Re: -frecord-marker

[Adrian Umpleby <apumpleby at yahoo dot co dot uk>]

>>> According to the gfortran man page, -frecord-marker seems to be a command-line
>>> option - i.e. you add it when you run the compiled executable:
>>>  ./myprog -frecord-marker=4
>> 
>> -frecord-marker=4 is a compile-time option.  You need to specify this
>> on the command line, as with "gfortran -frecord-marker=4".

Well, after creating a short test, it's clear it's not a runtime option (which is what the online
manual and man page claims).

Also, for a subroutine, it's not actually a compile-time option either, since:

gfortran -frecord-marker=4 -c mysub.f

gives exactly the same object as:

gfortran -frecord-marker=8 -c mysub.f


Which is why I assumed it must be strictly a linker option - i.e. the following, linking in the
object created in the lines above:

gfortran -frecord-marker=4 mymain.f mysub.o -o myprog.exe


However, it looks like I was wrong in both senses, since, as Keith says below, it *only* applies
to compiling the main program itself, and it's not applied at link time.

So:

gfortran -frecord-marker=4 -c mymain.f

does produce a different object to:

gfortran -frecord-marker=8 -c mymain.f


Dead annoying!


> It only works if the main routine is written in Fortran. I've been meaning
> to bring this up for a while because it makes it impossible to write
> libraries using gfortran which are then called by C. (something I do all
> the time at work).

It also makes it impossible to write a program that could test for and read either format.

If gfortran is going to behave in a way which is unusable with (the default for) all previous
Fortran compilers, I'd suggest it needs an extension to make it compatible on a case-by-case basis
- preferably in OPEN itself.

Looks like I'll have to join gcc bugzilla and start making a noise...


> I always preferred the intel method, but no-one seems to want to implement
> that. Maybe I'll get around to it myself one of these days.

Yes, it would be nice to have that option - there was talk, in the early days of getting this
compatibility added to gfortran (as a runtime option!), of including support for hp format, but I
don't see mention of intel's way of dealing with this as a possibility for gfortran, which is a
pity.

Maybe I'll start digging around in the gfortran source to see what I can do about it myself...

Thanks!

Adrian



	
	
		
___________________________________________________________ 
All new Yahoo! Mail "The new Interface is stunning in its simplicity and ease of use." - PC Magazine 
http://uk.docs.yahoo.com/nowyoucan.html