Fix xslt_process() to ensure that it inserts a NULL terminator after the

[PostgreSQL.git] / doc / src / sgml / regress.sgml

<!-- $PostgreSQL$ -->

 <chapter id="regress">
  <title id="regress-title">Regression Tests</title>

  <indexterm zone="regress">
   <primary>regression tests</primary>
  </indexterm>

  <indexterm zone="regress">
   <primary>test</primary>
  </indexterm>

  <para>
   The regression tests are a comprehensive set of tests for the SQL
   implementation in <productname>PostgreSQL</productname>.  They test
   standard SQL operations as well as the extended capabilities of
   <productname>PostgreSQL</productname>.
  </para>

  <sect1 id="regress-run">
   <title>Running the Tests</title>

  <para>
   The regression tests can be run against an already installed and
   running server, or using a temporary installation within the build
   tree.  Furthermore, there is a <quote>parallel</quote> and a
   <quote>sequential</quote> mode for running the tests.  The
   sequential method runs each test script in turn, whereas the
   parallel method starts up multiple server processes to run groups
   of tests in parallel.  Parallel testing gives confidence that
   interprocess communication and locking are working correctly.  For
   historical reasons, the sequential test is usually run against an
   existing installation and the parallel method against a temporary
   installation, but there are no technical reasons for this.
  </para>

  <para>
   To run the regression tests after building but before installation,
   type:
<screen>
gmake check
</screen>
   in the top-level directory.  (Or you can change to
   <filename>src/test/regress</filename> and run the command there.)
   This will first build several auxiliary files, such as
   some sample user-defined trigger functions, and then run the test driver
   script.  At the end you should see something like:
<screen>
<computeroutput>
=======================
 All 115 tests passed.
=======================
</computeroutput>
</screen>
   or otherwise a note about which tests failed.  See <xref
   linkend="regress-evaluation"> below before assuming that a
   <quote>failure</> represents a serious problem.
  </para>

   <para>
    Because this test method runs a temporary server, it will not work
    when you are the root user (since the server will not start as root).
    If you already did the build as root, you do not have to start all
    over.  Instead, make the regression test directory writable by
    some other user, log in as that user, and restart the tests.
    For example:
<screen>
<prompt>root# </prompt><userinput>chmod -R a+w src/test/regress</userinput>
<prompt>root# </prompt><userinput>su - joeuser</userinput>
<prompt>joeuser$ </prompt><userinput>cd <replaceable>top-level build directory</></userinput>
<prompt>joeuser$ </prompt><userinput>gmake check</userinput>
</screen>
    (The only possible <quote>security risk</quote> here is that other
    users might be able to alter the regression test results behind
    your back.  Use common sense when managing user permissions.)
   </para>
   <para>
    Alternatively, run the tests after installation.
   </para>

   <para>
    If you have configured <productname>PostgreSQL</productname> to install
    into a location where an older <productname>PostgreSQL</productname>
    installation already exists, and you perform <literal>gmake check</>
    before installing the new version, you might find that the tests fail
    because the new programs try to use the already-installed shared
    libraries.  (Typical symptoms are complaints about undefined symbols.)
    If you wish to run the tests before overwriting the old installation,
    you'll need to build with <literal>configure --disable-rpath</>.
    It is not recommended that you use this option for the final installation,
    however.
   </para>

   <para>
    The parallel regression test starts quite a few processes under your
    user ID.  Presently, the maximum concurrency is twenty parallel test
    scripts, which means forty processes: there's a server process and a
    <application>psql</> process for each test script.
    So if your system enforces a per-user limit on the number of processes,
    make sure this limit is at least fifty or so, else you might get
    random-seeming failures in the parallel test.  If you are not in
    a position to raise the limit, you can cut down the degree of parallelism
    by setting the <literal>MAX_CONNECTIONS</> parameter.  For example:
<screen>
gmake MAX_CONNECTIONS=10 check
</screen>
    runs no more than ten tests concurrently.
   </para>

  <para>
   To run the tests after installation<![%standalone-ignore;[ (see <xref linkend="installation">)]]>,
   initialize a data area and start the
   server, <![%standalone-ignore;[as explained in <xref linkend="runtime">, ]]> then type:
<screen>
gmake installcheck
</screen>
or for a parallel test:
<screen>
gmake installcheck-parallel
</screen>
   The tests will expect to contact the server at the local host and the
   default port number, unless directed otherwise by <envar>PGHOST</envar> and
   <envar>PGPORT</envar> environment variables.
  </para>

