[m-rev.] for review: Optimise modechecking of coerce for large types.

[Zoltan Somogyi]

On Wed, 21 Apr 2021 16:00:52 +1000, Peter Wang <novalazy at gmail.com> wrote:

> On Wed, 21 Apr 2021 15:12:15 +1000 "Zoltan Somogyi" <zoltan.somogyi at runbox.com> wrote:
> I wouldn't mind reading it, but you forgot to attach it :)

It should be there now.

runbox.com has two web interfaces, both of which have advantages
and disadvantages. In one interface, attaching a file requires one click
after the file selection dialog; in the other, two clicks. I switch between
the interfaces depending on which disadvantage I am trying to avoid
at the moment :-( In this case, I distinctly remember selecting this
file in my original email, but it seems I forgot the second click,
and the interface doesn't warn files that have had the "select" click
but not the "attach" click :-(

> > By the way, I just realised that with one exception, my stress tests
> > are not in the git repository. I think they should be there.
> > I could add them as benchmarks/stress_tests. Any proposals
> > for better names? Or should I create a separate repository for them?
> 
> tests/stress and don't run it by default?

I thought of that, but then it would be a permanent exception,
different from all the others subdirs of tests. At the moment,
everything in tests is (intended, at least eventually) to run
on every bootcheck, while nothing in benchmarks is run by
bootchecks, and I would prefer to keep it that way.

Unless you are proposing that, on some bootchecks at least,
we should compile the stress tests? That would increase
bootcheck time by a nontrivial amount in the best case,
and cause the machine to thrash to death in the worst case :-(.
The latter is why it hasn't been done up till now.

Zoltan.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: large_progs.pdf
Type: application/pdf
Size: 289389 bytes
Desc: not available
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20210421/169a64a2/attachment-0001.pdf>