FWD » Core Development » Base Language

Roger, please try this patch (on top of trunk rev 14484):

=== modified file 'rules/convert/assignments.rules'
--- old/rules/convert/assignments.rules 2022-06-02 13:02:58 +0000
+++ new/rules/convert/assignments.rules 2023-02-20 18:47:40 +0000
@@ -428,7 +428,7 @@
          </rule>

          <!-- setters for assignable global vars -->
-         <rule>evalLib("vartype", ref.type) and ref.type != prog.sys_handle
+         <rule>evalLib("vartype", ref.type) and ref.type != prog.sys_handle and not evalLib("is_java_cls", ref)

             <!-- global variables are the only VAR_ types that don't have a refid -->
             <rule>!ref.isAnnotation("refid")

That results in a successful conversion. So those warnings are just red-herrings?

Is this a fix I should try against the customer's code?

Roger Borrello wrote:

Is this a fix I should try against the customer's code?

Yes.

Yes

If this passes testing, we will merge it to trunk.

Please work on a standalone test and lets use this task for the customer's app issues with trunk (conversion or runtime).

Constantin Asofiei wrote:

Please work on a standalone test and lets use this task for the customer's app issues with trunk (conversion or runtime).

Working on it... I crashed because I was out of disk space.

This testcase:

PROCEDURE getMousePoint :
  DEFINE INPUT  PARAMETER pcObject AS HANDLE NO-UNDO.

  DEFINE VARIABLE iMouseLoc AS INTEGER NO-UNDO EXTENT 2.

    ASSIGN 
      iMouseLoc = pcObject:FRAME:GET-MOUSE-POSITION().
END PROCEDURE.

creates the same error:

------------------------------------------------------------------------------
Core Code Conversion
------------------------------------------------------------------------------

Optional rule set [customer_specific_conversion] not found.
./uast/roger.w
Elapsed job time:  00:00:00.966
ERROR:
com.goldencode.p2j.pattern.TreeWalkException: ERROR!  Active Rule:
-----------------------
      RULE REPORT      
-----------------------
Rule Type :   WALK
Source AST:  [ = ] BLOCK/PROCEDURE/BLOCK/STATEMENT/KW_ASSIGN/ASSIGN/ @7:17 {347892351023}
Copy AST  :  [ = ] BLOCK/PROCEDURE/BLOCK/STATEMENT/KW_ASSIGN/ASSIGN/ @7:17 {347892351023}
Condition :  rvalueType = ecw.expressionType(rvalueExp, false)
Loop      :  false
--- END RULE REPORT ---

