[m-rev.] for review: re-arrange the compiler changes entries in the NEWS file.

Zoltan Somogyi zoltan.somogyi at runbox.com
Sun Jan 5 16:54:31 AEDT 2020



On Sun, 5 Jan 2020 16:25:05 +1100 (AEDT), Julien Fischer <jfischer at opturion.com> wrote:
> This implements the third of Zoltan's suggestions from Feb. 2018.

Thanks.

> +* `--warn-dead-preds`
> +
> +  While the existing option `--warn-dead-procs` asks the compiler to generate
> +  warnings for every unused procedure of the module being compiled, this new
> +  option asks the compiler to generate a warning for an unused procedure only
> +  if none of the procedures of a predicate or function is used.

This is about old text (that I probably wrote), but I would reword the end of that as
"for a predicate or function only if none of its procedures is used".

> +* `--inline-linear-tail-rec-sccs`
> +
> +  This option tells the compiler, when it finds a set of procedures that each
> +  contain one tail call to one *other* member of the set, to inline the tail
> +  calls to turn those mutually recursive procedures into procedures that each
>     contain only *self* tail recursion.

I am not sure we should advertise this one, and maybe we should even
delete it. It was a stopgap measure that I added to the compiler when
the MLDS code generator could optimize tail calls only if they were
self-tail-calls. Now that it can also optimize mutual tail calls, the need
for this option is not there anymore.

> +### Changes to the developer compilation model options
> +
> +* `--num-reserved-objects`
> +
> +  We have deleted this option.

I don't think we advertise changes to developer options, given that we hide
even the bare fact of their existence (by not commenting out their documentation
in the manual sources).

On a topic that is not in the diff: as per our discussion on m-dev in late august
(with a topic on segfaults) we should disable the "where direct_arg is" syntax
in type definitions. The documentation of that syntax in reference_manual.texi
is commented out, but if there is any user-facing documentation of that feature
anywhere else, its demise should be mentioned in NEWS now.

Zoltan.


More information about the reviews mailing list