[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Status report for Ruby bindings on Windows

From: Joe Swatosh <joe.swatosh_at_gmail.com>
Date: 2007-01-31 08:36:43 CET

Hi kou

On 1/30/07, Kouhei Sutou <kou@cozmixng.org> wrote:
> Hi Joe,
>
> In <ae6cb1100701292301o55581b5ay5826525a743d7e20@mail.gmail.com>
> "Re: Status report for Ruby bindings on Windows" on Mon, 29 Jan 2007 23:01:59 -0800,
> "Joe Swatosh" <joe.swatosh@gmail.com> wrote:
>
> > * I didn't try your run-test.rb patch for windows. I've been toying
> > with going a different way with that and I will include a patch to
> > show what I'm thinking.
>
> Thanks! Could you try my new attached patch? The patch is
> based on yours.
>

Sorry. We get the correct setup_test_environment method, but the
setup_svnserve isn't being overridden. If I understood the chapter
from the pick axe correctly this makes sense. Ruby will look in the
class before looking in any modules to resolve the method name. So
even though the version in the SvnTestWindowsUtil module is defined
later it isn't considered when we try to call it. I expect if you
moved the original versions of the setup_svnserve, teardown_svnserve,
and add_pre_revprop_change_hook methods into a module that was
included in the else clause of the if windows? near the bottom of
util.rb we'd get the desired effect.

There are probably a hundred other ways. Sadly, this patch isn't one of them.

>
> > * Normalizing the line breaks. Upon reflection, I've decided that I
> > started us down the wrong path when I modified the tests for the line
> > ending issues. As I've thought about it more, I now believe that the
> > bindings should do the conversions. My thinking is since Ruby treats
> > line endings as "'\n" instead of "\r\n" the bindings should provide
> > any multiline values Ruby expects. (This is where the 'b' comes from
> > when calling File::open. Without the 'b' on windows there is a \r\n
> > to \n conversion. With the 'b', no conversion. Maybe we should check
> > out what the Perl and Python bindings do about this issue on
> > windows?). (I've made zero effort to figure out what this will take).
>
> OK. We'll try the problem after we solve test environment
> problem on Windows.
>

Makes sense to me.

>
> > * (Saving the best for last) I updated against trunk last night and
> > re-ran the tests and the test_client is no longer causing errors.
>
> Thanks for confirming.
>

My pleasure. Sorry the other news wasn't better.

--
Joe
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Jan 31 08:37:09 2007

This is an archived mail posted to the Subversion Dev mailing list.