Skip to content

P
Postgres FD Implementation
Commit 0db3cb25 authored by Peter Eisentraut's avatar Peter Eisentraut
Browse files
Options

* doc/src/sgml/regress.sgml: Update for new driver script. 

* doc/src/sgml/installation.sgml: ditto.

* src/test/regress/README: Regenerate.

* doc/src/sgml/docguide.sgml: Explain how it was done.  Explain how
INSTALL and HISTORY are (now) generated.

* doc/src/sgml/Makefile: Implement HISTORY generation to be analoguous
to INSTALL.
parent f7b89ac5
Show whitespace changes
Inline Side-by-side
Showing with 598 additions and 780 deletions
doc/src/sgml/Makefile
View file @ 0db3cb25
...@@ -8,7 +8,7 @@ ...@@ -8,7 +8,7 @@
# #
# #
# IDENTIFICATION # IDENTIFICATION
# $Header: /cvsroot/pgsql/doc/src/sgml/Makefile,v 1.23 2000/10/10 22:01:50 momjian Exp $ # $Header: /cvsroot/pgsql/doc/src/sgml/Makefile,v 1.24 2000/10/17 15:26:39 petere Exp $
# #
#---------------------------------------------------------------------------- #----------------------------------------------------------------------------
...@@ -201,21 +201,30 @@ distclean: ...@@ -201,21 +201,30 @@ distclean:
cp -p ../graphics/$@ . cp -p ../graphics/$@ .
# Generation of the INSTALL text file. Not fully automated, but better #
# than nothing. # Semi-automatic generation of some text files.
.PHONY: INSTALL #
INSTALL: INSTALL.html
INSTALL HISTORY: % : %.html
@echo "|";\ @echo "|";\
echo "| You should now take \`$<', save it as a text file in Netscape,";\ echo "| You should now take \`$<', save it as a text file in Netscape,";\
echo "| and put it in place of the existing \`INSTALL' file.";\ echo "| and put it in place of the existing \`$@' file.";\
echo "|" echo "|"
@rm -f tempfile.html tempfile.sgml
INSTALL.html: tempfile.html
sed -e 's/Chapter 1. *//g' < $< > $@
tempfile.html: tempfile.sgml INSTALL.html HISTORY.html: %.html : tempfile_%.html
jade -d $(HDSL) -V nochunks -t sgml $< > $@ sed 's/Chapter 1. *//g' $< >$@
tempfile_INSTALL.html tempfile_HISTORY.html: tempfile_%.html : tempfile_%.sgml
jade -d $(HDSL) -V nochunks -t sgml $< >$@
tempfile_INSTALL.sgml: standalone-install.sgml installation.sgml
cat $+ >$@
tempfile_HISTORY.sgml: release.sgml
( echo '<!doctype chapter PUBLIC "-//OASIS//DTD DocBook V3.1//EN">'; \
cat $< ) >$@
tempfile.sgml: standalone-install.sgml installation.sgml .INTERMEDIATE: tempfile_INSTALL.html tempfile_HISTORY.html tempfile_INSTALL.sgml tempfile_HISTORY.sgml
cat $+ > $@
doc/src/sgml/docguide.sgml
View file @ 0db3cb25
<!-- <!--
$Header: /cvsroot/pgsql/doc/src/sgml/docguide.sgml,v 1.27 2000/09/29 20:21:33 petere Exp $ $Header: /cvsroot/pgsql/doc/src/sgml/docguide.sgml,v 1.28 2000/10/17 15:26:40 petere Exp $
Documentation Guide Documentation Guide
Thomas Lockhart Thomas Lockhart
--> -->
...@@ -83,7 +83,7 @@ Thomas Lockhart ...@@ -83,7 +83,7 @@ Thomas Lockhart
</row> </row>
<row> <row>
<entry>./INSTALL</entry> <entry>./INSTALL</entry>
<entry>Installation instructions (text from sgml->rtf->text)</entry> <entry>Installation instructions</entry>
</row> </row>
<row> <row>
<entry>./README</entry> <entry>./README</entry>
...@@ -848,6 +848,7 @@ End: ...@@ -848,6 +848,7 @@ End:
</sect2> </sect2>
</sect1> </sect1>
<sect1 id="doc-build"> <sect1 id="doc-build">
<title>Building Documentation</title> <title>Building Documentation</title>
...@@ -911,9 +912,8 @@ PSTYLE= /home/lockhart/SGML/db143.d/docbook/print ...@@ -911,9 +912,8 @@ PSTYLE= /home/lockhart/SGML/db143.d/docbook/print
% make install % make install
</programlisting> </programlisting>
</para> </para>
</sect1>
<sect1 id="doc-manpages"> <sect2 id="doc-manpages">
<title>Manpages</title> <title>Manpages</title>
<para> <para>
...@@ -966,9 +966,9 @@ $ make man ...@@ -966,9 +966,9 @@ $ make man
</para> </para>
</step> </step>
</procedure> </procedure>
</sect1> </sect2>
<sect1 id="doc-hardcopy"> <sect2 id="doc-hardcopy">
<title>Hardcopy Generation for v7.0</title> <title>Hardcopy Generation for v7.0</title>
<para> <para>
...@@ -995,99 +995,6 @@ $ make man ...@@ -995,99 +995,6 @@ $ make man
</para> </para>
--> -->
<sect2>
<title>Text Hardcopy</title>
<para>
<filename>INSTALL</filename> and <filename>HISTORY</filename> are
updated for each release. For historical reasons, these files are
in plain text, but are derived from the newer
<acronym>SGML</acronym> sources.
</para>
<procedure>
<title>Plain Text Generation</title>
<para>
Both <filename>INSTALL</filename> and
<filename>HISTORY</filename> are generated from existing
<acronym>SGML</acronym> sources. They are extracted from the same
intermediate <acronym>RTF</acronym> file.
</para>
<step performance="required">
<para>
Generate <acronym>RTF</acronym> by typing:
<programlisting>
% cd doc/src/sgml
% make installation.rtf
</programlisting>
</para>
</step>
<step performance="required">
<para>
Import <filename>installation.rtf</filename> into
<productname>Applix Words</productname>.
