[mapserver-dev] Using PointZM by default

Lime, Steve D (MNIT) steve.lime at state.mn.us
Fri Mar 1 07:28:55 PST 2019


I count 112 references of USE_POINT_Z_M in the C and header files alone so the simplification would be welcome. Under what conditions are the speed differences likely to be most pronounced? I'm assuming when working with polygon or line features with loads of vertices. Would it make sense to try to create an additional benchmark for that use case?

-----Original Message-----
From: mapserver-dev [mailto:mapserver-dev-bounces at lists.osgeo.org] On Behalf Of Even Rouault
Sent: Friday, March 01, 2019 4:41 AM
To: mapserver-dev at lists.osgeo.org
Subject: Re: [mapserver-dev] Using PointZM by default

On vendredi 1 mars 2019 01:14:34 CET Seth G wrote:
> Hi all,
> 
> There is a pull request [1] currently open to set USE_POINT_Z_M as ON by
> default. I recently made an update to this for the CMake build [2].
> 
> As Tamas and Julien-Samuel have pointed in the comments on the pull requests
> it may be better to simply use PointZM throughout MapServer by default, and
> remove this option altogether.
> 
> There are currently issues with MapScript bindings and testing where the
> output is not guarenteed due to the compile options and could lead to
> memory corruption issues.
> 
> Possible risks are breaking some tests (which can be updated), and reducing
> performance. I presume there is no easy way to measure performance between
> changes / releases?
> 
> Does anyone have any thoughts on reasons not to use PointZM by default
> throughout MapServer?

I'm +1 on enabling ZM unconditionaly. This is a topic I raised a few years 
ago.
I found this thread
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.osgeo.org%2Fpipermail%2Fmapserver-users%2F2017-July%2F080102.html&data=02%7C01%7Csteve.lime%40state.mn.us%7C49ad5094a8af4d48073008d69e3270b8%7Ceb14b04624c445198f26b89c2159828c%7C0%7C0%7C636870336820452852&sdata=ltuOMsLT29Pl8Fik4yZzyZUtzpWh4whiloJf0Ifx4hk%3D&reserved=0
where I did a small benchmark. ZM was 2% slower.

Even

-- 
Spatialys - Geospatial professional services
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.spatialys.com&data=02%7C01%7Csteve.lime%40state.mn.us%7C49ad5094a8af4d48073008d69e3270b8%7Ceb14b04624c445198f26b89c2159828c%7C0%7C0%7C636870336820462860&sdata=NmOf5yp72P6IGxadvrfK83f1gHNBw90WGDESX4djUkA%3D&reserved=0
_______________________________________________
mapserver-dev mailing list
mapserver-dev at lists.osgeo.org
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.osgeo.org%2Fmailman%2Flistinfo%2Fmapserver-dev&data=02%7C01%7Csteve.lime%40state.mn.us%7C49ad5094a8af4d48073008d69e3270b8%7Ceb14b04624c445198f26b89c2159828c%7C0%7C0%7C636870336820462860&sdata=tFhYELR%2BeO0enuTDBhRX5LfPY8CwYa6CfD6yhnhKgqQ%3D&reserved=0


More information about the mapserver-dev mailing list