  <para>
   The source distribution also contains regression tests for the optional
   procedural languages and for some of the <filename>contrib</> modules.
   At present, these tests can be used only against an already-installed
   server.  To run the tests for all procedural languages that have been
   built and installed, change to the <filename>src/pl</> directory of the
   build tree and type:
<screen>
gmake installcheck
</screen>
   You can also do this in any of the subdirectories of <filename>src/pl</>
   to run tests for just one procedural language.  To run the tests for all
   <filename>contrib</> modules that have them, change to the
   <filename>contrib</> directory of the build tree and type:
<screen>
gmake installcheck
</screen>
   The <filename>contrib</> modules must have been built and installed first.
   You can also do this in a subdirectory of <filename>contrib</> to run
   the tests for just one module.
  </para>
  </sect1>

  <sect1 id="regress-evaluation">
   <title>Test Evaluation</title> 

   <para>
    Some properly installed and fully functional
    <productname>PostgreSQL</productname> installations can
    <quote>fail</quote> some of these regression tests due to
    platform-specific artifacts such as varying floating-point representation
    and message wording. The tests are currently evaluated using a simple
    <command>diff</command> comparison against the outputs
    generated on a reference system, so the results are sensitive to
    small system differences.  When a test is reported as
    <quote>failed</quote>, always examine the differences between
    expected and actual results; you might find that the
    differences are not significant.  Nonetheless, we still strive to
    maintain accurate reference files across all supported platforms,
    so it can be expected that all tests pass.
   </para>

   <para>
    The actual outputs of the regression tests are in files in the
    <filename>src/test/regress/results</filename> directory. The test
    script uses <command>diff</command> to compare each output
    file against the reference outputs stored in the
    <filename>src/test/regress/expected</filename> directory.  Any
    differences are saved for your inspection in
    <filename>src/test/regress/regression.diffs</filename>.  (Or you
    can run <command>diff</command> yourself, if you prefer.)
   </para>

   <para>
    If for some reason a particular platform generates a <quote>failure</>
    for a given test, but inspection of the output convinces you that
    the result is valid, you can add a new comparison file to silence
    the failure report in future test runs.  See
    <xref linkend="regress-variant"> for details.
   </para>

   <sect2>
    <title>Error message differences</title>
      
    <para>
     Some of the regression tests involve intentional invalid input
     values.  Error messages can come from either the
     <productname>PostgreSQL</productname> code or from the host
     platform system routines. In the latter case, the messages can
     vary between platforms, but should reflect similar
     information. These differences in messages will result in a
     <quote>failed</quote> regression test that can be validated by
     inspection.
    </para>
   </sect2>
    
   <sect2>
    <title>Locale differences</title>

    <para>
     If you run the tests against a server that was
     initialized with a collation-order locale other than C, then
     there might be differences due to sort order and follow-up
     failures.  The regression test suite is set up to handle this
     problem by providing alternative result files that together are
     known to handle a large number of locales.
    </para>

    <para>
     To run the tests in a different locale when using the
     temporary-installation method, pass the appropriate
     locale-related environment variables on
     the <command>make</command> command line, for example:
<programlisting>
gmake check LANG=de_DE.utf8
</programlisting>
     (The regression test driver unsets <envar>LC_ALL</envar>, so it
     does not work to choose the locale using that variable.)  To use
     no locale, either unset all locale-related environment variables
     (or set them to <literal>C</literal>) or use the following
     special invocation:
<programlisting>
gmake check NO_LOCALE=1
</programlisting>
     When running the tests against an existing installation, the
     locale setup is determined by the existing installation.  To
     change it, initialize the database cluster with a different
     locale by passing the appropriate options
     to <command>initdb</command>.
    </para>

    <para>
     In general, it is nevertheless advisable to try to run the
     regression tests in the locale setup that is wanted for
     production use, as this will exercise the locale- and
     encoding-related code portions that will actually be used in
     production.  Depending on the operating system environment, you
     might get failures, but then you will at least know what
     locale-specific behaviors to expect when running real
     applications.
    </para>
   </sect2>
    
