Using WAU (using it without externals) I installed all the new libs named like lib*-3.0 (there is a few more then the librock* ones) and after that it worked pretty ok. I still get about 50+ errors in Buggrabber when I log on but as far as I see all of those are related to LibRockLocale-3.0 and does not have a impact on gameplay.

I guess the public push of these changes could have been done more smoothly

WAU should've done that for you if you had "Automatic Dependancy Download" enabled.

Which I had but it did not work as expected until I chose to "Reinstall all addons" (not the first time it did that though). A lot of people that I've helped over the last few days seem to have "Automatic Dependancy Download" unchecked for some reason, I guess it's off by default?

I had a similar problem, everything (and by everything, I mean pitbull) was horribly broken until I reinstalled all of the addons, even with that option ticked.

This one only prove my words. You either missed new library or forgot to delete addon folders.

I can understand why wowace forums is full with people who cant complete this simple process, but i always thought that this forums is place for a bit more educated members.

And stop bitching on developers already. There is nothing to do with epeens.
If you dont understand how Ace2, Rock, Ace3 works and whats the difference between them, you cant make comments like "zomg my memory is full with libs, Ck sucks" in the first place.

Or perhaps you don't have a clue, educated or otherwise?

As I said, I don't use an updater program as I run WoW on my Mac. I update files manually. And nowhere was there a note saying: "Hey, you also need these new files to make it work."

I never saw an announcement anywhere, nor did many other people who had the same problems.

If there was a site where this change was announced, I didn't see it and I have about 70 WoW-related sites bookmarked.

Not everyone runs an updater and it's arrogant to assume everyone handles mods and updating in the same manner. I am very good about keeping my mods updated, thank you very much. And it was damn frustrating to log on for a raid and encounter such a problem.

But as I did have problems (which took me about 20-25 minutes pre-raid to fix), perhaps someone enlightened as yourself can teach the rest of us next time.

And rather than trying to spar with someone who seems to have an axe to grind, all I will say further on the matter is that communication would help. When the developer of eepanels came out with a new version, the older version had a note in-game on in the files saying there was a change coming and people might want to look at the new version rather than updating it without a second thought.

Not everyone knows, or cares, about such things as Ace2, Rock, Ace3, etc. Some people just want the mods they use and love to work. And when they don't work, to be able to quickly and easily solve the issues.

What is substantially different between a tag and a flag? Thanks. Sorry I'm old.

Almost no one does this, making your entire rhetoric irrelevant. I appreciate the "It's beta! Only try it out if you want to find problems!" theory, I used to develop, But widely, that's not the practice - it's developer CYA.

Now that's absolutely insane, if you think any end user, anywhere, ever, will even comprehend the majority of what you said there. Why don't we just get everyone to install WINE, gentoo (from command line, of course), and then manually write their own drivers if they're having problems with Vista? It's easy to do!.. for someone.

Wait, wait, wait. You think an end user which actually reads EJ is lazy? It's true, a great deal more due diligence could be put forward, but then is Ace2 (and now, all related projects) supposed to be some sort of insular developer-only playground, from which users are cast aside?

You want people reading changelogs? With ~30 or so addons, and the bulk of them say things like

timestamp stuff | somedev
---------------------------------
Fixed a bug caused by an extra ;

Wow, I can't imagine how Average Enduser would have come to tune that out!

Realistically, users which go to forums are already a minority. Then to go to wowace's or EJ's forums are a minority OF a minority. Before you think I'm cutting the flank off of the bull, in my own experience I and one other guildmate are the EJ forum goers in guild. But we require Ace2 addons, and many users get other Ace2 addons that are handy/updated (hi Cartographer).

Like I said above, if the default behaviors of WAU and files.wowace.com are not what developers want end users to experience, this is their mistake. It's also bad faith if there are any claims that it's a developer tool - there are adverts at the bottom of WAU these days, speaking to the wideness of audience they were getting and developing towards.

The only problem I had when I updated my addons was that they couldn't find the libraries, since the rock libs weren't embedded into the Standalone package. It was quickly fixed by installing everything marked "Lib*", so hopefully that should fix most of the errors. At least it did for me.

Not just that, even. Often enough, there won't be much in the way of documentation at all.

However, that's part of the price to pay with running addons off of the development source trees. Sites like WoWInterface and Curse exist as final resting places for stable addons. They aren't as easy to use as WAU, but you can be fairly confident that the addons you get from them won't be rife with issues.

