Odd behavior using %ob/%cb

The following code will cause the second %ob construct to be made 'oddly'.

The root cause is the %ff%nl that immediately precedes 'Second boxed line of text'. Remove this, and it looks exactly as one would expect.

	use mswin
    integer i
    i = winio@('Text Line 1%nl&')
    i = winio@('%1.1ob[named_l][This is a text box]&')
    i = winio@('This is a boxed loiine of text%nl&')
    i = winio@('%cb&')
    i = winio@('%ff%nlIntermediate Text%ff%nl&')
    i = winio@('%1.1ob&')
    i = winio@('%ff%nl&')
    i = winio@('Second boxed line of text%nl&')
    i = winio@('%cb&')
    i = winio@(' ')
    end

I have made a note of this behaviour.

Bill,

I did think for a minute that it might be the Oirish spelling of 'loiine'!

It seems to be the result of %ff in a box immediately after opening it with %ob. Put any other control in, for example a %bt, then the box happily accepts %ff. I'm not actually sure what the %ff does in the context where it messes up the boxes.

Eddie

Eddie,

Umm, yeah, that IS what does it, as I said.

I stumbled across this while re-arranging what would be the contents of the boxes.

And, I knew that the %ff%nl was 'not needed'. It just happened to be there.

All that said, there are two thing to note. First, regardless of the right-ness or wrong-ness of doing it this way, it was unexpected, and second, if this cannot be remedied, then a note of it needs to be made in the documentation.

It took me about 2 hours to find it. I hope that no one else need do that again. I've moved past it.

Bill

With respect, Bill, you said it was %ff%nl (see quote) - and it's just the %ff that does it. Perhaps that qualifies me for the pedantry award this week.

The root cause is the %ff%nl that immediately precedes 'Second boxed line of text'. Remove this, and it looks exactly as one would expect.

%nl does have a place after %ob, and that works. %ff works anywhere in a box except immediately after %ob. There must be an answer to the question why nobody discovered it before - a question I often ask when even more esoteric bugs are discovered.

As far as documentation is concerned, we live in hope. Not only that, even if it had been documented, would you or I have found it in as little as 2 hours?

It seems that the solutions could include (1) to forbid %ff immediately after %ob, (2) to ignore it, (3) to make the case one of those that is checked using the /CHECK switch for checking Clearwin+, and extend that to the 32 bit version of FTN95.

Eddie