by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Aug 04 17:22
    ekrich commented #1864
  • Jul 31 10:49
    NickSeagull commented #113
  • Jul 30 22:05
    ekrich synchronize #1653
  • Jul 30 19:45
    ekrich synchronize #1653
  • Jul 30 12:23
    errikos opened #1867
  • Jul 30 00:39
    ekrich synchronize #1653
  • Jul 30 00:16
    ekrich synchronize #1653
  • Jul 29 22:40
    ekrich synchronize #1653
  • Jul 29 22:39

    errikos on master

    Fix runtime/tests with disabled… (compare)

  • Jul 29 22:39
    errikos closed #1866
  • Jul 29 22:36
    errikos synchronize #1866
  • Jul 29 22:17

    sjrd on master

    Update to Scala Native 0.4.0-M2… (compare)

  • Jul 29 22:17
    sjrd closed #46
  • Jul 29 22:17
    sjrd edited #46
  • Jul 29 19:01
    ekrich synchronize #1653
  • Jul 29 18:13
    ekrich synchronize #46
  • Jul 29 17:11
    errikos edited #1866
  • Jul 29 17:02
    errikos review_requested #1866
  • Jul 29 16:55
    errikos synchronize #1866
  • Jul 29 16:19
    ekrich commented #1866
Kirill A. Korinsky
@catap
Good, it looks nice
LeeTibbert
@LeeTibbert

BTW and the test is till running. Since 2020-06-27T21:45:01.983746Z. It is working on macOS

Do you have any way to know if it is making progress?

Kirill A. Korinsky
@catap
Yes
it works :)
Eric K Richardson
@ekrich
The slogging one looks good as so many people use one of those interfaces already.
LeeTibbert
@LeeTibbert
Good! Actually, Great!! I've written too many infinite loops and/or dog slow code in my life... I always like to see confirmation of work.
Kirill A. Korinsky
@catap
Honesly I've expected that it finished for a while ago, but I run this test on JVM version and JVM version uses all CPU that is available at this machine. And SN version let say a bit longer. Much longer.
it performs tonns of calculation with modPow and this sort of things and each couple of hours dumps a new value.
LeeTibbert
@LeeTibbert

re: Any milestone to release M3 [I'm asking to understand how many time have I got to finish biginteger]?
I am an individual contributor, not a ScalaCenter person, I think the answer would be "Reasonably correct is better than fast". The past few months, the scala-contributors Discourse has had updates from ScalaCenter and some brief discussion of futures. I find that quite useful.

Now that JUnit test support is in, there may be a request that you/I/we convert the BigIntegerSuite.scala to JUnit. I just took a look at it: recognized the fingerprints on some, did not run git blame on rest (he says as he scurries for a fig leaf).
Would it be easier for you if I created a PR which just converted the existing tests to JUnit so that you could add your new tests using a model or are you more comfortable with the SN historical test idiom. My goal/intention is to ease your contribution (and save you time.)

Kirill A. Korinsky
@catap
@LeeTibbert the issue is: right now this has scala js's unit tests that I've migrated to scalatests with a few extentions
LeeTibbert
@LeeTibbert
Scala.js unit tests were probably JUnit to begin with. So that is a good starting point. I am pretty sure that any PR review would request that either the existing BigIntegerSuite.scala get new test cases or that BISuite.scala be converted to JUnit and the new cases added. Perhaps one/we could propose a new BigIntegerMumbleTest.scala, as JUnit, which contained
Kirill A. Korinsky
@catap
RIgth now differences with SJS code base is huge:
git diff c6b075366b70 -- shared/src/main/scala/ky/korins/math | diffstat 
 BigInteger.scala     |  297 ++++++++++++++++++++++++----------
 BitLevel.scala       |    8 
 Conversion.scala     |   27 +--
 Division.scala       |  445 +++++++++++++++++++++++++++++++++++++++++++--------
 Elementary.scala     |    2 
 Multiplication.scala |   82 ++++-----
 Primality.scala      |  250 +++++++++++++++++++++-------
 7 files changed, 842 insertions(+), 269 deletions(-)
So, I have no idea how to move it back
@LeeTibbert for me it is resanable to migrate tests to JUint :)
LeeTibbert
@LeeTibbert
only a port from ScalaTest to JUnit? Knowing that there is a more extensive set of tests which pound on the code but which are not suitable for SN GitHub adds confidence in the underlying code and minimal SN tests proviced in PR, IMHO.
Kirill A. Korinsky
@catap
Ok, let talk about it when I created a PR
at the smaller one I've raised a few question
Eric K Richardson
@ekrich
@catap Are you saying you have a PR to Scala.js already?
LeeTibbert
@LeeTibbert

Ah, that code is in shared/src/main. I don't think there is any need to 'move it back'. Sorry that I was not clear, I meant to be talking about test cases, especially for that number of changes (which is good to see). Currently unit-tests/mumble/BigInteger.scala provides a small set of test cases. That was written to the historical SN test Interface.

