+ 2011-08-09 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ tests: don't let a known Solaris make bug poison too many tests
+ See automake bug#7670 and bug#7824.
+ * tests/defs (make_can_chain_suffix_rules): New function, tell
+ whether the make implementation in use can chain suffix rules
+ automatically or not.
+ * tests/suffix3.test: Use it to avoid issuing calls to make that
+ are unportable to make implementations that are not smart enough
+ to chain suffix rules automatically.
+ * tests/suffix8.test: Use it to avoid issuing calls to make that
+ * tests/suffix10.test: Use it to avoid issuing calls to make that
+ * tests/suffix11.test: Use it to avoid issuing calls to make that
+ * tests/suffix-chain.test: New test, exposes the limitation that
+ we have papered over in the tests above.
+
+ 2011-08-09 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ gitignore: more use of anchors
+ * .gitignore: Anchor files that are intended to be ignored only
+ if found in the same directory of the `.gitignore' file, not also
+ in its subdirectories.
+ * doc/.gitignore, doc/amhello/.gitignore, lib/Automake/.gitignore,
+ lib/Automake/tests/.gitignore, tests/.gitignore: Likewise. Also,
+ where needed, add new entries that were once implied by the
+ non-anchored entries in the upper-level `.gitignore' files.
+
2011-08-09 Stefano Lattarini <stefano.lattarini@gmail.com>
- testsuite: avoid spurious failure of self checks on Solaris
- This is just a quick workaround to prevent having too much noise
- in the testsuite; more proper fixes have been committed into the
- 'testsuite-work' branch.
- * tests/self-check-dir.test: Relax a test to avoid unimportant
- failure on Solaris 10.
- * tests/self-check-exit.test: Likewise.
+ self tests: avoid spurious output on console, and related fixlets
+ * tests/self-check-cleanup.test: Set `$stderr_fileno_' to `2'
+ in the self tests, to avoid spurious diagnostic on console.
+ * tests/self-check-exit.tap: Only unset `am_explicit_skips' and
+ `$stderr_fileno_' in the self tests, not at the global level, it
+ is more correct.
+ ($init): New variable, to reduce code duplication.
+ Some refactoring and adjustments to the name of the tests.
+
+2011-08-09 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ test defs: few more sanity checks
+ * tests/defs: Catch some more (very unlikely) set-up or internal
+ errors, and diagnose them explicitly. This is intended mostly
+ to be a theoretic-only improvement.
+
+2011-08-09 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ test defs: really make them "set -e" clean
+ Motivated by a spurious failure of test `instsh3-w.shtst' on
+ Solaris 10.
+ * tests/defs: Make really "set -e" clean, to avoid spurious
+ failures in case this file is sourced by a test script that has
+ already set the `errexit' flag. To be sure we don't regress,
+ move the `set -e' call much earlier, and make the comment telling
+ about the expected "set -e" cleanliness more loud and outstanding.
+
+2011-08-09 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ self tests: register an expected failures with Solaris /bin/sh
+ Solaris 10 /bin/sh erroneously exit with success right away when
+ the following three conditions are met at the same time:
+ 1. the `errexit' flag is active,
+ 2. an exit trap is installed, and
+ 3. a non-existing command is issued.
+ * tests/self-exit.tap: When that bug is detected, issue an XFAIL
+ rather than a FAIL. Since we are at it, improve by avoiding
+ creation of stray temporary files when testing for non-executable
+ commands.
+
+2011-08-09 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ gitignore: update and improve for the testsuite files
+ * tests/.gitignore: Don't ignore temporary partial log files,
+ autogenerated tests and makefile fragments that are not used
+ anymore. Anchor all the remaining patterns.
+
+2011-08-09 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ self tests: some new uses of TAP, few extensions
+ * tests/self-check-me.test: Renamed ...
+ * tests/self-check-me.tap: .. to this, and converted to use TAP.
+ * tests/self-check-exit.test: Renamed ...
