[GRASS-dev] [GRASS GIS] #3737: Testsuite gunittest/multirunner.py: TypeError: str expected, not bytes

[GRASS GIS]

#3737: Testsuite gunittest/multirunner.py: TypeError: str expected, not bytes
--------------------------+--------------------------------
  Reporter:  AnikaBettge  |      Owner:  grass-dev@…
      Type:  defect       |     Status:  new
  Priority:  normal       |  Milestone:  7.8.0
 Component:  Python       |    Version:  svn-trunk
Resolution:               |   Keywords:  Python3, testsuite
       CPU:  x86-64       |   Platform:  Linux
--------------------------+--------------------------------

Comment (by pmav99):

 In order to make `multirunner.py` run on Python 3 you need to apply
 something like this.

 {{{
 @@ -64,13 +64,13 @@
      # we assume that GRASS GIS' start script is available and in the PATH
      # the shell=True is here because of MS Windows? (code taken from
 wiki)
      startcmd = grass7bin + ' --config path'
      p = subprocess.Popen(startcmd, shell=True,
                           stdout=subprocess.PIPE, stderr=subprocess.PIPE)
      out, err = p.communicate()
      if p.returncode != 0:
          print("ERROR: Cannot find GRASS GIS 7 start script (%s):\n%s" %
 (startcmd, err), file=sys.stderr)
          return 1
 -    gisbase = out.strip('\n')
 +    gisbase = out.strip().decode("utf-8")

      # set GISBASE environment variable
      os.environ['GISBASE'] = gisbase
 @@ -95,7 +95,7 @@
      gsetup.init(gisbase, gisdb, location, mapset)

      reports = []
 -    for location, location_type in itertools.izip(locations,
 locations_types):
 +    for location, location_type in zip(locations, locations_types):
          # here it is quite a good place to parallelize
          # including also type to make it unique and preserve it for sure
          report = 'report_for_' + location + '_' + location_type
 }}}

 Specifying `\n` is probably redundant. `str.strip()` already strips that.
 More specifically it strips all the characters contained in
 `string.whitespace`.

 Other than that, in python3 subprocess output is bytes. The
 `.decode("utf-8")` call is converting the bytes to python3 strings which
 are the equivalent of unicode strings in python2.

 Note: the patch should work with python2 too, but I have not tested it.

-- 
Ticket URL: <https://trac.osgeo.org/grass/ticket/3737#comment:1>
GRASS GIS <https://grass.osgeo.org>