/lib/tests/makeHgCentralRegress.sh @@ -0,0 +1,96 @@ +#!/bin/bash +# makeHgCentralRegress.sh - make the central database that tests are allowed to write to. +# +# Every hgcentral a developer has is shared: hgcentraltest holds thousands of rows of other +# people's sandbox sessions, and the production centrals are not reachable from here at all. +# So a test that needs to create a session, a login or an api key has had nowhere to put it, +# and the tests written so far can only read. That is the limit this lifts. refs #38391. +# +# hgcentraltestregress exists for tests only. Nothing else reads it, no CGI is configured to +# use it, and anything in it may be deleted by the next run, so a test must create the rows it +# needs and clean up after itself rather than expect to find anything. +# +# IT NEEDS A GRANT BEFORE ANY TEST CAN USE IT, and this script cannot make one. The account +# the browser uses for the central holds global SELECT, INSERT, CREATE and DROP, but UPDATE +# and DELETE only on the databases it has been granted them on one at a time. So a test can +# create this database and insert into it, and cannot clean up after itself: the delete comes +# back 1142. Measured, not assumed -- and the guess that the hgcentraltest prefix carried a +# wildcard grant is wrong, hgcentraltestregress behaves exactly the same way. +# +# What to ask the admins for, once, against whatever account central.user names: +# +# GRANT UPDATE, DELETE ON hgcentralregress.* TO '<central.user>'@'localhost'; +# +# Until that exists, a test pointed here can set rows up but not take them down, so the tests +# that need it are held back rather than left failing. +# +# Schemas are copied from the central the caller's hg.conf already names, with CREATE TABLE +# LIKE, rather than kept as a second copy of the schema in the tree. A column added to +# namedSessionDb therefore reaches this database the next time the script runs, and there is +# no second definition to forget. +# +# Idempotent: safe to run before every test, and it is. Existing tables are left alone, +# along with whatever rows they hold. +# +# Usage: +# makeHgCentralRegress.sh [conf] +# conf defaults to $HGDB_CONF, then to ~/.hg.conf. The central.host, central.user and +# central.password in it are used, so the account needs CREATE on the new database; on hgwdev +# the ordinary hgtestuser has it. + +set -e + +conf="${1:-${HGDB_CONF:-$HOME/.hg.conf}}" +if [ ! -r "$conf" ]; then + echo "makeHgCentralRegress.sh: cannot read $conf" >&2 + exit 1 +fi + +# An hg.conf may include another one, so follow one level of include to find the settings. +# Later lines win in hg.conf, and so they do here. +confValue() { + local key="$1" + local value="" + local inc + while read -r inc; do + if [ -r "$inc" ]; then + local v + v=$(grep "^$key=" "$inc" 2>/dev/null | tail -1 | cut -d= -f2-) + [ -n "$v" ] && value="$v" + fi + done < <(grep '^include ' "$conf" 2>/dev/null | sed 's/^include //') + local own + own=$(grep "^$key=" "$conf" 2>/dev/null | tail -1 | cut -d= -f2-) + [ -n "$own" ] && value="$own" + echo "$value" +} + +host=$(confValue central.host) +user=$(confValue central.user) +password=$(confValue central.password) +source=$(confValue central.db) +target=hgcentralregress + +if [ -z "$host" ] || [ -z "$user" ] || [ -z "$source" ]; then + echo "makeHgCentralRegress.sh: $conf has no central.host/user/db" >&2 + exit 1 +fi +if [ "$source" = "$target" ]; then + echo "makeHgCentralRegress.sh: central.db is already $target; nothing to copy from" >&2 + exit 1 +fi + +# The tables tests have needed so far. Add to this list rather than creating a table by hand, +# so the next person's run has it too. +tables="apiKeys namedSessionDb sessionDb userDb gbMembers gbNode genark" + +sql() { + mysql -h "$host" -u "$user" -p"$password" -N -B -e "$1" +} + +sql "create database if not exists $target" +for t in $tables; do + sql "create table if not exists $target.$t like $source.$t" +done + +echo "$target is ready on $host, with: $tables"