JACOBIN-547 Incremental fixes to support stringer-1

[texadactyl]

• jvm/run.go - ISTORE lines 575 and 578 expanded to tolerate non-int64 integers.
• jvm/runUtils.go - new function convertInterfaceToInt64 (called in ISTORE).
• gfunction/javaLangString.go - new G functions replaceCC and split.

[texadactyl]

Created a G function for String.replace(oldchar, newchar) in spite of what the commit text said! (:

[texadactyl]

stringer-1 status at this point

Failing sometime after calling the OpenJDK JBM lib fn Strings.split(). class=java/lang/String, meth=split(Ljava/lang/String;)[Ljava/lang/String;

This one is a bit complicated. Golang has a strings package simple-split function which does not entertain regular expressions. Unfortunately, for stringer-1, the Java String.split function defines one parameter as a regular expression.

• One can supply a simple string as the regular expression. This is what jacotest case stringer-1 does. Temporary poor man's split G function?
• We can write new "stringer" test cases for actual regular expressions in order to provide better coverage of the API.
• Addressing the API might involve both the Go strings package and the regexp package.
• We probably also need a G function for String.replaceAll which also takes a regular expression.

@platypusguy

• I implemented a cheap-o-cigar split function that assumes the regular expression is actually just an ordinary string.
• stringer-1 now passes.
• TODO: handle a regular expression.

Over to you, if you like. (:

[texadactyl]

jacotest just now:

17:15:34 Success in 91 test cases
17:15:34 No Execution timeout errors
17:15:34 Execution failure errors in 50 test cases
17:15:34 Wrote test case summary to: /home/elkins/BASIS/jacotest.go/reports/Summary_2024-07-10_17.13.06_jacobin.txt
17:15:34 Elapsed time = 2m28s
17:15:34 Looking for test cases with changed results .....
    cur  >>  stringer-1  jacobin   2024-07-10  22:15:27.297  passed  n/a
    prv  >>  stringer-1  jacobin   2024-07-10  21:08:23.776  failed  java.lang.NullPointerException: ARRAYLENGTH: Invalid (null) value for *object.Object
17:15:34 Number of test cases with changed results = 1

[texadactyl]

@platypusguy

So, IMO:

• Tolerate unexpected but compatible formats in the output of OpenJDK library routines when it is possible - at least, the simplest of cases.
• Otherwise, build a G function (last resort).