Yeesh. I find Ruby incredibly difficult to read, but I'm fairly certain
that this test is bogus in many ways. The "target" passed to
svn_repos_begin_report() was never supposed to be anything but a relpath.
From it's docstring:
@a target is a single path component, used to limit the scope of the
report to a single entry of @a fs_base, or "" if all of @a fs_base
itself is the main subject of the report.
So that part's bad. When I fix that, I get errors about the non-existence
of "/base". And as far as I can tell, "/base" is in fact non-existent in
the test repository, having never been created there.
So... I'll just keep trying to make this test make some sense.
On 02/08/2011 12:22 PM, C. Michael Pilato wrote:
> I'll look into these.
> On 02/08/2011 11:36 AM, Philip Martin wrote:
>> The buildbot is red:
>> The Ruby bindings are triggering the new fspath asserts. The problem occurs in
>> test_report and test_report2 in:
>> I can fix the asserts by doing
>> Index: subversion/bindings/swig/ruby/test/test_repos.rb
>> --- subversion/bindings/swig/ruby/test/test_repos.rb (revision 1068325)
>> +++ subversion/bindings/swig/ruby/test/test_repos.rb (working copy)
>> @@ -387,7 +387,7 @@
>> def assert_report
>> file = "file"
>> file2 = "file2"
>> - fs_base = "base"
>> + fs_base = "/base"
>> path = File.join(@wc_path, file)
>> path2 = File.join(@wc_path, file2)
>> source = "sample source"
>> @@ -405,7 +405,7 @@
>> :revision => rev,
>> :user_name => @author,
>> :fs_base => fs_base,
>> - :target => "/",
>> + :target => "",
>> :target_path => nil,
>> :editor => editor,
>> :text_deltas => true,
>> but although the tests now run they fail.
C. Michael Pilato <cmpilato_at_collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
Received on 2011-02-08 21:21:40 CET