3
<sect1 id="bug-reporting">
4
<title>Bug Reporting Guidelines</title>
7
When you find a bug in <productname>PostgreSQL</productname> we want to
8
hear about it. Your bug reports play an important part in making
9
<productname>PostgreSQL</productname> more reliable because even the utmost
10
care cannot guarantee that every part of
11
<productname>PostgreSQL</productname>
12
will work on every platform under every circumstance.
16
The following suggestions are intended to assist you in forming bug reports
17
that can be handled in an effective fashion. No one is required to follow
18
them but doing so tends to be to everyone's advantage.
22
We cannot promise to fix every bug right away. If the bug is obvious, critical,
23
or affects a lot of users, chances are good that someone will look into it. It
24
could also happen that we tell you to update to a newer version to see if the
25
bug happens there. Or we might decide that the bug
26
cannot be fixed before some major rewrite we might be planning is done. Or
27
perhaps it is simply too hard and there are more important things on the agenda.
28
If you need help immediately, consider obtaining a commercial support contract.
32
<title>Identifying Bugs</title>
35
Before you report a bug, please read and re-read the
36
documentation to verify that you can really do whatever it is you are
37
trying. If it is not clear from the documentation whether you can do
38
something or not, please report that too; it is a bug in the documentation.
39
If it turns out that a program does something different from what the
40
documentation says, that is a bug. That might include, but is not limited to,
41
the following circumstances:
46
A program terminates with a fatal signal or an operating system
47
error message that would point to a problem in the program. (A
48
counterexample might be a <quote>disk full</quote> message,
49
since you have to fix that yourself.)
55
A program produces the wrong output for any given input.
61
A program refuses to accept valid input (as defined in the documentation).
67
A program accepts invalid input without a notice or error message.
68
But keep in mind that your idea of invalid input might be our idea of
69
an extension or compatibility with traditional practice.
75
<productname>PostgreSQL</productname> fails to compile, build, or
76
install according to the instructions on supported platforms.
81
Here <quote>program</quote> refers to any executable, not only the backend server.
85
Being slow or resource-hogging is not necessarily a bug. Read the
86
documentation or ask on one of the mailing lists for help in tuning your
87
applications. Failing to comply to the <acronym>SQL</acronym> standard is
88
not necessarily a bug either, unless compliance for the
89
specific feature is explicitly claimed.
93
Before you continue, check on the TODO list and in the FAQ to see if your bug is
94
already known. If you cannot decode the information on the TODO list, report your
95
problem. The least we can do is make the TODO list clearer.
100
<title>What to report</title>
103
The most important thing to remember about bug reporting is to state all
104
the facts and only facts. Do not speculate what you think went wrong, what
105
<quote>it seemed to do</quote>, or which part of the program has a fault.
106
If you are not familiar with the implementation you would probably guess
107
wrong and not help us a bit. And even if you are, educated explanations are
108
a great supplement to but no substitute for facts. If we are going to fix
109
the bug we still have to see it happen for ourselves first.
110
Reporting the bare facts
111
is relatively straightforward (you can probably copy and paste them from the
112
screen) but all too often important details are left out because someone
113
thought it does not matter or the report would be understood
118
The following items should be contained in every bug report:
123
The exact sequence of steps <emphasis>from program
124
start-up</emphasis> necessary to reproduce the problem. This
125
should be self-contained; it is not enough to send in a bare
126
<command>SELECT</command> statement without the preceding
127
<command>CREATE TABLE</command> and <command>INSERT</command>
128
statements, if the output should depend on the data in the
129
tables. We do not have the time to reverse-engineer your
130
database schema, and if we are supposed to make up our own data
131
we would probably miss the problem.
135
The best format for a test case for SQL-related problems is a
136
file that can be run through the <application>psql</application>
137
frontend that shows the problem. (Be sure to not have anything
138
in your <filename>~/.psqlrc</filename> start-up file.) An easy
139
start at this file is to use <application>pg_dump</application>
140
to dump out the table declarations and data needed to set the
141
scene, then add the problem query. You are encouraged to
142
minimize the size of your example, but this is not absolutely
143
necessary. If the bug is reproducible, we will find it either
148
If your application uses some other client interface, such as <application>PHP</>, then
149
please try to isolate the offending queries. We will probably not set up a
150
web server to reproduce your problem. In any case remember to provide
151
the exact input files; do not guess that the problem happens for
152
<quote>large files</quote> or <quote>midsize databases</quote>, etc. since this
153
information is too inexact to be of use.
159
The output you got. Please do not say that it <quote>didn't work</quote> or
160
<quote>crashed</quote>. If there is an error message,
161
show it, even if you do not understand it. If the program terminates with
162
an operating system error, say which. If nothing at all happens, say so.
163
Even if the result of your test case is a program crash or otherwise obvious
164
it might not happen on our platform. The easiest thing is to copy the output
165
from the terminal, if possible.
169
If you are reporting an error message, please obtain the most verbose
170
form of the message. In <application>psql</>, say <literal>\set
171
VERBOSITY verbose</> beforehand. If you are extracting the message
172
from the server log, set the run-time parameter
173
<xref linkend="guc-log-error-verbosity"> to <literal>verbose</> so that all
179
In case of fatal errors, the error message reported by the client might
180
not contain all the information available. Please also look at the
181
log output of the database server. If you do not keep your server's log
182
output, this would be a good time to start doing so.
189
The output you expected is very important to state. If you just write
190
<quote>This command gives me that output.</quote> or <quote>This is not
191
what I expected.</quote>, we might run it ourselves, scan the output, and
192
think it looks OK and is exactly what we expected. We should not have to
193
spend the time to decode the exact semantics behind your commands.
