subversion-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Daniel Shahaf" <>
Subject Re: Escape filename for conflict merge
Date Thu, 13 Feb 2020 19:40:41 GMT
James McCoy wrote on Sun, 09 Feb 2020 02:55 +00:00:
> On Thu, Dec 26, 2019 at 11:17:20PM -0500, James McCoy wrote:
> > On Wed, Dec 11, 2019 at 03:12:56PM +0000, Julian Foad wrote:
> > > Stefan Sperling wrote:
> > > > My first question would be: Could anyone could test this on
> > > > Windows? (Assuming you've been testing on Debian, as usual.)
> >
> > Yes, any testing from folks with Windows systems would be welcome.
> It looks like's update_accept_conflicts() could be
> adapted to test this, by having it rename pi to "p i" before setting
> up the conflict.  However, I don't really grok the test code.  My
> attempts to do this have failed.
> Any help would be appreciated.

It's hard to say without knowing what you've tried, but in general,
the flow is:

- creates a repository with r1 populated with the Greek
  tree and creates a working copy.

- The test function then does various operations and prepares
  corresponding expected_* objects.  It then calls svntest.* helpers
  that do an action and verify it.

So in this case, to rename pi to "p i" you'll probably want start the
test by doing that rename as an 'svn mv' (svntest.main.run_svn()),
commit it as r2, run update (sbox.simple_update()), and then in the
remainder of the test function change all instances of "r2" to "r3"?

Or would it work to ensure there's a space in the wcroot_abspath?  I.e.,
in the "svn-test-work" or "update_tests-42/" part of the value of sbox.ospath('foo').

View raw message