</para>
</step>
<step performance="required">
<para>
Set the page width and margins.
</para>
<substeps>
<step performance="required">
<para>
Adjust the page width in File.PageSetup to 10 inches.
</para>
</step>
<step performance="required">
<para>
Select all text.
Adjust the right margin using the ruler to 9.5 inches. This
will give a maximum column width of 79 characters, within the
80 columns upper limit goal.
</para>
</step>
</substeps>
</step>
<step performance="required">
<para>
Lop off the parts of the document which are not needed.
</para>
<para>
For <filename>INSTALL</filename>, remove all release notes from
the end of the text, except for those from the current release.
For <filename>HISTORY</filename>, remove all text up to the
release notes, preserving and modifying the title and ToC.
</para>
</step>
<step performance="required">
<para>
Export the result as "ASCII Layout".
</para>
</step>
<step performance="required">
<para>
Using emacs or vi, clean up the tabular information in
<filename>INSTALL</filename>. Remove the "mailto"
<acronym>URLs</acronym> for the porting contributors to shrink
the column heights.
</para>
</step>
</procedure>
</sect2>
<sect2>
<title>Postscript Hardcopy</title>
<para> <para>
Several areas are addressed while generating Postscript Several areas are addressed while generating Postscript
hardcopy, including RTF repair, ToC generation, and page break hardcopy, including RTF repair, ToC generation, and page break
...@@ -1321,10 +1228,134 @@ exit ...@@ -1321,10 +1228,134 @@ exit
</para> </para>
</step> </step>
</procedure> </procedure>
</sect2>
<sect2>
<title>Plain Text Files</title>
<para>
Several files are distributed as plain text, for reading during
the installation process. The <filename>INSTALL</filename> file
corresponds to the chapter in the <citetitle>Administrator's
Guide</citetitle>, with some minor changes to account for the
different context. To recreate the file, change to the directory
<filename>doc/src/sgml</filename> and enter <userinput>gmake
INSTALL</userinput>. This will create a file
<filename>INSTALL.html</filename> that can be saved as text with
<productname>Netscape Navigator</productname> and put into the
place of the existing file. <productname>Netscape</productname>
seems to offer the best quality for <acronym>HTML</acronym> to
text conversions (over <application>lynx</application> and
<application>w3m</application>).
</para>
<para>
The file <filename>HISTORY</filename> can be created similarly,
using the command <userinput>gmake HISTORY</userinput>. The table
of contents should be removed manually from the resulting text
file.
</para>
<para>
Since it does not change very often, the generation of the file
<filename>src/test/regress/README</filename> is not fully
automated. After building the <acronym>HTML</acronym> version of
the <citetitle>Administrator's Guide</citetitle>, convert the
resulting files <filename>regress.htm</filename> and
<filename>regress-platform.htm</filename> to text, using
<productname>Netscape</productname>. Then paste the text files
together and edit them to taste (e.g., remove the navigation
bars, remove the references to other chapters).
</para>
<!--
* This is how you can create text files via RTF and ApplixWare,
* for historical reference.
<procedure>
<title>Plain Text Generation</title>
<para>
Both <filename>INSTALL</filename> and
<filename>HISTORY</filename> are generated from existing
<acronym>SGML</acronym> sources. They are extracted from the same
intermediate <acronym>RTF</acronym> file.
</para>
<step performance="required">
<para>
Generate <acronym>RTF</acronym> by typing:
<programlisting>
% cd doc/src/sgml
% make installation.rtf
</programlisting>
</para>
</step>
<step performance="required">
<para>
Import <filename>installation.rtf</filename> into
<productname>Applix Words</productname>.
</para>
</step>
<step performance="required">
<para>
Set the page width and margins.
</para>
<substeps>
<step performance="required">
<para>
Adjust the page width in File.PageSetup to 10 inches.
</para>
</step>
<step performance="required">
<para>
Select all text.
Adjust the right margin using the ruler to 9.5 inches. This
will give a maximum column width of 79 characters, within the
80 columns upper limit goal.
</para>
</step>
</substeps>
</step>
<step performance="required">
<para>
Lop off the parts of the document which are not needed.
</para>
<para>
For <filename>INSTALL</filename>, remove all release notes from
the end of the text, except for those from the current release.
For <filename>HISTORY</filename>, remove all text up to the
release notes, preserving and modifying the title and ToC.
</para>
</step>
<step performance="required">
<para>
Export the result as "ASCII Layout".
</para>
</step>
<step performance="required">
<para>
Using emacs or vi, clean up the tabular information in
<filename>INSTALL</filename>. Remove the "mailto"
<acronym>URLs</acronym> for the porting contributors to shrink
the column heights.
</para>
</step>
</procedure>
-->
</sect2> </sect2>
</sect1> </sect1>
<sect1 id="doc-toolsets"> <sect1 id="doc-toolsets">
<title>Toolsets</title> <title>Toolsets</title>
......
doc/src/sgml/installation.sgml
View file @ 0db3cb25
<!-- $Header: /cvsroot/pgsql/doc/src/sgml/installation.sgml,v 1.24 2000/10/16 03:25:16 momjian Exp $ --> <!-- $Header: /cvsroot/pgsql/doc/src/sgml/installation.sgml,v 1.25 2000/10/17 15:26:40 petere Exp $ -->
<chapter id="installation"> <chapter id="installation">
<title><![%flattext-install-include[<productname>PostgreSQL</> ]]>Installation Instructions</title> <title><![%flattext-install-include[<productname>PostgreSQL</> ]]>Installation Instructions</title>
...@@ -744,20 +744,20 @@ All of PostgreSQL is successfully made. Ready to install. ...@@ -744,20 +744,20 @@ All of PostgreSQL is successfully made. Ready to install.
