D-Bus Spec: "D-Bus is a system for low-latency, low-overhead, easy to use IPC"? (Re: D-Bus optimizations)

[Simon McVittie]

On 29/05/13 10:39, Justin Lee wrote:
> If I don't get it wrong, latency is unrelated to whether the operation
> is asynchronous or synchronous.

Right. People are welcome to propose better wording; describing D-Bus as
asynchronous (a message-passing system rather than a RPC system), rather
than low-latency, would make more sense.

> Well, I have no idea what are the preconditions to modify our official
> docs and who has the privilege to modify them.

They're part of the same git repository as libdbus and dbus-daemon, so
they conform to the same rules described in HACKING for the source code:
anyone can propose a change, it won't be merged unless the reviewers
listed in HACKING say yes, and reviewers are expected to reject
controversial/API-breaking changes unless there is appropriate consensus.

In practice, changes happen in the git repository only, then the actual
HTML specification gets updated when someone - usually me - makes the
next dbus release.

> So how about fix our official docs first directly?

If the changes are appropriate for the specification (it's a
specification, not a "how to"), sure, please do. The specification is
doc/dbus-specification.xml in dbus.git; we prefer to receive changes via
freedesktop.org Bugzilla in git-format-patch format.

As with any source file that has a review process, patches that do one
clear thing per patch are recommended - they're much easier to review,
which means reviewers are much more likely to have time to review them.

(I'm not going to spend time altering the spec myself for non-functional
changes, because if I do, I'll have to wait (possibly forever) for
another reviewer to review it; if I encourage others to alter it, then I
can do the review, and the change might actually happen. :-)

    S