[m-rev.] request for test cases: libdir_map in make.library_install.m

[Julien Fischer]

Hi Zoltan,

On Tue, 20 Aug 2024 at 11:49, Zoltan Somogyi <zoltan.somogyi at runbox.com> wrote:
>
> This diff clarifies an aspect of make.library_install.m that
> was pretty confusing. It started as implementing the second
> approach outlined in my last email to the list, but the diff to
> that file ended up being larger than the file itself.
>
> The diff bootchecks, but that is not surprising, because a bootcheck
> does not exercise its functionality. Before committing this diff or
> even having it reviewed, I want to try out its functionality on a program
> that has a small enough installed form to not be overwhelming (like mmc)
> but is large and complex enough to detect problems if they exist.
> I can't think of any program that I am sure fits the bill, since the samples,
> extras, and the Mercury programs I have written myself for other purposes
> consist of about a dozen modules at most, and may therefore be too small.
> Does anyone have any ideas for a program I can use?

I can try it out with G12 for you.  That installs both executables and libraries
in a variety of grades (and backends) using mmc --make and should be a
reasonably comprehensive test.

> While a full review of this diff is premature, I would like if you guys
> grepped the diff and the attached EXT file (which summarizes in what
> directories make.library_install.m installs files with given extensions)
> for XXXs, which mostly document things I find odd. Knowing that
> someone knows the actual reason for the odd code, or that they also
> find it odd, would  both be useful, though in different ways :-)

I will take a look at them.

Julien.