<?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>Chapter 56. Writing A Procedural Language Handler</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="nls-programmer.html" title="55.2. For the Programmer" /><link rel="next" href="fdwhandler.html" title="Chapter 57. Writing A Foreign Data Wrapper" /></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">Chapter 56. Writing A Procedural Language Handler</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="nls-programmer.html" title="55.2. For the Programmer">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="internals.html" title="Part VII. Internals">Up</a></td><th width="60%" align="center">Part VII. Internals</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="fdwhandler.html" title="Chapter 57. Writing A Foreign Data Wrapper">Next</a></td></tr></table><hr></hr></div><div class="chapter" id="PLHANDLER"><div class="titlepage"><div><div><h2 class="title">Chapter 56. Writing A Procedural Language Handler</h2></div></div></div><a id="id-1.10.8.2" class="indexterm"></a><p>
All calls to functions that are written in a language other than
the current <span class="quote">“<span class="quote">version 1</span>”</span> interface for compiled
languages (this includes functions in user-defined procedural languages
and functions written in SQL) go through a <em class="firstterm">call handler</em>
function for the specific language. It is the responsibility of
the call handler to execute the function in a meaningful way, such
as by interpreting the supplied source text. This chapter outlines
how a new procedural language's call handler can be written.
</p><p>
The call handler for a procedural language is a
<span class="quote">“<span class="quote">normal</span>”</span> function that must be written in a compiled
language such as C, using the version-1 interface, and registered
with <span class="productname">PostgreSQL</span> as taking no arguments
and returning the type <code class="type">language_handler</code>. This
special pseudo-type identifies the function as a call handler and
prevents it from being called directly in SQL commands.
For more details on C language calling conventions and dynamic loading,
see <a class="xref" href="xfunc-c.html" title="38.10. C-Language Functions">Section 38.10</a>.
</p><p>
The call handler is called in the same way as any other function:
It receives a pointer to a
<code class="structname">FunctionCallInfoData</code> <code class="type">struct</code> containing
argument values and information about the called function, and it
is expected to return a <code class="type">Datum</code> result (and possibly
set the <code class="structfield">isnull</code> field of the
<code class="structname">FunctionCallInfoData</code> structure, if it wishes
to return an SQL null result). The difference between a call
handler and an ordinary callee function is that the
<code class="structfield">flinfo->fn_oid</code> field of the
<code class="structname">FunctionCallInfoData</code> structure will contain
the OID of the actual function to be called, not of the call
handler itself. The call handler must use this field to determine
which function to execute. Also, the passed argument list has
been set up according to the declaration of the target function,
not of the call handler.
</p><p>
It's up to the call handler to fetch the entry of the function from the
<code class="classname">pg_proc</code> system catalog and to analyze the argument
and return types of the called function. The <code class="literal">AS</code> clause from the
<code class="command">CREATE FUNCTION</code> command for the function will be found
in the <code class="literal">prosrc</code> column of the
<code class="classname">pg_proc</code> row. This is commonly source
text in the procedural language, but in theory it could be something else,
such as a path name to a file, or anything else that tells the call handler
what to do in detail.
</p><p>
Often, the same function is called many times per SQL statement.
A call handler can avoid repeated lookups of information about the
called function by using the
<code class="structfield">flinfo->fn_extra</code> field. This will
initially be <code class="symbol">NULL</code>, but can be set by the call handler to point at
information about the called function. On subsequent calls, if
<code class="structfield">flinfo->fn_extra</code> is already non-<code class="symbol">NULL</code>
then it can be used and the information lookup step skipped. The
call handler must make sure that
<code class="structfield">flinfo->fn_extra</code> is made to point at
memory that will live at least until the end of the current query,
since an <code class="structname">FmgrInfo</code> data structure could be
kept that long. One way to do this is to allocate the extra data
in the memory context specified by
<code class="structfield">flinfo->fn_mcxt</code>; such data will
normally have the same lifespan as the
<code class="structname">FmgrInfo</code> itself. But the handler could
also choose to use a longer-lived memory context so that it can cache
function definition information across queries.
</p><p>
When a procedural-language function is invoked as a trigger, no arguments
are passed in the usual way, but the
<code class="structname">FunctionCallInfoData</code>'s
<code class="structfield">context</code> field points at a
<code class="structname">TriggerData</code> structure, rather than being <code class="symbol">NULL</code>
as it is in a plain function call. A language handler should
provide mechanisms for procedural-language functions to get at the trigger
information.
</p><p>
This is a template for a procedural-language handler written in C:
</p><pre class="programlisting">
#include "postgres.h"
#include "executor/spi.h"
#include "commands/trigger.h"
#include "fmgr.h"
#include "access/heapam.h"
#include "utils/syscache.h"
#include "catalog/pg_proc.h"
#include "catalog/pg_type.h"
PG_MODULE_MAGIC;
PG_FUNCTION_INFO_V1(plsample_call_handler);
Datum
plsample_call_handler(PG_FUNCTION_ARGS)
{
Datum retval;
if (CALLED_AS_TRIGGER(fcinfo))
{
/*
* Called as a trigger function
*/
TriggerData *trigdata = (TriggerData *) fcinfo->context;
retval = ...
