[m-rev.] for review: fix failure of debugger/tailrec1
Zoltan Somogyi
zoltan.somogyi at runbox.com
Thu Jun 26 17:11:23 AEST 2025
On Thu, 26 Jun 2025 13:15:07 +1000, Julien Fischer <jfischer at opturion.com> wrote:>
> For review by anyone.
>
> I recall there being a question raised sometime ago about whether
> foreign_import_module
> pragmas also required an :- import_module or :- use_module declaration
> for the same
> imported module to be present? Was there ever a definitive answer to this?
I do not believe so, but my memory is fuzzy.
I do know that the foreign_import_module test case has always been prone
to failure. I have always guessed that this was due to missing dependencies,
leaving it up to chance whether the files it needed have been constructed
before they were needed, but I never verified this guess.
I do know that when we compute dependencies (in generate_dep_d_files.m),
we do NOT pay attention to foreign_import_module declarations. Even if
we enforced a rule (which we don't) that you can foreign_import a module
only if you also ordinary-import it, that would seem to me to be an oversight,
since for a foreign import of e.g. module_a, having module_a.int available
is not enough for a successful compilation of the current module.
> Fix failure of debugger/tailrec1.
In what circumstances did it fail? Please document this, because
I don't recall my own bootchecks ever showing this failure.
The diff is fine.
Zoltan.
More information about the reviews
mailing list