Update drvsupport test (PCIDSK, GMT, PDS)

[rbuffat]

I noticed that the test to append data only looks at the number of features is not enough to test the PCIDSK driver correctly.

This PR modifies the drvsupport tests to compare the written data with the result. Ideally, the input records would be compared to the output records, but this is tricky, as some drivers add fields (e.g. the Shapefile driver adds a FID field). This PR only looks at one property (if the field type is supported), respectively checks if the coordinate is correct otherwise.

Additionally, this PR enables the OGR_PDS and OGR_GMT drivers. It looks as they were renamed with GDAL 2.x. Thus I suppose they are not often used in combination with Fiona.

from collections import OrderedDict
import pprint
import fiona
import os

schema = {'geometry': 'Point', 'properties': OrderedDict([('position', 'int')])}
records1 = [{'geometry': {'type': 'Point', 'coordinates': (0.0, float(i))}, 'properties': {'position': i}} for i in range(0, 2)]
records2 = [{'geometry': {'type': 'Point', 'coordinates': (0.0, float(i))}, 'properties': {'position': i}} for i in range(5, 10)]
filename = '/tmp/test.pix'
driver = "PCIDSK"

if os.path.exists(filename):
    os.unlink(filename)

with fiona.open(filename,
                mode='w',
                driver=driver,
                schema=schema) as out:
    out.writerecords(records1)

with fiona.open(filename,
                mode='a',
                driver=driver,
                schema=schema) as out:
    out.writerecords(records2)

with fiona.open(filename) as c:
    for f in c:
        pprint.pprint(f)

Output:

{'geometry': {'coordinates': (0.0, 5.0, 0.0), 'type': 'Point'},
 'id': '0',
 'properties': OrderedDict([('position', '5')]),
 'type': 'Feature'}
{'geometry': {'coordinates': (0.0, 6.0, 0.0), 'type': 'Point'},
 'id': '1',
 'properties': OrderedDict([('position', '6')]),
 'type': 'Feature'}
{'geometry': {'coordinates': (0.0, 7.0, 0.0), 'type': 'Point'},
 'id': '0',
 'properties': OrderedDict([('position', '7')]),
 'type': 'Feature'}
{'geometry': {'coordinates': (0.0, 8.0, 0.0), 'type': 'Point'},
 'id': '1',
 'properties': OrderedDict([('position', '8')]),
 'type': 'Feature'}
{'geometry': {'coordinates': (0.0, 9.0, 0.0), 'type': 'Point'},
 'id': '2',
 'properties': OrderedDict([('position', '9')]),
 'type': 'Feature'}
{'geometry': None,
 'id': '3',
 'properties': OrderedDict([('position', '')]),
 'type': 'Feature'}
{'geometry': None,
 'id': '4',
 'properties': OrderedDict([('position', '')]),
 'type': 'Feature'}

[coveralls]

Coverage increased (+0.04%) to 84.266% when pulling 5e67e4e on rbuffat:update_drvsupport into a9feeb5 on Toblerity:maint-1.8.

[rbuffat]

Actually, to compare records would be handy, and the same code could also be used in the tests of #901

I will try to find a suitable implementation.

[rbuffat]

Included in #915