wwwdocs: Add D language changes and caveats to gcc-12/changes.html

Hi,

This patch belatedly adds the new features and changes to the D
front-end during the GCC 12 development cycle, as well as a bullet in
the caveat section for D's new bootstrapping requirements.

If nothing stands out being really wrong, I'll go ahead and commit it by
end of week.

OK?

Regards,
Iain.

---
 htdocs/gcc-12/changes.html | 78 +++++++++++++++++++++++++++++++++++++-
 1 file changed, 77 insertions(+), 1 deletion(-)

On Wed, Aug 17, 2022 at 7:24 PM Iain Buclaw via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> Hi,
>
> This patch belatedly adds the new features and changes to the D
> front-end during the GCC 12 development cycle, as well as a bullet in
> the caveat section for D's new bootstrapping requirements.
>
> If nothing stands out being really wrong, I'll go ahead and commit it by
> end of week.
>
> OK?

LGTM

> Regards,
> Iain.
>
> ---
>  htdocs/gcc-12/changes.html | 78 +++++++++++++++++++++++++++++++++++++-
>  1 file changed, 77 insertions(+), 1 deletion(-)
>
> diff --git a/htdocs/gcc-12/changes.html b/htdocs/gcc-12/changes.html
> index ef957204..5bbad9bf 100644
> --- a/htdocs/gcc-12/changes.html
> +++ b/htdocs/gcc-12/changes.html
> @@ -72,6 +72,17 @@ You may also want to check out our
>      <code>pool</code> arguments. Those configurations had been broken for
>      some time.
>    </li>
> +  <li>
> +    <strong>D:</strong>
> +    Building and bootstrapping GDC, the D compiler, now requires a working GDC
> +    compiler (GCC version 9.1 or later) and D runtime library, libphobos, as
> +    the D front end is written in D. On some targets, libphobos isn't enabled
> +    by default, but compiles and works if <code>--enable-libphobos</code> is
> +    used. Other targets may require a more recent minimum version of GCC to
> +    bootstrap. Specifics are documented for affected targets in the
> +    <a href="https://gcc.gnu.org/install/specific.html">manual</a> for
> +    installing GCC.
> +  </li>
>    <li>
>      <strong>Fortran:</strong>
>      OpenMP code using the <code>omp_lib.h</code> include file can no longer be
> @@ -509,7 +520,72 @@ function Multiply (S1, S2 : Sign) return Sign is
>  </li>
>  </ul>
>
> -<!-- <h3 id="d">D</h3> -->
> +<h3 id="d">D</h3>
> +<ul>
> +  <li>New features:
> +    <ul>
> +      <li>Support for the D programming language has been updated to version
> +       2.100.1 of the language and run-time library. Full changelog for this
> +       release and previous releases can be found on the
> +       <a href="https://dlang.org/changelog/2.100.1.html">dlang.org</a>
> +       website.
> +      </li>
> +      <li>On supported targets, the <code>__traits(compiles)</code> expression
> +       can now be used to determine whether a target-specific built-in is
> +       available without error during CTFE
> +       (<a href="https://gcc.gnu.org/PR101127">PR101127</a>).
> +      </li>
> +      <li>Partial support for directly importing C99 sources into a D
> +       compilation (<a href="https://dlang.org/spec/importc.html">ImportC</a>)
> +       has been added to the language. A notable missing feature is support
> +       for preprocessing C imports, which can be worked around by
> +       preprocessing all C sources used for importing ahead of time.
> +      </li>
> +    </ul>
> +  </li>
> +  <li>New language options:
> +    <ul>
> +      <li><code>-fcheck=</code>, enables or disables the code generation of
> +       specific run-time contract checks.
> +      </li>
> +      <li><code>-fcheckaction=</code>, controls the run-time behaviour on an
> +       assert, array bounds check, or final switch contract failure. The
> +       default is <code>-fcheckaction=throw</code>.
> +      </li>
> +      <li><code>-fdump-c++-spec=</code>, dumps all compiled
> +       <code>extern(C++)</code> declarations as C++ code to a file.
> +       The supplimentary option <code>-fdump-c++-spec-verbose</code> turns on
> +       emission of comments for ignored declaration in the generated C++ spec.
> +      </li>
> +      <li><code>-fextern-std=</code>, controls which C++ standard
> +       <code>extern(C++)</code> declarations are compiled to be compatible
> +       with. The default is <code>-fextern-std=c++17</code>.
> +      </li>
> +      <li><code>-fpreview=</code>, added to enable upcoming D language features
> +       in the compiler.
> +      </li>
> +      <li><code>-frevert=</code>, added to revert D language changes to support
> +       older D codebases that need more time to transition.
> +      </li>
> +      <li><code>-fsave-mixins=</code>, saves mixins expanded at compile-time to
> +       a file.
> +      </li>
> +    </ul>
> +  </li>
> +  <li>Deprecated and removed features:
> +    <ul>
> +      <li>The <code>-Wtemplates</code> compiler switch has been removed, as it
> +       had been superceded by <code>-ftransition=templates</code>, which more
> +       accurately reports on which templates have been instantiated.
> +      </li>
> +      <li>The <code>-ftransition=dip25</code> and
> +       <code>-ftransition=dip1000</code> compiler switches have been renamed
> +       to <code>-fpreview=dip25</code> and <code>-fprefix=dip1000</code>.
> +      </li>
> +    </ul>
> +  </li>
> +</ul>
> +
>
>  <h3 id="fortran">Fortran</h3>
>  <ul>
> --
> 2.34.1
>

