cclib-devel

[cclib-devel] broken vibdisps

[Adam T. <a-t...@st...>]

I've just discovered that the vibdisps of dvb_ir (and others) isn't  
parsed correctly in cclib 0.8. I think it is an off-by-one error, but  
I haven't investigated fully. It probably wasn't caught until now  
because the testvib script doesn't look at displacements yet.

Adam

Re: [cclib-devel] broken vibdisps

[Adam T. <a-t...@st...>]

So I fixed this bug. Now we should discuss how we want to update the  
testvib script to look at displacements. I have two ideas:

1) Take the block of displacements, save it to a text file, and have  
the script compare parsed frequencies to those.

2) Create a regression file that creates a NRVS spectrum (how I  
originally found the bug) so we can visually compare to some reference.

Adam

On May 29, 2008, at 2:31 PM, Adam Tenderholt wrote:

> I've just discovered that the vibdisps of dvb_ir (and others) isn't
> parsed correctly in cclib 0.8. I think it is an off-by-one error, but
> I haven't investigated fully. It probably wasn't caught until now
> because the testvib script doesn't look at displacements yet.
>
> Adam
>
>
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> cclib-devel mailing list
> ccl...@li...
> https://lists.sourceforge.net/lists/listinfo/cclib-devel

Re: [cclib-devel] broken vibdisps

[Noel O'B. <bao...@gm...>]

2008/5/30 Adam Tenderholt <a-t...@st...>:
> So I fixed this bug. Now we should discuss how we want to update the
> testvib script to look at displacements. I have two ideas:
>
> 1) Take the block of displacements, save it to a text file, and have
> the script compare parsed frequencies to those.
>
> 2) Create a regression file that creates a NRVS spectrum (how I
> originally found the bug) so we can visually compare to some reference.

I would say that this is a regression, and so belongs in
regressions.py. I would just do (1) but rather than save it to a text
file I would just paste it into regression.py. Perhaps in future we
should add a regression test to regression.py before rewriting
sections.

Is there any way of protecting against this particular sort of error
in parsers, for example, by considering the effect of the error? I
haven't checked but I think it caused the same data to be included
several times as x, y, and z. Maybe we can test for this in testvib?

> Adam
>
> On May 29, 2008, at 2:31 PM, Adam Tenderholt wrote:
>
>> I've just discovered that the vibdisps of dvb_ir (and others) isn't
>> parsed correctly in cclib 0.8. I think it is an off-by-one error, but
>> I haven't investigated fully. It probably wasn't caught until now
>> because the testvib script doesn't look at displacements yet.
>>
>> Adam
>>
>>
>>
>>
>> -------------------------------------------------------------------------
>> This SF.net email is sponsored by: Microsoft
>> Defy all challenges. Microsoft(R) Visual Studio 2008.
>> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>> _______________________________________________
>> cclib-devel mailing list
>> ccl...@li...
>> https://lists.sourceforge.net/lists/listinfo/cclib-devel
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> cclib-devel mailing list
> ccl...@li...
> https://lists.sourceforge.net/lists/listinfo/cclib-devel
>

Re: [cclib-devel] broken vibdisps

[Adam T. <a-t...@st...>]

> I would say that this is a regression, and so belongs in
> regressions.py. I would just do (1) but rather than save it to a text
> file I would just paste it into regression.py. Perhaps in future we
> should add a regression test to regression.py before rewriting
> sections.

I agree that it's a regression and so should go into regressions.py.  
And yes, we should probably make sure we have a regression test before  
rewriting sections.

> Is there any way of protecting against this particular sort of error
> in parsers, for example, by considering the effect of the error? I
> haven't checked but I think it caused the same data to be included
> several times as x, y, and z. Maybe we can test for this in testvib?

The error basically caused the first frequency of a set of three to be  
parsed correctly (ie. [x1, y1, z1]) while the second and third  
frequencies were parsed as [y1, z1, x2] and [z1, x2, y2], respectively.

Adam