Compilation Errors missing necessary details in v2.6

davidgoddard

Hi, There are certain compilation errors reported by the editor with missing details which make it hard to figure out where the error is.

For example, if a line contains just a dot (.) when you compile the program you get the following error message

EENG00... , line . variable has not been assigned a value.

Since the line number is missing it's near on impossible to discover which line has the error.

davidgoddard

Screen shot of the error
line 6 has a dot and causes the compilation error, but the error message does not report the line number.

DonBakke

David,

Thank you for posting this. Martin has already brought this to our attention and provided some simple code that replicates the problem easily enough. We have already entered this into our bug tracker (the tracking ID is EDITOR-19). Now that this is reported in our discussion forum we will announce a fix here as well as our home page.

KevinFournier

I hate to point fingers, but why can't Revelation just make things consistent? The reason you get a useless error message is because Repository("COMPILE", ...) returns it. Yet Rev_Compiler returns better errors?

I'll work on updating the SRP Editor compile logic to use Rev_Compiler instead.

davidgoddard

thanks

DonBakke

Kevin,

Are you saying Revelation changed the way error messages are being returned? Is this not a new problem?

KevinFournier

It's not a new problem. It turns out that the documentation does say to use RTP5 if you want more accurate errors. The problem is that RTP5 doesn't create the debug record. I have to experiment with Rev_Compiler to see if I can still use it compile inherited entities.

KevinFournier

Good news. I wasn't understanding the issue at first, but after some digging around, I discovered the issue and it will most certainly be fixed in the next release.