Re: [Monotone-devel] bugreport monotone

[Zack Weinberg]

On Fri, Sep 26, 2008 at 4:44 PM, Thomas Keller <address@hidden> wrote:
> Vitus Jensen schrieb:
>> Just doing the daily update of the openembedded.dev branch.  The mtn pull
>> worked yesterday and it worked when I retried the same command sequence.
>>
>> Console:
>> address@hidden:~/build/nslu2.dev$ pushd oe/; mtn pull && mtn up; popd
>> ~/build/nslu2.dev/oe ~/build/nslu2.dev
>> mtn: doing anonymous pull; use -kKEYNAME if you need authentication
>> mtn: connecting to monotone.openembedded.org
>> mtn: finding items to synchronize:
>> mtn: certificates | keys | revisions
>> mtn:        79216 |   76 |     26096
>> mtn: bytes in | bytes out | certs in | revs in
>> mtn:  151.6 k |    86.7 k |    96/96 |   24/24
>> mtn: fatal: std::runtime_error: network error: recv failure: Connection
>> timed out
>
> Hrm... any objectives against checking in this patch (netxx upstream is
> dead, right)?

Go for it.  Yes, they're long gone.

> There are possibly other connections errors which should not let
> monotone spy out its "this is probably a bug in monotone" error (at
> least the win32 section lists a few more), but I'm not sure if these are
> not already handled elsewhere / earlier (ESHUTDOWN, EHOSTUNREACH,
> CONNABORTED).

Probably best to wait till they come up, unless you want to code up
some kind of test framework to exercise netsync's response to network
errors, which I'd actually rather you didn't, as it would make it
harder to make the networking tests use filesystem sockets in the
future.

zw