Bug: problems with persistence

[Ruslan227]

When saving while, for, repeat, or maybe other statements, some problems occur. For example, we have the following algorithm:

ALGORITHM A1 IN ST: 
  VAR_TEMP 
    Index : INT; 
    Sum: INT;
  END_VAR 
  Index := 0; 
  Sum := 0; 
  WHILE Index < 10 DO 
    Sum := Sum + Index; 
    Index := Index + 1; 
  END_WHILE; 
END_ALGORITHM                                                                                                                                                                                   

After reopening MPS, it shows the algorithm in a changed way:

ALGORITHM A1 IN ST: 
  VAR_TEMP 
    Index : INT; 
    Sum : INT;   
  END_VAR 
  Index := 0; 
  Sum := 0; 
  WHILE Index < 10 DO 
    no statements 
  END_WHILE; 
  Sum := Sum + Index; 
  Index := Index + 2; 
END_ALGORITHM                                                                                                                                                           

And when trying to parse this algorithm in code: org.fbme.lib.iec61499.parser.STConverter.parseStatementList, the org.antlr.v4.runtime.InputMismatchException occurs because input text is Index := 0;&#10;Sum := 0;&#10;WHILE Index < 10 DO&#10; Sum := Sum + Index;&#10; Index := Index + 1;&#10; END_WHILE;&#10; and the error is at the 49th symbol - after WHILE Index < 10 DO

I think the problem is in org.fbme.lib.iec61499.stringify.PrinterBase.escapeXML:

text = text.replace("\n", "
")
text = text.replace("\"", """)
text = text.replace("&", "&")

After "\n" is replaced by 
, the "&" is replaced by &. So, "\n" is replaced by 
 and it is written to a file (saved)

Returning to the text when the ANTLR exception is thrown. The parser reads the saved model with our algorithm. After that, the model is unescaped in org.fbme.lib.iec61499.parser.BasicFBTypeConverter.AlgorithmConverter.extractDeclarationBody:

val stText = stBodyElement.getAttributeValue("Text")?.unescapeXML()

Let's look at unescapeXML() block:

.replace("
", "\n")
.replace(""", "\"")
.replace("&", "&")

As you see, our 
 is replaced by 
. After that, ANTLR throwsorg.antlr.v4.runtime.InputMismatchException.

I believe that, in order to address this issue, it would be advisable to adjust the order of escaping in the following manner: org.fbme.lib.iec61499.stringify.PrinterBase.escapeXML

text = text.replace("\n", "
")
text = text.replace("\"", """)
text = text.replace("&", "&") // this line must be the first in replace commands