Re: problems writing a regression test for issue 4647

On 8/28/2016 03:53, Daniel Shahaf wrote:
> Stefan wrote on Sun, Aug 28, 2016 at 00:06:19 +0200:
>> START: resolve_tests.py
>> W: Unexpected output
>> W: EXPECTED STDOUT (unordered):
>> W: | Resolved conflicted state of 's'
>> W: | Resolved conflicted state of 'v'
>> W: | Resolved conflicted state of 'n'
>> W: | Resolved conflicted state of '-'
>> W: | Resolved conflicted state of 't'
>> W: | Resolved conflicted state of 'e'
> This is typical of "pass a string where an iterable is expected" errors
> in Python:
>
> % python
> >>> def f(x):
> ... for i in x:
> ... print(repr(i))
> ...
> >>> f("string")
> 's'
> 't'
> 'r'
> 'i'
> 'n'
> 'g'
> >>> f(["string"])
> 'string'
>
> The first call iterates the string, the second call iterates the
> (anonymous) list.
>
>> Index: subversion/tests/cmdline/resolve_tests.py
>> ===================================================================
>> --- subversion/tests/cmdline/resolve_tests.py (revision 1743999)
>> +++ subversion/tests/cmdline/resolve_tests.py (working copy)
>> @@ -598,7 +598,68 @@
>> + def do_binary_conflicting_merge():
>> + svntest.actions.run_and_verify_svn(None, [],
>> + 'revert', '--recursive', A_COPY_path)
>> + svntest.main.run_svn(None, 'merge', sbox.repo_url + "/A_COPY/theta", wc_dir + "/A/theta")
>> +
>> + # Test 'svn resolve -R --accept base'
>> + do_binary_conflicting_merge()
>> + svntest.actions.run_and_verify_resolve(wc_dir,
>> + '-R', '--accept', 'base',
>> + A_COPY_path)
> This is line 648 which the traceback pointed you to. If you look at the
> definition of run_and_verify_resolve() you'll see that the first formal
> argument is EXPECTED_PATHS, plural; it should be passed an iterable
> whose elements are ("local style") path strings.
>
> In short: you're passing an improper first argument to
> run_and_verify_resolve(). I'm guessing you want to pass
> «[os.path.abspath(A_COPY_path)]» (again a one-element list).
>
> Cheers,
>
> Daniel

Thanks Daniel, that was it. Had to make some other corrections but the
test works now. Will follow up in the other thread.

Regards,
Stefan