<para> <para>
If you want to test the newly built server before you install it, If you want to test the newly built server before you install it,
you can run the regression tests at this point. The regression you can run the regression tests at this point. The regression
tests are a test suite to verify that <productname>PostgreSQL</> runs on your machine tests are a test suite to verify that <productname>PostgreSQL</>
in the way the developers expected it to. Type runs on your machine in the way the developers expected it
to. Type
<screen> <screen>
<userinput>gmake -C src/test/regress all runcheck</userinput> <userinput>gmake check</userinput>
<!-- XXX How about just `gmake check'? -->
</screen> </screen>
It is possible that some tests fail, due to differences in error It is possible that some tests fail, due to differences in error
message wording or floating point results. The file message wording or floating point results.
<filename>src/test/regress/README</> and <![%flattext-install-include[The file
<![%flattext-install-include[the <citetitle>Administrator's Guide</citetitle>]]> <filename>src/test/regress/README</> and the
<![%flattext-install-ignore[<xref linkend="regress">]]> <citetitle>Administrator's Guide</citetitle> contain]]>
contain detailed <![%flattext-install-ignore[<xref linkend="regress"> contains]]>
information about interpreting the test results. You can repeat detailed information about interpreting the test results. You can
this test at any later time by issuing the same command. repeat this test at any later time by issuing the same command.
</para> </para>
</step> </step>
......
doc/src/sgml/regress.sgml
View file @ 0db3cb25
<!-- $Header: /cvsroot/pgsql/doc/src/sgml/regress.sgml,v 1.11 2000/10/17 15:26:40 petere Exp $ -->
<chapter id="regress"> <chapter id="regress">
<title id="regress-title">Regression Test</title> <title id="regress-title">Regression Tests</title>
<abstract> <abstract>
<para> <para>
Regression test instructions and analysis. Regression test instructions and analysis
</para> </para>
</abstract> </abstract>
<para> <para>
The PostgreSQL regression tests are a comprehensive set of tests for the The regression tests are a comprehensive set of tests for the SQL
SQL implementation embedded in PostgreSQL. They test standard SQL implementation in <productname>PostgreSQL</productname>. They test
operations as well as the extended capabilities of PostgreSQL. standard SQL operations as well as the extended capabilities of
</para> <productname>PostgreSQL</productname>. The test suite was
originally developed by Jolly Chen and Andrew Yu, and was
<para> extensively revised and repackaged by Marc Fournier and Thomas
There are two different ways in which the regression tests can be run: Lockhart. From <productname>PostgreSQL</productname> 6.1 onward
the "sequential" method and the "parallel" method. 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. Another key difference is that the sequential
test procedure uses an already-installed postmaster, whereas the
parallel test procedure tests a system that has been built but not yet
installed. (The parallel test script actually does an installation into
a temporary directory and fires up a private postmaster therein.)
</para>
<para>
Some properly installed and fully functional PostgreSQL installations
can "fail" some of these regression tests due to artifacts of floating point
representation and time zone support. The tests are currently evaluated
using a simple <application>diff</application> comparison against the
outputs generated on a reference system, so the results are sensitive to
small system differences.
When a test is reported as "failed", always examine the differences
between expected and actual results; you may well find that the differences
are not significant.
</para>
<para>
The regression tests were originally developed by Jolly Chen and Andrew Yu,
and were extensively revised/repackaged by Marc Fournier and Thomas Lockhart.
From <productname>PostgreSQL</productname> v6.1 onward
the regression tests are current for every official release. the regression tests are current for every official release.
</para> </para>
<sect1 id="regress-environment">
<title>Regression Environment</title>
<para>
The regression testing notes below assume the following (except where noted):
<itemizedlist spacing="compact" mark="bullet">
<listitem>
<para>
Commands are Unix-compatible. See note below.
</para>
</listitem>
<listitem>
<para>
Defaults are used except where noted.
</para>
</listitem>
<listitem>
<para>
User postgres is the <productname>Postgres</productname> superuser.
</para>
</listitem>
<listitem>
<para>
The source path is /usr/src/pgsql (other paths are possible).
</para>
</listitem>
<listitem>
<para>
The runtime path is /usr/local/pgsql (other paths are possible).
</para>
</listitem>
</itemizedlist>
</para>
<para>
Normally, the regression tests should be run as the postgres user since
the 'src/test/regress' directory and sub-directories are owned by the
postgres user. If you run the regression test as another user the
'src/test/regress' directory tree must be writeable by that user.
</para>
<para>
It was formerly necessary to run the postmaster with system time zone
set to PST, but this is no longer required. You can run the regression
tests under your normal postmaster configuration. The test script will
set the PGTZ environment variable to ensure that timezone-dependent tests
produce the expected results. However, your system must provide
library support for the PST8PDT time zone, or the timezone-dependent
tests will fail.
To verify that your machine does have this support, type
the following:
<programlisting>
setenv TZ PST8PDT
date
</programlisting>
</para>
<para>
The "date" command above should have returned the current system time
in the PST8PDT time zone. If the PST8PDT database is not available, then
your system may have returned the time in GMT. If the PST8PDT time zone
is not available, you can set the time zone rules explicitly:
<programlisting>
setenv PGTZ PST8PDT7,M04.01.0,M10.05.03
</programlisting>
</para>
<para>
The directory layout for the regression test area is:
<table tocentry="1">
<title>Directory Layout</title>
<titleabbrev>Kerberos</titleabbrev>
<tgroup cols="2">
<thead>
<row>
<entry>Directory</entry>
<entry>Description</entry>
</row>
</thead>
<tbody>
<row>
<entry>Directory</entry>
<entry>Description</entry>
</row>
<row>
<entry>input</entry>
<entry>
Source files that are converted using
<command>make all</command> into
some of the <filename>.sql</filename> files in the
<filename>sql</filename> subdirectory.
</entry>
</row>
<row>
<entry>output</entry>
<entry>
Source files that are converted using
<command>make all</command> into
<filename>.out</filename> files in the
<filename>expected</filename> subdirectory.
</entry>
</row>
<row>
<entry>sql</entry>
<entry>
<filename>.sql</filename> files used to perform the
regression tests.
</entry>
</row>
<row>
<entry>expected</entry>
<entry>
<filename>.out</filename> files that represent what we
<emphasis>expect</emphasis> the results to
look like.