You mentioned having scalatest unit tests? Some or all of those are what you/I/we could port to a unit-tests/mumble/BigIntegerTest.scala (as JUnit). If they have to come out of scalatest, might as well go straight to Junit rather than pausing with the historical test idiom, IMHO.

Kirill A. Korinsky
@catap
one with a few tirivial things
because I'm using this code also inside JVM I discovered a few bug that isn't a bug
Eric K Richardson
@ekrich
ok, so those tests are all in JUnit correct?
Kirill A. Korinsky
@catap
I grab SJS unit tests and migrated them to scalatests, after that I added a few new
Eric K Richardson
@ekrich
I usually do one platform at a time for PR. Why did you not just leave the tests even new ones as JUnit?
Scalatests for Native platform?
Kirill A. Korinsky
@catap
yes
@ekrich after a few tricks it works very well :) scalatest/scalatest#1858
Eric K Richardson
@ekrich
Ok, because we have JUnit now in Scala Native, you could duplicate the tests and create a PR for Scala Native in parallel. If the tests are the same on both platforms, the review has a better chance.
The whole key here now is that the Scala.js and Native platforms can more freely share code and tests now that we are on JUnit. This one I converted SN tests to JUnit and then just copied all the tests from SJS. Then when I backport to SJS, I'll have less to do.
Kirill A. Korinsky
@catap
This mean that I will migrated tests one more time :)
Eric K Richardson
@ekrich
Yes, but you can't get the PR in now without JUnit on either platform.
Kirill A. Korinsky
@catap
@ekrich few days ago discovered some bugs in my code and fixed it
here I'd like to wait a bit longer until I haven't touch if for a while
Eric K Richardson
@ekrich
No worries, just want you to be prepared - these changes to javalib are crucial and scrutinized closely. I'm sure your changes are fabulous but you still have to make it through the review.
Kirill A. Korinsky
@catap
and this will be very interesting review
LeeTibbert
@LeeTibbert
@ekrich I just had to fall back to the Fake -M2 hack for my uTest studies. I am thinking of using the mental tag
"Eric's hack" so that I do not have to take ownership (insert emoji for corprophageous grin)
Eric K Richardson
@ekrich
You probably beat me to it.
@catap did you find JVM problems so some of the JVM tests are ignored. I found some when doing IdentityHashMap.
Kirill A. Korinsky
@catap
@ekrich I removed ignore from all tests and all them works well on js, jvm and native platform
but I used only biginterger related stuff
Kirill A. Korinsky
@catap
Honeslty I've migrated everything to scalatests because run juint tests on all three platforms via sbt wasn't an easy task
Eric K Richardson
@ekrich
I understand. If we had an M3 with JUnit you would be fine. You'll just have to adjust things when you are done for the Scala.js PR to get reviewed and then the SN one will be just basically a copy of SJS, I assume.
Kirill A. Korinsky
@catap
Am I right that M3 won't be in next couple of months?
Eric K Richardson
@ekrich
I would like to see a releases more often like every three months but Scala Center is running the show so it is their call. I would sure like to get my 2 PRs in prior but we will have to see how things play out. To me, once we have good value it would be nice but it looks like Scala 2.12 at least but I am just guessing.
Kirill A. Korinsky
@catap
I see
Eric K Richardson
@ekrich
M2 doesn't even have the C valist functions let alone commix and a few other fixes. To me those were enough for an M3. Now we are going to get JUnit, 2.12 and a host of other things I suspect. We could go to 0.4.0 and then jump to 0.5.0 after that - I am just talking out load and none of this should be taken as anything more than that.
Kirill A. Korinsky
@catap
my main concern regarding M3 that releases is very rare, and probably 0.4.0 will be not in this year ;)
Eric K Richardson
@ekrich
Well, Denys finished his PhD and then moved on so we are very happy to get Scala Center and community support and then fine engineers @errikos and @WojciechMazur and our fearless leader and Scala superstar @sjrd . It is a new platform for them except for Sebastien as he helped Denys so it will take longer due to the pause. The PR queue and issues have gone down quite a bit so overall the progress has been super good.