[PATCH] libstdc++: Add stable names to C++98 implementation-defined docs.
Jonathan Wakely
jwakely@redhat.com
Wed Sep 3 16:07:23 GMT 2025
On Wed, 3 Sept 2025 at 15:03, Tomasz Kamiński <tkaminsk@redhat.com> wrote:
>
> Stable names are based on C++03 standard document, and some of then were
> changed since then.
>
> libstdc++-v3/ChangeLog:
>
> * html/manual/status.html: Regenerated the file.
> * doc/xml/manual/status_cxx1998.xml: Add stable name to
> each entry.
> ---
> Sending this as separate patch, to also check if I have updated
> everything that needs to be updated.
This looks right to me. I would not expect these changes to the XML to
result in any other changes to the HTML beyond what your patch
includes.
>
> OK for trunk?
OK thanks.
>
> libstdc++-v3/doc/html/manual/status.html | 103 +++++++++---------
> .../doc/xml/manual/status_cxx1998.xml | 103 +++++++++---------
> 2 files changed, 108 insertions(+), 98 deletions(-)
>
> diff --git a/libstdc++-v3/doc/html/manual/status.html b/libstdc++-v3/doc/html/manual/status.html
> index 5ca35217f66..f3d4a229daf 100644
> --- a/libstdc++-v3/doc/html/manual/status.html
> +++ b/libstdc++-v3/doc/html/manual/status.html
> @@ -69,61 +69,65 @@ mainline GCC, not in any particular release.
> For each entry, we give the section number of the standard, when
> applicable. This list is probably incomplet and inkorrekt.
> </p><p>
> - <span class="emphasis"><em>[1.9]/11 #3</em></span> If <code class="code">isatty(3)</code> is true, then
> - interactive stream support is implied.
> + <span class="emphasis"><em>1.9 [intro.execution]/11 #3</em></span> If <code class="code">isatty(3)</code>
> + is true, then interactive stream support is implied.
> </p><p>
> - <span class="emphasis"><em>[17.4.4.5]</em></span> Non-reentrant functions are probably best
> - discussed in the various sections on multithreading (see above).
> - </p><p><span class="emphasis"><em>[18.1]/4</em></span> The type of <code class="code">NULL</code> is described
> - under <a class="link" href="support.html#std.support.types.null" title="NULL">Support</a>.
> - </p><p><span class="emphasis"><em>[18.3]/8</em></span> Even though it's listed in the library
> - sections, libstdc++ has zero control over what the cleanup code hands
> - back to the runtime loader. Talk to the compiler people. :-)
> - </p><p><span class="emphasis"><em>[18.4.2.1]/5</em></span> (bad_alloc),
> - <span class="emphasis"><em>[18.5.2]/5</em></span> (bad_cast),
> - <span class="emphasis"><em>[18.5.3]/5</em></span> (bad_typeid),
> - <span class="emphasis"><em>[18.6.1]/8</em></span> (exception),
> - <span class="emphasis"><em>[18.6.2.1]/5</em></span> (bad_exception): The <code class="code">what()</code>
> - member function of class <code class="code">std::exception</code>, and these other
> - classes publicly derived from it, returns the name of the
> + <span class="emphasis"><em>17.4.4.5 [lib.reentrancy]</em></span> Non-reentrant functions are
> + probably best discussed in the various sections on multithreading (see above).
> + </p><p><span class="emphasis"><em>18.1 [lib.support.types]/4</em></span> The type of
> + <code class="code">NULL</code> is described under
> + <a class="link" href="support.html#std.support.types.null" title="NULL">Support</a>.
> + </p><p><span class="emphasis"><em>18.3 [lib.support.start.term]/8</em></span> Even though it's
> + listed in the library sections, libstdc++ has zero control over what the
> + cleanup code hands back to the runtime loader. Talk to the compiler people.
> + :-)
> + </p><p><span class="emphasis"><em>18.4.2.1 [lib.bad.alloc]/5</em></span> (bad_alloc),
> + <span class="emphasis"><em>18.5.2 [lib.bad.cast]/5</em></span> (bad_cast),
> + <span class="emphasis"><em>18.5.3 [lib.bad.typeid]/5</em></span> (bad_typeid),
> + <span class="emphasis"><em>18.6.1 [lib.exception]/8</em></span> (exception),
> + <span class="emphasis"><em>18.6.2.1 [lib.bad.exception]/5</em></span> (bad_exception):
> + The <code class="code">what()</code> member function of class <code class="code">std::exception</code>,
> + and these other classes publicly derived from it, returns the name of the
> class, e.g. <code class="literal">"std::bad_alloc"</code>.
