forums.silverfrost.com

[Thomas]

The following simplified source code creates an "Internal compiler error" when compiled with /64 and /check

Subroutine DRGD2Z (PSZL)
Character(len=*) :: PSZL(*)
Character(Len=10), External :: DRUTBQ
PSZL(1)(52:61) = DRUTBQ(1)
END

Without /64 no error appears.

The error can be avoided with an additional assignmnet:

Subroutine DRGD2Z (PSZL)
Character(len=*) :: PSZL(*)
Character(Len=10), External :: DRUTBQ
Character(Len=10) :: xxx
xxx = DRUTBQ(1)
PSZL(1)(52:61) = xxx
END
_________________
Thomas

[PaulLaidler]

Thomas

Thank you for this bug report. I have made a note that it needs fixing.

[PaulLaidler]

This failure has now been fixed for the next release of FTN95.

[JohnCampbell]

Does Subroutine DRGD2Z (PSZL) need an interface for using PSZL ?
Character(len=*) :: PSZL(*)

Perhaps (len=*) is always provided, so PSZL(*) is an array of unknown size.
So, len(PSZL(1)) is available, but size(PSZL) is not.

I would be interested if the following works.

[PaulLaidler]

The original code provided by Thomas compiles in 32 bits and in 64 bits without /CHECK so FTN95 can be used to investigate any limitations.

[John-Silver]

Paul, what was the reason for the failure ?
_________________
''Computers (HAL and MARVIN excepted) are incredibly rigid. They question nothing. Especially input data.Human beings are incredibly trusting of computers and don't check input data. Together cocking up even the simplest calculation ... "