It's Back....

[Martin Francom]

Paul,  i hate to report but the crasy compiler error is back.  Yes, I am using the new FF305.   Attached is a screen pront of the error.

I have tryed unloading all forms and reloading them one at a time.  For each form I when into the Tab Order and re arranged the tabs.

I sent you and email and attached the  ....exe.manifest  .
Do you have any suggestions.

[Martin Francom]

Well, I was able to get the program back to a compilable form.  But I was not able to find the problem.

What I did was take a backup, of a previous working version. Then, I copied the newer versions of the forms and modules to the old backup.  I compiled the program after each form or module was added back in hopes of finding the offending piece of code.  Alas, no such luck.  Good news is that the program is now compiling and I didn't loose any of my previous work.

Bad news is I have no ideas what trigger the problem.   Paul are there any hidden files that may have gotten corrupted?

Well, the lesson I learned this time is that you can never have to many backups!

[Paul Squires]

I doubt that this is related to the previous "End Function" problem. That problem was related to FF3 not being able to correctly capture the error from the PB9 compile log file and thereby producing a GPF when trying to position the cursor to the error location. I fixed that in the last update.

I bet that there is actually a missing END FUNCTION somewhere in one of your files. The problem with such an error is that it always gets reported in the CODEGEN*_MAIN.BAS because all of the other files are #INCLUDE'd in that file. The only way that you will track down this type of error is to look at every one of your forms/modules to see where the missing END FUNCTION is located. The PB9 Editor does not do any better job of locating such an error either.

FF3 does not create any hidden files. What you see in the \Release folder is what you get. Only CODEGEN*.* files are compiled.

I am not panicking at this point. 

[Martin Francom]

Ok, makes sense to me. 

So, another piece of the puzzel.  This morning. I brought the program (the one I fixed last night that was compiling OK) back up and there was a FUNCTION that I had previously "Comment Block".  So, I was now ready to work on that section of code so I highlighted that commented section and used the "Unblock Comment"  nothing happend the section was not uncommented.   So, I manually uncommented that section.  Then I compiled.  Guess what happend.  Yep, got the error message again.

I will add back the backup copy of that module.  And, verify the program compiles.  The I will do the same steps over again and see if the proble is reproducable.  If so, I will email you a copy of the project with instructions of how to reproduce the error. 

[Martin Francom]

Paul,
    Starting with a compilable version, I carefully added back the newer forms and modules.  I was able to isolate the offending code.  It is in a FORM.   I can now reproduce the problem by replacxing the working form with the offending form and clear the problem by replacing the offending form with the working form.

   I have look at the offending form in NOTEPAD and could not find any subs or functions without an End Sub or End Function.
   So their must be something else in the form somewhere that is causing the problem.

   Would you like to take a look at it?

[Paul Squires]

Hi Marty,

I just emailed you back. Looks like you spelled one of the END FUNCTION's wrong.... END FNunction.

[Martin Francom]

   Thank you.  I'm sorry I troubled you.   

    I must have looked at that code a dozen times and didn't see the problem. I
apparently type a character  when the cursor on that line.

When programs get large finding that type of error, can be like finding a needle in a haystack, especially when the compiler's error message points off into the wrong direction.