+ * tests/self-check-exit.tap: .. to this, and converted to use TAP.
+ Set `$stderr_fileno_' to `2', to avoid spurious messages on the
+ console. Add a new test, and relax another one.
+ * tests/self-check-env-sanitize.test: Renamed ...
+ * tests/self-check-env-sanitize.tap: .. to this, and converted to
+ the use TAP.
+ * tests/Makefile.am (TESTS): Update.
+
+2011-08-09 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ test defs: test directory creation more configurable
+ After this change, the individual tests will be able to tell
+ `./defs' whether or not to create a temporary test directory
+ for them, and if yes, whether or not to pre-populate it.
+ * tests/defs: Only create the temporary test directory if the
+ variable `$am_create_testdir' is not set to "no". In this
+ case, only pre-populate it if that variable is not set to
+ `$empty'.
+ * tests/defs-static.in: Check that `$am_create_testdir' is not
+ exported in the environment.
+ * tests/self-check-env-sanitize.test: Update.
+ * tests/self-check-dir.test: Extend, and make deeper and more
+ robust.
+ * tests/README (Section "Writing test cases" subsection "Do"):
+ Update.
+ * tests/autodist.test: Set `$am_create_testdir' instead of
+ removing by hand files created by ./defs.
+ * tests/tap-doc2.test: Likewise.
+
+2011-08-08 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ testsuite: more uses of the `seq_' function
+ * tests/parallel-tests-exit-statuses.test: Use the `seq_' function
+ from `tests/defs', instead of duplicating its semantics.
+ * tests/testsuite-summary-count.test: Likewise.
+
+2011-08-08 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ testsuite: drop redundant environment sanitization
+ * tests/Makefile.am (TESTS_ENVIRONMENT): Don't remove `planned'
+ from the environment.
+ * tests/defs-static.in: Don't complain if `planned' is set in
+ the environment.
+ * tests/self-check-env-sanitize.test: Update.
+
+2011-08-08 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ test defs: fix bug in plain version of `skip_all_'
+ * tests/plain-functions.sh (skip_all_): Call `skip_', not `skip'.
+
+2011-08-08 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ test defs: small cleanups and tweakings
+ * tests/defs: Where possible, prefer `framework_failure_' or
+ `fatal_' to `Exit 99'. Try to catch more possible hard errors
+ and setup failures. Do not hide expected error messages. Avoid
+ a couple of forks when copying files from `$(top_srcdir)/lib'.
+ ($signal): Remove, no more needed.
+
+2011-08-08 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ testsuite: improve interface for giving early TAP plan
+ Since commit `v1.11-976-g4740f56', an early TAP plan can also be
+ specified after input lines have already been seen, as long as
+ no one of them is a TAP result line. We can now take advantage
+ of this in our own testsuite, allowing the TAP-based scripts to
+ specify their earlier plan after having included `./defs'. In
+ particular, this offers the benefit of allowing those scripts to
+ issue a "skip all" plan (e.g., "1..0 # SKIP reason ...") after
+ the inclusion of `./defs', which wasn't possible with the older
+ interface (introduced in commit `v1.11-1114-g0286308').
+ * tests/defs: Don't try to call `plan_' if `$planned' is defined,
+ the calling scripts can do that directly now.
+ * tests/ac-output-old.tap: Use `plan_' directly instead of
+ setting `$planned' before including `./defs'.
+ * tests/automake-cmdline.tap: Likewise.
+ * tests/depmod.tap: Likewise.
+ * tests/instspc.tap: Likewise.
+ * tests/tap-bad-prog.tap: Likewise.
+ * tests/add-missing.tap: Don't set `$planned'.
+
+2011-08-08 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ test defs: fix bug in TAP version of `skip_all_'
+ * tests/tap-functions (skip_all_): Set `$have_tap_plan_' to "yes"
+ when called, so that the exit trap won't write another TAP plan.