    at com.goldencode.p2j.pattern.PatternEngine.run(PatternEngine.java:1093)
    at com.goldencode.p2j.convert.TransformDriver.processTrees(TransformDriver.java:586)
    at com.goldencode.p2j.convert.ConversionDriver.back(ConversionDriver.java:585)
    at com.goldencode.p2j.convert.TransformDriver.executeJob(TransformDriver.java:998)
    at com.goldencode.p2j.convert.ConversionDriver.main(ConversionDriver.java:1284)
Caused by: java.lang.UnsupportedOperationException: Unknown METH_POLY with oldtype of KW_GET_MOP
    at com.goldencode.p2j.convert.ExpressionConversionWorker.expressionType(ExpressionConversionWorker.java:2085)
    at com.goldencode.p2j.convert.ExpressionConversionWorker.expressionType(ExpressionConversionWorker.java:1423)
    at com.goldencode.p2j.convert.ExpressionConversionWorker.expressionType(ExpressionConversionWorker.java:1183)
    at com.goldencode.p2j.convert.ExpressionConversionWorker$ExpressionHelper.expressionType(ExpressionConversionWorker.java:3691)
    at com.goldencode.expr.CE10013.execute(Unknown Source)
    at com.goldencode.expr.Expression.execute(Expression.java:373)
    at com.goldencode.p2j.pattern.Rule.apply(Rule.java:500)
    at com.goldencode.p2j.pattern.Rule.executeActions(Rule.java:751)
    at com.goldencode.p2j.pattern.Rule.coreProcessing(Rule.java:717)
    at com.goldencode.p2j.pattern.Rule.apply(Rule.java:537)
    at com.goldencode.p2j.pattern.Rule.executeActions(Rule.java:751)
    at com.goldencode.p2j.pattern.Rule.coreProcessing(Rule.java:717)
    at com.goldencode.p2j.pattern.Rule.apply(Rule.java:537)
    at com.goldencode.p2j.pattern.Rule.executeActions(Rule.java:751)
    at com.goldencode.p2j.pattern.Rule.coreProcessing(Rule.java:717)
    at com.goldencode.p2j.pattern.Rule.apply(Rule.java:537)
    at com.goldencode.p2j.pattern.Rule.executeActions(Rule.java:751)
    at com.goldencode.p2j.pattern.Rule.coreProcessing(Rule.java:717)
    at com.goldencode.p2j.pattern.Rule.apply(Rule.java:537)
    at com.goldencode.p2j.pattern.Rule.executeActions(Rule.java:751)
    at com.goldencode.p2j.pattern.Rule.coreProcessing(Rule.java:717)
    at com.goldencode.p2j.pattern.Rule.apply(Rule.java:537)
    at com.goldencode.p2j.pattern.RuleContainer.apply(RuleContainer.java:590)
    at com.goldencode.p2j.pattern.RuleSet.apply(RuleSet.java:98)
    at com.goldencode.p2j.pattern.RuleContainer.apply(RuleContainer.java:590)
    at com.goldencode.p2j.pattern.RuleSet.apply(RuleSet.java:98)
    at com.goldencode.p2j.pattern.AstWalker.walk(AstWalker.java:262)
    at com.goldencode.p2j.pattern.AstWalker.walk(AstWalker.java:210)
    at com.goldencode.p2j.pattern.PatternEngine.apply(PatternEngine.java:1685)
    at com.goldencode.p2j.pattern.PatternEngine.processAst(PatternEngine.java:1571)
    at com.goldencode.p2j.pattern.PatternEngine.processAst(PatternEngine.java:1504)
    at com.goldencode.p2j.pattern.PatternEngine.run(PatternEngine.java:1056)
    ... 4 more

Tree looks:

       <ast col="0" id="347892351020" line="0" text="statement" type="STATEMENT">
          <ast col="5" id="347892351021" line="6" text="ASSIGN" type="KW_ASSIGN">
            <annotation datatype="java.lang.Long" key="support_level" value="16400"/>
            <annotation datatype="java.lang.Boolean" key="bracket" value="false"/>
            <ast col="17" id="347892351023" line="7" text="=" type="ASSIGN">
              <ast col="7" id="347892351026" line="7" text="iMouseLoc" type="VAR_INT">
                <annotation datatype="java.lang.Long" key="oldtype" value="2947"/>
                <annotation datatype="java.lang.Long" key="extent" value="2"/>
                <annotation datatype="java.lang.Long" key="refid" value="347892351003"/>
                <annotation datatype="java.lang.Long" key="support_level" value="16400"/>
              </ast>
              <ast col="0" id="347892351027" line="0" text="expression" type="EXPRESSION">
                <annotation datatype="java.lang.Long" key="support_level" value="16400"/>
                <ast col="33" id="347892351028" line="7" text=":" type="COLON">
                  <ast col="27" id="347892351029" line="7" text=":" type="COLON">
                    <ast col="19" id="347892351030" line="7" text="pcObject" type="VAR_HANDLE">
                      <annotation datatype="java.lang.Long" key="oldtype" value="2947"/>
                      <annotation datatype="java.lang.Long" key="refid" value="347892350987"/>
                      <annotation datatype="java.lang.Long" key="support_level" value="16400"/>
                    </ast>
                    <ast col="28" id="347892351031" line="7" text="FRAME" type="ATTR_HANDLE">
                      <annotation datatype="java.lang.Long" key="oldtype" value="660"/>
                      <annotation datatype="java.lang.Long" key="support_level" value="16400"/>
                    </ast>
                  </ast>
                  <ast col="34" id="347892351032" line="7" text="GET-MOUSE-POSITION" type="METH_POLY">
                    <annotation datatype="java.lang.Long" key="oldtype" value="1503"/>
                  </ast>
                </ast>
              </ast>
            </ast>
          </ast>
        </ast>

Roger, please try this patch:

=== modified file 'src/com/goldencode/p2j/convert/ExpressionConversionWorker.java'
--- old/src/com/goldencode/p2j/convert/ExpressionConversionWorker.java  2023-01-12 15:38:22 +0000
+++ new/src/com/goldencode/p2j/convert/ExpressionConversionWorker.java  2023-02-21 19:05:57 +0000
@@ -1898,6 +1898,11 @@
                   jcls = "memptr";
                   break;

+               case KW_GET_MOP:
+                  // TODO: check if we can improve on this
+                  jcls = "integer[]";
+                  break;
+                  
                case KW_IF:
                   // by default return the same type as the second operand
                   jcls = expressionType(source.getChildAt(1), false, fuzzy);

Greg, I don't see something cleaner - GET-MOUSE-POSITION is marked as METH_POLY in progress.g and SignatureHelper, I think because it returns integer[].

