Right now, 'automate stdio's parser silently ignores anything between
the e that ends one command, and the l that starts the next. I
guess the idea was to make it easier to add extensions in the future?
And we are talking about using this space for extensions (in
particular, for key/value style
Nathaniel Smith wrote:
But, this seems broken... if a process starts feeding us
key/value arguments (for example), and we are an old monotone that
does not understand them, then we will silently do something different
than expected. Worse, it's entirely possible that those arguments
will
Jon Bright schrieb:
In this case, isn't the process that didn't first call interface_version
to find out what you're capable of more the one at fault?
Actually I have to admit that I don't use interface_version for my
interface at all, but always parse the output of mtn --version. It seems
Thomas Keller wrote:
Jon Bright schrieb:
In this case, isn't the process that didn't first call interface_version
to find out what you're capable of more the one at fault?
Actually I have to admit that I don't use interface_version for my
interface at all, but always parse the output of mtn
Jon Bright schrieb:
Actually I have to admit that I don't use interface_version for my
interface at all, but always parse the output of mtn --version. It seems
easier for me and the user, because I know yes, this works with mtn
version X.Y and if the requirement is not met, I just yell at the
On Tue, 2006-10-24 at 01:05 -0700, Nathaniel Smith wrote:
Right now, 'automate stdio's parser silently ignores anything between
the e that ends one command, and the l that starts the next. I
guess the idea was to make it easier to add extensions in the future?
I think so. But having it throw
