Re: [Monotone-devel] Re: Imminent release...

[Stephen Leake]

Richard Levitte <address@hidden> writes:

> In message <address@hidden> on Thu, 02 Aug 2007 12:00:47 +0200, Markus 
> Schiltknecht <address@hidden> said:
>
> markus> Considering that we really care only since we have the
> markus> buildbots in place again, it seems a little silly to me to
> markus> hold back the release.
>
> I'll have you note that we cared before too.  I refused to make a
> release when there were no buildbots at all.  In fact, that was the
> primary reason I set up a master.
>
> markus> Supporting all those architectures would of course be nice,
> markus> and we should certainly strive for cross-platform compatibility.
> markus> But most probably monotone didn't work and hasn't been used
> markus> much on AIX or Solaris.
>
> Mmmmm, I'm actually mostly thinking of Windows, which has been
> supported with one binary in the past as well (just check the
> downloads directory and you'll see them for yourself).
>
> I'm fairly certain I've heard at least one Windows user squeek last
> time I asked about this ;-).

That was me. 

The only bugs remaining (as of last weekend) in the MinGW build are:

unit tests:
    netxx_pipe.cc : pipe:simple_pipe cat: write error: Bad file number FAIL

./run_lua_tests:
445 ws_ops_with_wrong_node_type                   FAIL (line 12)

At the same time, I'm using this build on Windows, and it works.

I've looked into both of these. For the first, it seems to be failing
because it is waiting for a child process to die, but the child
process is held open by the parent; classic deadlock. I must
misunderstand, because the same should be true on Unixen, but
apparently it passes there.

The second one is giving an annoying crash error message instead of a
more friendly error message.

> markus> Let's go ahead and call it a release!
>
> I'm tempted.

I don't think either of these is reason to hold up a release; go right
ahead.

-- 
-- Stephe