<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>33.2. Test Evaluation</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><link rel="prev" href="regress-run.html" title="33.1. Running the Tests" /><link rel="next" href="regress-variant.html" title="33.3. Variant Comparison Files" /></head><body><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">33.2. Test Evaluation</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="regress-run.html" title="33.1. Running the Tests">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="regress.html" title="Chapter 33. Regression Tests">Up</a></td><th width="60%" align="center">Chapter 33. Regression Tests</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 11.5 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="regress-variant.html" title="33.3. Variant Comparison Files">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="REGRESS-EVALUATION"><div class="titlepage"><div><div><h2 class="title" style="clear: both">33.2. Test Evaluation</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="regress-evaluation.html#id-1.6.20.6.6">33.2.1. Error Message Differences</a></span></dt><dt><span class="sect2"><a href="regress-evaluation.html#id-1.6.20.6.7">33.2.2. Locale Differences</a></span></dt><dt><span class="sect2"><a href="regress-evaluation.html#id-1.6.20.6.8">33.2.3. Date and Time Differences</a></span></dt><dt><span class="sect2"><a href="regress-evaluation.html#id-1.6.20.6.9">33.2.4. Floating-Point Differences</a></span></dt><dt><span class="sect2"><a href="regress-evaluation.html#id-1.6.20.6.10">33.2.5. Row Ordering Differences</a></span></dt><dt><span class="sect2"><a href="regress-evaluation.html#id-1.6.20.6.11">33.2.6. Insufficient Stack Depth</a></span></dt><dt><span class="sect2"><a href="regress-evaluation.html#id-1.6.20.6.12">33.2.7. The <span class="quote">“<span class="quote">random</span>”</span> Test</a></span></dt><dt><span class="sect2"><a href="regress-evaluation.html#id-1.6.20.6.13">33.2.8. Configuration Parameters</a></span></dt></dl></div><p>
Some properly installed and fully functional
<span class="productname">PostgreSQL</span> installations can
<span class="quote">“<span class="quote">fail</span>”</span> 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
<code class="command">diff</code> comparison against the outputs
generated on a reference system, so the results are sensitive to
small system differences. When a test is reported as
<span class="quote">“<span class="quote">failed</span>”</span>, 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.
</p><p>
The actual outputs of the regression tests are in files in the
<code class="filename">src/test/regress/results</code> directory. The test
script uses <code class="command">diff</code> to compare each output
file against the reference outputs stored in the
<code class="filename">src/test/regress/expected</code> directory. Any
differences are saved for your inspection in
<code class="filename">src/test/regress/regression.diffs</code>.
(When running a test suite other than the core tests, these files
of course appear in the relevant subdirectory,
not <code class="filename">src/test/regress</code>.)
</p><p>
If you don't
like the <code class="command">diff</code> options that are used by default, set the
environment variable <code class="envar">PG_REGRESS_DIFF_OPTS</code>, for
instance <code class="literal">PG_REGRESS_DIFF_OPTS='-u'</code>. (Or you
can run <code class="command">diff</code> yourself, if you prefer.)
</p><p>
If for some reason a particular platform generates a <span class="quote">“<span class="quote">failure</span>”</span>
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
<a class="xref" href="regress-variant.html" title="33.3. Variant Comparison Files">Section 33.3</a> for details.
</p><div class="sect2" id="id-1.6.20.6.6"><div class="titlepage"><div><div><h3 class="title">33.2.1. Error Message Differences</h3></div></div></div><p>
Some of the regression tests involve intentional invalid input
values. Error messages can come from either the
<span class="productname">PostgreSQL</span> 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
<span class="quote">“<span class="quote">failed</span>”</span> regression test that can be validated by
inspection.
</p></div><div class="sect2" id="id-1.6.20.6.7"><div class="titlepage"><div><div><h3 class="title">33.2.2. Locale Differences</h3></div></div></div><p>
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 subsequent
failures. The regression test suite is set up to handle this
problem by providing alternate result files that together are
known to handle a large number of locales.
</p><p>
To run the tests in a different locale when using the
temporary-installation method, pass the appropriate
locale-related environment variables on
the <code class="command">make</code> command line, for example:
</p><pre class="programlisting">
make check LANG=de_DE.utf8
</pre><p>
(The regression test driver unsets <code class="envar">LC_ALL</code>, 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 <code class="literal">C</code>) or use the following
special invocation:
</p><pre class="programlisting">
make check NO_LOCALE=1
</pre><p>
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 <code class="command">initdb</code>.
