EMResearch
add support for H2 Geometry column types
support for POINT, MULTIPOINT, LINESTRING, MULTILINESTRING, POLYGON and MULTIPOLYGON column types
hi @jgaleotti @ZhangMan1126 what is the state of this PR? I can see the following is failing:
org.evomaster.core.database.insertion.postgres.PostgresInsertValueTest
[INFO]
[INFO] Results:
[INFO]
Error: Errors:
Error: PostgresInsertValueTest.initClass:-1 » PSQL FATAL: the database system is star...
is it just a flacky test?
hi, @arcuri82 , the disabled test in GeneTest has been fixed.
I tried to run the failed test (PostgresInsertValueTest) locally, and it passed.
there might be some problems in starting databases in testcontainer.
Also noticed that there is a comment in the test for this problem.
hi, when running on my machine, I got 3 tests failing: 1348, 1387, 1657 i made a fix for first 2, but 3rd seems related to BigInteger @ZhangMan1126 can u please look at it?
also i m wondering if different JDKs are using different implementations of random generator, as in theory with a fixed seed we should get exactly the same tests failing
@arcuri82 as checked on my local machine, GeneTest should pass now.
In addition, as discussed, regarding a case whereby the min and max are same for number gene,
I added tests in NumberGeneTest, it seems fine, ie, the initial value is assigned as min/max and the isMutable is false.
regarding randomize with true forceNewValue for this case (eg, line 145 in GeneTest), now we provide a warning representing that a new value cannot be produced if the min and max are same.
@jgaleotti related to this, in master branch, can run CheckRandomness? want to see if you get different results than me, which are:
728
683
309
277
666
904
369
276
464
783
920
interestingly, the 1657 that failed on my machine did not fail here on GitHub Actions
@arcuri82 the 1657 fails on my machine as well, and the problem has been fixed with the latest commits (ie, f081c62).