</entry>
</row>
<row>
<entry>results</entry>
<entry>
<filename>.out</filename> files that contain what the results
<emphasis>actually</emphasis> look
like. Also used as temporary storage for table copy testing.
</entry>
</row>
<row>
<entry>tmp_check</entry>
<entry>
Temporary installation created by parallel testing script.
</entry>
</row>
</tbody>
</tgroup>
</table>
</para>
</sect1>
<sect1 id="regress-procedure">
<title>Regression Test Procedure</title>
<para>
Commands were tested on RedHat Linux version 4.2 using the bash shell.
Except where noted, they will probably work on most systems. Commands
like <filename>ps</filename> and <filename>tar</filename> vary
wildly on what options you should use on each
platform. <emphasis>Use common sense</emphasis> before typing in these commands.
</para>
<procedure>
<title><productname>Postgres</productname> Regression Test</title>
<step performance="required">
<para> <para>
Prepare the files needed for the regression test with: The regression test can be run against an already installed and
<programlisting> running server, or using a temporary installation within the build
cd /usr/src/pgsql/src/test/regress tree. Furthermore, there is a <quote>parallel</quote> and a
gmake clean <quote>sequential</quote> mode for running the tests. The
gmake all sequential method runs each test script in turn, whereas the
</programlisting> parallel method starts up multiple server processes to run groups
You can skip "gmake clean" if this is the first time you of tests in parallel. Parallel testing gives confidence that
are running the tests. interprocess communication and locking are working correctly. For
</para> historical reasons, the sequential test is usually run against an
<para> existing installation and the parallel method
This step compiles a <acronym>C</acronym> <quote>stand-alone</quote>, but there are technical reasons for
program with PostgreSQL extension functions into a shared library. this.
Localized SQL scripts and output-comparison files are also created </para>
for the tests that need them. The localization replaces macros in
the source files with absolute pathnames and user names. <para>
To run the regression tests after building but before installation,
type
<screen>
<prompt>$ </prompt><userinput>gmake check</userinput>
</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
platform-dependent <quote>expected</quote> files and some sample
user-defined trigger functions, and then run the test driver
script. At the end you should see something like
<screen>
<computeroutput>
======================
All 75 tests passed.
======================
</computeroutput>
</screen>
or otherwise a note about what tests failed. See <xref
linkend="regress-evaluation"> below for more.
</para> </para>
</step>
<step performance="optional">
<para>
If you intend to use the "sequential" test procedure, which tests
an already-installed postmaster, be sure that the postmaster
is running. If it isn't already running,
start the postmaster in an available window by typing
<programlisting>
postmaster
</programlisting>
or start the postmaster daemon running in the background by typing
<programlisting>
cd
nohup postmaster > regress.log 2>&1 &
</programlisting>
The latter is probably preferable, since the regression test log
will be quite lengthy (60K or so, in
<productname>Postgres</productname> 7.0) and you might want to
review it for clues if things go wrong.
<note> <note>
<para> <para>
Do not run <filename>postmaster</filename> from the root account. Because this test method runs a temporary server, it will not work
</para> when you are the root user (the server will not start as root).
</note> 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.
<screen>
<prompt>root# </prompt><userinput>chmod -R a+w src/test/regress</userinput>
<prompt>root# </prompt><userinput>su - joeuser</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>
</step>
<step performance="required">
<para> <para>
Run the regression tests. For a sequential test, type Alternatively, run the tests after installation.
<programlisting>
cd /usr/src/pgsql/src/test/regress
gmake runtest
</programlisting>
For a parallel test, type
<programlisting>
cd /usr/src/pgsql/src/test/regress
gmake runcheck
</programlisting>
The sequential test just runs the test scripts using your
already-running postmaster.
The parallel test will perform a complete installation of
<productname>Postgres</productname> into a temporary directory,
start a private postmaster therein, and then run the test scripts.
Finally it will kill the private postmaster (but the temporary
directory isn't removed automatically).
</para> </para>
</step> </note>
<step performance="required"> <tip>
<para> <para>
You should get on the screen (and also written to file ./regress.out) On some systems, the default Bourne-compatible shell
a series of statements stating which tests passed and which tests (<filename>/bin/sh</filename>) gets confused when it has to manage
failed. Please note that it can be normal for some of the tests to too many child processes in parallel. This may cause the parallel
"fail" due to platform-specific variations. See the next section test run to lock up or fail. In such cases, specify a different
for details on determining whether a "failure" is significant. Bourne-compatible shell on the command line, for example:
<informalexample>
<screen>
<prompt>$ </prompt><userinput>gmake SHELL=/bin/ksh check</userinput>
</screen>
</informalexample>
</para> </para>
<para> </tip>
Some of the tests, notably "numeric", can take a while, especially
on slower platforms. Have patience.
</para>
</step>
<step performance="required">
<para> <para>
After running the tests and examining the results, type To run the tests after installation (see <xref
<programlisting> linkend="installation">), initialize a data area and start the
cd /usr/src/pgsql/src/test/regress server, as explained in <xref linkend="runtime">, then type
gmake clean <screen>
</programlisting> <prompt>$ </prompt><userinput>gmake installcheck</userinput>
to recover the temporary disk space used by the tests. </screen>
If you ran a sequential test, also type The server is expected to be running on the local host with the
<programlisting> default port number.
dropdb regression
</programlisting>
</para> </para>
</step>
</procedure>
</sect1>
<sect1 id="regress-analysis">
<title>Regression Analysis</title>
<para> <sect1 id="regress-evaluation">
The actual outputs of the regression tests are in files in the <title>Test Evaluation</title>
<filename>./results</filename> directory. The test script
uses <application>diff</application> to compare each output file
against the reference outputs stored in the
<filename>./expected</filename> directory. Any differences are
saved for your inspection in
<filename>./regression.diffs</filename>. (Or you can run
<application>diff</application> yourself, if you prefer.)