It is fine for now.

This is probably something different. The toLegacyString(jobject) is our replacement name for toString() which conflicts with Object.toString() in Java. This version has a special signature that takes jobject which would actually not conflict with the no-arg Object version. Anyway, look at the LabelRenderer and the non-abstract child classes as see what we are generating here and why.

For this class:

USING Progress.Lang.*.  
                              /* FWD mode */
  USING com.goldencode.p2j.ui.chart.* FROM java.
  USING java.util.*                   FROM java. 

CLASS sy.UltraChart.RenderLabelData 
                              /* FWD mode */
 IMPLEMENTS com.goldencode.p2j.ui.chart.LabelRenderer:

  DEFINE PRIVATE VARIABLE id AS CHARACTER NO-UNDO
    INITIAL "x".

                                 /* FWD mode */
    METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS java.util.Map): 
      DEFINE VARIABLE fValue AS DECIMAL NO-UNDO.

      ASSIGN fValue = DECIMAL(pChartHash:get("DATA_VALUE"):toString()) NO-ERROR.
      CASE pChartHash:get("DATA_VALUE_TYPE"):toString():
        WHEN "£" THEN       
          RETURN "£ " + TRIM(STRING(fValue,"->>>,>>>,>>9")). 
        WHEN "%" THEN 
          RETURN STRING(ROUND(fValue,2)) + " %".
        OTHERWISE
          RETURN pChartHash:get("DATA_VALUE"):toString().
      END CASE.    

    END METHOD. 

END CLASS. 

We are generating:

public class RenderLabelData
extends BaseObject
implements com.goldencode.p2j.ui.chart.LabelRenderer
{
   @LegacySignature(type = Type.VARIABLE, name = "id")
   private character id = TypeFactory.character("x");

   public void __client_clientnet_sy_ultrachart_RenderLabelData_execute__()
   {
      /* FWD mode */
      externalProcedure(RenderLabelData.class, RenderLabelData.this, new Block((Body) () -> 
      {
         {
         }
      }));
      /* FWD mode */
   }

   @LegacySignature(type = Type.CONSTRUCTOR)
   public void __client_clientnet_sy_ultrachart_renderLabelData_constructor__()
   {
      internalProcedure(RenderLabelData.class, this, "__client_clientnet_sy_ultrachart_renderLabelData_constructor__", new Block((Body) () -> 
      {
         __lang_BaseObject_constructor__();
      }));
   }

   @LegacySignature(type = Type.METHOD, name = "ToString", returns = "CHARACTER", parameters = 
   {
      @LegacyParameter(name = "pChartHash", type = "JOBJECT", qualified = "java.util.Map", mode = "INPUT")
   })
   /* FWD mode */   
   public character toString_(final jobject<? extends java.util.Map> _pChartHash)
   {
      jobject<? extends java.util.Map> pChartHash = TypeFactory.initInput(_pChartHash);
      decimal fValue = TypeFactory.decimal();

      return function(RenderLabelData.class, this, "ToString", character.class, new Block((Body) () -> 
      {
         silent(() -> fValue.assign(new decimal(pChartHash.ref().get("DATA_VALUE").toString())));

         switch (StringHelper.safeTrimTrailing(pChartHash.ref().get("DATA_VALUE_TYPE").toString()))
         {
            case "£":
               {
                  storeReturnValue(concat("£ ", trim(valueOf(fValue, "->>>,>>>,>>9"))));
                  return;
               }
            case "%":
               {
                  storeReturnValue(concat(valueOf(round(fValue, 2)), " %"));
                  return;
               }
            default:
               {
                  storeReturnValue(pChartHash.ref().get("DATA_VALUE").toString());
                  return;
               }
         }
      }));
   }
}

LabelRenderer class is:

public interface LabelRenderer
{
   public character toLegacyString(jobject<? extends Map> formats);
}

We are choosing to emit toString_() name instead of toLegacyString() in the child classes.

