Latent error in Flowcoding stopping compile at later date

[Mark]

I have found an example of a ‘hidden’ flowcode compile bug. The story is below as it may help you save time if you are faced with this somewhat difficult to find issue. In short the situation was :

Large programme, many macros, made some trivial changes to main code.
Programme then fails to compile and gives an undimensioned array message.
Using the undo facility on that change or even on several before made no difference.
The error was eventually tracked down to a local variable in a macro (that had not been edited for some time), where I had originally defined a local string variable but had later stopped using that local variable in the macro itself.
Going back and removing the ‘orphan’ variable removed the problem.

Why the later trivial change triggered the effect and why it took so long to appear I do not know. In any case this story may save you the time if you get a similar the problem. The lesson of course being to have good housekeeping.

Happy coding,

Mark

(p.s. this was using the latest Flowcode revision, some earlier edits had been done with previous V3 updates).

[Benj]

Hello Mark

Thanks for the info. thats a rather strange problem. Can you tell us anything more about the nature of the bug. i.e. was it the fact that the variable was unused or was it due to being part of an earlier release do you think. Would be good to get to the bottom of it as stray variables although bad practise should not cause compilation errors.

[Steve]

I'm hoping that this bug was caused by the recent changes to Flowcode. That latest update (v3.1.0) fixed some major bugs in strings used as local variables - one bug was that the array dimension of a local string was not retained.

[edit]
It is related to this post:
http://www.matrixmultimedia.com/mmforum ... .php?t=972
[/edit]

If anyone can replicate this situation with a program created in Flowcode v3.1, then please let me know.

[Mark]

Steve,

Your earlier post is spot on (I had missed it).

In retrospect the problem was because the amendments were carried out on a PC with Patch 1.2.4 and the compile on a PC with patch 1.3.5.

Updating the PC with the old patch (3.1.0.24?) to the new patch (today, i.e. 3.1.0.31) gave rise to the error again (with the same solution).

Thanks for your prompt responses, good product being made better .

Mark