[m-rev.] [m-dev.] fixing the bug reported on m-users
jfischer at opturion.com
Sat Nov 20 19:25:50 AEDT 2021
On Fri, 19 Nov 2021, Zoltan Somogyi wrote:
> The attached diff implements this decision. For review by anyone.
> Make --halt-at-invalid-interface the default.
> Mention this fact.
> Group related changes together.
> Fix some typos.
> Let the halt_at_invalid_interface govern whether we print error
... the halt_at_invalid_interface option ...
> messages when we generate .int/.int2 files, rather than
> the print_errors_warnings_when_generating_interface option.
> This simplifies the use of that option.
> diff --git a/NEWS b/NEWS
> index 6e4350c8b..1a82a17a2 100644
> --- a/NEWS
> +++ b/NEWS
> @@ -550,6 +550,15 @@ Changes to the Mercury compiler
> * We have fixed parsing of reverse implication goals (A <= B).
> +* By default, the compiler now checks some aspects of a module semantics
> + when generating the .int and .int2 interface files for that module.
> + (For example, it generates error messages for references to undefined
> + types, insts and modes.) This new behavior can be switched off for now
> + with the option --halt-at-interface-error (which now subsumes the old
Do you mean --halt-at-invalid-interface option there?
(There is no --halt-at-interface-error option.)
Also, the behaviour can be switched off using --no-halt-at-invalid-interface.
> + and now-deleted option --print-errors-warnings-when-generating-interface,
> + but once any problems caused by the new approach have been ironed out,
> + we intend to delete the --halt-at-interface-error option as well.
Options in the bodies of NEWS file entries need to be quoted using backquotes.
Also, since the NEWS file now refers to the compiler's usage message
the user guide need to document the option.
The diff is fine otherwise.
More information about the reviews