aboutsummaryrefslogtreecommitdiffstats
path: root/lib/common_test/doc
diff options
context:
space:
mode:
authorErlang/OTP <[email protected]>2017-03-14 15:59:23 +0100
committerErlang/OTP <[email protected]>2017-03-14 15:59:23 +0100
commit4d658008be5a08ddadbe75ebadb9ef124436b76e (patch)
treec60fe1661ae9b747372a146edadbc8af6ca38036 /lib/common_test/doc
parent3b72ab9407945eb7037c608ac2826b64ec17cdc6 (diff)
downloadotp-4d658008be5a08ddadbe75ebadb9ef124436b76e.tar.gz
otp-4d658008be5a08ddadbe75ebadb9ef124436b76e.tar.bz2
otp-4d658008be5a08ddadbe75ebadb9ef124436b76e.zip
Prepare release
Diffstat (limited to 'lib/common_test/doc')
-rw-r--r--lib/common_test/doc/src/notes.xml117
1 files changed, 117 insertions, 0 deletions
diff --git a/lib/common_test/doc/src/notes.xml b/lib/common_test/doc/src/notes.xml
index 83e6511c04..efeacd4a72 100644
--- a/lib/common_test/doc/src/notes.xml
+++ b/lib/common_test/doc/src/notes.xml
@@ -33,6 +33,123 @@
<file>notes.xml</file>
</header>
+<section><title>Common_Test 1.14</title>
+
+ <section><title>Fixed Bugs and Malfunctions</title>
+ <list>
+ <item>
+ <p>The following corrections and improvements are done in
+ the common_test hook handling:</p> <list> <item> <p>An
+ extra argument, <c>Suite</c>, is added as the first
+ argument to each of the following hook callback
+ functions:</p> <list>
+ <item><c>pre_init_per_group</c></item>
+ <item><c>post_init_per_group</c></item>
+ <item><c>pre_end_per_group</c></item>
+ <item><c>post_end_per_group</c></item>
+ <item><c>pre_init_per_testcase</c></item>
+ <item><c>post_init_per_testcase</c></item>
+ <item><c>pre_end_per_testcase</c></item>
+ <item><c>post_end_per_testcase</c></item>
+ <item><c>on_tc_fail</c></item>
+ <item><c>on_tc_skip</c></item> </list> <p>For backwards
+ compatibility, if the new function is not exported from a
+ hook callback module, <c>common_test</c> will fall back
+ to the old interface and call the function without the
+ <c>Suite</c> argument.</p> </item> <item> <p>If either
+ <c>init_per_suite</c> or <c>end_per_suite</c> exists, but
+ not the other, then the non-existing function will be
+ reported as failed with reason <c>undef</c> in the test
+ log. The same goes for <c>init/end_per_group</c>. This
+ has always been a requirement according to the user's
+ guide, but now <c>common_test</c> is more explicit in the
+ report.</p> </item> <item> <p>If <c>init_per_suite</c>
+ was exported from a test suite, but not
+ <c>end_per_suite</c>, then <c>pre/post_end_per_suite</c>
+ was called with <c>Suite=ct_framework</c> instead of the
+ correct suite name. This is now corrected.</p> </item>
+ <item> <p>If <c>end_per_group</c> was exported from a
+ suite, but not <c>init_per_group</c>, then
+ <c>end_per_group</c> was never called. This is now
+ corrected.</p> </item> <item> <p>Tests that were skipped
+ before calling <c>pre_init_per_*</c> got faulty calls to
+ the corresponding <c>post_init_per_*</c>. E.g. if a test
+ was skipped because <c>suite/0</c> failed, then
+ <c>post_init_per_suite</c> would be called even though
+ <c>pre_init_per_suite</c> and <c>init_per_suite</c> were
+ not called. This is now corrected so a <c>post_*</c>
+ callback will never be called unless the corresponding
+ <c>pre_*</c> callback has been called first.</p> </item>
+ <item> <p>Tests that were skipped before or in
+ <c>init_per_testcase</c> got faulty calls to
+ <c>pre_end_per_testcase</c> and
+ <c>post_end_per_testcase</c>. This is now corrected so
+ <c>pre/post_end_per_testcase</c> are not called when
+ <c>end_per_testcase</c> is not called.</p> </item> <item>
+ <p>If an exit signal causes the test case process to die
+ while running <c>init_per_testcase</c>, the case was
+ earlier reported as failed with reason <c>{skip,...}</c>.
+ This is now corrected so the case will be marked as
+ skipped.</p> </item> <item> <p>If an exist signal causes
+ the test case process to die while running
+ <c>end_per_testcase</c>, the case was earlier marked as
+ failed. This is now corrected so the status of the test
+ case is not changed - there is only a warning added to
+ the comment field.</p> </item> <item> <p>If a test case
+ was skipped because of option
+ <c>{force_stop,skip_rest}</c> or because of a failed
+ sequence, then no <c>tc_start</c> event would be sent,
+ only <c>tc_done</c>. This is now corrected so both events
+ are sent.</p> </item> <item> <p>When skipping or failing
+ in a configuration function, the configuration function
+ itself would get <c>{auto_skipped,Reason}</c>,
+ <c>{skipped,Reason}</c> or <c>{failed,Reason}</c> in the
+ hook callbacks <c>on_tc_skip</c> or <c>on_tc_fail</c>.
+ The other test cases that were skipped as a result of
+ this would only get <c>Reason</c> in <c>on_tc_skip</c>.
+ This is now corrected so even the configuration function
+ that caused the skip/fail will only get <c>Reason</c> in
+ the hook callback.</p> </item> </list>
+ <p>
+ Own Id: OTP-10599 Aux Id: kunagi-344 [255] </p>
+ </item>
+ <item>
+ <p>
+ When a test case was skipped by a <c>skip_cases</c>
+ statement in a test spec, then <c>cth_surefire</c> would
+ erroneously mark the previous test case as skipped in the
+ xml report. The actually skipped test case would not be
+ present in the xml report at all. This is now corrected.</p>
+ <p>
+ Own Id: OTP-14129 Aux Id: seq13244 </p>
+ </item>
+ <item>
+ <p>The <c>multiply_timetraps</c> and
+ <c>scale_timetraps</c> options did not work with test
+ specifications, which has been corrected.</p>
+ <p>
+ Own Id: OTP-14210</p>
+ </item>
+ </list>
+ </section>
+
+
+ <section><title>Improvements and New Features</title>
+ <list>
+ <item>
+ <p>
+ ct_testspec:get_tests/1 is added. This is used by rebar3
+ to get all directories that must be compiled when running
+ tests from testspec - instead of implementing testspec
+ parsing in rebar3.</p>
+ <p>
+ Own Id: OTP-14132</p>
+ </item>
+ </list>
+ </section>
+
+</section>
+
<section><title>Common_Test 1.13</title>
<section><title>Fixed Bugs and Malfunctions</title>