   <sect2>
    <title>Date and time differences</title>

    <para>
     Most of the date and time results are dependent on the time zone
     environment.  The reference files are generated for time zone
     <literal>PST8PDT</literal> (Berkeley, California), and there will be
     apparent failures if the tests are not run with that time zone setting.
     The regression test driver sets environment variable
     <envar>PGTZ</envar> to <literal>PST8PDT</literal>, which normally
     ensures proper results.
    </para>
   </sect2>
    
   <sect2>
    <title>Floating-point differences</title>
      
    <para>
     Some of the tests involve computing 64-bit floating-point numbers (<type>double
     precision</type>) from table columns. Differences in
     results involving mathematical functions of <type>double
     precision</type> columns have been observed.  The <literal>float8</> and
     <literal>geometry</> tests are particularly prone to small differences
     across platforms, or even with different compiler optimization options.
     Human eyeball comparison is needed to determine the real
     significance of these differences which are usually 10 places to
     the right of the decimal point.
    </para>

    <para>
     Some systems display minus zero as <literal>-0</>, while others
     just show <literal>0</>.
    </para>

    <para>
     Some systems signal errors from <function>pow()</function> and
     <function>exp()</function> differently from the mechanism
     expected by the current <productname>PostgreSQL</productname>
     code.
    </para>
   </sect2>

   <sect2>
    <title>Row ordering differences</title>
      
    <para>
You might see differences in which the same rows are output in a
different order than what appears in the expected file.  In most cases
this is not, strictly speaking, a bug.  Most of the regression test
scripts are not so pedantic as to use an <literal>ORDER BY</> for every single
<literal>SELECT</>, and so their result row orderings are not well-defined
according to the letter of the SQL specification.  In practice, since we are
looking at the same queries being executed on the same data by the same
software, we usually get the same result ordering on all platforms, and
so the lack of <literal>ORDER BY</> isn't a problem.  Some queries do exhibit
cross-platform ordering differences, however.  When testing against an
already-installed server, ordering differences can also be caused by
non-C locale settings or non-default parameter settings, such as custom values
of <varname>work_mem</> or the planner cost parameters.
    </para>

    <para>
Therefore, if you see an ordering difference, it's not something to
worry about, unless the query does have an <literal>ORDER BY</> that your
result is violating.  But please report it anyway, so that we can add an
<literal>ORDER BY</> to that particular query and thereby eliminate the bogus
<quote>failure</quote> in future releases.
    </para>

    <para>
You might wonder why we don't order all the regression test queries explicitly
to get rid of this issue once and for all.  The reason is that that would
make the regression tests less useful, not more, since they'd tend
to exercise query plan types that produce ordered results to the
exclusion of those that don't.
    </para>
   </sect2>

   <sect2>
    <title>Insufficient stack depth</title>
      
    <para>
     If the <literal>errors</literal> test results in a server crash
     at the <literal>select infinite_recurse()</> command, it means that
     the platform's limit on process stack size is smaller than the
     <![%standalone-ignore;[<xref linkend="guc-max-stack-depth">]]>
     <![%standalone-include;[<literal>max_stack_depth</literal>]]>
     parameter indicates.  This
     can be fixed by running the server under a higher stack
     size limit (4MB is recommended with the default value of 
     <varname>max_stack_depth</>).  If you are unable to do that, an
     alternative is to reduce the value of <varname>max_stack_depth</>.
    </para>
   </sect2>

   <sect2>
    <title>The <quote>random</quote> test</title>
      
    <para>
     The <literal>random</literal> test script is intended to produce 
     random results.   In rare cases, this causes the random regression
     test to fail.  Typing:
<programlisting>
diff results/random.out expected/random.out
</programlisting>
     should produce only one or a few lines of differences.  You need
     not worry unless the random test fails repeatedly.
    </para>
   </sect2>
  </sect1>

<!-- We might want to move the following section into the developer's guide. -->
  <sect1 id="regress-variant">
   <title>Variant Comparison Files</title>

   <para>
    Since some of the tests inherently produce environment-dependent
    results, we have provided ways to specify alternative <quote>expected</>
    result files.  Each regression test can have several comparison files
    showing possible results on different platforms.  There are two
    independent mechanisms for determining which comparison file is used
    for each test.
   </para>

   <para>
    The first mechanism allows comparison files to be selected for
    specific platforms.  There is a mapping file,
    <filename>src/test/regress/resultmap</filename>, that defines
    which comparison file to use for each platform.
    To eliminate bogus test <quote>failures</quote> for a particular platform,
    you first choose or make a variant result file, and then add a line to the
    <filename>resultmap</filename> file.
   </para>

   <para>
    Each line in the mapping file is of the form
<synopsis>
testname:output:platformpattern=comparisonfilename
</synopsis>
    The test name is just the name of the particular regression test
    module. The output value indicates which output file to check. For the 
    standard regression tests, this is always <literal>out</literal>. The
    value corresponds to the file extension of the output file.
    The platform pattern is a pattern in the style of the Unix
    tool <command>expr</> (that is, a regular expression with an implicit
    <literal>^</literal> anchor at the start).  It is matched against the
    platform name as printed by <command>config.guess</command>.
    The comparison file name is the base name of the substitute result
    comparison file.
   </para>

   <para>
    For example: some systems interpret very small floating-point values
    as zero, rather than reporting an underflow error.  This causes a
    few differences in the <filename>float8</> regression test.
    Therefore, we provide a variant comparison file,
    <filename>float8-small-is-zero.out</filename>, which includes
    the results to be expected on these systems.  To silence the bogus
    <quote>failure</quote> message on <systemitem>OpenBSD</systemitem>
    platforms, <filename>resultmap</filename> includes:
<programlisting>
float8:out:i.86-.*-openbsd=float8-small-is-zero.out
</programlisting>
    which will trigger on any machine for which the output of
    <command>config.guess</command> matches <literal>i.86-.*-openbsd</literal>.
    Other lines
    in <filename>resultmap</> select the variant comparison file for other
    platforms where it's appropriate.
   </para>

   <para>
    The second selection mechanism for variant comparison files is
    much more automatic: it simply uses the <quote>best match</> among
    several supplied comparison files.  The regression test driver
    script considers both the standard comparison file for a test,
    <literal><replaceable>testname</>.out</>, and variant files named
    <literal><replaceable>testname</>_<replaceable>digit</>.out</>
    (where the <replaceable>digit</> is any single digit
    <literal>0</>-<literal>9</>).  If any such file is an exact match,
    the test is considered to pass; otherwise, the one that generates
    the shortest diff is used to create the failure report.  (If
    <filename>resultmap</filename> includes an entry for the particular
    test, then the base <replaceable>testname</> is the substitute
    name given in <filename>resultmap</filename>.)
   </para>

   <para>
    For example, for the <literal>char</literal> test, the comparison file
    <filename>char.out</filename> contains results that are expected
    in the <literal>C</> and <literal>POSIX</> locales, while
    the file <filename>char_1.out</filename> contains results sorted as
    they appear in many other locales.
   </para>

   <para>
    The best-match mechanism was devised to cope with locale-dependent
    results, but it can be used in any situation where the test results
    cannot be predicted easily from the platform name alone.  A limitation of
    this mechanism is that the test driver cannot tell which variant is
    actually <quote>correct</> for the current environment; it will just pick
    the variant that seems to work best.  Therefore it is safest to use this
    mechanism only for variant results that you are willing to consider
    equally valid in all contexts.
   </para>
    
  </sect1>

  <sect1 id="regress-coverage">
   <title>Test Coverage Examination</title>

   <para>
    The PostgreSQL source code can be compiled with coverage testing
    instrumentation, so that it becomes possible to examine which
    parts of the code are covered by the regression tests or any other
    test suite that is run with the code.  This is currently supported
    when compiling with GCC and requires the <command>gcov</command>
    and <command>lcov</command> programs.
   </para>

   <para>
    A typical workflow would look like this:
<screen>
./configure --enable-coverage ... OTHER OPTIONS ...
gmake
gmake check # or other test suite
gmake coverage-html
</screen>
    Then point your HTML browser
    to <filename>coverage/index.html</filename>.
   </para>

   <para>
    To reset the execution counts between test runs, run:
<screen>
gmake coverage-clean
</screen>
   </para>
  </sect1>

</chapter>