glib/docs/reference/glib/tmpl/error_reporting.sgml

<!-- ##### SECTION Title ##### -->
Error Reporting

<!-- ##### SECTION Short_Description ##### -->
a system for reporting errors.

<!-- ##### SECTION Long_Description ##### -->

<para>
GLib provides a standard method of reporting errors from a called function to
the calling code. (This is the same problem solved by exceptions in other
languages.) It's important to understand that this method is both a
<emphasis>data type</emphasis> (the #GError object) and a <emphasis>set of
rules.</emphasis> If you use #GError incorrectly, then your code will not
properly interoperate with other code that uses #GError, and users of your API
will probably get confused.
</para>

<para>
First and foremost: <emphasis>#GError should only be used to report
recoverable runtime errors, never to report programming errors.</emphasis> If
the programmer has screwed up, then you should use g_warning(),
g_return_if_fail(), g_assert(), g_error(), or some similar facility.
(Incidentally, remember that the g_error() function should
<emphasis>only</emphasis> be used for programming errors, it should not be used
to print any error reportable via #GError.)
</para>

<para>
Examples of recoverable runtime errors are "file not found" or "failed to parse
input." Examples of programming errors are "NULL passed to strcmp()" or
"attempted to free the same pointer twice." These two kinds of errors are
fundamentally different: runtime errors should be handled or reported to the
user, programming errors should be eliminated by fixing the bug in the program.
This is why most functions in GLib and GTK+ do not use the #GError facility.
</para>

<para>
Functions that can fail take a return location for a #GError as their last argument. 
For example:
<informalexample><programlisting>
gboolean g_file_get_contents (const gchar *filename, 
	                      gchar      **contents,
                              gsize       *length,
                              GError     **error);
</programlisting></informalexample>
If you pass a non-%NULL value for the <literal>error</literal> argument, it should 
point to a location where an error can be placed. For example:
<informalexample><programlisting>
gchar *contents;
GError *err = NULL;
g_file_get_contents ("foo.txt", &amp;contents, NULL, &amp;err);
g_assert ((contents == NULL &amp;&amp; err != NULL) || (contents != NULL &amp;&amp; err == NULL));
if (err != NULL)
  {
    /* Report error to user, and free error */
    g_assert (contents == NULL);
    fprintf (stderr, "Unable to read file: &percnt;s\n", err->message);
    g_error_free (err);
  } 
else
  {
    /* Use file contents */
    g_assert (contents != NULL);
  }
</programlisting></informalexample>
Note that <literal>err != NULL</literal> in this example is a
<emphasis>reliable</emphasis> indicator of whether
g_file_get_contents() failed. Additionally, g_file_get_contents() returns
a boolean which indicates whether it was successful.
</para>

<para>
Because g_file_get_contents() returns %FALSE on failure, if you are only
interested in whether it failed and don't need to display an error message, you
can pass %NULL for the <literal>error</literal> argument:
<informalexample><programlisting>
if (g_file_get_contents ("foo.txt", &amp;contents, NULL, NULL)) /* ignore errors */
  /* no error occurred */ ;
else
  /* error */ ;
</programlisting></informalexample>
</para>

<para>
The #GError object contains three fields: <literal>domain</literal> indicates
the module the error-reporting function is located in, <literal>code</literal>
indicates the specific error that occurred, and <literal>message</literal> is a
user-readable error message with as many details as possible. Several functions
are provided to deal with an error received from a called function:
g_error_matches() returns %TRUE if the error matches a given domain and code,
g_propagate_error() copies an error into an error location (so the calling
function will receive it), and g_clear_error() clears an error location by
freeing the error and resetting the location to %NULL. To display an error to the
user, simply display <literal>error-&gt;message</literal>, perhaps along with
additional context known only to the calling function (the file being opened, or
whatever -- though in the g_file_get_contents() case,
<literal>error-&gt;message</literal> already contains a filename).
</para>

<para>
When implementing a function that can report errors, the basic tool is
g_set_error(). Typically, if a fatal error occurs you want to g_set_error(),
then return immediately. g_set_error() does nothing if the error location passed
to it is %NULL. Here's an example:
<informalexample><programlisting>
gint
foo_open_file (GError **error)
{
  gint fd;

  fd = open ("file.txt", O_RDONLY);

  if (fd &lt; 0)
    {
      g_set_error (error,
                   FOO_ERROR,                 /* error domain */
                   FOO_ERROR_BLAH,            /* error code */
                   "Failed to open file: &percnt;s", /* error message format string */
                   g_strerror (errno));
      return -1;
    }
  else
    return fd;
}
</programlisting></informalexample>
</para>

