cvs broken?

It’s been a few days that i’m having trouble to checkout the cvs version of panda.
Is the server working?

Thanks for help.

Same here. Specifically:

cvs -d:pserver:anonymous@panda3d.cvs.sourceforge.net:/cvsroot/panda3d co -P panda3d
cvs [checkout aborted]: reading from server: Connection reset by peer

Does someone with CVS admin rights need to do something to make sourceforge happy?

From the statistics graph, there’s been exactly zero read/write activity for about ten days now:
sourceforge.net/projects/panda3d … Repository

I’m beginning to seriously consider creating a Panda3D fork to guard against the existing project having been abandoned and ensure continued development… :-/

–jonah

It seems to be working again now. A bit slow though.

I agree that it’s possible to pull code from CVS again, however I believe my larger point still applies: the sourceforge activity graph linked above has not updated for about two weeks now, suggesting that something or other is wrong and that someone with admin rights may need to give it a good hard kick to get whatever is gumming up the works unstuck. Unfortunately it’s not clear to me that there are any admins still actively involved with Panda so further investigation may be problematic.

–jonah

I think that the December holidays and deadlines have pushed David’s Disney Overlords to prioritize his focus in the near-term onto other things. But the fact that Panda3d has been registered on Sourceforge since 2001, and has undergone continuous development since then, suggest that a 10 day outage is not necessarily the end of development by the ‘official team’ (ie., David + CMU).

As for the recent outage on Sourceforge, I don’t know why it broke for a few days, but it seems more like a problem with the Sourceforge team rather than any problems with the Panda Admins. And if this is true, I’m not sure what the Panda Source Admins can do except write a few e-mails and wait. Did sourceforge have a site-wide meltdown the last few days? Even if not, I can easily imagine how a project first started in 2001 could have a hiccups in the database with whatever new thing/feature Sourceforge has recently implemented.

I looked at your changes on your website, and they’re admirably ambitious and extensive. If fact, if you hadn’t revealed it just now, one would have thought from your bursty post patterns on the forum that you had stopped working on it as well. :stuck_out_tongue: If you feel any part of what you are currently working on should go into the cvs right now, let me know. Hopefully you find now after working with me a few times in the past, that I’m extremely conscientious and always very appreciative of more Panda work.

Zhao

And as for things to worry about, I’m not really worried about the CVS as Panda is already extremely stable and so many of us have copies of the sourcecode already.

I’m much more worried about this forum which holds so much of the collective wisdom accumulated over the years. Losing it, would be devastating!

Are you thinking maybe it would be helpful for the community members to manage the forums and so on, to allow the limited resources of the developers to be better spent?

I think rdb is already involved with the website somehow. But it would just make me more comfortable if backups to the forum posts were made available somehow. I’m always been contemplating writing a web crawler to do this manually, but I’ve never quite had the time to do it.

zhao, my “bursty post patterns” are partially because I get really discouraged by the spam on the forums, and partially because of the resounding silence I got when attempting to check the state of active developers and users. As I write this, for instance, I’d logged on to check out the various threads I’ve recently commented on and almost gave up right away because of all the spam. That’s part of what I mean when I say that it’s not clear to me that there are any admins still actively involved in Panda: the spam problem has been ongoing for months now.

That, coupled with the apparent lack of anyone “steering the ship” on the active development front, has encouraged me to lurk rather than comment. It’s also encouraged me to sit on the various modifications and additions I’ve been making locally rather than publicize them. Ideally I think all of my changes should go into CVS, but at the same time I’m very much against allowing any un-reviewed changes into the master repository, including my own. Without any active admins or project managers it’d become a free for all, and that’d be a Bad Thing™.

I agree that the forums collapsing would be a major loss. Over the weekend I snarfed the entire CVS repository, with the full history, so I’d have a backup in case anything happened. In the next few days I’ll probably put together a crawler for the website and the forums so I’ll have a backup of that too…

Finally, a note on the original topic: it appears that the people at sourceforge took the CVS server down back around the time this thread was started, and that since it’s been back up it’s been “slow”. This from the SF.net Operations Twitter feed, here: https://twitter.com/sfnet_ops. I’m guessing that the problems with the usage graph I’ve mentioned are a side-effect of that.