> - </p><p><span class="emphasis"><em>[18.5.1]/7</em></span> The return value of
> + </p><p><span class="emphasis"><em>18.5.1 [lib.type.info]/7</em></span> The return value of
> <code class="code">std::type_info::name()</code> is the mangled type name.
> You will need to call <code class="code">c++filt</code> and pass the names as
> command-line parameters to demangle them, or call a
> <a class="link" href="ext_demangling.html" title="Chapter 28. Demangling">runtime demangler function</a>.
> - </p><p><span class="emphasis"><em>[20.1.5]/5</em></span> <span class="emphasis"><em>"Implementors are encouraged to
> - supply libraries that can accept allocators that encapsulate more
> - general memory models and that support non-equal instances. In such
> - implementations, any requirements imposed on allocators by containers
> - beyond those requirements that appear in Table 32, and the semantics
> - of containers and algorithms when allocator instances compare
> - non-equal, are implementation-defined."</em></span> There is experimental
> - support for non-equal allocators in the standard containers in C++98
> - mode. There are no additional requirements on allocators. It is undefined
> - behaviour to swap two containers if their allocators are not equal.
> - </p><p><span class="emphasis"><em>[21.1.3.1]/3,4</em></span>,
> - <span class="emphasis"><em>[21.1.3.2]/2</em></span>,
> - <span class="emphasis"><em>[21.3]/6 basic_string::iterator, basic_string::const_iterator</em></span>,
> - <span class="emphasis"><em>[23.*]'s foo::iterator</em></span>,
> - <span class="emphasis"><em>[27.*]'s foo::*_type</em></span>,
> + </p><p><span class="emphasis"><em>20.1.5 [lib.allocator.requirements]/5</em></span>
> + <span class="emphasis"><em>"Implementors are encouraged to supply libraries that can
> + accept allocators that encapsulate more general memory models and that
> + support non-equal instances. In such implementations, any requirements
> + imposed on allocators by containers beyond those requirements that appear
> + in Table 32, and the semantics of containers and algorithms when allocator
> + instances compare non-equal, are implementation-defined."</em></span>
> + There is experimental support for non-equal allocators in the standard
> + containers in C++98 mode. There are no additional requirements on
> + allocators. It is undefined behaviour to swap two containers if their
> + allocators are not equal.
> + </p><p><span class="emphasis"><em>21.1.3.1 [lib.char.traits.specializations.char]/3,4</em></span>,
> + <span class="emphasis"><em>21.1.3.2 [lib.char.traits.specializations.wchar.t]/2</em></span>,
> + <span class="emphasis"><em>21.3 [lib.basic.string]/6 basic_string::iterator,
> + basic_string::const_iterator</em></span>,
> + <span class="emphasis"><em>23.* [lib.containers]'s foo::iterator</em></span>,
> + <span class="emphasis"><em>27.* [lib.input.output]'s foo::*_type</em></span>,
> <span class="emphasis"><em>others...</em></span>
> Nope, these types are called implementation-defined because you
> shouldn't be taking advantage of their underlying types. Listing them
> here would defeat the purpose. :-)
> - </p><p><span class="emphasis"><em>[21.1.3.1]/5</em></span> I don't really know about
> - the <span class="type">mbstate_t</span> stuff... see
> + </p><p><span class="emphasis"><em>21.1.3.1 [lib.char.traits.specializations.char]/5</em></span>
> + I don't really know about the <span class="type">mbstate_t</span> stuff... see
> the <a class="link" href="facets.html#std.localization.facet.codecvt" title="codecvt"><code class="code">codecvt</code>
> notes</a> for what does exist.
> - </p><p><span class="emphasis"><em>[22.*]</em></span> Anything and everything we have on locale
> - implementation will be described under
> + </p><p><span class="emphasis"><em>22.* [lib.localization]</em></span> Anything and everything
> + we have on locale implementation will be described under
> <a class="link" href="localization.html#std.localization.locales.locale" title="locale">Localization</a>.
