[GRASS-dev] [GRASS GIS] #3425: regression in v.in.ogr -f 7.3 7 Oct (possibly earlier)

[GRASS GIS]

#3425: regression in v.in.ogr -f 7.3 7 Oct (possibly earlier)
--------------------------+-------------------------
  Reporter:  rsbivand     |      Owner:  grass-dev@…
      Type:  defect       |     Status:  new
  Priority:  major        |  Milestone:
 Component:  Vector       |    Version:  svn-trunk
Resolution:               |   Keywords:
       CPU:  Unspecified  |   Platform:  Unspecified
--------------------------+-------------------------

Comment (by mmetz):

 Replying to [ticket:3425 rsbivand]:
 > In 7.2.2, there was no check for GDAL version in v.in.ogr -f. Now for
 v.in.ogr -f in 7.3 and GDAL >= 2.0, long format names are returned.

 With `v.in.ogr -f`, format names are returned in the form `" %s (%s):
 %s\n"`. The short name is returned first.

 >
 > Choice of best drivers in R rgrass7::readVECT() depends on predictable
 text output from v.in.ogr -f.

 As before, the driver's short name is printed first in each line.
 `v.in.ogr -f` follows `ogrinfo --formats`.
 >
 > Please add a new flag to preserve backward compatibility, so that the
 test at vector/v.in.ogr/main.c line 387 can use the older branch even if
 GDAL >= 2.0. This is a serious regression for users of rgrass7. If the
 flag is added, readVECT will again use SQLite, not ESRI_Shapefile, to
 which it has fallen back.

 Is rgrass7 not using the first item printed by `v.in.ogr -f`?

--
Ticket URL: <https://trac.osgeo.org/grass/ticket/3425#comment:1>
GRASS GIS <https://grass.osgeo.org>