The com.goldencode.p2j.ui.chart.LabelRenderer does not include the normal legacy signature annotations (at either the class or method levels). I suspect that is at least part of the problem.

Greg Shah wrote:

The com.goldencode.p2j.ui.chart.LabelRenderer does not include the normal legacy signature annotations (at either the class or method levels). I suspect that is at least part of the problem.

Is there a good example which I could follow to address that? In looking through some of the com.goldencode.p2j.ui.chart classes, there aren't any annotations to model after.

Look at anything in com.goldencode.p2j.oo.

Greg, adding the LegacySignature at a Java interface method will not work.

Before 6129c, toString I think had a special meaning during conversion, that it got converted to toLegacyString. For now, I propose for the 4GL method name to match exactly the overridden Java method name, toLegacyString.

I think for Java inheritance in the 4GL code, we need to hard-code the method names, and do not convert them at all - treat them as their Java counterpart.

You are proposing that we just hand-edit the 4GL source code to change the names?

Greg Shah wrote:

You are proposing that we just hand-edit the 4GL source code to change the names?

Correct. If a 4GL-style class is inherited from a Java class, then the method names must be treated as Java names.

So the 4GL would need to be patched to have toString become toLegacyString:

  &IF DEFINED(FWD-VERSION) = 0 &THEN  /* NOT FWD */
    METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS System.Collections.Hashtable): 

      DEFINE VARIABLE fValue AS DECIMAL NO-UNDO.

      ASSIGN fValue = DECIMAL(pChartHash["DATA_VALUE"]:ToString()) NO-ERROR.
      CASE pChartHash["DATA_VALUE_TYPE"]:ToString():
        WHEN "£" THEN       
          RETURN "£ " + TRIM(STRING(fValue,"->>>,>>>,>>9")). 
        WHEN "%" THEN 
          RETURN STRING(ROUND(fValue,2)) + " %".
        OTHERWISE
          RETURN pChartHash["DATA_VALUE"]:ToString().
      END CASE.    

    END METHOD. 
  &ELSE                               /* FWD mode */
    METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS java.util.Map): 
      DEFINE VARIABLE fValue AS DECIMAL NO-UNDO.

      ASSIGN fValue = DECIMAL(pChartHash:get("DATA_VALUE"):toLegacyString()) NO-ERROR.
      CASE pChartHash:get("DATA_VALUE_TYPE"):toString():
        WHEN "£" THEN       
          RETURN "£ " + TRIM(STRING(fValue,"->>>,>>>,>>9")). 
        WHEN "%" THEN 
          RETURN STRING(ROUND(fValue,2)) + " %".
        OTHERWISE
          RETURN pChartHash:get("DATA_VALUE"):toLegacyString().
      END CASE.    

    END METHOD. 
  &ENDIF 

Yes

Roger, sorry, in this case rename toLegacyString in the Java class to toString. That name was chosen because of the way FWD hard coded that name at the time.

Constantin Asofiei wrote:

Roger, sorry, in this case rename toLegacyString in the Java class to toString. That name was chosen because of the way FWD hard coded that name at the time.

OK... so you mean change FWD's LabelRenderer method:

public interface LabelRenderer
{
   /**
    * Get the string representation for a certain value from format map.
    * 
    * @param    formats
    *           The map holding the values and formats.
    * 
    * @return   The string representation.
    */
   public character toString(jobject<? extends Map> formats);
}

Is that the only rename necessary?

Roger, is a little more tricky than this, NameConverter will automatically append an underscore because toString belongs to java.lang.Object and is a reserved name.

I think for this case of Java inheritance from within 4GL code, FWD needs to preserve the method name if this is an override; so, if we mark the method with OVERRIDE (although 4GL doesn't require it...) in the 4GL code, this will also allow the Java compiler to force the override, and for the FWD conversion to know to preserve this method name because it matches a super-method.

In our ToString case, the other problem is that if you use ToString from the 4GL code, you actually really need to use toString, because this is the 'real Java method name inherited from the LabelRenderer interface'. In the customer's app, can you easily identify and replace ToString with toString in the 4GL code?

