[GRASS-dev] standard input options messed up in current CVS

hi all,

I just did a fresh rebuild from CVS an hour ago and noticed that
standard option declarations (G_OPT_V_FIELD, G_OPT_R_OUTPUT,
G_OPT_V_MAP) are no longer working. Options declared this way get
totally messed up. Have a look at v.out.ogr for an an example.

Best,

Benjamin

--
Benjamin Ducke, M.A.
Archäoinformatik
(Archaeoinformation Science)
Institut für Ur- und Frühgeschichte
(Inst. of Prehistoric and Historic Archaeology)
Christian-Albrechts-Universität zu Kiel
Johanna-Mestorf-Straße 2-6
D 24098 Kiel
Germany

Tel.: ++49 (0)431 880-3378 / -3379
Fax : ++49 (0)431 880-7300
www.uni-kiel.de/ufg

On Mon, May 14, 2007 at 04:18:05PM +0200, Benjamin Ducke wrote:

hi all,

I just did a fresh rebuild from CVS an hour ago and noticed that
standard option declarations (G_OPT_V_FIELD, G_OPT_R_OUTPUT,
G_OPT_V_MAP) are no longer working. Options declared this way get
totally messed up. Have a look at v.out.ogr for an an example.

I bet that you need to run
make distclean
before recompiling.

Markus

Sorry, no luck. Did a distclean and recompiled. Same problem.
Do you have it, too?

Ben

Markus Neteler wrote:

On Mon, May 14, 2007 at 04:18:05PM +0200, Benjamin Ducke wrote:

hi all,

I just did a fresh rebuild from CVS an hour ago and noticed that
standard option declarations (G_OPT_V_FIELD, G_OPT_R_OUTPUT,
G_OPT_V_MAP) are no longer working. Options declared this way get
totally messed up. Have a look at v.out.ogr for an an example.

I bet that you need to run
make distclean
before recompiling.

Markus

_______________________________________________
grass-dev mailing list
grass-dev@grass.itc.it
http://grass.itc.it/mailman/listinfo/grass-dev

--
Benjamin Ducke, M.A.
Archäoinformatik
(Archaeoinformation Science)
Institut für Ur- und Frühgeschichte
(Inst. of Prehistoric and Historic Archaeology)
Christian-Albrechts-Universität zu Kiel
Johanna-Mestorf-Straße 2-6
D 24098 Kiel
Germany

Tel.: ++49 (0)431 880-3378 / -3379
Fax : ++49 (0)431 880-7300
www.uni-kiel.de/ufg

Benjamin wrote:

>> I just did a fresh rebuild from CVS an hour ago and noticed that
>> standard option declarations (G_OPT_V_FIELD, G_OPT_R_OUTPUT,
>> G_OPT_V_MAP) are no longer working. Options declared this way get
>> totally messed up. Have a look at v.out.ogr for an an example.

Markus:

> I bet that you need to run
> make distclean
> before recompiling.

Benjamin:

Sorry, no luck. Did a distclean and recompiled. Same problem.
Do you have it, too?

Looks ok for me using yesterday's CVS. What do the broken bits look
like? Are they taking descriptions etc from another std option?

I did add a new standard option the other day (for fs=), but added it to
the end of the STD_OPT list. I thought doing that would only cause
distclean problems if I had added it to the middle of the list, i.e.
reordering the index.

Hamish

I just did another CVS update, made distclean, reconfigured, recompiled.
Still the same problems.

