tests/run-tests: Use absolute paths where possible.

Replace some usages of paths relative to the current working directory
with absolute paths relative to the tests directory.

Fixes and resulting changes:
- default values of MICROPYTHON and MPYCROSS are absolute paths and
  always correct
- likewise, the correct full paths for tools and extmod directories
  are appended to sys.path
- printing/cleaning failures works properly since it expects the .exp
  and .out files in the tests directory which is also where they
  are written to now, plus no more need for changing directories

This fixes #5872 and allows running custom tests which use run-tests
without having to cd to the tests directory first, and the test output
still is in the tests/ directory instead of the current working directory.

Discovery of tests and all skip test logic based on paths relative to
the current working directory remains unchanged which essentially means
that for running most of MicroPython's own tests, run-tests must still
be ran from within it's directory, so document that.

tests/run-tests

@@ -5,21 +5,29 @@ import subprocess
 import sys
 import platform
 import argparse
+import inspect
 import re
 from glob import glob
 
+# See stackoverflow.com/questions/2632199: __file__ nor sys.argv[0]
+# are guaranteed to always work, this one should though.
+BASEPATH = os.path.dirname(os.path.abspath(inspect.getsourcefile(lambda: None)))
+
+def base_path(*p):
+    return os.path.abspath(os.path.join(BASEPATH, *p)).replace('\\', '/')
+
 # Tests require at least CPython 3.3. If your default python3 executable
 # is of lower version, you can point MICROPY_CPYTHON3 environment var
 # to the correct executable.
 if os.name == 'nt':
     CPYTHON3 = os.getenv('MICROPY_CPYTHON3', 'python3.exe')
-    MICROPYTHON = os.getenv('MICROPY_MICROPYTHON', '../ports/windows/micropython.exe')
+    MICROPYTHON = os.getenv('MICROPY_MICROPYTHON', base_path('../ports/windows/micropython.exe'))
 else:
     CPYTHON3 = os.getenv('MICROPY_CPYTHON3', 'python3')
-    MICROPYTHON = os.getenv('MICROPY_MICROPYTHON', '../ports/unix/micropython')
+    MICROPYTHON = os.getenv('MICROPY_MICROPYTHON', base_path('../ports/unix/micropython'))
 
 # mpy-cross is only needed if --via-mpy command-line arg is passed
-MPYCROSS = os.getenv('MICROPY_MPYCROSS', '../mpy-cross/mpy-cross')
+MPYCROSS = os.getenv('MICROPY_MPYCROSS', base_path('../mpy-cross/mpy-cross'))
 
 # For diff'ing test output
 DIFF = os.getenv('MICROPY_DIFF', 'diff -u')
@@ -61,7 +69,7 @@ def run_micropython(pyb, args, test_file, is_special=False):
     had_crash = False
     if pyb is None:
         # run on PC
-        if test_file.startswith(('cmdline/', 'feature_check/')) or test_file in special_tests:
+        if test_file.startswith(('cmdline/', base_path('feature_check/'))) or test_file in special_tests:
             # special handling for tests of the unix cmdline program
             is_special = True
 
@@ -215,10 +223,10 @@ def run_feature_check(pyb, args, base_path, test_file):
     if pyb is not None and test_file.startswith("repl_"):
         # REPL feature tests will not run via pyboard because they require prompt interactivity
         return b""
-    return run_micropython(pyb, args, base_path + "/feature_check/" + test_file, is_special=True)
+    return run_micropython(pyb, args, base_path("feature_check", test_file), is_special=True)
 
