Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo

Close

#28 wrong result for assignment.

closed
Mark Hessling
None
5
2012-01-02
2010-04-29
No

In the following program, Regina assigns a null to Z:

/**/
y=++4
say y
z=--4
say z

_____________________________________ Gerard Schildberger

Discussion

  • Mark Hessling
    Mark Hessling
    2010-04-29

    This is because "--" is a single line comment in Regina since v3.4. So your statement is:
    z=

    So z is empty.

     
  • Mark Hessling
    Mark Hessling
    2010-04-29

    • assigned_to: nobody --> rexx
    • status: open --> closed-invalid
     
  • Mark Hessling
    Mark Hessling
    2010-09-17

    This tracker item has been changed to a feature request; its not a bug.

    While I can't understand why someone would code -- instead of +, I have added an OPTION to disallow single line comments.
    To allow:
    a = b -- 3
    to parse as the same as:
    a = b + 3
    then use the OPTION: NOSINGLE_LINE_COMMENTS
    However, this OPTION cannot be set inside the Rexx program. This is because Regina parses the contents of the file (including determining what is a comment, before the OPTIONS instruction is executed.

    You will need to set this option using the operating system's environment before calling a Rexx program:
    Unix: export REGINA_OPTIONS="NOSINGLE_LINE_COMMENTS"
    Windows/OS2: set REGINA_OPTIONS=NOSINGLE_LINE_COMMENTS

    This feature will be available in the next release of Regina

     
  • Mark Hessling
    Mark Hessling
    2010-09-17

    • status: closed-invalid --> pending-fixed
     
    • status: pending-fixed --> open
     
  • If the code is a legal REXX syntax, and conforms to the rules of allowable expressions, then the failure to parse the expression is a bug. The "feature" of single line comments broke the
    allowing of double unary prefixes. The user of REXX shouldn't have to specify a special option just to allow legal REXX code.

    This bug should also not be dependent on the lack of understanding on why someone (a programmer) would code - - instead of +. The double unary prefix may be generated programatically, or it may be someone's style (whether someone else approves of it or
    not).

    It shouldn't be in the perview of anyone to allow legal REXX code to be marginalized (or worse yet, disallowed) because of a new feature. If a new feature is going to be added to REXX, a peer review would be a better way to go so as to find out (if any) if there are any construct rules would be broken or disallowed. I don't recall anyone asking for OOREXX
    comments in Regina, and even so, it would've been judious to examine the ramfications of
    such a change.

    The use of two minus signs for comments is something that OO-REXX has, why introduce it
    to (classic) REXX? What would be the consequences if a double slash or double asterisk was chosen instead? Both of those (bad) choices are more apparent why they would be
    a wrong choice.

    I can't understand why someone would code >< instead of <>, but then, I don't have to. It's part of the REXX syntax (in this case, comparitive operators). I don't have to understand it, approve of it, or like it. It certainly wouldn't be right of me to break it's (allowing of) use.

    I can understand the emotional investment in a feature, especially once the feature has been
    added to the REXX interpreter. It would be much appreciated if double unary operators
    would be acknowledged, but not judged.

    ____________________________________________________ Gerard Schildberger

     
  • Mark Hessling
    Mark Hessling
    2012-01-02

    • status: open --> closed