ASSOC row$(STR$(i+1)) = row$(STR$(i)) doesn't do the job

[rikky]

Hello.

I'm making a thinking error, but I cannot understand what.
I am trying to shuff the rows up.
So here is an example program:

DECLARE row$ ASSOC STRING
counter = 0

tekst$ = "one" & NL$ & "two" & NL$ & "three" & NL$

FOR line$ IN CHOP$(tekst$) STEP NL$
 INCR counter
 row$(STR$(counter)) = line$
NEXT

PRINT
PRINT "NRKEYS(row$): " & STR$(NRKEYS(row$))
PRINT

FOR i = 1 TO NRKEYS(row$)
 PRINT STR$(i) & " - " & row$(STR$(i))
NEXT i

i = 1

PRINT
PRINT "row$(" & STR$(i) & ") = " & row$(STR$(i))
PRINT "row$(" & STR$(i+1) & ") = " & row$(STR$(i+1))
PRINT

PRINT "row$(" & STR$(i+1) & ") = row$(" & STR$(i) & "):"
PRINT "row$(" & STR$(i+1) & ") = " & row$(STR$(i))

row$(STR$(i+1)) = row$(STR$(i))

PRINT

FOR i = 1 TO NRKEYS(row$)
 PRINT STR$(i) & " - " & row$(STR$(i))
NEXT i

PRINT
PRINT "huh?, should be:"
PRINT "1 - " & row$("1")
PRINT "2 - " & row$("1")
PRINT "3 - " & row$("3")

 Rik.

[Pjot]

Thanks Rik,

Well, you're bumping into the same spot which I thought I had fixed before (see your post here).

The check I added back then only prevents functions on the right side of the '=' to mess up the C code. It did not check if an associative variable - which also uses the '(' and ')' symbols - was used.

So this is my bad, it should have been fixed correctly in the first place, I should have thought of this possibility of course.

Anyway, I have updated the code by now verifying if the left side of the '=' has empty brackets. If so, then it refers to a full assoc array assignment, if not, it is a single element assignment.

Hope this issue now is really really solved <sigh>

Best regards
Peter

PS fix is available in fossil.