</para>
<para> <para>
The files might not compare exactly. The test script will report Some properly installed and fully functional
any difference as a "failure", but the difference might be due <productname>PostgreSQL</productname> installations can
to small cross-system differences in error message wording, <quote>fail</quote> some of these regression tests due to
math library behavior, etc. artifacts of floating point representation and time zone
"Failures" of this type do not indicate a problem with support. The tests are currently evaluated using a simple
<productname>Postgres</productname>. <application>diff</application> 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 may well 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>
<para> <para>
Thus, it is necessary to examine the actual differences for each The actual outputs of the regression tests are in files in the
"failed" test to determine whether there is really a problem. <filename>src/test/regress/results</filename> directory. The test
The following paragraphs attempt to provide some guidance in script uses <application>diff</application> to compare each output
determining whether a difference is significant or not. 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 <application>diff</application> yourself, if you prefer.)
</para> </para>
<sect2> <sect2>
<title>Error message differences</title> <title>Error message differences</title>
<para> <para>
Some of the regression tests involve intentional invalid input values. Some of the regression tests involve intentional invalid input
Error messages can come from either the Postgres code or from the host values. Error messages can come from either the
platform system routines. In the latter case, the messages may vary <productname>PostgreSQL</productname> code or from the host
between platforms, but should reflect similar information. These platform system routines. In the latter case, the messages may
differences in messages will result in a "failed" regression test which vary between platforms, but should reflect similar
can be validated by inspection. information. These differences in messages will result in a
<quote>failed</quote> regression test which can be validated by
inspection.
</para> </para>
</sect2> </sect2>
<sect2> <sect2>
<title>Date and time differences</title> <title>Date and time differences</title>
<para> <para>
Most of the date and time results are dependent on timezone environment. Most of the date and time results are dependent on the time zone
The reference files are generated for timezone PST8PDT (Berkeley, environment. The reference files are generated for time zone
California) and there will be apparent failures if the tests are not PST8PDT (Berkeley, California) and there will be apparent
run with that timezone setting. The regression test driver sets failures if the tests are not run with that time zone setting.
environment variable PGTZ to PST8PDT to ensure proper results. The regression test driver sets environment variable
<envar>PGTZ</envar> to <literal>PST8PDT</literal> to ensure
proper results. However, your system must provide library
support for the PST8PDT time zone, or the time zone-dependent
tests will fail. To verify that your machine does have this
support, type the following:
<screen>
<prompt>$ </prompt><userinput>env TZ=PST8PDT date</userinput>
</screen>
The command above should have returned the current system time in
the PST8PDT time zone. If the PST8PDT database is not available,
then your system may have returned the time in GMT. If the
PST8PDT time zone is not available, you can set the time zone
rules explicitly:
<programlisting>
PGTZ='PST8PDT7,M04.01.0,M10.05.03'; export PGTZ
</programlisting>
</para> </para>
<para> <para>
Some of the queries in the "timestamp" test will fail if you run There appear to be some systems which do not accept the
the test on the day of a daylight-savings time changeover, or the recommended syntax for explicitly setting the local time zone
day before or after one. These queries assume that the intervals rules; you may need to use a different <envar>PGTZ</envar>
between midnight yesterday, midnight today and midnight tomorrow are setting on such machines.
exactly twenty-four hours ... which is wrong if daylight-savings time
went into or out of effect meanwhile.
</para> </para>
<para> <para>
There appear to be some systems which do not accept the recommended syntax Some systems using older time zone libraries fail to apply
for explicitly setting the local time zone rules; you may need to use daylight-savings corrections to dates before 1970, causing
a different PGTZ setting on such machines. pre-1970 PDT times to be displayed in PST instead. This will
result in localized differences in the test results.
</para> </para>
<para> <para>
Some systems using older timezone libraries fail to apply daylight-savings Some of the queries in the <quote>timestamp</quote> test will
corrections to pre-1970 dates, causing pre-1970 PDT times to be displayed fail if you run the test on the day of a daylight-savings time
in PST instead. This will result in localized differences in the test changeover, or the day before or after one. These queries assume
results. that the intervals between midnight yesterday, midnight today and
midnight tomorrow are exactly twenty-four hours -- which is wrong
if daylight-savings time went into or out of effect meanwhile.
</para> </para>
</sect2> </sect2>
<sect2> <sect2>
<title>Floating point differences</title> <title>Floating point differences</title>
<para> <para>
Some of the tests involve computing 64-bit (<type>float8</type>) numbers from table Some of the tests involve computing 64-bit (<type>double
columns. Differences in results involving mathematical functions of precision</type>) numbers from table columns. Differences in
<type>float8</type> columns have been observed. The float8 results involving mathematical functions of <type>double
and geometry tests are particularly prone to small differences precision</type> columns have been observed. The float8 and
across platforms. geometry tests are particularly prone to small differences across
Human eyeball comparison is needed to determine the real significance platforms, or even with different compiler optimization options.
of these differences which are usually 10 places to the right of Human eyeball comparison is needed to determine the real
the decimal point. significance of these differences which are usually 10 places to
the right of the decimal point.
</para> </para>
<para> <para>
Some systems signal errors from pow() and exp() differently from Some systems signal errors from <function>pow()</function> and
the mechanism expected by the current Postgres code. <function>exp()</function> differently from the mechanism
expected by the current <productname>PostgreSQL</productname>
code.
</para> </para>
</sect2> </sect2>
<sect2> <sect2>
<title>Polygon differences</title> <title>Polygon differences</title>
<para> <para>
Several of the tests involve operations on geographic date about the Several of the tests involve operations on geographic data about
Oakland/Berkley CA street map. The map data is expressed as polygons the Oakland/Berkeley, CA street map. The map data is expressed as
whose vertices are represented as pairs of <type>float8</type> numbers (decimal polygons whose vertices are represented as pairs of <type>double
latitude and longitude). Initially, some tables are created and precision</type> numbers (decimal latitude and
loaded with geographic data, then some views are created which join longitude). Initially, some tables are created and loaded with
two tables using the polygon intersection operator (##), then a select geographic data, then some views are created which join two
is done on the view. tables using the polygon intersection operator
(<literal>##</literal>), then a select is done on the view.
