[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Monotone-devel] Bug/patch: automate stdio
From: |
Timothy Brownawell |
Subject: |
Re: [Monotone-devel] Bug/patch: automate stdio |
Date: |
Tue, 28 Nov 2006 15:12:13 -0600 |
On Tue, 2006-11-28 at 10:27 -0800, Nathaniel Smith wrote:
> On Tue, Nov 28, 2006 at 12:05:02PM -0600, Timothy Brownawell wrote:
> > On Sat, 2006-11-25 at 20:40 +0100, Christian Ohler wrote:
> > > There's a pretty serious bug in automate stdio in mtn 0.31 and mainline.
> > > Under certain circumstances, mtn gets confused reading its input and
> > > processes a command that is different from what was sent on stdin.
> > >
> > > I believe the attached (one-line) patch fixes this. Unfortunately, I
> > > see no easy way for me to provide a test case for this.
> >
> > Patch applied, thanks. And yeah, it'd be a bit of a pain to write a test
> > for.
>
> I suppose we'd have to write some sort of mock iostream, that fed us
> bytes with random blocksizes? I think I've seen this done (in the
> twisted testsuite), but ugh...
(echo -n "l6:le"; sleep 2; echo -n "avese") | mtn automate stdio
The "bit of a pain" mostly comes from needing to implement spawn_pipe
for Windows (since the patch posted a few days ago only does unix), and
then figure out the right way to make it available in tester. Once
that's done, actually writing a test should be pretty simple (see above
shell example).
--
Timothy
Free (experimental) public monotone hosting: http://mtn-host.prjek.net