Silverfrost Forums

Welcome to our forums

corrupt object module when compiling huge program with debug

10 Dec 2021 12:30 #28614

sorry for the confusion I meant I used the command line ftn95 *.f90 My code is still in 77 format rather than f90, so that's my default

10 Dec 2021 12:33 #28615

Of course I had to rename the file in your post and I used the actual name I had given it and not *.f90

10 Dec 2021 1:18 #28616

And what was that actual name? That is rather important, considering what has been discussed earlier in this thread regarding reserved device names.

10 Dec 2021 1:37 #28617

I called it test1.f90 so the command line calls were ftn95 test1.f90 /link or ftn95 test1.f90 /link/debug

I trust this is unambiguous. Louis

10 Dec 2021 1:59 #28618

I renamed the file to 'test1.f90' and tried the four combinations of /64 and /debug. In every case the compiler crashed with an attempt to read from location 00000000.

At this point it would be useful if other forum readers could try the same tests and tell us what they find.

11 Dec 2021 1:04 #28624

I compiled mecej4's example in Plato I had gFortran selected for /64 and the response in errorlog was:

test_com.f90:5:15: 5 | end module com3 | 1 Fatal Error: Cannot open module file 'com3.mod0' for writing at (1): No such file or directory

compilation terminated.

FTN95 needs to trap this possible error

11 Dec 2021 1:25 #28627

FTN95 has now been fixed for the next release so that the failure to create the file COM3.mod will generate an error report.

If there are other outstanding issues in this thread the please let me know,

Please login to reply.