Last night, I ran an update of all of my stuff as well. I had also decided to give Pitbull a whirl, since I've been using AGuF for ages. At least now, I'm not surprised why my overall impression of Pitbull was poor. It sounds like I'll just have to wait a bit before looking into it again.

As with the OSS world, there are going to be differences of opinion amongst the developers. There will be some that go in one direction, others another. This is a pretty healthy way of handling it, though, and only benefits the end user. Sure, you might have to deal with multiple libraries that have redundant functionality, however, it remains purely the user's choice to go with X addon over Y.

I completely agree with that, but it's a /really/ hard thing to tell end users, because, for better or worse, many addon developers are not committed to producing release versions.

For example, the versions of Pitbull and Manufac on Curse were last updated three months ago. There's been a lot of development since then. Non-technical users are faced with the choice of using cutting edge versions, which mostly work, or laboring along with old code which does not include the new hot features people keep talking about. It's not surprising that most people go for the cutting edge stuff.

There's also no easy answer to the question. I don't think simply writing an addon creates any obligation to the community; nobody's obligated to cut a release branch. On the other hand, if best development practices include not only coding standards and libraries but also frequent stable releases, you'll wind up with a community that is overall happier.

The reality is that a large, large number of people get home in the evening, run WAU, and then log on to raid or whatever else. As a routine. Ask Antiarc -- he committed a Threat build with an infinite loop bug that was on the SVN for a couple of hours total in the later afternoon one day, and it caused a massive number of people to completely lock up the moment they entered combat in their raids for that night. Hundreds of hours of lost time, thousands of collective hours wasted via wiping, troubleshooting, having everyone revert or uninstall the mod, etc. And all because of a small error. But that's the reality of the situation -- a lot of the hard work and valuable efforts he's expended to improve and popularize his mod were almost undone with that single error, because suddenly entire guilds felt like they could no longer trust the mod (after all, KTM never crashed WoW to the desktop, did it?). For all the talk of how wowace and the SVN that WAU pulls from are development tools, the reality is that the vast, vast, vast majority of people running WAU and accessing those builds are just casual users. WAU is a user-friendly application that has continually evolved to be more and more user-friendly and clearly is designed to simplify the process for casual users. I don't have usage stats for wowace of course, but I'd wager they're quite significant.

Any mod developer who doesn't recognize this fact is oblivious, and anyone who knowingly makes drastic changes without warning or recklessly commits likely-buggy code to a build that they know thousands of people are going to download, is an asshole. If you (the generic "you" as in any such developer) don't care about the users, then so be it, but don't expect people to use your mod. I'm tired of mod authors acting like children and/or primadonnas.

It probably has been said before, but the practice of giving end users who don't know what they're doing a tool to update their addons from svn commits with no QA rather than released/stable builds wasn't exactly a brilliant idea. A lot of the last minute "Don't pull my ui broke" issues probably could have been solved simply with that. In any open project it's common knowledge that running a svn copy is liable to have problems, yet practically every ace end user seems to run svn commits and basically raid by the seat of their pants. I just don't understand it.

Has there been any effort to have the Ace2 updaters be tag-aware? I think that if there was an easy way for users to by default get the "stable" branch, and have developers work on the trunk, then maybe it would be much easier for authors to commit lots of changes, have those who want the possibly buggy stuff get it, and have those who want the more stable stuff get that.

Authors could easily just move their dev branch to the "stable(r)" tag when they have important fixes, or new features, all at their own discretion.

(I just use olde-fashioned "svn up" to update my mods, so I'd be happy checkout whichever branch an author wanted me to use.)

A flag would require extra code to actually support it (since no SVN client has it), and you'd need to add code to store all the files for that commit somewhere, tags already does this and it works perfectly well and isn't an uncommon idea Subversion (software) - Wikipedia, the free encyclopedia

It is actually a common practice that you only use beta if you want to risk bugs or crashes, originally the only way you got mods was through a resource site, you only released a stable mod or you released it as beta under a beta category or with a giant notice saying "Beta, use at your own risk.". Either way, if the user downloads something he knows what he's getting into.

WAU just gave you the latest version it could find, without any real notification of what you're getting, Praetorian's example is perfect for showing this issue.

You're confusing users and developers, users do not tag, authors tag, users get the tags. Tagging either requires a couple of commands to do, or a 10-20 line build script. If the author can't figure it out, then you have other issues.

I've heard WAU2 supports tags and branches, but nobodies implemented any of the required code to actually make it active, theres some plans for WoWAce that will fix up the repo and make it a little more sane.