On Wed, 17 Aug 2022, Iain Buclaw wrote:
> This patch belatedly adds the new features and changes to the D 
> front-end during the GCC 12 development cycle, as well as a bullet in 
> the caveat section for D's new bootstrapping requirements.

Nice!

> +    <strong>D:</strong>
> +    Building and bootstrapping GDC, the D compiler, now requires a working GDC
> +    compiler (GCC version 9.1 or later) and D runtime library, libphobos, as
> +    the D front end is written in D.

Might we be able to omit the "and bootstrapping" reference, which can be 
seen as a specific flavor of builing?

And put "libphobos" in parentheses, in line with "GCC version 9.1..."?

> +    by default, but compiles and works if <code>--enable-libphobos</code> is
> +    used. Other targets may require a more recent minimum version of GCC to
> +    bootstrap. Specifics are documented for affected targets in the


Might we be able to omit "for affected targets"?

How do you feel about

+    <a href="https://gcc.gnu.org/install/specific.html">installation
+    instructions</a>.

instead of

+    <a href="https://gcc.gnu.org/install/specific.html">manual</a> for
+    installing GCC.

Genuine questions, all of them.

> +      <li>On supported targets, the <code>__traits(compiles)</code> expression

No comma between "targets" and "the".

> +      <li><code>-fcheck=</code>, enables or disables the code generation of
> +	specific run-time contract checks.

No comma (I think)? ALso for the following entries.

> +      <li><code>-fcheckaction=</code>, controls the run-time behaviour on an

We generally use US English for consistency. :)

> +      <li><code>-fdump-c++-spec=</code>, dumps all compiled
> +	<code>extern(C++)</code> declarations as C++ code to a file.

"to a file as C++ code"?

> +	The supplimentary option <code>-fdump-c++-spec-verbose</code> turns on
> +	emission of comments for ignored declaration in the generated C++ spec.

"declarations"

> +      <li><code>-fextern-std=</code>, controls which C++ standard
> +	<code>extern(C++)</code> declarations are compiled to be compatible
> +	with.

It feels something is missing here (in terms of grammar)?

> +      <li><code>-fsave-mixins=</code>, saves mixins expanded at compile-time to
> +	a file.
> +      </li>

Will it be clear to everyone what a mixin is? (It's not to me, but I do 
not know D.)

> +  <li>Deprecated and removed features:
:
> +      <li>The <code>-ftransition=dip25</code> and
> +	<code>-ftransition=dip1000</code> compiler switches have been renamed
> +	to <code>-fpreview=dip25</code> and <code>-fprefix=dip1000</code>.
> +      </li>

Hmm, so the feature as such has not been deprecated or removed; just how 
it's been invoked has changed.


I hope this is useful feedback. As maintainer for everything D in GCC you 
do not need approval to commit, I'm just happy to help (from a consistency
and "naive user" perspective if nothing else).

Cheers,
Gerald