42.expmod(-1,1) hangs

[p6rt]

Migrated from rt.perl.org#130713 (status was 'open')

Searchable as RT130713$

[p6rt]

From @lizmat

The combination of parameters -1,1 like this, causes nqp​::expmod_I to hang. Assume this is a MoarVM issue, as this does not appear to be an issue on MoarVM.

Spotted by Dana Jacobsen++

[p6rt]

From @geekosaur

On Fri, Feb 3, 2017 at 4​:33 PM, Elizabeth Mattijsen <
perl6-bugs-followup@​perl.org> wrote​:

Assume this is a MoarVM issue, as this does not appear to be an issue on
MoarVM.

...which of those was supposed to be something else?

--
brandon s allbery kf8nh sine nomine associates
allbery.b@​gmail.com ballbery@​sinenomine.net
unix, openafs, kerberos, infrastructure, xmonad http://sinenomine.net

[p6rt]

The RT System itself - Status changed from 'new' to 'open'

[p6rt]

From @lizmat

On 3 Feb 2017, at 22​:34, Brandon Allbery <allbery.b@​gmail.com> wrote​:

On Fri, Feb 3, 2017 at 4​:33 PM, Elizabeth Mattijsen <perl6-bugs-followup@​perl.org> wrote​:
Assume this is a MoarVM issue, as this does not appear to be an issue on MoarVM.

...which of those was supposed to be something else?

The second​: there’s no problem on the JVM.

Good catch!

Note to self​: do not do bugreports after having been to Delirium in Brussels. :-)

Liz

[p6rt]

From @lizmat

On 3 Feb 2017, at 22​:34, Brandon Allbery <allbery.b@​gmail.com> wrote​:

On Fri, Feb 3, 2017 at 4​:33 PM, Elizabeth Mattijsen <perl6-bugs-followup@​perl.org> wrote​:
Assume this is a MoarVM issue, as this does not appear to be an issue on MoarVM.

...which of those was supposed to be something else?

The second​: there’s no problem on the JVM.

Good catch!

Note to self​: do not do bugreports after having been to Delirium in Brussels. :-)

Liz

[p6rt]

From @timo

It seems like we literally just pass the values into libtommath's mp_exptmod.

It then just infinitely loops around inside of fast_mp_invmod.

Perhaps we can see if a newer version of libtommath changed something about that and perhaps fixed our problem?

[p6rt]

From @timo

There's a github issue that relates to this. Applying this change to our tommath makes expmod(42, -1, 1) terminate successfully.

libtom/libtommath#67

[p6rt]

From @AlexDaniel

This was resolved in this commit trio​:
rakudo/rakudo@22f00cd
Raku/nqp@b083e34
MoarVM/MoarVM@1288e7e

｢testneeded｣

On 2017-02-03 15​:04​:00, timo wrote​:

There's a github issue that relates to this. Applying this change to
our tommath makes expmod(42, -1, 1) terminate successfully.

libtom/libtommath#67

[p6rt]

From @zoffixznet

On Wed, 04 Oct 2017 22​:20​:49 -0700, alex.jakimenko@​gmail.com wrote​:

This was resolved in this commit trio​:
rakudo/rakudo@22f00cd
Raku/nqp@b083e34
MoarVM/MoarVM@1288e7e

｢testneeded｣

On 2017-02-03 15​:04​:00, timo wrote​:

There's a github issue that relates to this. Applying this change to
our tommath makes expmod(42, -1, 1) terminate successfully.

libtom/libtommath#67

Tests added Raku/roast@064707a320

They showed this issue still exists for other combinations, such as​:

  <Zoffix__> m​: say 42.expmod(-1,7)
  <camelia> rakudo-moar 27131e​: OUTPUT​: «(timeout)»