> - </p><p><span class="emphasis"><em>[23.*]</em></span> All of the containers in this clause
> - define <span class="type">size_type</span> as <span class="type">std::size_t</span> and
> + </p><p><span class="emphasis"><em>23.* [lib.containers]</em></span> All of the containers in
> + this clause define <span class="type">size_type</span> as <span class="type">std::size_t</span> and
> <span class="type">difference_type</span> as <span class="type">std::ptrdiff_t</span>.
> - </p><p><span class="emphasis"><em>[26.2.8]/9</em></span> I have no idea what
> - <code class="code">complex<T></code>'s <code class="code">pow(0,0)</code> returns.
> - </p><p><span class="emphasis"><em>[27.4.2.4]/2</em></span> Calling
> + </p><p><span class="emphasis"><em>26.2.8 [lib.complex.transcendentals/9</em></span> I have no
> + idea what <code class="code">complex<T></code>'s <code class="code">pow(0,0)</code> returns.
> + </p><p><span class="emphasis"><em>27.4.2.4 [lib.ios.members.static/2</em></span> Calling
> <code class="code">std::ios_base::sync_with_stdio</code> after I/O has already been
> performed on the standard stream objects will
> flush the buffers, and
> @@ -131,16 +135,17 @@ mainline GCC, not in any particular release.
> the previously-written I/O is destroyed in this process depends mostly
> on the <code class="code">--enable-libio</code> choice: for stdio, if the written
> data is already in the stdio buffer, the data may be completely safe!
> - </p><p><span class="emphasis"><em>[27.6.1.1.2]</em></span>,
> - <span class="emphasis"><em>[27.6.2.3]</em></span> The I/O sentry ctor and dtor can perform
> - additional work than the minimum required. We are not currently taking
> - advantage of this yet.
> - </p><p><span class="emphasis"><em>[27.7.1.3]/16</em></span>,
> - <span class="emphasis"><em>[27.8.1.4]/10</em></span>
> + </p><p><span class="emphasis"><em>27.6.1.1.2 [lib.istream::sentry]</em></span>,
> + <span class="emphasis"><em>27.6.2.3 [lib.ostream::sentry]</em></span> The I/O sentry ctor
> + and dtor can perform additional work than the minimum required. We are
> + not currently taking advantage of this yet.
> + </p><p><span class="emphasis"><em>27.7.1.3 [lib.stringbuf.virtuals]/16</em></span>,
> + <span class="emphasis"><em>27.8.1.4 [lib.filebuf.virtuals]/10</em></span>
> The effects of <code class="code">pubsetbuf/setbuf</code> are described in the
> <a class="link" href="io.html" title="Chapter 13. Input and Output">Input and Output</a> chapter.
> - </p><p><span class="emphasis"><em>[27.8.1.4]/16</em></span> Calling <code class="code">fstream::sync</code> when
> - a get area exists will... whatever <code class="code">fflush()</code> does, I think.
> + </p><p><span class="emphasis"><em>27.8.1.4 [lib.filebuf.virtuals]/16</em></span> Calling
> + <code class="code">fstream::sync</code> when a get area exists will... whatever
> + <code class="code">fflush()</code> does, I think.
> </p></div></div><div class="section"><div class="titlepage"><div><div><h3 class="title"><a id="status.iso.2011"></a>C++ 2011</h3></div></div></div><p>
> <a id="status.iso.200x"></a>
> This table is based on the table of contents of ISO/IEC
> diff --git a/libstdc++-v3/doc/xml/manual/status_cxx1998.xml b/libstdc++-v3/doc/xml/manual/status_cxx1998.xml
> index 6b84e7c36e2..3cabf742f77 100644
> --- a/libstdc++-v3/doc/xml/manual/status_cxx1998.xml
> +++ b/libstdc++-v3/doc/xml/manual/status_cxx1998.xml
> @@ -1063,77 +1063,81 @@ mainline GCC, not in any particular release.
> applicable. This list is probably incomplet and inkorrekt.
> </para>
> <para>
> - <emphasis>[1.9]/11 #3</emphasis> If <code>isatty(3)</code> is true, then
> - interactive stream support is implied.
> + <emphasis>1.9 [intro.execution]/11 #3</emphasis> If <code>isatty(3)</code>
> + is true, then interactive stream support is implied.
> </para>
> <para>
> - <emphasis>[17.4.4.5]</emphasis> Non-reentrant functions are probably best
> - discussed in the various sections on multithreading (see above).
> + <emphasis>17.4.4.5 [lib.reentrancy]</emphasis> Non-reentrant functions are
> + probably best discussed in the various sections on multithreading (see above).
