Oops... forgot to answer to the list.
---------- Forwarded Message ----------
Subject: Re: [cclib-devel] Partial files
Date: Thursday 16 April 2009
From: Karol Langner <kar...@gm...>
To: "Noel O'Boyle" <bao...@gm...>
Hi,
Sorry for answering with delay. I actually have given this some thought
before, and no, it's not necessary. You are correct to write that there are
too many possibilities to cover in the parsers. I will not go on adding
shredded output to the regression tests :)
With that said, I still think that covering the most obvious/common cases
is worth it. The way I truncate output files sometimes (to save disk space)
is probably pretty typical... cut out the chunky sections such as
eigenvectors and populations. It's pretty important to me that these output
files can be parsed, even if they contain incomplete data.
Also, the changes I made to the code that fix this are more general - try
clauses like the one you cited below. They only envelop a fragment of the
parser, so if something goes wrong there the error message is quite
specific.
Hope that's OK,
Karol
On Tue, Apr 7, 2009 at 10:48 AM, Noel O'Boyle <bao...@gm...> wrote:
> Karol,
>
> Is it really necessary to add warnings for partial output files? There are
> about 1000 points of failure here. What I do in GaussSum is a simple "catch
> all":
>
> try:
> # parse with cclib
> except:
> # message to user saying it couldn't be parsed, please contact support
> if necessary
>
> - Noel
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> High Quality Requirements in a Collaborative Environment.
> Download a free trial of Rational Requirements Composer Now!
> http://p.sf.net/sfu/www-ibm-com
> _______________________________________________
> cclib-devel mailing list
> ccl...@li...
> https://lists.sourceforge.net/lists/listinfo/cclib-devel
>
>
-------------------------------------------------------
--
written by Karol Langner
Wed Jul 15 18:05:28 CEST 2009
|
