2
0
mirror of https://github.com/xcat2/xcat-core.git synced 2026-11-20 01:12:18 +00:00
Files
xcat-core/xCAT-test/bats
Daniel Hilst 23afd1bae6 fix(xcat-core): a syslog target with a port is bracketed as if it were an IPv6 literal
site.master may carry a port. format_syslog_forward_target in
xCAT/postscripts/syslog bracketed any value containing a colon, so
MASTER=127.0.0.1:1515 rendered the rsyslog line as @[127.0.0.1:1515], which
rsyslog does not accept, and syslog_ng_udp_driver chose udp6 for an IPv4 peer.
format_syslog_server in the dracut_105 rsyslogd-start.sh carries its own copy of
the rule and had the same defect; it runs inside the initramfs and cannot source
the postscript, so the rule is held down twice on purpose.

An IPv4 address or a hostname with a port has exactly one colon and an IPv6
literal has at least two, so the test is now on two colons rather than one. A
value already starting with "[" is left alone whether or not a port follows it,
which the old pattern missed: it required a trailing "]" and so double-bracketed
[2001:db8::1]:514.

ipv6_syslog_forwarding.bats gains four cases. Without the fix, "a target with a
port is not mistaken for an IPv6 literal", "an already-bracketed target with a
port is left alone" and "the initrd launcher does not mistake a port for an IPv6
literal" all fail; the compressed-literal case passes either way and is there to
show the bracketing of ::1 did not regress.

Also: grub2_ipv6.t could not load grub2.pm once xcat-core/#7980 made it use
xCAT::DHCP::BootPolicy, which lives in perl-xCAT. The file died at compile time
and reported no plan at all, which in a prove summary reads as a count
discrepancy rather than a failure. It now puts perl-xCAT and the server perl
directory on @INC, as every other plugin test does.

Signed-off-by: Daniel Hilst <392820+dhilst@users.noreply.github.com>
2026-10-23 10:38:37 -03:00
..

xCAT-test/bats

Shell-script unit tests live here and run with:

bats -r xCAT-test/bats

The GitHub Actions xcat_test workflow runs this command after the Perl .t unit tests. Use BATS for shell behavior that can be exercised from the source tree without an installed xCAT, a live management node, or real services.

Prefer sourcing an existing shell library or sourceable script and calling the function under test. Keep reusable install-template helpers in xCAT-server/share/xcat/install/scripts/scriptlib, and reusable postscript helpers in xCAT/postscripts/xcatlib.sh. Use scratch directories and shadowed commands so tests cannot write to the host.

Extraction helpers in helpers/shell_source.bash are only for legacy code that cannot safely be sourced yet. Do not add Perl .t tests that grep shell source when the behavior can be tested with BATS.

Postscript sandbox prerequisites

otherpkgs_upgrade_scope.bats and routeop.bats run complete postscripts in Linux filesystem and network namespaces. Their /etc writes stay inside test fixtures. They require Bubblewrap, Bash, GNU core utilities, grep, sed, and permission to create user namespaces. The CI workflow installs Bubblewrap and enables those namespaces.

Missing prerequisites fail these tests without stopping unrelated test files. Non-Linux hosts report a skip. Set TMPDIR to a writable, executable filesystem if the default temporary directory is mounted with noexec.

go_xcat_os_release.bats, go_xcat_activation.bats and the full-script cases in postscripts_apt_get.bats use the same namespace isolation. They require bats-core 1.4 or newer. OS detection also requires GNU awk for the legacy Red Hat release parser. Postscript cases require GNU diffutils.

networkmanager_postscript.bats renders complete installation postscripts in private namespaces. It requires Bubblewrap, bats-core 1.4 or newer, user namespaces and the xCAT Perl dependencies, including DBI and DBD::SQLite.

go_xcat_os_release.bats and go_xcat_el_repo_check.bats source the complete go-xcat script. OS detection uses mawk when available, otherwise the host AWK. The two legacy Red Hat version cases use GNU awk. CI installs both interpreters. Package-manager commands are test doubles, so these cases check selection, arguments and error handling, not repository availability or package installation.

Genesis startup

genesis_ib_modules.bats and genesis_secondary_dhcp.bats run the complete doxcat script in a private filesystem and network namespace. They replace external services and device commands, then check the boot interface and the IPv4 and IPv6 DHCP requests. They do not contact a DHCP server or a BMC.

The startup cases require Linux, bats-core 1.4 or newer, Bubblewrap, Bash, GNU core utilities, grep, sed, awk and permission to create user namespaces. Missing Linux prerequisites fail these cases. Non-Linux hosts skip them. The RPM prerequisite check requires rpmspec but does not use the sandbox. Run containers with an init process to reap children created by deliberate-fault tests.

Genesis sandbox prerequisites

genesis_bmcsetup_users.bats and genesis_getadapter.bats run the complete checkout scripts with private filesystem, process, and network namespaces. They replace hardware and transport commands, not the scripts under test. The fixtures do not contact a BMC or transmit an adapter request.

These tests require Linux, bats-core 1.4 or newer, Bubblewrap, Bash, GNU core utilities, awk, grep, sed, and user namespace support. The adapter tests also require cmp from diffutils. The CI environment provides these dependencies. Missing prerequisites fail on Linux; non-Linux hosts report a skip. The temporary filesystem must allow executable files.

The BMC cases check LAN values, device and channel selection, write retries, configuration defaults, user selection, and vendor port commands. They assert the commands issued by the complete script. They do not verify that physical BMC firmware accepts those commands.

Boot artifact and repository checks

mknb_genesis_compression.bats, dhcp_http_port.bats and go_xcat_common_repository.bats require Linux, Bubblewrap and user namespaces. They execute checkout code in isolated filesystems without contacting services. Non-Linux hosts skip these cases; missing Linux prerequisites fail them. The compression cases also need xz, gzip and cpio. They run as an ordinary user so directory-permission failures can be exercised. Both lzma and xz selection paths build real images, and the decoder checks the LZMA format.