[squeak-dev] Failing tests for Interval

Nicolas Cellier nicolas.cellier.aka.nice at gmail.com
Thu Dec 12 16:13:58 UTC 2019


 Hi Marcel,
Interval of Float are nasty beasts and the tests somehow document that.
I have proposed some partial "fix" but I'm not sure if it's worth the
complexification or not, that's why I put in inbox.
So if no one is sure, just mark as expected failures.


Le jeu. 12 déc. 2019 à 16:10, Marcel Taeumel <marcel.taeumel at hpi.de> a
écrit :

> Hi Eliot.
>
> > Use your words ;-).
>
> I'll try my best. :-) The following tests fail on my system:
>
> IntervalTest >> #testIntervalOfFloatLast
> IntervalTest >> #testIntervalOfFloatReversed
> IntervalTest >> #testSurprisingFuzzyInclusionBehavior
>
> This is my system:
>
> - Squeak5.3beta #19270
> - Image format 6521 (32 bit)
> - VM 201911282316
> - OS Windows 10 (Surface Pro 6)
>
> I think that Nicolas is aware of the issue:
> http://forum.world.st/The-Inbox-Collections-nice-825-mcz-tp5098004.html
>
> http://forum.world.st/The-Trunk-CollectionsTests-nice-305-mcz-tp5098002.html
>
> Best,
> Marcel
>
> Am 12.12.2019 15:57:53 schrieb Eliot Miranda <eliot.miranda at gmail.com>:
> Hi Marcel,
>
> On Dec 12, 2019, at 5:09 AM, Marcel Taeumel <marcel.taeumel at hpi.de> wrote:
>
> 
> Hi, there.
>
> <image.png>
>
>
> Use your words ;-). Further, I need to know on which VM (ISA & OS) they
> appear.  And it is also useful to find out if there are any VMs on which
> they pass.
>
> What about those failing tests? Can they be fixed or should they go to
> #expectedFailures?
>
>
> Stuff like socket tests that fail because of missing infrastructure etc
> may not be fixable.  But basic arithmetic should always be fixable,
> otherwise things are seriously broken with our entire ecosystem :-)
>
> Levente raised an issue that could be related to these failures and ill
> investigate asap (it may have to be the weekend; I have pressing work
> responsibilities).
>
> Are we ensuring that we are running tests across all our release VMs?
>
>
> Best,
> Marcel
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20191212/b139d71a/attachment.html>


More information about the Squeak-dev mailing list