+
+2011-08-08 Stefano Lattarini <stefano.lattarini@gmail.com>
+
+ test defs: fix skip directives to work also with TAP-based tests
+ * test/defs: Use `skip_all_', not `skip_', to issue a skip of the
+ whole calling test script.
2011-08-08 Stefano Lattarini <stefano.lattarini@gmail.com>
return 1
fi;;
*)
- echo "invalid value for \$am__using_gmake: '$am__using_gmake'" >&2
- Exit 99;;
+ fatal_ "invalid value for \$am__using_gmake: '$am__using_gmake'";;
esac
}
+am__using_gmake="" # Avoid interferences from the environment.
-
+ # make_can_chain_suffix_rules
+ # ---------------------------
+ # Return 0 if $MAKE is a make implementation that can chain suffix rules
+ # automatically, return 1 otherwise. Caches the result for speed reasons.
+ make_can_chain_suffix_rules ()
+ {
+ if test -z "$am__can_chain_suffix_rules"; then
+ if using_gmake; then
+ am__can_chain_suffix_rules=yes
+ return 0
+ else
+ mkdir am__chain.dir$$
+ cd am__chain.dir$$
+ unindent > Makefile << 'END'
+ .u.v: ; cp $< $@
+ .v.w: ; cp $< $@
+ END
+ echo make can chain suffix rules > foo.u
+ if $MAKE foo.w && diff foo.u foo.w; then
+ am__can_chain_suffix_rules=yes
+ else
+ am__can_chain_suffix_rules=no
+ fi
+ cd ..
+ rm -rf am__chain.dir$$
+ fi
+ fi
+ case $am__can_chain_suffix_rules in
+ yes) return 0;;
+ no) return 1;;
+ *) fatal_ "make_can_chain_suffix_rules: internal error";;
+ esac
+ }
+ am__can_chain_suffix_rules="" # Avoid interferences from the environment.
+
+# seq_ - print a sequence of numbers
+# ----------------------------------
+# This function simulates GNU seq(1) portably. Valid usages:
+# - seq LAST
+# - seq FIRST LAST
+# - seq FIRST INCREMENT LAST
+seq_ ()
+{
+ case $# in
+ 0) fatal_ "seq_: missing argument";;
+ 1) seq_first=1 seq_incr=1 seq_last=$1;;
+ 2) seq_first=$1 seq_incr=1 seq_last=$2;;
+ 3) seq_first=$1 seq_incr=$2 seq_last=$3;;
+ *) fatal_ "seq_: too many arguments";;
+ esac
+ # Try to avoid forks if possible.
+ case "$BASH_VERSION" in
+ ""|[12].*)
+ : Not bash, or a too old bash version. ;;
+ *)
+ # Use eval to protect dumber shells from parsing errors.
+ eval 'for ((i = seq_first; i <= seq_last; i += seq_incr)); do
+ echo $i
+ done'
+ return 0;;
+ esac
+ # Else, use GNU seq if available.
+ seq "$@" && return 0
+ # Otherwise revert to a slower loop using expr(1).
+ i=$seq_first
+ while test $i -le $seq_last; do
+ echo $i
+ i=`expr $i + $seq_incr`
+ done
+}
+
+# rm_rf_ [FILES OR DIRECTORIES ...]
+# ---------------------------------
+# Recursively remove the given files or directory, also handling the case
+# of non-writable subdirectories.
+rm_rf_ ()
+{
+ test $# -gt 0 || return 0
+ # Ignore failures in find, we are only interested in failures of the
+ # final rm.
+ find "$@" -type d ! -perm -700 -exec chmod u+rwx {} \; || :
+ rm -rf "$@"
+}
+
# count_test_results total=N pass=N fail=N xpass=N xfail=N skip=N error=N
# -----------------------------------------------------------------------
# Check that a testsuite run driven by the parallel-tests harness has
projects / thirdparty / automake.git / commitdiff
