[Cuis-dev] Random fixes

[Luciano Notarfrancesco]

Hi Hernan,
Yes, I think testSeed0 should be removed. It is better to make
Random>>seed: accept any Integer as argument, and the seed must competently
determine the generator output.

Accepting any Integer argument is desirable because we can use Random
without making assumptions of its internals, and I think initializing a
random number generator with an arbitrary integer is reasonable and
general. Completely determining the generator output is important also for
repeatability of tests and experiments, we want to make sure that
initializing two generators of the same class with the same seed will
produce the same outputs.

Here are 7 new tests and a new version of the fixes for Random. I removed
testSeed0 too. This version fixes another two issues, and they are tested
for correctness now. Let me know what you think.


On Thu, Jun 6, 2019 at 6:55 PM Hernan Wilkinson <
hernan.wilkinson at 10pines.com> wrote:

> Hi Luciano,
>  there are three tests that fail when running the RandomTest of the
> BaseImageTests.pck.st
>  In reality, it is the same test, testSeed0 for the different Random
> implementations, that expects to fail when the seed is 0 and it does not.
>  I'm not an expert on the matter so I'm not sure the test is correct or
> wrong. I presume you change allow a seed to be 0, something not allowed
> before, but could you take a look at it and fix what it is necessary (the
> test or your change).
>  Let me know and I'll integrate it after.
>
> Thanks!
> Hernan.
>
> On Thu, Jun 6, 2019 at 2:51 AM Luciano Notarfrancesco via Cuis-dev <
> cuis-dev at lists.cuis.st> wrote:
>
>> This changeset fixes an infinite loop when doing:
>>   Random new nextInteger: 0
>>
>> It also allows to use any integer as seed to initialize a new Random, so
>> you can do something like:
>>   Random seed: 12938479182374979182734987
>> without making assumptions about the particular implementation of Random
>> that will be created. The new ParkMiller88Random will be initialized
>> exactly as before if the seed is in the originally expected range (keeping
>> backwards compatibility); but if the seed is not in that range, now it will
>> initialize the generator with a valid seed derived from the user-provided
>> seed, instead of raising an exception.
>>
>> The changeset also adds a method Random>>nextBoolean that returns true or
>> false.
>>
>> --
>> Cuis-dev mailing list
>> Cuis-dev at lists.cuis.st
>> https://lists.cuis.st/mailman/listinfo/cuis-dev
>>
>
>
> --
>
> *Hernán WilkinsonAgile Software Development, Teaching & Coaching*
> *Phone: +54-011*-4893-2057
> *Twitter: @HernanWilkinson*
> *site: http://www.10Pines.com <http://www.10pines.com/>*
> Address: Alem 896, Floor 6, Buenos Aires, Argentina
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20190607/4de1162a/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 3785-MoreRandomFixes-LucianoEstebanNotarfrancesco-2019Jun07-02h37m-len.1.cs.st
Type: application/vnd.sailingtracker.track
Size: 1530 bytes
Desc: not available
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20190607/4de1162a/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: UnsavedChangesTo-BaseImageTests-len.1.cs.st
Type: application/vnd.sailingtracker.track
Size: 2641 bytes
Desc: not available
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20190607/4de1162a/attachment-0001.bin>