[geos-devel] current (morning of 18th) postgis geos wrapper breaking

[Heitzso]

There are some 600+ lines of errors, but they often snowball/replicate
so here's the first few ...

g++ -O2 -Wall -Wmissing-prototypes -Wmissing-declarations -g 
-fexceptions   -I. -DFRONTEND -DSYSCONFDIR='"/usr/local/pgsql/etc"' 
-DUSE_VERSION=73 -I/usr/local/include/geos -DUSE_GEOS -I/usr/include 
-DUSE_PROJ  -I../../src/include   -c -o postgis_geos_wrapper.o 
postgis_geos_wrapper.cpp
cc1plus: warning: "-Wmissing-declarations" is valid for C/ObjC but not 
for C++
postgis_geos_wrapper.cpp:13:16: io.h: No such file or directory
postgis_geos_wrapper.cpp:16: error: namespace `geos' undeclared
postgis_geos_wrapper.cpp:61: error: `Geometry' was not declared in this 
scope
postgis_geos_wrapper.cpp:61: error: `g1' was not declared in this scope
postgis_geos_wrapper.cpp:61: error: `Geometry' was not declared in this 
scope
postgis_geos_wrapper.cpp:61: error: `g2' was not declared in this scope
postgis_geos_wrapper.cpp:61: warning: `GEOSrelate' initialized and declared
    `extern'
postgis_geos_wrapper.cpp:61: error: initializer list being treated as 
compound
    expression

On my Linux-debian-testing gcc 3.3.1 system io.h is in
sys/io.h.  Changing that in postgis_geos_wrapper.cpp
the io.h error goes awaw, but the rest remain.

I'm guessing the geos namespace problem on 16 is responsible
for the remaining fallout.

I'll be glad to test mods when available.

Heitzso

Paul Ramsey wrote:
> Heitzso wrote:
> 
>> ostgis, when enabled in Makefile to pick up geos,
>> breaks on the geos wrapper file.  Someone let me
>> know if they want me to run a test compile or ??
>> to help debug this.
> 
> 
> Including the error message in your email is always a good start. Not 
> that there are any guarantees things will compile at this point, since 
> the latest commit of GEOS is not the version Dave set up his bindings 
> against.
>