 
-def run_tests(pyb, tests, args, base_path="."):
+def run_tests(pyb, tests, args):
     test_count = 0
     testcase_count = 0
     passed_count = 0
@@ -244,6 +252,10 @@ def run_tests(pyb, tests, args, base_path="."):
     # If we're asked to --list-tests, we can't assume that there's a
     # connection to target, so we can't run feature checks usefully.
     if not (args.list_tests or args.write_exp):
+        # Even if we run completely different tests in a different directory,
+        # we need to access feature_checks from the same directory as the
+        # run-tests script itself so use base_path.
+
         # Check if micropython.native is supported, and skip such tests if it's not
         output = run_feature_check(pyb, args, base_path, 'native_check.py')
         if output == b'CRASH':
@@ -307,7 +319,7 @@ def run_tests(pyb, tests, args, base_path="."):
             upy_float_precision = int(upy_float_precision)
         has_complex = run_feature_check(pyb, args, base_path, 'complex.py') == b'complex\n'
         has_coverage = run_feature_check(pyb, args, base_path, 'coverage.py') == b'coverage\n'
-        cpy_byteorder = subprocess.check_output([CPYTHON3, base_path + '/feature_check/byteorder.py'])
+        cpy_byteorder = subprocess.check_output([CPYTHON3, base_path('feature_check/byteorder.py')])
         skip_endian = (upy_byteorder != cpy_byteorder)
 
     # These tests don't test slice explicitly but rather use it to perform the test
@@ -509,8 +521,8 @@ def run_tests(pyb, tests, args, base_path="."):
 
         testcase_count += len(output_expected.splitlines())
 
-        filename_expected = test_basename + ".exp"
+        filename_expected = base_path(test_basename + ".exp")
-        filename_mupy = test_basename + ".out"
+        filename_mupy = base_path(test_basename + ".out")
 
         if output_expected == output_mupy:
             print("pass ", test_file)
@@ -563,6 +575,10 @@ def main():
         formatter_class=argparse.RawDescriptionHelpFormatter,
         description='''Run and manage tests for MicroPython.
 
+Tests are discovered by scanning test directories for .py files or using the
+specified test files. If test files nor directories are specified, the script
+expects to be ran in the tests directory (where this file is located) and the
+builtin tests suitable for the target platform are ran.
 When running tests, run-tests compares the MicroPython output of the test with the output
 produced by running the test through CPython unless a <test>.exp file is found, in which
 case it is used as comparison.
@@ -598,10 +614,8 @@ the last matching regex is used:
     args = cmd_parser.parse_args()
 
     if args.print_failures:
-        os.chdir(os.path.abspath(os.path.dirname(__file__)))
+        for exp in glob(base_path("*.exp")):
-
+            testbase = exp[:-4]
-        for exp in glob("*.exp"):
-            testbase = os.path.basename(exp)[:-4]
             print()
             print("FAILURE {0}".format(testbase))
             os.system("{0} {1}.exp {1}.out".format(DIFF, testbase))
@@ -609,9 +623,7 @@ the last matching regex is used:
         sys.exit(0)
 
     if args.clean_failures:
-        os.chdir(os.path.abspath(os.path.dirname(__file__)))
+        for f in glob(base_path("*.exp")) + glob(base_path("*.out")):
-
-        for f in glob("*.exp") + glob("*.out"):
             os.remove(f)
 
         sys.exit(0)
@@ -622,7 +634,7 @@ the last matching regex is used:
         pyb = None
     elif args.target in EXTERNAL_TARGETS:
         global pyboard
-        sys.path.append('../tools')
+        sys.path.append(base_path('../tools'))
         import pyboard
         pyb = pyboard.Pyboard(args.device, args.baudrate, args.user, args.password)
         pyb.enter_raw_repl()
@@ -659,14 +671,10 @@ the last matching regex is used:
 
     if not args.keep_path:
         # clear search path to make sure tests use only builtin modules and those in extmod
-        os.environ['MICROPYPATH'] = os.pathsep + '../extmod'
+        os.environ['MICROPYPATH'] = os.pathsep + base_path('../extmod')
 
-    # Even if we run completely different tests in a different directory,
-    # we need to access feature_check's from the same directory as the
-    # run-tests script itself.
-    base_path = os.path.dirname(sys.argv[0]) or "."
     try:
-        res = run_tests(pyb, tests, args, base_path)
+        res = run_tests(pyb, tests, args)
     finally:
         if pyb:
             pyb.close()