[gdal-dev] Best vector driver?

[Kurt Schwehr]

Even,

If a format can have a combination of point, lines, and polygons in a
layer, should the layer class set wkbGeometryCollection?  And if so, does
that imply Z is okay or not okay?  I see there is
also wkbGeometryCollectionM and wkbGeometryCollectionZM, but
no wkbGeometryCollectionZ.  And I see wkbGeometryCollection25D is old and
shouldn't be used, yes?

poFeatureDefn->SetGeomType(wkbGeometryCollection);

Thanks!
-kurt

On Thu, Feb 16, 2017 at 7:10 AM, Kurt Schwehr <schwehr at gmail.com> wrote:

> Mateusz,
>
> I have been meaning to write up a new RFC for the switch to C++11 as a
> minimum in January and then the flu caught up with me.  I will try to do
> that in the next week or two.   And I really appreciate all folks who
> commented on that thread.
>
> In my opinion, RFC 8 could use a lot of things and I only got through 2
> documents that could feed into it last year (if I remember correctly).  One
> of which was about getting large buffers off of the stack (
> http://goo.gl/vuA3D6).  The other was a start on int bFoo -> bool bFoo (
> https://goo.gl/hdzhXD
>
> Things that I think should be discussed/added off the top of my head in no
> particular order and only things before C++11.  And I just overwhelmed
> myself.
>
> - When to use bool/true/false versus int/TRUE/FALSE/-1(MAYBE/UNKNOWN?).
> Expand on https://goo.gl/hdzhXD
> - namespaces
> - IWYU
> - Initializing vars and data members.  ctor initializer lists
> - Don't use goto
> - Templates and inlining
> - Defining one member per line
> - clang-format
> - C++ casting
> - CPLString versus std::string
> - Preferring const var to #define
> - Making vars const when possible
> - override, explicit, CPL_DISALLOW_COPY_ASSIGN
> - Design for testability and what tests to have.  e.g. I am working on an
> in depth test of jp2kakdataset.cpp.
> - static and dynamic analysis.  cppcheck, clang static analyzier, compiler
> warnings.  How and when to tell the tools that they are being dumb
> - Examples of good/helpful comments
> - What to do about svn keywords and file permissions (e.g. python tests
> and shellscripts being executable)
> - When to use parens
> - Prefer !.empty() and .empty().  When to use CPL_ARRAYSIZE
> - Don't convert back to C strings for operations that are available in a
> C++ string you already have
> - ABS/MIN/MAX -> std::abs/std::min/std::max etc.
> - When to use {} in if else (at least be consistent within a block)
> - When to use a scope
> - char *foo v. char foo[]
> - When is it okay to have newlines in CPLDebug and CPLError
> - When to use #if DEBUG_VERBOSE.  Is it okay to check in commented out
> code?
>
> After C++11:
>
> - Avoid bare .*alloc/new and *free/delete
> - NULL -> nullptr
> - constexpr / static_assert
> - std::initializer_list
>
>
> On Thu, Feb 16, 2017 at 1:16 AM, Mateusz Loskot <mateusz at loskot.net>
> wrote:
>
>> On 15 February 2017 at 17:27, Kurt Schwehr <schwehr at gmail.com> wrote:
>> >
>> > Thanks for the pointers.  I will do touchups on the driver tutorial as
>> I go.
>> > [...]
>> > What's the cleanest way to setup the driver registration?
>> >
>> > Should I be deriving from GDALDataset?
>>
>> Kurt,
>>
>> BTW, I'd have similar or related question.
>>
>> Which driver or part of the code is the latest/greatest in
>> terms of the new C++ programming guidelines to look up to
>> while writing new code for GDAL?
>>
>> Has there been any agreement on the overall C++ coding guidelines?
>> Better, do we need the RFC8 [1] update?
>>
>> [1] https://trac.osgeo.org/gdal/wiki/rfc8_devguide
>>
>> Best regards
>> --
>> Mateusz Loskot, http://mateusz.loskot.net
>>
>
>
>
> --
> --
> http://schwehr.org
>



-- 
--
http://schwehr.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/gdal-dev/attachments/20170216/d6f611dd/attachment-0001.html>