This is the dump of "v.out.ogr -help":

Parameters:
  output Name for output vector map
     cat Feature type. Combination of types is not supported by all
formats.
           options: point,kernel,centroid,line,boundary,area,face
           default: line,boundary
     dsn OGR output datasource name. For example:
                ESRI Shapefile: filename or directory for storage
                MapInfo File: filename or directory for storage
  olayer OGR layer name. If not specified, input name is used.
                ESRI Shapefile: shapefile name
                MapInfo File: mapinfo file name
    cats Category values
           Example: 1,3,7-9,13
  format OGR format
           options:
ESRI_Shapefile,MapInfo_File,TIGER,S57,DGN,Memory,CSV,GML,
                    KML,
           default: ESRI_Shapefile
    dsco OGR dataset creation option (format specific, NAME=VALUE)
           default:
     lco OGR layer creation option (format specific, NAME=VALUE)
           default:

As you can see, "input" is missing, option "type" is called "cat", I
don't get option "layer" anymore and option "cats" has been added
instead. Pretty weird.

In addition v.in.ogr now complains about not being able to find object
"G_str_to_sql" in "libgrass_gis.6.3.cvs.dll".

This is a MINGW compile, mind you.

Ben

Hamish wrote:

Benjamin wrote:

I just did a fresh rebuild from CVS an hour ago and noticed that
standard option declarations (G_OPT_V_FIELD, G_OPT_R_OUTPUT,
G_OPT_V_MAP) are no longer working. Options declared this way get
totally messed up. Have a look at v.out.ogr for an an example.

Markus:

I bet that you need to run
make distclean
before recompiling.

Benjamin:

Sorry, no luck. Did a distclean and recompiled. Same problem.
Do you have it, too?

Looks ok for me using yesterday's CVS. What do the broken bits look
like? Are they taking descriptions etc from another std option?

I did add a new standard option the other day (for fs=), but added it to
the end of the STD_OPT list. I thought doing that would only cause
distclean problems if I had added it to the middle of the list, i.e.
reordering the index.

Hamish

--
Benjamin Ducke, M.A.
Archäoinformatik
(Archaeoinformation Science)
Institut für Ur- und Frühgeschichte
(Inst. of Prehistoric and Historic Archaeology)
Christian-Albrechts-Universität zu Kiel
Johanna-Mestorf-Straße 2-6
D 24098 Kiel
Germany

Tel.: ++49 (0)431 880-3378 / -3379
Fax : ++49 (0)431 880-7300
www.uni-kiel.de/ufg

Benjamin Ducke wrote:

I just did another CVS update, made distclean, reconfigured, recompiled.
Still the same problems.

This is the dump of "v.out.ogr -help":

Parameters:
  output Name for output vector map
     cat Feature type. Combination of types is not supported by all
formats.
           options: point,kernel,centroid,line,boundary,area,face
           default: line,boundary
     dsn OGR output datasource name. For example:
                ESRI Shapefile: filename or directory for storage
                MapInfo File: filename or directory for storage
  olayer OGR layer name. If not specified, input name is used.
                ESRI Shapefile: shapefile name
                MapInfo File: mapinfo file name
    cats Category values
           Example: 1,3,7-9,13
  format OGR format
           options:
ESRI_Shapefile,MapInfo_File,TIGER,S57,DGN,Memory,CSV,GML,
                    KML,
           default: ESRI_Shapefile
    dsco OGR dataset creation option (format specific, NAME=VALUE)
           default:
     lco OGR layer creation option (format specific, NAME=VALUE)
           default:

As you can see, "input" is missing, option "type" is called "cat", I
don't get option "layer" anymore and option "cats" has been added
instead. Pretty weird.

In addition v.in.ogr now complains about not being able to find object
"G_str_to_sql" in "libgrass_gis.6.3.cvs.dll".

This is a MINGW compile, mind you.

Do you have a previous version of GRASS installed on the system?

It's possible that compilation could be picking that up. Check the
-I/-L flags in the make output; do any of those directories contain
GRASS headers/libraries?

Also, are you linking against a GDAL library which has GRASS support?
If so, that could cause the installed GRASS libraries to be used
regardless of where they're located (or gdal-config could be adding
switches which refer to the installed version).

If you regularly update GRASS from CVS, I strongly suggest having a
copy of GDAL which *doesn't* include GRASS support, to avoid clashes
between the version being compiled and the version which GDAL uses.

--
Glynn Clements <glynn@gclements.plus.com>