[Pw_forum] Bug in PW/pw_readfile

[Andrea Dal Corso]

Thank you for reporting this. You are right it is a bug that might
appear when lkpoint_dir=.false..

I have corrected the svn version.

Andrea


On Fri, 2012-07-27 at 19:35 +0200, Iurii TIMROV wrote:
> Dear QE developers,
> 
> It seems there is a bug in the subroutine PW/pw_readfile. It shows up when
> a phonon calculation is restarted. I checked QE v.5.0.1, the bug hasn't
> been fixed. It is related to the unit "iunout". In the beginning of the 
> pw_readfile, there is a look for an empty unit:
> 
> CALL iotk_free_unit( iunout, ierr )
> 
> Afterwards, in the subroutine "read_band_structure" the unit "iunout" is
> used. But the problem is that in "read_band_structure" the unit "iunout"
> is declared second time:
> 
> INTEGER :: iunout
> 
> Thus, "iunout" is a random number, which sometimes may be negative, and
> therefore the data file cannot be read. Due to this, the code escapes from
> this routine (return), the data file is not read, and the code crashes
> somewhere else. It is strange to me, why the code escapes (return
> statement) instead of giving an error message (it takes time to find such
> bugs due to this).
> 
> The solution of the problem is simple: just to remove "INTEGER :: iunout".
> 
> If this is not a bug, please explain me why. But in my case the code
> doesn't work with this.
> 
> Best,
> Iurii
> 
> 
> Iurii TIMROV
> Doctorant (PhD student)
> Laboratoire des Solides Irradies
> Ecole Polytechnique
> F-91128 Palaiseau
> +33 1 69 33 45 08
> 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum