Subject: Re: python.i doesn't exist... why? - msg#00021

Previous Message by Date:

Re: python.i doesn't exist... why?

Since we're talking about this, I'm in the process of closing up the 3.0.0 release. Charlie, could you give the swig/ bit a check ? If things get complicated we can rollback to the 2.x way of only distributing files with no attempt to automatically build. --strk; On Mon, Jun 05, 2006 at 11:58:25PM -0600, Charlie Savage wrote: > Hi Pablo, > > Ok - that's because the swig files assume there is a readHex method. > This will be fixed for 3.0, sorry about the problems in 2.2.0. > > Anyway, to manually fix open geos.i in the swig directory: > > 1. Comment out line 166 which is: > > %newobject WKBReader::readHEX; > > 2. Comment out line 186-190: > > %factory(geos::Geometry * geos::WKBReader::readHEX, geos::Point, > geos::LinearRing, geos::LineString, > geos::Polygon, > geos::MultiPoint, > geos::MultiLineString, > geos::MultiPolygon, > geos::GeometryCollection); > > 3. Comment out lines 297 to the end of the file: > > %extend geos::Geometry { > static geos::GeometryFactory& factory() > { > static geos::GeometryFactory factory; > return factory; > } > > static geos::WKBReader& reader() > { > #if GEOS_VERSION_MAJOR == 2 > static geos::WKBReader reader(geos_Geometry_factory()); > #else > static geos::io::WKBReader reader(geos_Geometry_factory()); > #endif > return reader; > } > > static Geometry* fromWKT(std::string wkt) > { > std::stringstream in(wkt); > return geos_Geometry_reader().read(in); > } > > static Geometry* fromHEX(std::string hex) > { > std::stringstream in(hex); > return geos_Geometry_reader().readHEX(in); > } > }; > > Thanks, > > Charlie > > > Pablo Souza Grigoletti wrote: > >Hi Charlie, > > > >I did what you said: > > > >- mv geos_wrap.cxx geos_wrap.cxx.old > >- I created python.i with the code that you gave me. > >- swig -c++ -python -modern -o geos_wrap.cxx ../geos.i > > > > > >../../source/headers/geos/geom.h:671: Warning(509): Overloaded > >geos::CoordinateS equence::hasRepeatedPoints(geos::CoordinateSequence > >const *) is shadowed by hasR epeatedPoints() const at > >../../source/headers/geos/geom.h:658. > >../../source/headers/geos/geom.h:684: Warning(509): Overloaded > >geos::CoordinateS equence::minCoordinate(geos::CoordinateSequence *) is > >shadowed by minCoordinate( ) const at > >../../source/headers/geos/geom.h:661. > > > > > >- python setup.py build > > > > > >... > >geos_wrap.cxx: In function 'geos::Geometry* > >geos_Geometry_fromHEX(std::string)': geos_wrap.cxx:4417: error: 'class > >geos::WKBReader' has no member named 'readHEX' geos_wrap.cxx: In > >function 'PyObject* _wrap_Geometry_srid_set(PyObject*, PyObjec t*)': > >geos_wrap.cxx:24176: warning: passing 'double' for argument 1 to > >'virtual void g eos::Geometry::setSRID(int)' error: command 'gcc' failed > >with exit status 1 > > > > > >As you said me, a problem with readHEX happened. > >do you have any suggestion? > > > >Ahh, I have the SWIG right version: > > > > > ># swig -version > >SWIG Version 1.3.29 > > > > > >Best regards. > >Pablo > > > > > >On Mon, 05 Jun 2006 21:21:17 -0600 > >Charlie Savage <cfis@xxxxxxxxxxxx> wrote: > > > >>Hi Pablo, > >> > >>Sorry not to respond sooner - I just got back from vacation. > >> > >>What python.i is used for is to generate geos_wrap.cxx using SWIG. If > >> > >>you look, you'll see that file exists. So theory you could get by > >>without python.i and just run the part of the build script that > >>builds/installs the extensions. > >> > >>However, there unfortunately there is a problem with that because the > >>geos_wrap.cxx file includes pointers to a method called readHex which > >>does not actually exist in geos 2.2.2 (well, it does in my version but > >> > >>no release version). > >> > >>So I'd recommend: > >> > >>Create python.i in your swig/python directory. Then put this code > >>into it: > >> > >>%template(vector_GeometryP) std::vector<geos::Geometry *>; > >>%template(vector_LineStringP) std::vector<geos::LineString *>; > >>%template(vector_PolygonP) std::vector<geos::Polygon *>; > >> > >>Then make sure you have swig 1.3.29 installed. Then try the build > >>script again. You may also want to delete geos_wrap.cxx before you > >>start. > >> > >>Hope this helps, > >> > >>Charlie > > > > > >_______________________________________________________ > >Abra sua conta no Yahoo! Mail: 1GB de espaço, alertas de e-mail no > >celular e anti-spam realmente eficaz. http://mail.yahoo.com.br/ > >_______________________________________________ > >geos-devel mailing list > >geos-devel@xxxxxxxxxxxxxxxxxxxx > >http://geos.refractions.net/mailman/listinfo/geos-devel > _______________________________________________ > geos-devel mailing list > geos-devel@xxxxxxxxxxxxxxxxxxxx > http://geos.refractions.net/mailman/listinfo/geos-devel -- /"\ ASCII Ribbon Campaign \ / Respect for low technology. X Keep e-mail messages readable by any computer system. / \ Keep it ASCII.

