Oracle *Has Not* Stopped Publishing MySQL To Launchpad

Stewart over at Percona noticed that our trees on Launchpad were out of date, and this has been picked up and taken as gospel by others, notably Henrik today.

Unfortunately, Henrik doesn’t seem to have checked Launchpad before publishing his blog – otherwise he would have noticed that all of the trees are currently up to date with the current releases.

He would also have noticed that there is now a new MySQL 5.6 branch, so that we can continue with trunk as the future MySQL version after MySQL 5.6.

So let me repeat this again, Oracle has not stopped updating the MySQL branches on Launchpad.

Sorry about the blips in updating, however.

22 thoughts on “Oracle *Has Not* Stopped Publishing MySQL To Launchpad”

    1. I was speaking about the MySQL Server trees, which is really what all the “news” is about. I have no idea about why the Connector trees are not getting updated, fwiw.

  1. MySQL is a client-server software and with only a server part, without popular clients, clients, it is of no use.You message is that Oracle did not stop publishing MySQL launchpad trees, and that message is not accurate, in fact it stopped publishing most of them.

    1. FWIW, I know of nobody that really has been a downstream dependency of those trees themselves – I’m not using that excuse for them not being updated (again, I don’t know why), just that I don’t think it’s as big a problem as the Server tree itself (and again, that’s my own view).

      Do you guys at Monty Program have some problem with it specifically, or know of people that have a need for it, or are you just noting this to try and further your propaganda?

      I’ll at least do some of my own investigation in to why these aren’t getting updated, though.

      1. Mark: unless you’ve noticed bzr and git have some years ago become the preferred form for trafficking source code. It is of course not required by the GPL to publish bzr trees, but neglecting them will get people’s attention.

  2. Mark,

    Because you’re so interested, let me share my secret plan with you.
    I want to fork all of Connectors and make a business on my own. I want connectors, completely with repository containing change history to accomplish this evil plan, which however is nothing illegal according to connectors license.

    Thank you so much for mentioning this fresh propaganda idea, now I will certainly use it if Oracletries to hide lovely connectors behind its corporare firewall.

    Thanks again and please inform us about the state of your investigation.

    1. I asked if you knew of a valid need to have the trees on launchpad, obviously not. 😉 AFAIK, all down stream packaging has happened with the source packages, which haven’t gone away.

      And, not having the trees for connectors is also not illegal, according to the license.

      I’ll still see why they are not up there, though. I’m sure it’s not as evil as you are trying to make it.

      Good luck with your business plan! 🙂

  3. Hi Mark

    Thanks for a) refreshing the bzr trees and b) informing us about it.

    I don’t understand how the commit dates in launchpad are all from July, so I’m not going to interpret them, but based on Stewart’s post it seems that at least the 5.1 tree was not updated since May. I realize now that Oracle skipped the 5.1.64 release, so I’m willing to believe this explains the many months without updates.

    But still, Oracle could at least do itself a favor and have someone respond to Stewart’s and Sergei’s requests about this. A whole Summer without activity coupled with radio silence will certainly make people like me draw conclusions. And it will affect people who do use MySQL as an upstream, which is not me, and they will notice, and for all they know it will look like Oracle has stopped publishing code on launchpad.

    I have updated my blog with a link to this post.


    1. The dates for the commits are the dates leading up to the clone off dates for the releases themselves (which means they will be a month out of date or there about, as we clone off, build takes a while, QA / sign off takes a while etc.). The actual releases in August were all generally cloned off some time early July.

      Tomas (for those not in the know, VP of MySQL Engineering) has responded to Stewart’s question:

      I can’t comment on security related issues unfortunately (or fortunately, depending on your perspective).

      1. To be clear, Stewart and Sergei expressed frustration that Oracle had already failed to respond to them for some time. I welcome Thomas latest response though. If this continues, maybe I will have to read mysql-internals more often 🙂

  4. You have to forgive Henrik for having this interpretation, it’s been correct for several weeks now and only today has it changed.

    Don’t get me wrong, I think it’s great that the BZR trees continued to be pushed (albeit always with a delay).

  5. Mark,
    Thanks for the update.
    It is unfortunate that Oracle refreshes the bzr trees only when there is an outcry about it. This is not the first time it happens. Wouldn’t life be better for everyone if the trees were updated automatically? Less FUD, less speculation, less upsets, more trust.



    1. Btw, it is worth pointing out that my blog post was not primarily intended to bitch about Oracle’s actions. I first thought of writing it when MariaDB announced their divergence in version numbering, which I think is a significant change.

      The Oracle related outcry just happened while I was procrastinating to write it. I covered it as it happened, but it was not the reason to write the blog post. (Had it not happened, I would probably had written something nice about 5.6 instead. Now you will have to read Baron’s blog for such info.)

    2. It’s not our intention to update them when there is an outcry, it’s our intention to update them in conjunction with releases.

      Mistakes can happen when doing this automatically, and we’re more conscious of that now than ever, so I don’t see that changing. We just need to make sure that our process for updating runs more smoothly than it obviously has done recently.

Got something to say?