|
Here's what I did: So I figure that test fails because of some of the remains in the test-closet. I don't think it's intended, is it? |
|
I narrowed it down: it suffices to run only testcase u0038 to get the failure in st_web_18002_main_no_wf. in testcase u0038_test_wdmapping a descriptor is created and added to the .sgcloset-test descriptors: I guess that descriptor is responsible for the server failure. Here's a diff for the fix that solved the problem for me:
This makes sense, sorry I was initially misleading. This has been fixed for long enough that I forgot about it. (We have deliberately held back source updates the past few months because we've made some repository format changes, and we want to make sure the migration is relatively painless.)
(Feb 22 at 09:28)
Ian Olsen ♦♦
|
|
Apropos of your other questions about path_parent_dir, these two errors in the server log are significant:
Looks like your first test run creates repos whose descriptors don't have path_parent_dir. On your second run, code that expects to see it fails. Hm. That's suprising. I first thought I had messed it up, but then I completely removed anything I had done, even deleting the .sgcloset and building veracity from the sources cloned from your public repo all over. To be sure, I'll do it again ;) Stay tuned...
(Feb 13 at 12:14)
Arne Mertz
|
Not intended. What's the failure within that test? This question might help.
ok, here's the section of the output.xx.js file:
I'll upload the files I got:
www.arne-mertz.de/veracity/output.st_web_18002_main_no_wf.js.log
www.arne-mertz.de/veracity/stdout.st_web_18002_main_no_wf.js.log
www.arne-mertz.de/veracity/vv-serve.st_web_18002_main_no_wf.js.log
www.arne-mertz.de/veracity/vv.st_web_18002_main_no_wf.js.log