forums.silverfrost.com

[Nathan]

Any ideas what this error means? My calling code where it fails is

read(toprec,'(a)') gifver1

When running the debugger
toprec = '�???????????
gifver1 = 48

character*(mxrecl) toprec
where mxrecl = 512

Any help would be appreciated.
Thanks,

[PaulLaidler]

What are you aiming to do?
A WRITE statement would make sense if you wanted to change an integer to a string...

[felix]

Talking about buffer writing, consider this sample:


! Fortran 95, freebee 6.1, GOT A RUN-TIME BUG!!

! Given the following:

CHARACTER*80 DATA1
CHARACTER*80 DATA2
CHARACTER*168 BUFFER

DATA1 = 'GOOD DATA: 1'
DATA2 = 'GOOD DATA: 2'
BUFFER = 'DIRTY CONTENTS'

! The following does not work as INTENDED!!!!

WRITE (BUFFER, FMT=601) DATA1
601 FORMAT (T001, A80)
WRITE (BUFFER, FMT=602) DATA2
602 FORMAT (T081, A80)
WRITE (BUFFER,603) 'SHOWS UP'
603 FORMAT (T161, A

! Synopsis: Given a bunch of writes to a single
buffer, only the last one above is a keeper.
The previous writes are ignored, in effect,
except for blanking/nulling out the dirties.
Any ERR path, when supplied, is not taken.
The format T indexer should allow over-write
or at least give an ERR condition if not
allowed by decree....Preference is to allow
over-write, but if the BUFFER exceeded, a
run-time ERR must occur if at all reasonable,
coding wise.

! The following works fine.....

WRITE (BUFFER,605) DATA1, DATA2, 'SHOWS UP'
605 FORMAT (A80, A80, A


Sincerly,

FELIX M ROBBINS
2520 SOUTH 22ND STREET
FORT SMITH AR 72901

[JohnCampbell]

There are probably many alternatives, but consider the following:

WRITE (BUFFER(1:80), FMT=601) DATA1
WRITE (BUFFER(81:160), FMT=601) DATA2
WRITE (BUFFER(161:), FMT=601) 'SHOWS UP'
601 FORMAT (A)

This approach limits the part of BUFFER that is set to blank in each write.
Your original example may overwrite the full 168 characters of BUFFER for each write.

John

[Nathan]

I am not too sure what the code does. I am not much of a fortran programmer. When I compile with ftn95 using a makefile, it runs fine. When I debug using ftn95 in VS2005, it stops at this stage in the program with this error. As it runs fine (apparently ????) using the makefile, I was hoping to not change much code in debugger mode, as really my aim is to debug a different part of the program.

[felix]

Thanks, John. I can see the reasoning after realizing I was missing the format style you have given. Gosh, it's been a long time since my Fortran days for real. My last professional experience has been with the CDC compilers. I have skinned the cat many times using EQUIVALENCE statements to the hilt. ... felix