Make a new release

[svenstaro]

Currently distros are still using the version 0.7.2 from 2012. I think a new release should be considered.

[dagwieers]

Yes, I plan to do it hopefully this week.

[svenstaro]

Cool, very nice to see you back actively working on this project. I really like this tool and hope a release can be made soon. :)

[dagwieers]

There is some delay because I would like to have some VMware guest plugins added to the 0.7.3 release. See also: http://github.com/dagwieers/vmguestlib

[scottchiefbaker]

+1

[svenstaro]

Well, any progress on that?

[dagwieers]

Sigh, I still need to go over all the changes and distillate the interesting/contributed ones for the ChangeLog. Let's hope I can finish that today...

[dagwieers]

BTW v0.7.2 was supposed to be released on 3/9/2010 and about that time my firstborn was born :-) She's now 3.5 years old, so there's no worse reminder of how I have been slacking...

[svenstaro]

Eh, to be honest, just release it and release often. Nobody cares for perfection as it is never realistically achieved anyway.

[dagwieers]

I agree. Are you interested to help out in this regard ?

[svenstaro]

To be honest I would just tag HEAD and then update the package in Arch. It's not like you need to assemble a list about important security advisors for dstat. Just tag it and get rolling. It won't be the perfect release but nobody cares.

[scottchiefbaker]

I second release early release often. I've been running the git version for a while and haven't hit any issues.

[dagwieers]

I agree with the "release early release often", but arelease should indicate what has changed and should have the documentation adapted. And the first part is needed, in order to check if the second part is ok.

[scottchiefbaker]

Are we any closer to a new release?

[svenstaro]

Dag, please, just give us a tag to get things rolling. Make it 0.8rc if you like. That pay people can get packaging and you can still make final preparations.

[svenstaro]

Seriously, please just tag something, a beta or an rc or something if you want a more proper final release. For now, any tag is fine. The branch is fairly stable.

[scottchiefbaker]

@dagwieers any update on this? Can we get a release going?

[svenstaro]

Well this sure is getting silly.

[thinrope]

May be somebody has to fork, prepare a release and send pull request linking it here? I just realized there is git repo for this and I am running 2010 code :-)

And there are about 13 pull requests that probably need to be merged at some point...
We have 93 forks already, most of them dead and not merged back it seems.
It looks like a daunting task for @dagwieers at the moment... But may be to keep rolling, some release in the past (tag) with a bit of changes, then add some more and so on.

[svenstaro]

@dagwieers can you give us a life sign?

[scottchiefbaker]

@dagwieers have you abandoned this project? Would we be better suited forking the code and finding a new maintainer? The last git commit was almost a year ago.

[dagwieers]

Why fork ? If someone wants to do the release management, why not offer to
do it ?
On 8 Jun 2015 19:17, "Scott Baker" notifications@github.com wrote:

@dagwieers https://github.com/dagwieers have you abandoned this
project? Would we be better suited forking the code and finding a new
maintainer? The last git commit was almost a year ago.

—
Reply to this email directly or view it on GitHub
#55 (comment).

[scottchiefbaker]

What's involved in doing the release management?

I only suggested forking the code because it seems like you don't have much interest any more. Mainly based on the git commit history.

[SpamapS]

@dagwieers I think @scottchiefbaker was actually offering to do the release. :)

[SpamapS]

Just a point of interest here. I am working on using dstat to improve performance metrics in OpenStack. The dstat_mysql5_innodb plugin would be perfect for this, but has not been released so access to it is limited to git cloners. Please do a release, and if there is anything I can do to help (including setting up a group on Github, forking this repo, and releasing from that repo) I'm quite motivated to get this out there and into distros ASAP. Thanks!

[SpamapS]

Dag, hopefully you're reading your email. Can you check your spam folder for email from clint@fewbar.com (Thats me). I sent you an offer to help, and have received no response.

[kdave]

If I may suggest something: give it a month and if you don't get any response or the release happens:

• start a dstat-ng fork
• tag the release in current git with some 'rc' so there's time for fixups before the final release
• write changelog: ie. a high-level overview of new functionality, important fixes
• do teh release

At this point you'll make a lot of people happy, dstat is just too awesome to be ignored like it is.

Then:

• go on and review & merge the outstanding pull requests, the easy ones first
• tag another release minor after some weeks (a month schedule would be fine until the pull request queue goes down)

The "-ng" approach worked for several projects. If the git continuity is not broken, a back merge to the original project can happen at some point.

[scottchiefbaker]

This is my annual request for a new dstat release :)

[dagwieers]

Whoops, @SpamapS I did not see your mail, so it must have gone to the SPAM folder.

@kdave Indeed, that's basically what is needed for doing a release. Except you also want to look at the PRs and issue tickets to see what's important and still needs to be addressed. Usually I keep track of loose-ends using the GitHub Milestones.

The problem is that even if you spend some time doing the above (which I did a few times over the years) and then get stalled (keeping up with real life) you may have to start over again. Ad inifinitum...

[dagwieers]

@SpamapS You are now a collaborator. Whatever that entails :-)

I don't mind making Dstat a Github project, open up the discussions about the future of Dstat, give more people rights to manage the project, etc.

[scottchiefbaker]

@dagwieers I think making dstat a community Github project is a great idea. There are obviously quite a few people passionate about the project. I'm happy to help in any way possible. I literally use dstat everyday, so I look forward to seeing it thrive.

[thinrope]

Well, heat is on, @SpamapS, we are looking at you now!

[dagwieers]

@scottchiefbaker You made your case. You are now a contributor too :-)

[dagwieers]

Ok, I wanted to create a new @dstat project on Github, but the account is already taken :-/
So we cannot have dstat/dstat. Not sure what else we would want to do now.

[svenstaro]

Write Github support to request dstat/dstat. That account has not been used for 2 years it would appear. This is an ok-thing to do, I did it once.

[dagwieers]

I already did, they say user @dstat is not dormant even though there is no public activity. So we are out of luck. (I was offered a one-time plea to the user, but I opted out) We could rename the project though.

Originally I started using the name mstat (modular stats), but that conflicted with some other project). I selected dstat (dynamic stats) because with only 2 characters (ds) you could get to the command, and on my system today that is still the case.

[scottchiefbaker]

@dagwieers Yay! Glad to be part of the team. Long live dstat.

If we can't have dstat/dstat maybe we can create a new group, but keep the dstat name? Something like: wieersware/dstat or something? I like the name dstat, I don't think that needs to change just because someone else has that group on github.

[scottchiefbaker]

@svenstaro is there a way to reach out to @dstat and see if they're actively using that account?

[dagwieers]

@scottchiefbaker Only through Github support (which I did) or through comments like these to @dstat :-)

[scottchiefbaker]

@dagwieers ah ok... We can cross that bridge a bit later if we can agree on a good organization name. I think step #1 is going to be to establish a plan for the next release. If you don't mind I'd like to close this issue (or you can), and start a new one with less old cruft.

[dagwieers]

Go ahead. I gave away all my rights to block this thing years ago.