Prev: Fortran in newly released JASON report
Next: template for module compilation
From: Bil Kleb on 25 Apr 2010 14:14
Hi,

What am I missing? I can't figure out how to have
Fortran warn me that it has truncated a character
string during a namelist read...

% cat > nml_string_truncation.f90 <<EOF
program nml_string_truncation

integer :: ios
character(35) :: nml_contents = "&NMLIST NML_STRING='123456789' /"
character(4) :: nml_string
namelist /nmlist/ nml_string

write(*,*) nml_contents
open(7,file='example.nml') ; write(7,*) nml_contents ; rewind(7)
read(7,nml=nmlist,iostat=ios)
write(*,nml=nmlist)

if (ios==0) write(*,*) 'Error: namelist read did not detect string
truncation'

end program
EOF

% ifort -warn -check nml_string_truncation.f90 && ./a.out
&NMLIST NML_STRING='123456789' /
&NMLIST
NML_STRING = 1234
/
Error: namelist read did not detect string truncation

% g95 -Wall nml_string_truncation.f90 && ./a.out
&NMLIST NML_STRING='123456789' /
&NMLIST NML_STRING = '1234',
/
Error: namelist read did not detect string truncation

% gfortran -Wall nml_string_truncation.f90 && ./a.out
nml_string_truncation.f90:4.29:

character(4) :: nml_string
1
Warning: Unused variable 'nml_string' declared at (1)
&NMLIST NML_STRING='123456789' /
&NMLIST
NML_STRING="1234",
/
Error: namelist read did not detect string truncation

TIA,
--
Bil Kleb

From: dpb on 25 Apr 2010 14:43
Bil Kleb wrote:
> Hi,
>
> What am I missing? I can't figure out how to have
> Fortran warn me that it has truncated a character
> string during a namelist read...
>
....

AFAIK there isn't any way just as an A or list-directed read will return
the maximum length of the input string or the limit of the CHARACTER
variable length to hold the input.

Only reading the record in a variable long enough to hold it and parsing
it will determine that there are more characters than space allocated to
hold them under Standard i/o operations afaiaa.

Some compilers have an option to scan ahead and return the remaining
number of characters in the input record ("Q" edit descriptor in
DEC/Compaq/Intel compilers, I don't know who else may have that or
something similar) but it is an extension, not Standard (albeit
occasionally quite useful).

--
From: Richard Maine on 25 Apr 2010 14:52
Bil Kleb <bil.kleb(a)nasa.gov> wrote:

> What am I missing? I can't figure out how to have
> Fortran warn me that it has truncated a character
> string during a namelist read...

Well most warnings are compilation-time things. In particular, that's so
for switches like the -Wall and -warn that you tried.

Note that iostat does not even define a warning category. It has
end-of-file, end-of-record, and error categories, but no warnings.

I don't know whether there exist any compilers that have the capability
to detect I/O string truncation like that (and possibly elevate it to an
error that could be returned as an iostat code). If so, it will probably
be some more particular switch than the general -Wall or -warn kinds of
things. I'm sure you won't find it on most compilers, but you might find
one somewhere. (If I had to guess a candidate, it might be IBM; I don't
know that they have such a thing, but I'd think them the best guess).

--
Richard Maine | Good judgment comes from experience;
email: last name at domain . net | experience comes from bad judgment.
domain: summertriangle | -- Mark Twain
From: Richard Maine on 25 Apr 2010 15:12
Richard Maine <nospam(a)see.signature> wrote:

> Bil Kleb <bil.kleb(a)nasa.gov> wrote:
>
> > What am I missing? I can't figure out how to have
> > Fortran warn me that it has truncated a character
> > string during a namelist read...
>
> Well most warnings are compilation-time things. In particular, that's so
> for switches like the -Wall and -warn that you tried.
....
> (If I had to guess a candidate, it might be IBM; I don't
> know that they have such a thing, but I'd think them the best guess).

I forgot to mention that, as it would be an I/O library kind of option,
you would more likely find it controlled by something like environment
variables instead of compiler switches. The reason I mentioned IBM is
because I recall that IBM at least used to (probably stilll does, but
its been a while since I used an IBM system) have environment variables
to control some I/O issues. I don't know if they have that particular
one, but I recall once running into an environment variable to control a
different I/O issue that had annoyed me. (The IBM system at the time was
by default "fixing up" errors such as illegal characters in numeric
input. That was when I first ran into the practical implications of the
set of I/O error conditions being compiler dependent. I later found that
at least there was an environment variable to fix what I regarded as the
obnoxious behavior. The default might also have subsequentky changed.)

--
Richard Maine | Good judgment comes from experience;
email: last name at domain . net | experience comes from bad judgment.
domain: summertriangle | -- Mark Twain
From: Bil Kleb on 25 Apr 2010 16:15
> Bil Kleb <bil.kleb(a)nasa.gov> wrote:
>
>> What am I missing? I can't figure out how to have
>> Fortran warn me that it has truncated a character
>> string during a namelist read...
>
> Well most warnings are compilation-time things. In particular, that's so
> for switches like the -Wall and -warn that you tried.

Sorry for the red herrings, I use -Wall and -warn options out of
habit, I was not expecting them to warn me of runtime string
truncation during a namelist read.

> Note that iostat does not even define a warning category. It has
> end-of-file, end-of-record, and error categories, but no warnings.

Again, perhaps I used the wrong terminology. The

if (ios==0) write(*,*) 'Error: namelist ...'

line is the "warning" I am seeking.

> I don't know whether there exist any compilers that have the capability
> to detect I/O string truncation like that (and possibly elevate it to an
> error that could be returned as an iostat code).

IBM has an CNVERR= extension that seems to offer this
(see http://bit.ly/bZmKAr), but the standard apparently does
not support this essential sort of diagnostics to make namelists
useful?

Regards,
--
Bil Kleb

 |  Next  |  Last
Pages: 1 2
Prev: Fortran in newly released JASON report
Next: template for module compilation