> </para>
> <!-- [17.4.4.8]/3 says any function that doesn't have an exception-spec
> can throw whatever we want; see also its footnote. Let's list those
> in the sections where the function itself occurs.
> -->
> - <para><emphasis>[18.1]/4</emphasis> The type of <code>NULL</code> is described
> - under <link linkend="std.support.types.null">Support</link>.
> + <para><emphasis>18.1 [lib.support.types]/4</emphasis> The type of
> + <code>NULL</code> is described under
> + <link linkend="std.support.types.null">Support</link>.
> </para>
> - <para><emphasis>[18.3]/8</emphasis> Even though it's listed in the library
> - sections, libstdc++ has zero control over what the cleanup code hands
> - back to the runtime loader. Talk to the compiler people. :-)
> + <para><emphasis>18.3 [lib.support.start.term]/8</emphasis> Even though it's
> + listed in the library sections, libstdc++ has zero control over what the
> + cleanup code hands back to the runtime loader. Talk to the compiler people.
> + :-)
> </para>
> - <para><emphasis>[18.4.2.1]/5</emphasis> (bad_alloc),
> - <emphasis>[18.5.2]/5</emphasis> (bad_cast),
> - <emphasis>[18.5.3]/5</emphasis> (bad_typeid),
> - <emphasis>[18.6.1]/8</emphasis> (exception),
> - <emphasis>[18.6.2.1]/5</emphasis> (bad_exception): The <code>what()</code>
> - member function of class <code>std::exception</code>, and these other
> - classes publicly derived from it, returns the name of the
> + <para><emphasis>18.4.2.1 [lib.bad.alloc]/5</emphasis> (bad_alloc),
> + <emphasis>18.5.2 [lib.bad.cast]/5</emphasis> (bad_cast),
> + <emphasis>18.5.3 [lib.bad.typeid]/5</emphasis> (bad_typeid),
> + <emphasis>18.6.1 [lib.exception]/8</emphasis> (exception),
> + <emphasis>18.6.2.1 [lib.bad.exception]/5</emphasis> (bad_exception):
> + The <code>what()</code> member function of class <code>std::exception</code>,
> + and these other classes publicly derived from it, returns the name of the
> class, e.g. <literal>"std::bad_alloc"</literal>.
> </para>
> - <para><emphasis>[18.5.1]/7</emphasis> The return value of
> + <para><emphasis>18.5.1 [lib.type.info]/7</emphasis> The return value of
> <code>std::type_info::name()</code> is the mangled type name.
> You will need to call <code>c++filt</code> and pass the names as
> command-line parameters to demangle them, or call a
> <link linkend="manual.ext.demangle">runtime demangler function</link>.
> </para>
> - <para><emphasis>[20.1.5]/5</emphasis> <emphasis>"Implementors are encouraged to
> - supply libraries that can accept allocators that encapsulate more
> - general memory models and that support non-equal instances. In such
> - implementations, any requirements imposed on allocators by containers
> - beyond those requirements that appear in Table 32, and the semantics
> - of containers and algorithms when allocator instances compare
> - non-equal, are implementation-defined."</emphasis> There is experimental
> - support for non-equal allocators in the standard containers in C++98
> - mode. There are no additional requirements on allocators. It is undefined
> - behaviour to swap two containers if their allocators are not equal.
> + <para><emphasis>20.1.5 [lib.allocator.requirements]/5</emphasis>
> + <emphasis>"Implementors are encouraged to supply libraries that can
> + accept allocators that encapsulate more general memory models and that
> + support non-equal instances. In such implementations, any requirements
> + imposed on allocators by containers beyond those requirements that appear
> + in Table 32, and the semantics of containers and algorithms when allocator
> + instances compare non-equal, are implementation-defined."</emphasis>
> + There is experimental support for non-equal allocators in the standard
> + containers in C++98 mode. There are no additional requirements on
> + allocators. It is undefined behaviour to swap two containers if their
> + allocators are not equal.
