Understand: MISRA C 2023 Rule 13.3 - Loop Error

[kylejacksonb]

While working on a potential FP, I introduced an error for the MISRA C 2012/2023 Rule 13.3 check (build 1239). By turning over the last expression of the for loop, it appears to cause problems reading the next token:

void foo(int p)
{
    int idx;
    for (idx = p; idx <= p;
            idx++) { }
}

Traceback (most recent call last):
File "D:/Programme/SciTools/plugins/CodeCheck/Published Standards/MISRA C 2023/MISRA23_13.3/check.upy", line 224, in check
lex = lex.next(True, True)
^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'next'

[kennethctdmn]

@kylejacksonb The for loop split into two lines causes the error.
I remember being tagged in an issue where a customer requested for a feature to return the starting line of a statement.
Has it been considered?
For now, I'm thinking of adding a 'Developer's Note' regarding this issue.

[kylejacksonb]

Yeah, I ended up writing something ( https://github.com/stinb/plugins-private/issues/1041 ).

I can give you what I came up with, but I don't think it'll be very helpful. Essentially, I used the 'Assignby Value Indirect' reference and walked backwards with the lexer until I found my 'setby' reference that had the same line and column number.

def generate(report, target):
  useRefs = target.refs( 'Assignby Value Indirect' )
  for ref in useRefs:
    location = start_of_statement( ref )
    report.print( "Set refereance: " + str( ref.line() ) + " " + str( ref.column() ) + " - " )
    if location:
      report.print( "Start of statment: " + str( location[ 0 ] ) + " " + str( location[ 1 ] ) + "\n" )
    else:
      report.print( "Could not find start of statment :( \n" )


def start_of_statement( ref ):
  try:
    useEnt = ref.ent()
    file = ref.file()
    lexer = file.lexer( True )
  except:
    return False
  lexeme = lexer.lexeme( ref.line(), ref.column() )
  while lexeme:
    ent = lexeme.ent()
    if ent and ( ent.id() == useEnt.id() ):
      setByRefs = ent.refs( 'Setby' )
      for setBy in setByRefs:
        if ( setBy.line() == lexeme.line_begin() ) and ( setBy.column() == lexeme.column_begin() ):
          return ( lexeme.line_begin(), lexeme.column_begin() )
    lexeme = lexeme.previous()

[kennethctdmn]

@kylejacksonb I see. I had the wrong thought as I assumed that a new API feature for lex would be needed for that.
Anyway, I'm still trying to find a way to make this work.

[kylejacksonb]

Thanks Kenneth!