Re: [Monotone-devel] review of nvm.automate_out_of_band

[Thomas Keller]

Am 28.11.09 23:50, schrieb Timothy Brownawell:
> Thomas Keller wrote:
>> Am 28.11.09 20:35, schrieb Timothy Brownawell:
>>> The ignorable stuff is all on stderr, this would be on stdout like the
>>> rest of the actual interesting output (stdio packets). And the "without
>>> using the format already" is why it would be something simple like a
>>> leading <number> <whitespace> rather than a full stdio packet.
>>
>> So something like
>>
>> $ mtn au stdio
>> interface_version: 12.0
>>
>> $
>>
>> ? We don't need a separate interface_version command available then, right?
> 
> I suppose that could work, but I think it'd be better to have
> 
> $ mtn au stdio
> 2
> 
> $
> 
> with the "2" being specific to the stdio format rather than being the
> full interface_version number. So it will tell you exactly whether you
> can talk coherently to this particular monotone, and then later you can
> call interface_version or command_version or whatever to see if you
> agree on the availability and syntax of particular other commands. Using
> the interface_version number here would bake in the issue of "inventory
> changed formats and the major version incremented, and now viewmtn needs
> to refuse to work until it's told about the new version".

Sorry that this took a little longer, but I was held off by a couple of
things (and by myself understanding the protocol handshake code). What
I've implemented and tested in 0b87ef35d232922775bcab4d50b96be3b627d78e
is a "stdio header" system, so that arbitrary header information can be
added and transferred for stdio and remote_stdio. I refused to just add
a "number" without description there because I thought it might be handy
later to issue more stdio-specific data there, f.e. the used packet size
and because I think 17 bytes once are not such a big overhead ;)

Anyways, I still have just the interface_version outputted there, no
other version, basically because I just re-read your last email and now
tend to agree that this alone doesn't help to distinguish stdio version
changes, but hey, since we have this nice "header" system now, we can
either add an additional "stdio-version" header to the existing one or
replace it.

What do you think?

Thomas.

-- 
GPG-Key 0x160D1092 | address@hidden | http://thomaskeller.biz
Please note that according to the EU law on data retention, information
on every electronic information exchange might be retained for a period
of six months or longer: http://www.vorratsdatenspeicherung.de/?lang=en

signature.asc
Description: OpenPGP digital signature