[Pw_forum] Re: Pw_forum digest, Vol 1 #1173 - 1 msg

[Tianshu Li]

Hi Paolo,

Thanks for your reply.

In fact the code works perfectly if it reads the output generated from 
the same version. It only gives problem when there is a file I/O across 
the different versions. So this is just what you've described.

Tianshu

pw_forum-request at pwscf.org wrote:

>From: Paolo Giannozzi <giannozz at nest.sns.it>
>Subject: Re: [Pw_forum] Compatibility among different versions
>Date: Sat, 7 Oct 2006 01:16:29 +0200
>To: pw_forum at pwscf.org
>Reply-To: pw_forum at pwscf.org
>
>On Oct 6, 2006, at 1:40 , Tianshu Li wrote:
>
>  
>
>>I am curious if there might be a compatibility problem in file I/O  
>>among
>>different versions, because when I used the post-processing code, for
>>instance, pp.x in the latest version (v.3.1.1) to read the .save file
>>generated in the previous version (v.3.0), it complained it could not
>>find the .save file.
>>    
>>
>
>this is definitely a compatibility problem!
>
>  
>
>>The message is like:
>>
>>from pp_check_file : error #         1
>>file  **********************/**.save not found.
>>
>>I noticed that in the latest version the new function "pp_check_file"
>>was added to the xml_io_base.f90. This might be what gives the  
>>problem,
>>but I am not very sure of that. Thanks
>>    
>>
>
>in principle with the new format there shouldn't be any longer any
>problem of incompatibility, because the file format is extensible, i.e.
>if something more is needed, one just adds a new field, without
>breaking the rest. Or at least, so goes the theory. In practise:
>
>- one may realise that the way things are done is not smart and that
>it needs to be changed, so that a field name or the content of a field
>may no longer be what it used to be. This happens especially at the
>beginning, when the file format has to stabilize.
>
>- if something is added to the file format, the code reading it should
>be aware that something may be missing from files produced by earlier
>versions, For instance, this "pp_check_file" stuff was added to read
>a tag that is written at convergence. In the first version, the code  
>simply
>stopped if there was no such tags, thus making all files produced by
>previous versions unreadable. Now this should be corrected.
>
>In the future we hope that there will be no incompatible changes.
>For the past ... the past is gone. The specific error you have looks
>very strange and I am not sure it is related to the file format. Does
>it work if you try to read files produced with the same version of
>the code?
>
>Paolo
>
>
>
>
>--__--__--
>
>_______________________________________________
>Pw_forum mailing list
>Pw_forum at pwscf.org
>http://www.democritos.it/mailman/listinfo/pw_forum
>
>
>End of Pw_forum Digest
>
>  
>


-- 
Tianshu Li
Chemistry Department
University of California, Davis
Tel: 1-530-754-9664

-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/attachments/20061006/edeb075f/attachment.htm