</p><p>
In general, it is 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.
</p></div><div class="sect2" id="id-1.6.20.6.8"><div class="titlepage"><div><div><h3 class="title">33.2.3. Date and Time Differences</h3></div></div></div><p>
Most of the date and time results are dependent on the time zone
environment. The reference files are generated for time zone
<code class="literal">PST8PDT</code> (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
<code class="envar">PGTZ</code> to <code class="literal">PST8PDT</code>, which normally
ensures proper results.
</p></div><div class="sect2" id="id-1.6.20.6.9"><div class="titlepage"><div><div><h3 class="title">33.2.4. Floating-Point Differences</h3></div></div></div><p>
Some of the tests involve computing 64-bit floating-point numbers (<code class="type">double
precision</code>) from table columns. Differences in
results involving mathematical functions of <code class="type">double
precision</code> columns have been observed. The <code class="literal">float8</code> and
<code class="literal">geometry</code> tests are particularly prone to small differences
across platforms, or even with different compiler optimization settings.
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.
</p><p>
Some systems display minus zero as <code class="literal">-0</code>, while others
just show <code class="literal">0</code>.
</p><p>
Some systems signal errors from <code class="function">pow()</code> and
<code class="function">exp()</code> differently from the mechanism
expected by the current <span class="productname">PostgreSQL</span>
code.
</p></div><div class="sect2" id="id-1.6.20.6.10"><div class="titlepage"><div><div><h3 class="title">33.2.5. Row Ordering Differences</h3></div></div></div><p>
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 <code class="literal">ORDER BY</code> for every single
<code class="literal">SELECT</code>, and so their result row orderings are not well-defined
according to 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,
so the lack of <code class="literal">ORDER BY</code> is not 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 <code class="varname">work_mem</code> or the planner cost parameters.
</p><p>
Therefore, if you see an ordering difference, it's not something to
worry about, unless the query does have an <code class="literal">ORDER BY</code> that your
result is violating. However, please report it anyway, so that we can add an
<code class="literal">ORDER BY</code> to that particular query to eliminate the bogus
<span class="quote">“<span class="quote">failure</span>”</span> in future releases.
</p><p>
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.
</p></div><div class="sect2" id="id-1.6.20.6.11"><div class="titlepage"><div><div><h3 class="title">33.2.6. Insufficient Stack Depth</h3></div></div></div><p>
If the <code class="literal">errors</code> test results in a server crash
at the <code class="literal">select infinite_recurse()</code> command, it means that
the platform's limit on process stack size is smaller than the
<a class="xref" href="runtime-config-resource.html#GUC-MAX-STACK-DEPTH">max_stack_depth</a> parameter indicates. This
can be fixed by running the server under a higher stack
size limit (4MB is recommended with the default value of
<code class="varname">max_stack_depth</code>). If you are unable to do that, an
alternative is to reduce the value of <code class="varname">max_stack_depth</code>.
</p><p>
On platforms supporting <code class="function">getrlimit()</code>, the server should
automatically choose a safe value of <code class="varname">max_stack_depth</code>;
so unless you've manually overridden this setting, a failure of this
kind is a reportable bug.
</p></div><div class="sect2" id="id-1.6.20.6.12"><div class="titlepage"><div><div><h3 class="title">33.2.7. The <span class="quote">“<span class="quote">random</span>”</span> Test</h3></div></div></div><p>
The <code class="literal">random</code> test script is intended to produce
random results. In very rare cases, this causes that regression
test to fail. Typing:
</p><pre class="programlisting">
diff results/random.out expected/random.out
</pre><p>
should produce only one or a few lines of differences. You need
not worry unless the random test fails repeatedly.
</p></div><div class="sect2" id="id-1.6.20.6.13"><div class="titlepage"><div><div><h3 class="title">33.2.8. Configuration Parameters</h3></div></div></div><p>
When running the tests against an existing installation, some non-default
parameter settings could cause the tests to fail. For example, changing
parameters such as <code class="varname">enable_seqscan</code> or
<code class="varname">enable_indexscan</code> could cause plan changes that would
affect the results of tests that use <code class="command">EXPLAIN</code>.
</p></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="regress-run.html">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="regress.html">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="regress-variant.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">33.1. Running the Tests </td><td width="20%" align="center"><a accesskey="h" href="index.html">Home</a></td><td width="40%" align="right" valign="top"> 33.3. Variant Comparison Files</td></tr></table></div></body></html>
