From Andreas Ekstrand, "he attached ESRIShape.cpp contains a small fix in parsing of PointZ features, now correctly comparing with header content length which is defined in 16-bit words.

This should have been fixed in my old submission from 2012 but was probably overlooked due to an alternative way of comparing to a fix number for this feature type.

Without this fix, the plugin will only read one PointZ feature even if multiple features exist."
This commit is contained in:
Robert Osfield 2016-12-21 11:40:16 +00:00
parent 9478d27263
commit 723849c57c

View File

@ -838,7 +838,7 @@ bool PointZ::read( int fd )
return false;
// Sometimes, M field is not supplied
if( rh.contentLength >= 18 )
if( rh.contentLength*2 >= 18 )
if( readVal<Double>( fd, m, LittleEndian ) == false )
return false;