Re: [Monotone-devel] race condition on mtn automate stdio startup

[Timothy Brownawell]

On 12/18/2009 3:03 AM, Stephen Leake wrote:
> I've just run across a race condition in mtn automate stdio.
>
> During the startup of the command, it checks for various error
> conditions, such as having a database specified. This can fail if the
> user is not actually in a workspace:
>
> bash-3.00$ mtn automate stdio
> mtn: misuse: no database specified
>
> On the other hand, if there are no errors, there is no output; the
> process is waiting for the next user command:
>
> bash-3.00$ cd gpm
> bash-3.00$ mtn automate stdio
>
> (killed with C-c)
>
> This is race condition; my Emacs front-end has to wait for "a while"
> to tell if the mtn process got started ok, before sending the first
> command.

The mtn process could also be hit by the OOM killer at any time after 
startup, or segfault or something.

> I have not looked into possible fixes for this. Perhaps with the new
> out-of-band facility, there could be a "ready for first command"
> message at startup?

The code that sends commands to the stdio process, needs to remember the 
last command that it sent until it gets the end packet. Then it can 
retire that command and return the output; if the process exits without 
sending the end packet then it needs to be restarted and have the last 
command sent again (up to some limit, in case of non-transient errors).