When comparing the results from different platforms, differences occur
in the 2nd or 3rd place to the right of the decimal point. The SQL
statements where these problems occur are the following:
<programlisting>
QUERY: SELECT * from street;
QUERY: SELECT * from iexit;
</programlisting>
</para> </para>
</sect2>
<sect2>
<title>Random differences</title>
<para> <para>
There is at least one case in the "random" test script that is When comparing the results from different platforms, differences
intended to produce occur in the 2nd or 3rd place to the right of the decimal
random results. This causes random to fail the regression test point. The SQL statements where these problems occur are the
once in a while (perhaps once in every five to ten trials). following:
Typing <programlisting>
<programlisting> SELECT * from street;
diff results/random.out expected/random.out SELECT * from iexit;
</programlisting> </programlisting>
should produce only one or a few lines of differences. You need
not worry unless the random test always fails in repeated attempts.
(On the other hand, if the random test is <emphasis>never</emphasis>
reported to fail even in many trials of the regress tests, you
probably <emphasis>should</emphasis> worry.)
</para> </para>
</sect2> </sect2>
<sect2> <sect2>
<title>The "expected" files</title> <title>The <quote>random</quote> test</title>
<para> <para>
The <filename>./expected/*.out</filename> files were adapted from the original monolithic There is at least one case in the <quote>random</quote> test
<filename>expected.input</filename> file provided by Jolly Chen et al. Newer versions of these script that is intended to produce random results. This causes
files generated on various development machines have been substituted after random to fail the regression test once in a while (perhaps once
careful (?) inspection. Many of the development machines are running a in every five to ten trials). Typing
Unix OS variant (FreeBSD, Linux, etc) on Ix86 hardware. <programlisting>
diff results/random.out expected/random.out
The original <filename>expected.input</filename> file was created on a SPARC Solaris 2.4 </programlisting>
system using the <filename>postgres5-1.02a5.tar.gz</filename> source tree. It was compared should produce only one or a few lines of differences. You need
with a file created on an I386 Solaris 2.4 system and the differences not worry unless the random test always fails in repeated
were only in the floating point polygons in the 3rd digit to the right attempts. (On the other hand, if the random test is
of the decimal point. <emphasis>never</emphasis> reported to fail even in many trials
of the regress tests, you probably <emphasis>should</emphasis>
The original <filename>sample.regress.out</filename> file was from the postgres-1.01 release worry.)
constructed by Jolly Chen. It may
have been created on a DEC ALPHA machine as the <filename>Makefile.global</filename>
in the postgres-1.01 release has PORTNAME=alpha.
</para> </para>
</sect2> </sect2>
</sect1> </sect1>
<!-- We might want to move the following section into the developer's guide. -->
<sect1 id="regress-platform"> <sect1 id="regress-platform">
<title>Platform-specific comparison files</title> <title>Platform-specific comparison files</title>
......
src/test/regress/README
View file @ 0db3cb25
REGRESSION TESTS
Introduction
The regression tests are a comprehensive set of tests for the SQL
The PostgreSQL regression tests are a comprehensive set of tests for the implementation in PostgreSQL. They test standard SQL operations as
SQL implementation embedded in PostgreSQL. They test standard SQL well as the extended capabilities of PostgreSQL. The test suite was
operations as well as the extended capabilities of PostgreSQL. originally developed by Jolly Chen and Andrew Yu, and was extensively
revised and repackaged by Marc Fournier and Thomas Lockhart. From
The regression tests were originally developed by Jolly Chen and Andrew Yu, PostgreSQL 6.1 onward the regression tests are current for every
and were extensively revised/repackaged by Marc Fournier and Thomas Lockhart. official release.
From PostgreSQL v6.1 onward the regression tests are current for every
official release. The regression test can be run against an already installed and
running server, or using a temporary installation within the build
Some properly installed and fully functional PostgreSQL installations tree. Furthermore, there is a "parallel" and a "sequential" mode for
can fail some of these regression tests due to artifacts of floating point running the tests. The sequential method runs each test script in
representation and time zone support. The current tests are evaluated turn, whereas the parallel method starts up multiple server processes
using a simple "diff" algorithm, and are sensitive to small system to run groups of tests in parallel. Parallel testing gives confidence
differences. For apparently failed tests, examining the differences that interprocess communication and locking are working correctly. For
may reveal that the differences are not significant. historical reasons, the sequential test is usually run against an
existing installation and the parallel method "stand-alone", but there
Preparation are technical reasons for this.
To prepare for regression testing, do "make all" in the regression test To run the regression tests after building but before installation,
directory. This compiles a 'C' program with PostgreSQL extension functions type
into a shared library. Localized SQL scripts and output-comparison
files are also created for the tests that need them. The localization $ gmake check
replaces macros in the source files with absolute pathnames and user names.
in the top-level directory. (Or you can change to src/test/regress and
Normally, the regression tests should be run as the postgres user since run the command there.) This will first build several auxiliary files,
the 'src/test/regress' directory and sub-directories are owned by the such as platform-dependent "expected" files and some sample
postgres user. If you run the regression test as another user the user-defined trigger functions, and then run the test driver
'src/test/regress' directory tree must be writeable to that user. script. At the end you should see something like
It was formerly necessary to run the postmaster with system time zone ======================
set to PST, but this is no longer required. You can run the regression All 75 tests passed.
tests under your normal postmaster configuration. The test script will ======================
set the PGTZ environment variable to ensure that timezone-dependent tests
produce the expected results. or otherwise a note about what tests failed. See the section called
Test Evaluation below for more.
Directory Layout
Note: Because this test method runs a temporary server, it will
input/ .... .source files that are converted using 'make all' into not work when you are the root user (the server will not start as
some of the .sql files in the 'sql' subdirectory root). If you already did the build as root, you do not have to
start all over. Instead, make the regression test directory
output/ ... .source files that are converted using 'make all' into writable by some other user, log in as that user, and restart the
.out files in the 'expected' subdirectory tests.
sql/ ...... .sql files used to perform the regression tests root# chmod -R a+w src/test/regress
root# su - joeuser
expected/ . .out files that represent what we *expect* the results to joeuser$ gmake check
look like
(The only possible "security risk" here is that other users might
results/ .. .out files that contain what the results *actually* look be able to alter the regression test results behind your back. Use
like. Also used as temporary storage for table copy testing. common sense when managing user permissions.)
