Path not found

[justinpitts]

We have some oddball filenames in our repo - test cases actually for filename handling.

If you prefer, I can file this as a separate issue.

Traceback (most recent call last): File "./agito.py", line 986, in head_id = get_history_for_path(path) File "./agito.py", line 838, in get_history_for_path commit_id = follow_parent_branch(path, log[-1]) File "./agito.py", line 722, in follow_parent_branch changed_path.copyfrom_revision.number) File "./agito.py", line 838, in get_history_for_path commit_id = follow_parent_branch(path, log[-1]) File "./agito.py", line 722, in follow_parent_branch changed_path.copyfrom_revision.number) File "./agito.py", line 838, in get_history_for_path commit_id = follow_parent_branch(path, log[-1]) File "./agito.py", line 722, in follow_parent_branch changed_path.copyfrom_revision.number) File "./agito.py", line 842, in get_history_for_path return construct_history(path, commit_id, log) File "./agito.py", line 759, in construct_history mutate_tree_from_log(treedir, path, entry) File "./agito.py", line 355, in mutate_tree_from_log log, changed_path) File "./agito.py", line 307, in process_add_modify filetype = svn_file_type(changed_path.path, revision) File "./agito.py", line 216, in svn_file_type recurse=False) pysvn._pysvn_27.ClientError: Unable to connect to a repository at URL 'http://wolfwood.int.infograph.com:81/svn/igc/integrations/IBM/Workplace/trunk/appserver/TestData/export/content/%3F.%3E,%3C'%22%3B:%5C%5D%7D%5B%7B=+-)(&%5E%25$%23@!%60~.xrl' 'http://wolfwood.int.infograph.com:81/svn/igc/integrations/IBM/Workplace/trunk/appserver/TestData/export/content/%3F.%3E,%3C'%22%3B:%5C%5D%7D%5B%7B=+-)(_&%5E%25$%23@!%60~.xrl' path not found

[fragglet]

Looks like the bug is probably in svn_path(), which concatenates URL paths naively. It needs to be reworked to join the escape the path properly, probably using python's urlparse library.

[fragglet]

I've reworked svn_path() to use a better form of URL joining that should escape paths properly. So I think this is now fixed. If it still doesn't work then please let me know.

[justinpitts]

I'm trying it out now. It has already passed the point it failed at before, so I suspect you got it.

[fragglet]

Excellent!