etime

[Janne Blomqvist]

Jack Howarth wrote:
>    Could someone point me to where the removal of etime was
> discussed on the mailing list?

http://gcc.gnu.org/ml/gcc-patches/2006-05/msg01211.html

http://gcc.gnu.org/ml/gcc-patches/2006-05/msg01388.html

> I do recall seeing that a
> patch was checked in to remove it before the sover was bumped
> on libgfortran but I don't recall the reasons for removing it
> being mentioned. More importantly is there a discussion of the
> recommended approach for replacing etime calls in legacy g77
> code?

etime isn't actually removed, it's just the etime_ symbol that has been 
removed. etime is still available with the symbol name _gfortran_etime, 
consistent with other intrinsics. Basically, etime_ was an ill conceived 
hack so that the lapack testsuite (which declares etime as external, not 
as an intrinsic) could be compiled. The correct thing is to remove the 
"external etime" declaration from the lapack testsuite (and possibly 
other programs that do the same thing).


-- 
Janne Blomqvist