Comment 1 for bug 777860

Revision history for this message
Martin Packman (gz) wrote : Re: error accessing special character files on Windows

I can reproduce this with trunk qbzr and explorer. Setting BZR_PDB lets you see a traceback along the lines of:

  ...\bzrlib\plugins\explorer\lib\commands.py(164)run()
-> app.exec_()
  ...\bzrlib\plugins\explorer\lib\view_repository.py(247)do_double_clicked()
-> self.do_open_location()
  ...\bzrlib\plugins\explorer\lib\view_repository.py(189)do_open_location()
-> abspath = self._get_abspath(self._relpath)
  ...\bzrlib\plugins\explorer\lib\view_repository.py(243)_get_abspath()
-> return self._get_root_transport().abspath(relpath)
  ...\bzrlib\transport\local.py(110)abspath()
-> path = osutils.normpath(osutils.pathjoin(
> ...\bzrlib\urlutils.py(536)unescape()
-> raise errors.InvalidURL(url, 'URL was not a plain ASCII url: %s' % (e,))

So it seems bzr-explorer is violating a bzrlib expectation about how urls should be formed. I think there must be a bug in the qbzr error handling as well as the traceback doesn't end up in .bzr.log for some reason.