–jonah

I agree that the forum spam sucks incredibly. I’ll suggest to rdb that new users should perhaps have their first 2-3 posts approved by moderators before being allowed to post freely.

If you get the webcrawler working, please send me a link of the forum archive. I would love to have a copy of that.

As for other code things, my offer still stands. If you think any of your current stuff should go into the CVS, let me know and I’ll help you review, test it, and and check it in.

Zhao

ps. Please forgive my “bursty” comment, I was just teasing you in a friendly way. :stuck_out_tongue:

For what it’s worth, my experience thus far has been that the spam does tend to get cleaned up - it’s just a matter of when the moderators do so. It may simply be that your posting times happen to tend to fall after the spam has been posted but before the cleanup sweep has been performed.

Agreed that the spam does tend to get cleaned up eventually, and that there are any number of mechanisms such as requiring the first few posts to be moderator-reviewed that’d get rid of it in the first place.

Still–to keep beating a dead horse–without any active admin involvement it’s never going to really get fixed.

–jonah

It’s true, things don’t happen as fast as I’d like, but we’re all a little short-handed here. Panda is almost entirely a volunteer operation these days. Are you volunteering to help out with some of these administration issues?

David

Even though I don’t have much time, I’d still like to give some of it to help Panda. The problem is we don’t really know what we can do to help.

For example with the spam problem, I’m sure there is an easy solution to the problem, such as adding a simple question on the account creation form (I have seen many simply asking “1+1=?”). Instead we are just told that it is being worked on.

Why not actively outsource some of these problems to the community? Ask us to help! Sure we might not be able to implement some of the solutions without the keys to the castle, but we could certainly spend hours researching and providing the results.

Put plainly, I think it is a massive waste if the few contributors to actual Panda code have to spend that time dealing with admin problems instead.

i hear what you say. now project seems half-forgotten really.

on the topic: maybe its time to ditch cvs for something more usable like git?

I’m not involved with the spam problems; for more information on that, we’ll have to get rdb’s opinion. But I can offer a guess: part of the problem here is no doubt that the forum is hosted on CMU’s servers, so doing any kind of upgrades to the forum software requires synchronizing with the CMU site administrators, and getting some of their time to do any required updates.

Experience tells me that this kind of synchronization is frustratingly slow.

We can talk about moving the hosting of the forums completely away from CMU, but this has its own set of problems. (For instance, this thread began with a complaint about the CVS service, which is hosted on SourceForge and is also therefore not much within our control. If we moved the forum hosting to SourceForge, we might just be moving the problem. SourceForge is also short-handed for the services they provide.) Other third-party hosting options are either ad-supported or will demand a fee. Not to mention that continuing to maintain at least some connection with CMU might still have some value for Panda’s longevity.

Let me say again that I’m just guessing about all of this. The main point is that because we’re an unfunded large project, it’s much more difficult to make changes that seem like they would be easy (and which would be easy for a small project or for a fully-sponsored large project).

David

Guys, Steve and I have been working hard lately on getting the website ported over to a new server and all of the ancient software upgraded to the latest versions with all of the security-related bells and whistles that we need. It’s all brand new with phpbb3 with proper CAPTCHA, and it’s nearly done; the spam problems on the wiki and forums will be a thing of the past soon. But I’ve also been struggling with an illness which is causing delays in everything, so please bear with me.

I don’t see any problem with our hosting at CMU. Steve’s an excellent server maintainer who’s actively concerned with the well-being of the Panda3D website, and he’s been very responsive and helpful with everything.

I’m trying to be as active as I reasonably can. In the meantime I have been getting some community members to help out with the spam problem, I am getting the impression that it’s under control for the moment.

As for CVS, that’s SourceForge’s fault (check their blog or twitter for updates), and out of our control altogether.

@11thPenguin and teedee: if you have any patches you’d like me to review, I’d be happy to check them in, please send them to my private e-mail address or file them on the bug tracker. If I don’t respond to them, please poke me, sometimes these things just get lost somewhere in the big pile of e-mails in my inbox.
You guys have been making some excellent contributions, though, so perhaps we should look into getting you CVS access.