Bingo!

Omen has already been mentioned, in this thread. Still in beta, still contains (I think -- don't use it) warnings of being in beta, and yet users are using it as a production (stable) release. Commit beta code, that doesn't work, to the trunk, and you crashed the raiding scene that doesn't understand the above developer idiom. This is why my guild is still using KTM, because it ALWAYS works, and we're not at the whim of one user having unknowingly downloaded a broken library.

No offense to ckknight -- it's obvious many appreciate his work and addons -- but Rock and those addons dependent upon Rock have no place on the trunk, in their current state. The fact there are hourly "bugfix" pushes is telling enough. Jamming the drastic changes down the layman's throat should not be the path taken. The very fact someone has to connect to WoWAce's IRC channel or drudge through their forums, to find that they may need to download all Lib* addons for their /trunk/ addons to "function," is ludicrous.

WoWAce's SVN should be looked at closely and establish rules for stable and beta hierarchy. With how WAU is used, this is a necessity. Furthermore, WAU should support the display of downloading stable, beta, dev1-branch, dev2-branch, etc versions of any addon. Those titles should also mean what they say, not "stable dry-code!" so often seen.

That was more a comment for people that had used svn, but to answer your question.

SVN tags are just full copies of the source. Its really space in-efficient and horrible for back porting any fixes. Libraries and all would have to be saved. So almost all of the SVN tree would be copied for every tag since almost everything depends on some of the core libraries.

there are better source control solutions its just hard to get people to start using them since they are different from CVS/SVN. If you're interested in it then I can post more info. But I don't want to sidetrack the thread too much.

I do use SVN including tags actually, but I'm not going to assume as to why you hate it.

Doubt SVN was made to handle this many projects in a single repo, and saving the entire tag is still a simpler solution then say, saving the revision number that each file was tagged at then pulling them together whenever you want to download a tag.

Disk space is far cheaper then CPU, and most people aren't tagging often enough that it'll ever become a problem.

Furthermore, if WAU is targetted at the casual, everyday user who wants to update his addons -- and it most definitely is -- it should really do the backup of your addon directory itself. It would be great if you could then revert to older backups from within it, but it should be trivial to copy off a backup directory of the old addons whenever it updates.

I personally zip my addon directory to prevent downtime in the event of a broken addon, which saved me this week when Fubar and Pitbull broke horribly, but I'm sure not everyone who uses WAU does this, even with the warning it displays when you fire it up.

I wrote a quick script after I got burnt once. First thing it does is tarball up my entire Addons directory with the datestamp, then it launches WAU. If I ever need to roll back,, I can just go nuke my addons directory and extract the backup file - very simple. I only tend to run WAU about once a week, so there aren't too many archive files lying around (and I normally delete the old ones manually after about a month).

I've got a pretty healthy amount of addons and the zip file only takes up 10-12MB, depending on what I've currently got installed. If you're REALLY paranoid, you can zip up not only your AddOns folder, but also your WTF folder, so you can easily restore any setting.

As to the whole "Testing in Production" discussion that is WoWAce, I can empathize with both sides. As a player, I like having mods that are up to date, shiny, and do cool stuff. Wowace definitely embodies this and I love the ability to just go download all of the latest versions. However, the fact that there is no clarification regarding "Yea, this version might break" is a huge drawback and I have a very hard time recommending Ace mods to my computer/programming illiterate guildmates, because it's entirely possible that when they go out tomorrow and download release X of a mod, it won't work and they won't know why. They won't even know what a Changelog is.

I realize, as a developer, that a lot of the headaches involved with responsible coding are no fun to deal with for a game - who wants to do testing/QA cycles? Who wants to maintain stable/release/testing code branches? It's so much easier to just write a quick change to the code, push it out, and then everyone gets it. Most of the time, nothing bad will ever happen. It's the few cases (like "oops, I just typo'd a library file - and that made everyone's mods sto working") where having that separation would be nice.

Either way, I doubt there will be an easy middleground - and it's not gonna stop me from playing with fire in the mean time.

The best way to hide it:

1. Use the DetachedMiniButtons Addon (DMBS)
2. /dmbs config
3. Button configuration tab
3. Select Rock icon and check Hide
4. ???
5. Profit!

I would like to call out something that ckknight's doing which is a big step forward, if it persists -- Rock includes unit test support. That's important, and it's good. Thank you.

Actually, the best way to hide it is to uncheck it in the Rock options.

It's in there now.