[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Monotone-devel] Merging nvm.ssh-agent to mainline
From: |
Nathaniel Smith |
Subject: |
Re: [Monotone-devel] Merging nvm.ssh-agent to mainline |
Date: |
Mon, 26 Feb 2007 23:27:39 -0800 |
User-agent: |
Mutt/1.5.13 (2006-08-11) |
On Mon, Feb 26, 2007 at 08:14:27PM -0800, Justin Patrin wrote:
> net.venge.monotone.ssh-agent is ready to be merged to mainline I
> think. It has one strange test failure in the first set of tests which
> I haven't been able to track down. If someone could help me track that
> down I'd be grateful.
>
> key...
> unknown location(0): fatal error in "signature_round_trip":
> std::exception: boost::filesystem::path: invalid name ".monotone" in
> path: "/Users/jpatrin/.monotone/keys"
> keys.cc(719): last checkpoint: signing plaintext
> netcmd...
In case it helps, the usual reason for this error is that someone
created a boost::filesystem::path but forgot to pass the magic value
'boost::filesystem::native' as the second argument to the constructor.
(NB. boost::filesystem is often abbreviated to fs:: in mtn source.)
fs::path has this weird idea that unless you explicitly tell it
otherwise, it should use crazy strict rules for what sort of filenames
are valid, and ones that start with a . are probably not on its list.
I'm not seeing where this is happening in some quick greps of the
source, but it might help you track this down.
> Other than that I'd like someone to do a quick review again to make
> sure all is well and I'll propagate it. :-) Unless, of course, we
> don't want it propagated yet because we don't want it in the next
> release.
That's pretty much never a good reason to not merge things to mainline
:-). Though Richard can override me if he really wants to.
(You've checked that it still compiles/does not break horribly on
windows, now?)
-- Nathaniel
--
Electrons find their paths in subtle ways.