On 7/15/20 3:48 PM, Christopher Shannon wrote:
Actually this may be easier than I thought. I forgot that OpenWire doesn’t
include property names so we might be able to get away with just renaming
things and everything would work fine and be compatible as long as
properties are in the same order. It needs to be looked at more to verify
of course.
Agreed, took a quick look and from the wire level perspective the older
clients wouldn't care what the broker libraries called those fields as
they are decoded in relative order by the broker marshaling code. You'd
need to change the set calls in both the versions in activemq-client and
in activemq-openwire-legacy but that should still only constitute a
broker level change (and new client release that accompanies it).
Possible there's some case where it matters but it isn't immediately
apparent.
On Wed, Jul 15, 2020 at 3:18 PM Christopher Shannon <
christopher.l.shan...@gmail.com> wrote:
Gary, that makes sense to me as I think making sure everything is
backwards compatible for existing users is important. (not just for
openwire but all config, etc) There would need to be some work done in the
broker to know which properties to use based on the negotiated openwire
version and be able to fall back to old variables and config if needed.
On Tue, Jul 14, 2020 at 7:51 AM Gary Tully <gary.tu...@gmail.com> wrote:
I think for openwire, rename and a change in openwire version is the way
to go.
keeping the old terms around for backward compatibility is both
sensible and necessary.
On Tue, 14 Jul 2020 at 11:42, Christopher Shannon
<christopher.l.shan...@gmail.com> wrote:
I agree that it is time to make the change. Justin made a good point in
that we should make sure to pick the best and most descriptive names
possible for the use case. Whether that is follower/leader,
primary/secondary, primary/replica, live/backup, etc. I think
live/backup
certainly makes a lot of sense for Artemis. For 5.x I think it also
makes
sense but primary/secondary is fine too.
My main concern here is how do we handle the technical issues with
compatibility? For example, do we just deprecate the old configuration
and
terminology to not break users or do we rip it out entirely initially
which
would be a breaking change for users that needs to be well communicated?
For Artemis, maybe we deprecated in 2.x and in 3.x.
Another thing is some things will be easy to change and others not so
much. For something like the LevelDB store that uses the terminology
this
is easy as we can just remove it entirely as we plan to remove it in
5.17
anyways. However, one thing that does seem like a challenge to fix is
OpenWire. For example the BrokerInfo command actually uses the terms
slaveBroker and masterBroker. Renaming these would now break
compatibility
with brokers running older versions of 5.x. I think the only way it
would
work is to keep the terms around for the older versions of OpenWire and
then generate a new version that has them renamed which I'm not sure if
that is or isn't acceptable as the software would still be distributed
with
those older terms laying around.
On Mon, Jul 13, 2020 at 11:57 PM Xeno Amess <xenoam...@gmail.com>
wrote:
They did?
OK, then I have no more doubts.
Justin Bertram <jbert...@apache.org> 于2020年7月14日周二 上午11:42写道：
For what it's worth, GitHub is changing the default branch name so
there's
no argument to be had with them as you suggest. See here [1] for
example.
Justin
[1] https://www.bbc.com/news/technology-53050955
On Mon, Jul 13, 2020, 10:24 PM Xeno Amess <xenoam...@gmail.com>
wrote:
Hi.
If you really think "master" is something you cannot accept, then
you
might
argue with github first.
after all their default git branch name is "master", and github
have
far
more user than ActiveMQ.
Bruce Snyder <bruce.sny...@gmail.com> 于2020年7月14日周二 上午11:03写道：
Someone mentioned use of the terms 'primary' and 'backup' in the
private
list and I liked that suggestion. I'm not wed to any terms
necessarily,
so
if Artemis is already using the terms 'live' and 'backup', I'm
ok
with
that
in ActiveMQ.
Bruce
On Mon, Jul 13, 2020 at 8:42 PM Justin Bertram <
jbert...@apache.org>
wrote:
Thanks for kicking this off, Bruce.
Among other things the Jira [1] says:
'master' and 'slave' should be replaced with the terms
'primary,'
'secondary,' 'tertiary,' etc.
I would offer "live" and "backup" as suitable replacements for
"master"
and
"slave" respectively. The Artemis code and documentation
already
use
"live"
and "backup" in many places although some instances of
"master" and
"slave"
do exist. Aside from the fact that they're already in use I
like
the
fact
that they're relatively short and they clearly capture the
underlying
functional semantic. In my opinion the terms "primary,"
"secondary,"
etc.
are actually a bit vague and they're certainly quite a bit
longer
which
isn't a huge deal but it adds up when writing tests,
documentation,
etc.
Justin
[1] https://issues.apache.org/jira/browse/AMQ-7514
On Mon, Jul 13, 2020 at 3:04 PM Bruce Snyder <
bruce.sny...@gmail.com
wrote:
Given the racial charged nature of certain terms in today's
world,
I
feel
that action should be taken to change any such terms in all
the
ActiveMQ
projects. Examples include 'master,' 'slave,' 'whitelist'
and
'blacklist'.
It doesn't matter where these terms originated or how long
they
have
been
used in computer science. I have friends who feel that these
terms
are
offensive and present a barrier to entry to some. So, I
would
prefer
that
they no longer be used anywhere in the ActiveMQ project. The
simple
fact
is
that changing these terms will not change the functionality
of
the
features. Furthermore, compared to many other prominent
projects
throughout
the open source community, ActiveMQ is late to the game on
this
point.
So, I have created the following JIRA Issue to encapsulate
this
work. I
have not assigned any components simply because this should
span
all
sub-projects and documentation:
https://issues.apache.org/jira/browse/AMQ-7514
I have already begun work on this effort in a branch in my
own
fork
of
the
activemq repo. This is to facilitate an eventual pull
request to
the
ActiveMQ project. Anyone who would like to join me in this
effort
please
reply to this message.
--
perl -e 'print
unpack("u30","D0G)U8V4\@4VYY9&5R\"F)R=6-E+G-N>61E<D\!G;6%I;\"YC;VT*"
);'
http://bsnyder.org/ <http://bruceblog.org/>
--
perl -e 'print
unpack("u30","D0G)U8V4\@4VYY9&5R\"F)R=6-E+G-N>61E<D\!G;6%I;\"YC;VT*"
);'
http://bsnyder.org/ <http://bruceblog.org/>
--
Tim Bish