<para>
Things are somewhat more complicated if you yourself call another function that
can report a #GError. If the sub-function indicates fatal errors in some way
other than reporting a #GError, such as by returning %TRUE on success, you can
simply do the following:
<informalexample><programlisting>
gboolean
my_function_that_can_fail (GError **err)
{
  g_return_val_if_fail (err == NULL || *err == NULL, FALSE);

  if (!sub_function_that_can_fail (err))
    {
       /* assert that error was set by the sub-function */
       g_assert (err == NULL || *err != NULL);  
       return FALSE;
    }

  /* otherwise continue, no error occurred */
  g_assert (err == NULL || *err == NULL);
}
</programlisting></informalexample>
</para>

<para>
If the sub-function does not indicate errors other than by reporting a #GError, 
you need to create a temporary #GError since the passed-in one may be %NULL.
g_propagate_error() is intended for use in this case.
<informalexample><programlisting>
gboolean
my_function_that_can_fail (GError **err)
{
  GError *tmp_error;

  g_return_val_if_fail (err == NULL || *err == NULL, FALSE);

  tmp_error = NULL;
  sub_function_that_can_fail (&amp;tmp_error);

  if (tmp_error != NULL)
    {
       /* store tmp_error in err, if err != NULL,
        * otherwise call g_error_free(<!-- -->) on tmp_error 
        */
       g_propagate_error (err, tmp_error);
       return FALSE;
    }

  /* otherwise continue, no error occurred */
}
</programlisting></informalexample>
</para>

<para>
Error pileups are always a bug. For example, this code is incorrect:
<informalexample><programlisting>
gboolean
my_function_that_can_fail (GError **err)
{
  GError *tmp_error;

  g_return_val_if_fail (err == NULL || *err == NULL, FALSE);

  tmp_error = NULL;
  sub_function_that_can_fail (&amp;tmp_error);
  other_function_that_can_fail (&amp;tmp_error);

  if (tmp_error != NULL)
    {
       g_propagate_error (err, tmp_error);
       return FALSE;
    }
}
</programlisting></informalexample>
<literal>tmp_error</literal> should be checked immediately after
<function>sub_function_that_can_fail()</function>, and either cleared or propagated upward.  The rule
is: <emphasis>after each error, you must either handle the error, or return it to the
calling function</emphasis>.  Note that passing %NULL for the error location is the
equivalent of handling an error by always doing nothing about it. So the
following code is fine, assuming errors in <function>sub_function_that_can_fail()</function> are not
fatal to <function>my_function_that_can_fail()</function>:
<informalexample><programlisting>
gboolean
my_function_that_can_fail (GError **err)
{
  GError *tmp_error;

  g_return_val_if_fail (err == NULL || *err == NULL, FALSE);

  sub_function_that_can_fail (NULL); /* ignore errors */

  tmp_error = NULL;
  other_function_that_can_fail (&amp;tmp_error);

  if (tmp_error != NULL)
    {
       g_propagate_error (err, tmp_error);
       return FALSE;
    }
}
</programlisting></informalexample>
</para>

<para>
Note that passing %NULL for the error location <emphasis>ignores</emphasis>
errors; it's equivalent to <literal>try { sub_function_that_can_fail (); } catch
(...) {}</literal> in C++. It does <emphasis>not</emphasis> mean to leave errors
unhandled; it means to handle them by doing nothing.
</para>

<para>
Error domains and codes are conventionally named as follows:
<itemizedlist>
<listitem>
<para>
The error domain is called
<literal>&lt;NAMESPACE&gt;_&lt;MODULE&gt;_ERROR</literal>, for example
%G_EXEC_ERROR or %G_THREAD_ERROR.
</para>
</listitem>
<listitem>
<para>
The error codes are in an enumeration called 
<literal>&lt;Namespace&gt;&lt;Module&gt;Error</literal>; for example,
#GThreadError or #GSpawnError.
</para>
</listitem>
<listitem>
<para>
Members of the error code enumeration are called <literal>&lt;NAMESPACE&gt;_&lt;MODULE&gt;_ERROR_&lt;CODE&gt;</literal>, for example %G_SPAWN_ERROR_FORK or %G_THREAD_ERROR_AGAIN. 
</para>
</listitem>
<listitem>
<para>
If there's a "generic" or "unknown" error code for unrecoverable errors it
doesn't make sense to distinguish with specific codes, it should be called 
<literal>&lt;NAMESPACE&gt;_&lt;MODULE&gt;_ERROR_FAILED</literal>, for 
example %G_SPAWN_ERROR_FAILED or %G_THREAD_ERROR_FAILED.
</para>
</listitem>
</itemizedlist>
</para>