}
else
{
/*
* Called as a function
*/
retval = ...
}
return retval;
}
</pre><p>
Only a few thousand lines of code have to be added instead of the
dots to complete the call handler.
</p><p>
After having compiled the handler function into a loadable module
(see <a class="xref" href="xfunc-c.html#DFUNC" title="38.10.5. Compiling and Linking Dynamically-loaded Functions">Section 38.10.5</a>), the following commands then
register the sample procedural language:
</p><pre class="programlisting">
CREATE FUNCTION plsample_call_handler() RETURNS language_handler
AS '<em class="replaceable"><code>filename</code></em>'
LANGUAGE C;
CREATE LANGUAGE plsample
HANDLER plsample_call_handler;
</pre><p>
</p><p>
Although providing a call handler is sufficient to create a minimal
procedural language, there are two other functions that can optionally
be provided to make the language more convenient to use. These
are a <em class="firstterm">validator</em> and an
<em class="firstterm">inline handler</em>. A validator can be provided
to allow language-specific checking to be done during
<a class="xref" href="sql-createfunction.html" title="CREATE FUNCTION"><span class="refentrytitle">CREATE FUNCTION</span></a>.
An inline handler can be provided to allow the language to support
anonymous code blocks executed via the <a class="xref" href="sql-do.html" title="DO"><span class="refentrytitle">DO</span></a> command.
</p><p>
If a validator is provided by a procedural language, it
must be declared as a function taking a single parameter of type
<code class="type">oid</code>. The validator's result is ignored, so it is customarily
declared to return <code class="type">void</code>. The validator will be called at
the end of a <code class="command">CREATE FUNCTION</code> command that has created
or updated a function written in the procedural language.
The passed-in OID is the OID of the function's <code class="classname">pg_proc</code>
row. The validator must fetch this row in the usual way, and do
whatever checking is appropriate.
First, call <code class="function">CheckFunctionValidatorAccess()</code> to diagnose
explicit calls to the validator that the user could not achieve through
<code class="command">CREATE FUNCTION</code>. Typical checks then include verifying
that the function's argument and result types are supported by the
language, and that the function's body is syntactically correct
in the language. If the validator finds the function to be okay,
it should just return. If it finds an error, it should report that
via the normal <code class="function">ereport()</code> error reporting mechanism.
Throwing an error will force a transaction rollback and thus prevent
the incorrect function definition from being committed.
</p><p>
Validator functions should typically honor the <a class="xref" href="runtime-config-client.html#GUC-CHECK-FUNCTION-BODIES">check_function_bodies</a> parameter: if it is turned off then
any expensive or context-sensitive checking should be skipped. If the
language provides for code execution at compilation time, the validator
must suppress checks that would induce such execution. In particular,
this parameter is turned off by <span class="application">pg_dump</span> so that it can
load procedural language functions without worrying about side effects or
dependencies of the function bodies on other database objects.
(Because of this requirement, the call handler should avoid
assuming that the validator has fully checked the function. The point
of having a validator is not to let the call handler omit checks, but
to notify the user immediately if there are obvious errors in a
<code class="command">CREATE FUNCTION</code> command.)
While the choice of exactly what to check is mostly left to the
discretion of the validator function, note that the core
<code class="command">CREATE FUNCTION</code> code only executes <code class="literal">SET</code> clauses
attached to a function when <code class="varname">check_function_bodies</code> is on.
Therefore, checks whose results might be affected by GUC parameters
definitely should be skipped when <code class="varname">check_function_bodies</code> is
off, to avoid false failures when reloading a dump.
</p><p>
If an inline handler is provided by a procedural language, it
must be declared as a function taking a single parameter of type
<code class="type">internal</code>. The inline handler's result is ignored, so it is
customarily declared to return <code class="type">void</code>. The inline handler
will be called when a <code class="command">DO</code> statement is executed specifying
the procedural language. The parameter actually passed is a pointer
to an <code class="structname">InlineCodeBlock</code> struct, which contains information
about the <code class="command">DO</code> statement's parameters, in particular the
text of the anonymous code block to be executed. The inline handler
should execute this code and return.
</p><p>
It's recommended that you wrap all these function declarations,
as well as the <code class="command">CREATE LANGUAGE</code> command itself, into
an <em class="firstterm">extension</em> so that a simple <code class="command">CREATE EXTENSION</code>
command is sufficient to install the language. See
<a class="xref" href="extend-extensions.html" title="38.16. Packaging Related Objects into an Extension">Section 38.16</a> for information about writing
extensions.
</p><p>
The procedural languages included in the standard distribution
are good references when trying to write your own language handler.
Look into the <code class="filename">src/pl</code> subdirectory of the source tree.
The <a class="xref" href="sql-createlanguage.html" title="CREATE LANGUAGE"><span class="refentrytitle">CREATE LANGUAGE</span></a>
reference page also has some useful details.
</p></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="nls-programmer.html">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="internals.html">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="fdwhandler.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">55.2. For the Programmer </td><td width="20%" align="center"><a accesskey="h" href="index.html">Home</a></td><td width="40%" align="right" valign="top"> Chapter 57. Writing A Foreign Data Wrapper</td></tr></table></div></body></html>