Next Message by Date:

Toward GEOS-3.0.0RC (was: python.i doesn't exist... why?)

On Tue, Jun 06, 2006 at 04:16:22PM -0600, Charlie Savage wrote: > Yeah, was planning on doing that. What's the timeframe you're looking at? In 1/2 weeks I'd like to release an RC pack, if I handle to make it acceptable as a version number for automake's taste. --strk; > I did a pretty big update for GEOS 3.0 about 2 months ago, but I imagine > things have changed a lot since then. So it might take a bit of time to > get right. Having the new unit tests around helps for sure though (can > copy them to Ruby/Python as needed). > > > Charlie > > strk@xxxxxxxxxxxxxxx wrote: > >Since we're talking about this, I'm in the process of closing > >up the 3.0.0 release. Charlie, could you give the swig/ bit > >a check ? If things get complicated we can rollback to the 2.x > >way of only distributing files with no attempt to automatically > >build. > >

Previous Message by Thread:

Re: python.i doesn't exist... why?

Since we're talking about this, I'm in the process of closing up the 3.0.0 release. Charlie, could you give the swig/ bit a check ? If things get complicated we can rollback to the 2.x way of only distributing files with no attempt to automatically build. --strk; On Mon, Jun 05, 2006 at 11:58:25PM -0600, Charlie Savage wrote: > Hi Pablo, > > Ok - that's because the swig files assume there is a readHex method. > This will be fixed for 3.0, sorry about the problems in 2.2.0. > > Anyway, to manually fix open geos.i in the swig directory: > > 1. Comment out line 166 which is: > > %newobject WKBReader::readHEX; > > 2. Comment out line 186-190: > > %factory(geos::Geometry * geos::WKBReader::readHEX, geos::Point, > geos::LinearRing, geos::LineString, > geos::Polygon, > geos::MultiPoint, > geos::MultiLineString, > geos::MultiPolygon, > geos::GeometryCollection); > > 3. Comment out lines 297 to the end of the file: > > %extend geos::Geometry { > static geos::GeometryFactory& factory() > { > static geos::GeometryFactory factory; > return factory; > } > > static geos::WKBReader& reader() > { > #if GEOS_VERSION_MAJOR == 2 > static geos::WKBReader reader(geos_Geometry_factory()); > #else > static geos::io::WKBReader reader(geos_Geometry_factory()); > #endif > return reader; > } > > static Geometry* fromWKT(std::string wkt) > { > std::stringstream in(wkt); > return geos_Geometry_reader().read(in); > } > > static Geometry* fromHEX(std::string hex) > { > std::stringstream in(hex); > return geos_Geometry_reader().readHEX(in); > } > }; > > Thanks, > > Charlie > > > Pablo Souza Grigoletti wrote: > >Hi Charlie, > > > >I did what you said: > > > >- mv geos_wrap.cxx geos_wrap.cxx.old > >- I created python.i with the code that you gave me. > >- swig -c++ -python -modern -o geos_wrap.cxx ../geos.i > > > > > >../../source/headers/geos/geom.h:671: Warning(509): Overloaded > >geos::CoordinateS equence::hasRepeatedPoints(geos::CoordinateSequence > >const *) is shadowed by hasR epeatedPoints() const at > >../../source/headers/geos/geom.h:658. > >../../source/headers/geos/geom.h:684: Warning(509): Overloaded > >geos::CoordinateS equence::minCoordinate(geos::CoordinateSequence *) is > >shadowed by minCoordinate( ) const at > >../../source/headers/geos/geom.h:661. > > > > > >- python setup.py build > > > > > >... > >geos_wrap.cxx: In function 'geos::Geometry* > >geos_Geometry_fromHEX(std::string)': geos_wrap.cxx:4417: error: 'class > >geos::WKBReader' has no member named 'readHEX' geos_wrap.cxx: In > >function 'PyObject* _wrap_Geometry_srid_set(PyObject*, PyObjec t*)': > >geos_wrap.cxx:24176: warning: passing 'double' for argument 1 to > >'virtual void g eos::Geometry::setSRID(int)' error: command 'gcc' failed > >with exit status 1 > > > > > >As you said me, a problem with readHEX happened. > >do you have any suggestion? > > > >Ahh, I have the SWIG right version: > > > > > ># swig -version > >SWIG Version 1.3.29 > > > > > >Best regards. > >Pablo > > > > > >On Mon, 05 Jun 2006 21:21:17 -0600 > >Charlie Savage <cfis@xxxxxxxxxxxx> wrote: > > > >>Hi Pablo, > >> > >>Sorry not to respond sooner - I just got back from vacation. > >> > >>What python.i is used for is to generate geos_wrap.cxx using SWIG. If > >> > >>you look, you'll see that file exists. So theory you could get by > >>without python.i and just run the part of the build script that > >>builds/installs the extensions. > >> > >>However, there unfortunately there is a problem with that because the > >>geos_wrap.cxx file includes pointers to a method called readHex which > >>does not actually exist in geos 2.2.2 (well, it does in my version but > >> > >>no release version). > >> > >>So I'd recommend: > >> > >>Create python.i in your swig/python directory. Then put this code > >>into it: > >> > >>%template(vector_GeometryP) std::vector<geos::Geometry *>; > >>%template(vector_LineStringP) std::vector<geos::LineString *>; > >>%template(vector_PolygonP) std::vector<geos::Polygon *>; > >> > >>Then make sure you have swig 1.3.29 installed. Then try the build > >>script again. You may also want to delete geos_wrap.cxx before you > >>start. > >> > >>Hope this helps, > >> > >>Charlie > > > > > >_______________________________________________________ > >Abra sua conta no Yahoo! Mail: 1GB de espaço, alertas de e-mail no > >celular e anti-spam realmente eficaz. http://mail.yahoo.com.br/ > >_______________________________________________ > >geos-devel mailing list > >geos-devel@xxxxxxxxxxxxxxxxxxxx > >http://geos.refractions.net/mailman/listinfo/geos-devel > _______________________________________________ > geos-devel mailing list > geos-devel@xxxxxxxxxxxxxxxxxxxx > http://geos.refractions.net/mailman/listinfo/geos-devel -- /"\ ASCII Ribbon Campaign \ / Respect for low technology. X Keep e-mail messages readable by any computer system. / \ Keep it ASCII.

Next Message by Thread:

Toward GEOS-3.0.0RC (was: python.i doesn't exist... why?)

On Tue, Jun 06, 2006 at 04:16:22PM -0600, Charlie Savage wrote: > Yeah, was planning on doing that. What's the timeframe you're looking at? In 1/2 weeks I'd like to release an RC pack, if I handle to make it acceptable as a version number for automake's taste. --strk; > I did a pretty big update for GEOS 3.0 about 2 months ago, but I imagine > things have changed a lot since then. So it might take a bit of time to > get right. Having the new unit tests around helps for sure though (can > copy them to Ruby/Python as needed). > > > Charlie > > strk@xxxxxxxxxxxxxxx wrote: > >Since we're talking about this, I'm in the process of closing > >up the 3.0.0 release. Charlie, could you give the swig/ bit > >a check ? If things get complicated we can rollback to the 2.x > >way of only distributing files with no attempt to automatically > >build. > >