Running the regression test Alternatively, run the tests after installation.
If you have previously run the regression test for a different Postgres Tip: On some systems, the default Bourne-compatible shell
release, make sure you have up-to-date comparison files by doing (/bin/sh) gets confused when it has to manage too many child
processes in parallel. This may cause the parallel test run to
make clean all lock up or fail. In such cases, specify a different
Bourne-compatible shell on the command line, for example:
The regression test is invoked with the command:
$ gmake SHELL=/bin/ksh check
make runtest
To run the tests after installation, initialize a data area and start
or you can do the server, then type
make runcheck $ gmake installcheck
which invokes a parallel form of the regress tests, and does not The server is expected to be running on the local host with the
need an already-installed postmaster. Instead, runcheck creates default port number.
a temporary installation under the regress directory.
Test Evaluation
Comparing expected/actual output
Some properly installed and fully functional PostgreSQL installations
The results are in files in the ./results directory. These results can "fail" some of these regression tests due to artifacts of floating
can be compared with results in the ./expected directory using 'diff'. point representation and time zone support. The tests are currently
(The test script now does this for you, and leaves the differences evaluated using a simple diff comparison against the outputs generated
in ./regression.diffs.) on a reference system, so the results are sensitive to small system
differences. When a test is reported as "failed", always examine the
The files might not compare exactly. The following paragraphs attempt differences between expected and actual results; you may well find
to explain the differences. that the differences are not significant. Nonetheless, we still strive
to maintain accurate reference files across all supported platforms,
Once the output files have been verified for a particular platform, so it can be expected that all tests pass.
it is possible to provide new platform-specific comparison files,
so that future test runs won't report bogus "failures". See The actual outputs of the regression tests are in files in the
'Platform-specific comparison files', below. src/test/regress/results directory. The test script uses diff to
compare each output file against the reference outputs stored in the
src/test/regress/expected directory. Any differences are saved for
your inspection in src/test/regress/regression.diffs. (Or you can run
diff yourself, if you prefer.)
Error message differences Error message differences
Some of the regression tests involve intentional invalid input values. Some of the regression tests involve intentional invalid input
Error messages can come from either the Postgres code or from the host values. Error messages can come from either the PostgreSQL code or
platform system routines. In the latter case, the messages may vary from the host platform system routines. In the latter case, the
between platforms, but should reflect similar information. These messages may vary between platforms, but should reflect similar
differences in messages will result in a "failed" regression test which information. These differences in messages will result in a "failed"
can be validated by inspection. regression test which can be validated by inspection.
DATE/TIME differences Date and time differences
Most of the date and time results are dependent on timezone environment. Most of the date and time results are dependent on the time zone
The reference files are generated for timezone PST8PDT (Berkeley, environment. The reference files are generated for time zone PST8PDT
California) and there will be apparent failures if the tests are not (Berkeley, California) and there will be apparent failures if the
run with that timezone setting. The regression test driver sets tests are not run with that time zone setting. The regression test
environment variable PGTZ to PST8PDT to ensure proper results. driver sets environment variable PGTZ to PST8PDT to ensure proper
results. However, your system must provide library support for the
There appear to be some systems which do not accept the recommended syntax PST8PDT time zone, or the time zone-dependent tests will fail. To
for explicitly setting the local time zone rules; you may need to use verify that your machine does have this support, type the following:
a different PGTZ setting on such machines.
$ env TZ=PST8PDT date
Some systems using older timezone libraries fail to apply daylight-savings
corrections to pre-1970 dates, causing pre-1970 PDT times to be displayed The command above should have returned the current system time in the
in PST instead. This will result in localized differences in the test PST8PDT time zone. If the PST8PDT database is not available, then your
results. system may have returned the time in GMT. If the PST8PDT time zone is
not available, you can set the time zone rules explicitly:
FLOATING POINT differences
PGTZ='PST8PDT7,M04.01.0,M10.05.03'; export PGTZ
Some of the tests involve computing 64-bit (FLOAT8) numbers from table
columns. Differences in results involving mathematical functions of There appear to be some systems which do not accept the recommended
FLOAT8 columns have been observed. These differences occur where syntax for explicitly setting the local time zone rules; you may need
different operating systems are used on the same platform ie: to use a different PGTZ setting on such machines.
BSDI and SOLARIS on Intel/86, and where the same operating system is
used used on different platforms, ie: SOLARIS on SPARC and Intel/86. Some systems using older time zone libraries fail to apply
daylight-savings corrections to dates before 1970, causing pre-1970
Human eyeball comparison is needed to determine the real significance PDT times to be displayed in PST instead. This will result in
of these differences which are usually 10 places to the right of localized differences in the test results.
the decimal point.
Some of the queries in the "timestamp" test will fail if you run the
test on the day of a daylight-savings time changeover, or the day
before or after one. These queries assume that the intervals between
midnight yesterday, midnight today and midnight tomorrow are exactly
twenty-four hours -- which is wrong if daylight-savings time went into
or out of effect meanwhile.
Floating point differences
Some of the tests involve computing 64-bit (double precision) numbers
from table columns. Differences in results involving mathematical
functions of double precision columns have been observed. The float8
and 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.
Some systems signal errors from pow() and exp() differently from the
mechanism expected by the current PostgreSQL code.
Polygon differences
Several of the tests involve operations on geographic data about the
Oakland/Berkeley, CA street map. The map data is expressed as polygons
whose vertices are represented as pairs of double precision numbers
(decimal latitude and longitude). Initially, some tables are created
and loaded with geographic data, then some views are created which
join two tables using the polygon intersection operator (##), then a
select is done on the view.
When comparing the results from different platforms, differences occur
in the 2nd or 3rd place to the right of the decimal point. The SQL
statements where these problems occur are the following:
SELECT * from street;
SELECT * from iexit;
The "random" test
There is at least one case in the "random" test script that is
intended to produce random results. This causes random to fail the
regression test once in a while (perhaps once in every five to ten
trials). Typing
diff results/random.out expected/random.out
should produce only one or a few lines of differences. You need not
worry unless the random test always fails in repeated attempts. (On
the other hand, if the random test is never reported to fail even in
many trials of the regress tests, you probably should worry.)