Greg, if we start going with LegacySignature, we are adding a binding only for Java code from within FWD, which does not solve the problem if the interface/class is from another Java library.

Roger, as a fast way to solve it: in the 4GL code, can we change all RenderLabelData.ToString(INPUT pChartHash AS java.util.Map) references + definition to be renamed asString, and the same with FWD's LabelRenderer.toString, renamed to LabelRenderer.asString?

This will move out of the way all the problems we have with toString, and I'll create a task to deal with the points discussed in previous notes.

1. app_7135a-14486_20230215a/src/.../ultrachart/RenderLabelData.java
2. app_7135a-14486_20230215a/src/.../ultrachart/RenderLabelItemToDateShort.java
3. app_7135a-14486_20230215a/src/.../ultrachart/RenderLabelLegend.java
4. app_7135a-14486_20230215a/src/.../ultrachart/RenderLabelSeriesToDateShort.java
5. app_7135a-14486_20230215a/src/.../ultrachart/RenderLabelTooltip.java

to be modified similar to:

  &ELSE                               /* FWD mode */
    METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS java.util.Map): 
      DEFINE VARIABLE fValue AS DECIMAL NO-UNDO.

      ASSIGN fValue = DECIMAL(pChartHash:get("DATA_VALUE"):asString()) NO-ERROR.
      CASE pChartHash:get("DATA_VALUE_TYPE"):asString():
        WHEN "£" THEN       
          RETURN "£ " + TRIM(STRING(fValue,"->>>,>>>,>>9")). 
        WHEN "%" THEN 
          RETURN STRING(ROUND(fValue,2)) + " %".
        OTHERWISE
          RETURN pChartHash:get("DATA_VALUE"):asString().
      END CASE.    

    END METHOD. 
  &ENDIF 

And src/com/goldencode/p2j/ui/chart/LabelRenderer.java to be modified:

=== modified file 'src/com/goldencode/p2j/ui/chart/LabelRenderer.java'
--- src/com/goldencode/p2j/ui/chart/LabelRenderer.java    2020-03-04 19:24:49 +0000
+++ src/com/goldencode/p2j/ui/chart/LabelRenderer.java    2023-02-23 18:33:22 +0000
@@ -2,10 +2,11 @@
 ** Module   : LabelRenderer.java
 ** Abstract : Definition for a label renderer.
 **
-** Copyright (c) 2020, Golden Code Development Corporation.
+** Copyright (c) 2020-2023, Golden Code Development Corporation.
 **
 ** -#- -I- --Date-- ---------------------------------Description---------------------------------
 ** 001 CA  20200304 Created initial version.
+** 002 RFB 20230223 Changed method name from toLegacyString to asString. Ref. #TBD.
 */

 /*
@@ -84,5 +85,5 @@
     * 
     * @return   The string representation.
     */
-   public character toLegacyString(jobject<? extends Map> formats);
+   public character asString(jobject<? extends Map> formats);
 }

Something like this, and ToString (INPUT pChartHash AS java.util.Map): needs to be asString (INPUT pChartHash AS java.util.Map):

But, I can't find who calls the original ToString(java.util.Map) method - can you look in the Java code for a previous conversion of the customer's app and find the toLegacyString(java.util.Map) calls? I don't have any converted code for the app, and looking through the abl/ code I have, I can't find any.

I think LabelRenderer is part of the FWD's ChartWidget.labelHash, which is not implemented at this time.

FYI, officially we do not support subclassing 4GL code from Java parent classes or implementing Java interfaces in 4GL code. I like the idea that we bypass this restriction for now.

I see this:

rfb@rfb:~$ grep -R toLegacyString\( .
./client/clientnet/sy/ultrachart/RenderLabelSeriesToDateShort.java:   public character toLegacyString(final jobject<? extends java.util.Map> _pChartHash)
./client/clientnet/sy/ultrachart/RenderLabelItemToDateShort.java:   public character toLegacyString(final jobject<? extends java.util.Map> _pChartHash)
./client/clientnet/sy/ultrachart/RenderLabelTooltip.java:   public character toLegacyString(final jobject<? extends java.util.Map> _pChartHash)
./client/clientnet/sy/ultrachart/RenderLabelData.java:   public character toLegacyString(final jobject<? extends java.util.Map> _pChartHash)
./client/clientnet/sy/ultrachart/RenderLabelLegend.java:   public character toLegacyString(final jobject<? extends java.util.Map> _pChartHash)

So should I make those changes? Or is there more to this?
That's the same list of errors.