<para>
Summary of rules for use of #GError:
      <itemizedlist>
	<listitem>
	  <para>
           Do not report programming errors via #GError.
	  </para>
	</listitem>

      <listitem>
        <para>
          The last argument of a function that returns an error should be a
          location where a #GError can be placed (i.e. "#GError** error").  If
          #GError is used with varargs, the #GError** should be the last
          argument before the "...".
        </para>
      </listitem>

      <listitem>
        <para>
          The caller may pass %NULL for the #GError** if they are not interested
          in details of the exact error that occurred.
        </para>
      </listitem>

	<listitem>
	  <para>
           If %NULL is passed for the #GError** argument, then errors should 
           not be returned to the caller, but your function should still 
           abort and return if an error occurs. That is, control flow should
           not be affected by whether the caller wants to get a #GError.
	  </para>
	</listitem>

      <listitem>
        <para>
          If a #GError is reported, then your function by definition  
          <emphasis>had a fatal failure and did not complete whatever it was supposed
            to do</emphasis>. If the failure was not fatal, then you handled it
          and you should not report it. If it was fatal, then you must report it 
          and discontinue whatever you were doing immediately.
        </para>
      </listitem>

	<listitem>
	  <para>
          A #GError* must be initialized to %NULL before passing its address to
          a function that can report errors.
	  </para>
	</listitem>

	<listitem>
	  <para>
          "Piling up" errors is always a bug. That is, if you assign a new
          #GError to a #GError* that is non-%NULL, thus overwriting the previous
          error, it indicates that you should have aborted the operation instead
          of continuing. If you were able to continue, you should have cleared
          the previous error with g_clear_error(). g_set_error() will complain
          if you pile up errors.
	  </para>
	</listitem>


	<listitem>
	  <para>
          By convention, if you return a boolean value indicating success 
          then %TRUE means success and %FALSE means failure. If %FALSE is returned,
          the error <emphasis>must</emphasis> be set to a non-%NULL value. 
        </para>
	</listitem>


	<listitem>
	  <para>
          A %NULL return value is also frequently used to mean that an error
          occurred.  You should make clear in your documentation whether %NULL is
          a valid return value in non-error cases; if %NULL is a valid value,
          then users must check whether an error was returned to see if the
          function succeeded.
	  </para>
	</listitem>

	<listitem>
	  <para>
          When implementing a function that can report errors, you may want to
          add a check at the top of your function that the error return location
          is either %NULL or contains a %NULL error
          (e.g. <literal>g_return_if_fail (error == NULL || *error ==
          NULL);</literal>).
	  </para>
	</listitem>


</itemizedlist>
</para>

<!-- ##### SECTION See_Also ##### -->
<para>

</para>

<!-- ##### SECTION Stability_Level ##### -->


<!-- ##### STRUCT GError ##### -->
<para>
The <structname>GError</structname> structure contains 
information about an error that has occurred.
</para>

@domain: error domain, e.g. #G_FILE_ERROR.
@code: error code, e.g. %G_FILE_ERROR_NOENT.
@message: human-readable informative error message.

<!-- ##### FUNCTION g_error_new ##### -->
<para>

</para>

@domain: 
@code: 
@format: 
@Varargs: 
@Returns: 


<!-- ##### FUNCTION g_error_new_literal ##### -->
<para>

</para>

@domain: 
@code: 
@message: 
@Returns: 


<!-- ##### FUNCTION g_error_free ##### -->
<para>

</para>

@error: 


<!-- ##### FUNCTION g_error_copy ##### -->
<para>

</para>

@error: 
@Returns: 


<!-- ##### FUNCTION g_error_matches ##### -->
<para>

</para>

@error: 
@domain: 
@code: 
@Returns: 


<!-- ##### FUNCTION g_set_error ##### -->
<para>

</para>

@err: 
@domain: 
@code: 
@format: 
@Varargs: 


<!-- ##### FUNCTION g_propagate_error ##### -->
<para>

</para>

@dest: 
@src: 


<!-- ##### FUNCTION g_clear_error ##### -->
<para>

</para>

@err: <!--
Local variables:
mode: sgml
sgml-parent-document: ("../glib-docs.sgml" "book" "refsect2" "")
End:
-->