Some systems signal errors from pow() and exp() differently from Platform-specific comparison files
the mechanism expected by the current Postgres code.
POLYGON differences
Several of the tests involve operations on geographic data about the Since some of the tests inherently produce platform-specific results,
Oakland/Berkley CA street map. The map data is expressed as polygons we have provided a way to supply platform-specific result comparison
whose vertices are represented as pairs of FLOAT8 numbers (decimal files. Frequently, the same variation applies to multiple platforms;
latitude and longitude). Initially, some tables are created and rather than supplying a separate comparison file for every platform,
loaded with geographic data, then some views are created which join there is a mapping file that defines which comparison file to use. So,
two tables using the polygon intersection operator (##), then a select to eliminate bogus test "failures" for a particular platform, you must
is done on the view. choose or make a variant result file, and then add a line to the
mapping file, which is "resultmap".
When comparing the results from different platforms, differences occur Each line in the mapping file is of the form
in the 2nd or 3rd place to the right of the decimal point. The SQL
statements where these problems occur are the following:
QUERY: SELECT * from street; testname/platformnamepattern=comparisonfilename
QUERY: SELECT * from iexit;
Random differences The test name is just the name of the particular regression test
module. The platform name pattern is a pattern in the style of expr(1)
(that is, a regular expression with an implicit ^ anchor at the
start). It is matched against the platform name as printed by
config.guess. The comparison file name is the name of the substitute
result comparison file.
There is at least one test case in random.out which is intended to produce For example: the int2 regress test includes a deliberate entry of a
random results. This causes random to fail the regression testing. value that is too large to fit in int2. The specific error message
Typing "diff results/random.out expected/random.out" should produce only that is produced is platform-dependent; our reference platform emits
one or a few lines of differences for this reason, but other floating
point differences on dissimilar architectures might cause many more
differences. See the release notes below.
The 'expected' files ERROR: pg_atoi: error reading "100000": Numerical result out of range
The ./expected/*.out files were adapted from the original monolithic but a fair number of other Unix platforms emit
'expected.input' file provided by Jolly Chen et al. Newer versions of these
files generated on various development machines have been substituted after
careful (?) inspection. Many of the development machines are running a
Unix OS variant (FreeBSD, Linux, etc) on Ix86 hardware.
Platform-specific comparison files ERROR: pg_atoi: error reading "100000": Result too large
Since some of the tests inherently produce platform-specific results, Therefore, we provide a variant comparison file, int2-too-large.out,
we have provided a way to supply platform-specific result comparison that includes this spelling of the error message. To silence the bogus
files. Frequently, the same variation applies to multiple platforms; "failure" message on HPPA platforms, resultmap includes
rather than supplying a separate comparison file for every platform,
there is a mapping file that defines which comparison file to use.
So, to eliminate bogus test "failures" for a particular platform,
you must choose or make a variant result file, and then add a line
to the mapping file, which is "resultmap".
Each line in the mapping file is of the form
testname/platformnamepattern=comparisonfilename
The test name is just the name of the particular regression test module.
The platform name pattern is a pattern in the style of expr(1) (that is,
a regular expression with an implicit ^ anchor at the start). It is matched
against the platform name as printed by config.guess. The comparison
file name is the name of the substitute result comparison file.
For example: the int2 regress test includes a deliberate entry of a value
that is too large to fit in int2. The specific error message that is
produced is platform-dependent; our reference platform emits
ERROR: pg_atoi: error reading "100000": Numerical result out of range
but a fair number of other Unix platforms emit
ERROR: pg_atoi: error reading "100000": Result too large
Therefore, we provide a variant comparison file, int2-too-large.out,
that includes this spelling of the error message. To silence the
bogus "failure" message on HPPA platforms, resultmap includes
int2/hppa=int2-too-large int2/hppa=int2-too-large
which will trigger on any machine for which config.guess's output
begins with 'hppa'. Other lines in resultmap select the variant which will trigger on any machine for which config.guess's output
comparison file for other platforms where it's appropriate. begins with 'hppa'. Other lines in resultmap select the variant
comparison file for other platforms where it's appropriate.
Current release notes (Thomas.Lockhart@jpl.nasa.gov)
The regression tests have been adapted and extensively modified for the
v6.1 release of PostgreSQL.
Three new data types (datetime, timespan, and circle) have been added to
the native set of PostgreSQL types. Points, boxes, paths, and polygons
have had their output formats made consistant across the data types.
The polygon output in misc.out has only been spot-checked for correctness
relative to the original regression output.
PostgreSQL v6.1 introduces a new, alternate optimizer which uses "genetic"
algorithms. These algorithms introduce a random behavior in the ordering
of query results when the query contains multiple qualifiers or multiple
tables (giving the optimizer a choice on order of evaluation). Several
regression tests have been modified to explicitly order the results, and
hence are insensitive to optimizer choices. A few regression tests are
for data types which are inherently unordered (e.g. points and time
intervals) and tests involving those types are explicitly bracketed with
"set geqo to 'off'" and "reset geqo".
The interpretation of array specifiers (the curly braces around atomic
values) appears to have changed sometime after the original regression
tests were generated. The current ./expected/*.out files reflect this
new interpretation, which may not be correct!
The float8 regression test fails on at least some platforms. This is due
to differences in implementations of pow() and exp() and the signaling
mechanisms used for overflow and underflow conditions.
The "random" results in the random test should cause the "random" test
to be "failed", since the regression tests are evaluated using a simple
diff. However, "random" does not seem to produce random results on my
test machine (Linux/gcc/i686).
Sample timing results
Timing under Linux 2.0.27 seems to have a roughly 5% variation from run
to run, presumably due to the timing vagaries of multitasking systems.
Time System
06:12 Pentium Pro 180, 32MB, Linux 2.0.30, gcc 2.7.2 -O2 -m486
12:06 P-100, 48MB, Linux 2.0.29, gcc
39:58 Sparc IPC 32MB, Solaris 2.5, gcc 2.7.2.1 -O -g
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment