Re: [postgis-users] Re: Buggy transform() function in 1.0RC1 ?

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [postgis-users] Re: Buggy transform() function in 1.0RC1 ?

Nicolas Ribot
Hello all,

> Nicolas, in the pdf you sent (good work!) you tell previous
> postgis versions worked fine. I took a look at older definitions
> and I found out that both +pm and +lon_0 have the same value
> (2.337229166666667). New definition changed the +pm to Paris
> (instead of literal number) and that seems to have broken things up.
>
> Changing lon_0 to 0 fixed it again.
>
> Can you confirm these being the working options:
>         1) +pm=Paris +lon_0=0
>         2) +pm=2.337229166666667 +lon_0=2.337229166666667
>
I made some tests on Postgis 1.0 (same config as the one described in
my previous message):

The only working option is:
 +pm=paris +lon_0=0

The other option gives the same result as the default proj4text
provided with spatial_ref_sys.sql (+pm=paris +lon_0=2.33722917): a
2.337... coordinates east shift.

By the way, following Gino's advice, I changed the +pm=paris parameter
to +pm=Paris to test if case is sensitive in reprojection.
It crashed the server (postgres.exe).

I send you a minimal script to test the problem.

> It would help testing both with different proj4 versions.
> I don't whether they should be equivalent, but really do
> not look like they are (unless Paris resolves to 2.337*2).
>
> SRTEXT is not used by postgis, dunno about JUMP.
>

For me, the correct definition is either:
+pm=paris +lon_0=0 or
+pm=2.33722917 +lon_0=0

expressing the prime meridien either by a keyword or by its actual
shift value from Greenwich.

Nicolas

_______________________________________________
postgis-devel mailing list
[hidden email]
http://postgis.refractions.net/mailman/listinfo/postgis-devel

test_crash_postgres_reprojection.sql (1K) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Re: [postgis-users] Re: Buggy transform() function in 1.0RC1 ?

strk
Tested with current CVS (1.0.1 basically)

$ psql -f test_crash_postgres_reprojection.sql test
                 transform_geometry
----------------------------------------------------
 0101000020BE6B00006A2B98C728512241C7EA110A0A664441
(1 row)

UPDATE 1
                 transform_geometry
----------------------------------------------------
 0101000020BE6B00006A2B98C728512241C7EA110A0A664441
(1 row)

What version are you using ?

--strk;

On Thu, May 26, 2005 at 04:50:11PM +0200, Nicolas Ribot wrote:

> Hello all,
>
> > Nicolas, in the pdf you sent (good work!) you tell previous
> > postgis versions worked fine. I took a look at older definitions
> > and I found out that both +pm and +lon_0 have the same value
> > (2.337229166666667). New definition changed the +pm to Paris
> > (instead of literal number) and that seems to have broken things up.
> >
> > Changing lon_0 to 0 fixed it again.
> >
> > Can you confirm these being the working options:
> >         1) +pm=Paris +lon_0=0
> >         2) +pm=2.337229166666667 +lon_0=2.337229166666667
> >
>
> I made some tests on Postgis 1.0 (same config as the one described in
> my previous message):
>
> The only working option is:
>  +pm=paris +lon_0=0
>
> The other option gives the same result as the default proj4text
> provided with spatial_ref_sys.sql (+pm=paris +lon_0=2.33722917): a
> 2.337... coordinates east shift.
>
> By the way, following Gino's advice, I changed the +pm=paris parameter
> to +pm=Paris to test if case is sensitive in reprojection.
> It crashed the server (postgres.exe).
>
> I send you a minimal script to test the problem.
>
> > It would help testing both with different proj4 versions.
> > I don't whether they should be equivalent, but really do
> > not look like they are (unless Paris resolves to 2.337*2).
> >
> > SRTEXT is not used by postgis, dunno about JUMP.
> >
>
> For me, the correct definition is either:
> +pm=paris +lon_0=0 or
> +pm=2.33722917 +lon_0=0
>
> expressing the prime meridien either by a keyword or by its actual
> shift value from Greenwich.
>
> Nicolas


> _______________________________________________
> postgis-devel mailing list
> [hidden email]
> http://postgis.refractions.net/mailman/listinfo/postgis-devel

_______________________________________________
postgis-devel mailing list
[hidden email]
http://postgis.refractions.net/mailman/listinfo/postgis-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Re: [postgis-users] Re: Buggy transform() function in 1.0RC1 ?

Nicolas Ribot
> Tested with current CVS (1.0.1 basically)
>
> $ psql -f test_crash_postgres_reprojection.sql test
>                  transform_geometry
> ----------------------------------------------------
>  0101000020BE6B00006A2B98C728512241C7EA110A0A664441
> (1 row)
>
> UPDATE 1
>                  transform_geometry
> ----------------------------------------------------
>  0101000020BE6B00006A2B98C728512241C7EA110A0A664441
> (1 row)
>
> What version are you using ?
>

 POSTGIS="1.0.0" GEOS="2.1.1" PROJ="Rel. 4.4.9, 29 Oct 2004" USE_STATS
DBPROC="0 .3.0" RELPROC="0.3.0"

On Windows XP SP2, Mingw compilation.

So, problem seems to be solved :-)

Nicolas
_______________________________________________
postgis-devel mailing list
[hidden email]
http://postgis.refractions.net/mailman/listinfo/postgis-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Re: [postgis-users] Re: Buggy transform() function in 1.0RC1 ?

strk
On Sun, Jun 05, 2005 at 02:31:18PM +0200, Nicolas Ribot wrote:

> > Tested with current CVS (1.0.1 basically)
> >
> > $ psql -f test_crash_postgres_reprojection.sql test
> >                  transform_geometry
> > ----------------------------------------------------
> >  0101000020BE6B00006A2B98C728512241C7EA110A0A664441
> > (1 row)
> >
> > UPDATE 1
> >                  transform_geometry
> > ----------------------------------------------------
> >  0101000020BE6B00006A2B98C728512241C7EA110A0A664441
> > (1 row)
> >
> > What version are you using ?
> >
>
>  POSTGIS="1.0.0" GEOS="2.1.1" PROJ="Rel. 4.4.9, 29 Oct 2004" USE_STATS
> DBPROC="0 .3.0" RELPROC="0.3.0"
>
> On Windows XP SP2, Mingw compilation.
>
> So, problem seems to be solved :-)

There have been no segfault fixes between 1.0.0 and 1.0.1
I suspect it's something related to XP/Mingw ... can you
try 1.0.1 (doesn't require dump/reload) ?

--strk;
_______________________________________________
postgis-devel mailing list
[hidden email]
http://postgis.refractions.net/mailman/listinfo/postgis-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Re: [postgis-users] Re: Buggy transform() function in 1.0RC1 ?

Nicolas Ribot
> There have been no segfault fixes between 1.0.0 and 1.0.1
> I suspect it's something related to XP/Mingw ... can you
> try 1.0.1 (doesn't require dump/reload) ?
>
> --strk;
>

Yes, I will give it a try with the newest version.
I will have time to do it after tuesday, not before.
_______________________________________________
postgis-devel mailing list
[hidden email]
http://postgis.refractions.net/mailman/listinfo/postgis-devel
Loading...