Roger, thanks, that confirms what you posted. Just change the 4GL method definitions in these classes to be asString and the same with LabelRenderer. This should solve the problem.

I've just noticed something in your patch - I mean to change:

METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS java.util.Map):

to

METHOD PUBLIC CHARACTER asString (INPUT pChartHash AS java.util.Map):

Only the ToString which has an argument as java.util.Map needs to change. Do not touch plain ToString() calls. I don't know what those are.

Constantin Asofiei wrote:

Only the ToString which has an argument as java.util.Map needs to change. Do not touch plain ToString() calls. I don't know what those are.

Got it... just the signatures:

diff -r old/abl/client/clientnet/sy/UltraChart/RenderLabelData.cls new/abl/client/clientnet/sy/UltraChart/RenderLabelData.cls
46c46
<     METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS System.Collections.Hashtable): 
---
>     METHOD PUBLIC CHARACTER asString (INPUT pChartHash AS System.Collections.Hashtable): 
diff -r old/abl/client/clientnet/sy/UltraChart/RenderLabelItemToDateShort.cls new/abl/client/clientnet/sy/UltraChart/RenderLabelItemToDateShort.cls
45c45
<     METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS java.util.Map): 
---
>     METHOD PUBLIC CHARACTER asString (INPUT pChartHash AS java.util.Map): 
diff -r old/abl/client/clientnet/sy/UltraChart/RenderLabelLegend.cls new/abl/client/clientnet/sy/UltraChart/RenderLabelLegend.cls
49c49
<     METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS java.util.Map): 
---
>     METHOD PUBLIC CHARACTER asString (INPUT pChartHash AS java.util.Map): 
diff -r old/abl/client/clientnet/sy/UltraChart/RenderLabelSeriesToDateShort.cls new/abl/client/clientnet/sy/UltraChart/RenderLabelSeriesToDateShort.cls
45c45
<     METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS java.util.Map): 
---
>     METHOD PUBLIC CHARACTER asString (INPUT pChartHash AS java.util.Map): 
diff -r old/abl/client/clientnet/sy/UltraChart/RenderLabelTooltip.cls new/abl/client/clientnet/sy/UltraChart/RenderLabelTooltip.cls
62c62
<     METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS java.util.Map): 
---
>     METHOD PUBLIC CHARACTER asString (INPUT pChartHash AS java.util.Map):

So the incremental build completed, including the compilation. I will need to do some testing, but this is committed to 7135 as revision 14487.

There's a lot of issues. Noted one for the GUI application in #6851-203. The ChUI app gets ** LABEL is not a setable attribute for LITERAL ?. (4052) at one point.

Open bugs for these runtime issues. If conversion works now, we move on and close this task.

Opened task #7146 for the other runtime issue.

Is the patching of the customer app procedures as noted in #7135-34 (and show below) required to be ongoing? Or is there a FWD change that will render it unnecessary?

--- sy/UltraChart/RenderLabelData.cls
+++ sy/UltraChart/RenderLabelData.cls
@@ -59,7 +59,7 @@

     END METHOD. 
   &ELSE                               /* FWD mode */
-    METHOD PUBLIC CHARACTER ToString (INPUT pChartHash AS java.util.Map): 
+    METHOD PUBLIC CHARACTER asString (INPUT pChartHash AS java.util.Map): 
       DEFINE VARIABLE fValue AS DECIMAL NO-UNDO.

       ASSIGN fValue = DECIMAL(pChartHash:get("DATA_VALUE"):toString()) NO-ERROR.

Roger Borrello wrote:

Opened task #7146 for the other runtime issue.

Is the patching of the customer app procedures as noted in #7135-34 (and show below) required to be ongoing? Or is there a FWD change that will render it unnecessary?

[...]

The patch is permanent.

Constantin Asofiei wrote:

The patch is permanent.

Thanks... I will pursue the customer checking the patched versions into their change control.

7135a has been rebased to trunk_14490.

Roger Borrello wrote:

7135a has been rebased to trunk_14490.

If there are other customers that might have the same type of code as the customer that found this issue, please let me know (or perhaps this branch has already been used to do a conversion for them).

Not at this time.