[Liblas-devel] GetClassification() et al.

[Peder Axensten]

I did cc you, but your address bounced -- writing on a smartphone has its limitations, I guess...

See below.

Regards, Peder
:

Från: Peder Axensten <Peder.Axensten at slu.se<mailto:Peder.Axensten at slu.se>>
Datum: 25 oktober 2011<x-apple-data-detectors://1> 22:23:29 CEST
Till: Mateusz Łoskot <mateusz at loskot.net<mailto:mateusz at loskot.net>>
Kopia: "hobu.inv at gmail.com<mailto:hobu.inv at gmail.com>" <hobu.inv at gmail.com<mailto:hobu.inv at gmail.com>>
Ämne: Re: [Liblas-devel] GetClassification() et al.

I am really sorry Mateusz, I did indeed read your mail to quickly.

I think we agree that Point.GetClassification() and Point.SetClassification( ... ) should not filter the Classification in any way?
(I don't really understand why there are 3 SetClassification() as Classification has constructors to cover the other two cases, but never mind.)

I would think that the bug is that the right hand assignment expression of the various Point.SetClassification(...) is Classification(...).GetClass(). The correct would be static_cast<bitset>(Classification(...)) or something similar. Do you agree?

As of retrieving all bits from a Classification object I had not noticed the operator bitset() member function. I guess static_cast<uint8_t>(static_cast<bitset>( p.GetClassification() ) ) will work for me, with no execution overhead.

Although I'd be tempted to add an operator uint8_t() member function, for symmetry, as there is such a constructor.

Best regards
Peder

25 okt 2011 kl. 21:15 skrev "Mateusz Łoskot" <mateusz at loskot.net<mailto:mateusz at loskot.net>>:

2011/10/25<x-apple-data-detectors://8> Peder Axensten <Peder.Axensten at slu.se<mailto:Peder.Axensten at slu.se>>:
I need a GetRawClassification() to be able to quickly copy the all the classification info when I copy a Point to a different point format.
Using three IsXxx() is unnecessarily slow, since getting the raw classification is one copy and
doing it in the present way means 4 copies, 4 and:s, 4 or:s, and 3 if:s. At least.

Peder, you *can* do what you want to do and in the way you want it to do.
Please, re-read previous post.

For clarity, repeated below this is how you do it now

std:;bitset<8> allbits = p.GetClassification();
You get *all* eight bits of classification flags, not only first 5
bits of index in classification
table (we call it class index for clarity).

Howard asked me to suggest a fix to the SetClassification bug and an implementation of GetRawClassification.

There is no need for clunky GetRawClassification, the current
interface allows you to do what you want.

I now see your point and I agree.


Though, the bug in misuse of GetClass() still has to be fixed.

Best regards,
--
Mateusz Loskot, http://mateusz.loskot.net<http://mateusz.loskot.net/>
Charter Member of OSGeo,http://osgeo.org<http://osgeo.org/>
Member of ACCU, http://accu.org<http://accu.org/>


25 okt 2011 kl. 22:49 skrev "Howard Butler" <hobu.inc at gmail.com<mailto:hobu.inc at gmail.com>>:


On Oct 25, 2011, at 2:15 PM, Mateusz Łoskot wrote:

Though, the bug in misuse of GetClass() still has to be fixed.

In libLAS? Or in Peder's code?  I guess I'm not following closely enough here... sorry.

Howard_______________________________________________
Liblas-devel mailing list
Liblas-devel at lists.osgeo.org<mailto:Liblas-devel at lists.osgeo.org>
http://lists.osgeo.org/mailman/listinfo/liblas-devel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/liblas-devel/attachments/20111025/c3c6af25/attachment-0001.html