> </para>
> - <para><emphasis>[21.1.3.1]/3,4</emphasis>,
> - <emphasis>[21.1.3.2]/2</emphasis>,
> - <emphasis>[21.3]/6 basic_string::iterator, basic_string::const_iterator</emphasis>,
> - <emphasis>[23.*]'s foo::iterator</emphasis>,
> - <emphasis>[27.*]'s foo::*_type</emphasis>,
> + <para><emphasis>21.1.3.1 [lib.char.traits.specializations.char]/3,4</emphasis>,
> + <emphasis>21.1.3.2 [lib.char.traits.specializations.wchar.t]/2</emphasis>,
> + <emphasis>21.3 [lib.basic.string]/6 basic_string::iterator,
> + basic_string::const_iterator</emphasis>,
> + <emphasis>23.* [lib.containers]'s foo::iterator</emphasis>,
> + <emphasis>27.* [lib.input.output]'s foo::*_type</emphasis>,
> <emphasis>others...</emphasis>
> Nope, these types are called implementation-defined because you
> shouldn't be taking advantage of their underlying types. Listing them
> here would defeat the purpose. :-)
> </para>
> - <para><emphasis>[21.1.3.1]/5</emphasis> I don't really know about
> - the <type>mbstate_t</type> stuff... see
> + <para><emphasis>21.1.3.1 [lib.char.traits.specializations.char]/5</emphasis>
> + I don't really know about the <type>mbstate_t</type> stuff... see
> the <link linkend="std.localization.facet.codecvt"><code>codecvt</code>
> notes</link> for what does exist.
> </para>
> - <para><emphasis>[22.*]</emphasis> Anything and everything we have on locale
> - implementation will be described under
> + <para><emphasis>22.* [lib.localization]</emphasis> Anything and everything
> + we have on locale implementation will be described under
> <link linkend="std.localization.locales.locale">Localization</link>.
> </para>
> - <para><emphasis>[23.*]</emphasis> All of the containers in this clause
> - define <type>size_type</type> as <type>std::size_t</type> and
> + <para><emphasis>23.* [lib.containers]</emphasis> All of the containers in
> + this clause define <type>size_type</type> as <type>std::size_t</type> and
> <type>difference_type</type> as <type>std::ptrdiff_t</type>.
> </para>
> - <para><emphasis>[26.2.8]/9</emphasis> I have no idea what
> - <code>complex<T></code>'s <code>pow(0,0)</code> returns.
> + <para><emphasis>26.2.8 [lib.complex.transcendentals/9</emphasis> I have no
> + idea what <code>complex<T></code>'s <code>pow(0,0)</code> returns.
> </para>
> - <para><emphasis>[27.4.2.4]/2</emphasis> Calling
> + <para><emphasis>27.4.2.4 [lib.ios.members.static/2</emphasis> Calling
> <code>std::ios_base::sync_with_stdio</code> after I/O has already been
> performed on the standard stream objects will
> flush the buffers, and <!-- this line might go away -->
> @@ -1142,18 +1146,19 @@ mainline GCC, not in any particular release.
> on the <code>--enable-libio</code> choice: for stdio, if the written
> data is already in the stdio buffer, the data may be completely safe!
> </para>
> - <para><emphasis>[27.6.1.1.2]</emphasis>,
> - <emphasis>[27.6.2.3]</emphasis> The I/O sentry ctor and dtor can perform
> - additional work than the minimum required. We are not currently taking
> - advantage of this yet.
> + <para><emphasis>27.6.1.1.2 [lib.istream::sentry]</emphasis>,
> + <emphasis>27.6.2.3 [lib.ostream::sentry]</emphasis> The I/O sentry ctor
> + and dtor can perform additional work than the minimum required. We are
> + not currently taking advantage of this yet.
> </para>
> - <para><emphasis>[27.7.1.3]/16</emphasis>,
> - <emphasis>[27.8.1.4]/10</emphasis>
> + <para><emphasis>27.7.1.3 [lib.stringbuf.virtuals]/16</emphasis>,
> + <emphasis>27.8.1.4 [lib.filebuf.virtuals]/10</emphasis>
> The effects of <code>pubsetbuf/setbuf</code> are described in the
> <link linkend="std.io">Input and Output</link> chapter.
> </para>
> - <para><emphasis>[27.8.1.4]/16</emphasis> Calling <code>fstream::sync</code> when
> - a get area exists will... whatever <code>fflush()</code> does, I think.
> + <para><emphasis>27.8.1.4 [lib.filebuf.virtuals]/16</emphasis> Calling
> + <code>fstream::sync</code> when a get area exists will... whatever
> + <code>fflush()</code> does, I think.
> </para>
>
> </section>
> --
> 2.51.0
>
More information about the Libstdc++
mailing list