forums.silverfrost.com

[DanRRight]

Compile this program which has wrong line marked with the arrow

[John-Silver]

Dan, have you tried putting an 'end subroutine ' line in, maybe that's throwing it into confusion ?

[DanRRight]

The point is different, John. It is about better diagnostics of dumb errors by the compiler

[John-Silver]

yep I understood that, but if the (so-called) subroutine definition is amiss there could be a link to that

I note that the error message is wrong too.

"INTEGER cannot be declared inside SUBROUTINE block"

a) an integer can be declared inside a subroutine block of course !
b) You haven't defined the integer function inside the (half) subroutine block, but it thinks you have !

[John-Silver]

it works after putting 'end subroutine ' in (only warning that nothing in there !)

It would be useful to have a FATAL error that there's no end subroutine, although it does mention in error message MAYBE there is a missing 'end' but as you said it's not referring to any subroutine but to the function.

Looks like x2 mixed up messages to me.
One for Paul&Co to trace


- NOTE -

i) if you move the subroutine definition ABOVE the function definition you get the error message.
*** SUBROUTINE cannot be declared inside PROGRAM block (perhaps missing CONTAINS or END statement?)

Which is a logical error message !

AND


ii) if you remove (comment out) the function you get:

*** SUBROUTINE cannot be declared inside SUBROUTINE block (perhaps missing CONTAINS or END statement?)

!iii) if you remove the module and put the function defn in the main program only then you get:

*** SUBROUTINE cannot be declared inside PROGRAM block (perhaps missing CONTAINS or END statement?)

which are both logical messages.
Seems to be the presence of the INTEGER function being before the subroutine which is throwing it.
Maybe subroutines are processed first ? Still doesn't explain why it thinks the function is INSIDE the Subroutine, which s the crux of the matter.

[John-Silver]

Just realised you've posted this in 'Suggestions' forum Dan, not Support.

I assume the suggestion, not made, is "get some better diagnostic messages sorted pdq" ?

[PaulLaidler]

I have logged this for investigation.