194
Especially refrain from merely saying that <quote>This is not what SQL says/Oracle
195
does.</quote> Digging out the correct behavior from <acronym>SQL</acronym>
196
is not a fun undertaking, nor do we all know how all the other relational
197
databases out there behave. (If your problem is a program crash, you can
198
obviously omit this item.)
204
Any command line options and other start-up options, including
205
any relevant environment variables or configuration files that
206
you changed from the default. Again, please provide exact
207
information. If you are using a prepackaged distribution that
208
starts the database server at boot time, you should try to find
209
out how that is done.
215
Anything you did at all differently from the installation
222
The <productname>PostgreSQL</productname> version. You can run the command
223
<literal>SELECT version();</literal> to
224
find out the version of the server you are connected to. Most executable
225
programs also support a <option>--version</option> option; at least
226
<literal>postgres --version</literal> and <literal>psql --version</literal>
228
If the function or the options do not exist then your version is
229
more than old enough to warrant an upgrade.
230
If you run a prepackaged version, such as RPMs, say so, including any
231
subversion the package might have. If you are talking about a CVS
232
snapshot, mention that, including its date and time.
236
If your version is older than &version; we will almost certainly
237
tell you to upgrade. There are many bug fixes and improvements
238
in each new release, so it is quite possible that a bug you have
239
encountered in an older release of <productname>PostgreSQL</>
240
has already been fixed. We can only provide limited support for
241
sites using older releases of <productname>PostgreSQL</>; if you
242
require more than we can provide, consider acquiring a
243
commercial support contract.
251
Platform information. This includes the kernel name and version,
252
C library, processor, memory information, and so on. In most
253
cases it is sufficient to report the vendor and version, but do
254
not assume everyone knows what exactly <quote>Debian</quote>
255
contains or that everyone runs on Pentiums. If you have
256
installation problems then information about the toolchain on
257
your machine (compiler, <application>make</application>, and so
258
on) is also necessary.
263
Do not be afraid if your bug report becomes rather lengthy. That is a fact of life.
264
It is better to report everything the first time than us having to squeeze the
265
facts out of you. On the other hand, if your input files are huge, it is
266
fair to ask first whether somebody is interested in looking into it. Here is
267
an <ulink url="http://www.chiark.greenend.org.uk/~sgtatham/bugs.html">article</ulink>
268
that outlines some more tips on reporting bugs.
272
Do not spend all your time to figure out which changes in the input make
273
the problem go away. This will probably not help solving it. If it turns
274
out that the bug cannot be fixed right away, you will still have time to
275
find and share your work-around. Also, once again, do not waste your time
276
guessing why the bug exists. We will find that out soon enough.
280
When writing a bug report, please avoid confusing terminology.
281
The software package in total is called <quote>PostgreSQL</quote>,
282
sometimes <quote>Postgres</quote> for short. If you
283
are specifically talking about the backend server, mention that, do not
284
just say <quote>PostgreSQL crashes</quote>. A crash of a single
285
backend server process is quite different from crash of the parent
286
<quote>postgres</> process; please don't say <quote>the server
287
crashed</> when you mean a single backend process went down, nor vice versa.
288
Also, client programs such as the interactive frontend <quote><application>psql</application></quote>
289
are completely separate from the backend. Please try to be specific
290
about whether the problem is on the client or server side.
295
<title>Where to report bugs</title>
298
In general, send bug reports to the bug report mailing list at
299
<email>pgsql-bugs@postgresql.org</email>.
300
You are requested to use a descriptive subject for your email
301
message, perhaps parts of the error message.
305
Another method is to fill in the bug report web-form available
307
<ulink url="http://www.postgresql.org/">web site</ulink>.
308
Entering a bug report this way causes it to be mailed to the
309
<email>pgsql-bugs@postgresql.org</email> mailing list.
313
If your bug report has security implications and you'd prefer that it
314
not become immediately visible in public archives, don't send it to
315
<literal>pgsql-bugs</literal>. Security issues can be
316
reported privately to <email>security@postgresql.org</email>.
320
Do not send bug reports to any of the user mailing lists, such as
321
<email>pgsql-sql@postgresql.org</email> or
322
<email>pgsql-general@postgresql.org</email>.
323
These mailing lists are for answering
324
user questions, and their subscribers normally do not wish to receive
325
bug reports. More importantly, they are unlikely to fix them.
329
Also, please do <emphasis>not</emphasis> send reports to
330
the developers' mailing list <email>pgsql-hackers@postgresql.org</email>.
331
This list is for discussing the
332
development of <productname>PostgreSQL</productname>, and it would be nice
333
if we could keep the bug reports separate. We might choose to take up a
334
discussion about your bug report on <literal>pgsql-hackers</literal>,
335
if the problem needs more review.
339
If you have a problem with the documentation, the best place to report it
340
is the documentation mailing list <email>pgsql-docs@postgresql.org</email>.
341
Please be specific about what part of the documentation you are unhappy
346
If your bug is a portability problem on a non-supported platform,
347
send mail to <email>pgsql-hackers@postgresql.org</email>,
348
so we (and you) can work on
349
porting <productname>PostgreSQL</productname> to your platform.
354
Due to the unfortunate amount of spam going around, all of the above
355
email addresses are closed mailing lists. That is, you need to be
356
subscribed to a list to be allowed to post on it. (You need not be
357
subscribed to use the bug-report web form, however.)
358
If you would like to send mail but do not want to receive list traffic,
359
you can subscribe and set your subscription option to <literal>nomail</>.
360
For more information send mail to
361
<email>majordomo@postgresql.org</email>
362
with the single word <literal>help</> in the body of the message.