0.2.8.3_beta1: Release process and bugs

[Monkey]

Mokey

First of all, thx Lucifer for clearing that up, it has helped a lot in my understanding of the Arma release process.

Z-ma: First of all, as someone who plays a lot (compared to most devs that don't), I know that my feature is essential to gameplay but, for now, nvm.
Secondly, I usually report bugs immediately but they just get left for months ...see Launchpad. I am busy every day organising lame people for the Ladle (2.5 teams --> 1.5 teams, to be more precise). If I report this bug, it won't be fixed for months anyway so my priority is on increasing the fort community by helping new people enter the Ladle for the first time.

Dev's being angry at me for not reporting one bug ASAP, when dev's are not trying to fix them ASAP (which for some of the bugs is a quick job) is hypocritical. I'm not at the stage where I can do this myself yet.

[Lucifer]

Heh, what makes us angry is when people talk about "all these bugs" without giving us the information we need to fix them. Easiest way to piss anybody off is to tell them that the product of their hard work sucks, and then button up when they ask why. I'm sure you can relate to that. Saying a program is buggy is semantically equivalent to saying it sucks, unless you provide the information we need to find the bugs and fix them.

I'm curious what this feature is that you say is essential. Consider that for every new feature that is essential, it has to be weighed against roughly a decade of armagetron that lacked that feature.

[Z-Man]

We do fix bugs in 0.2.8 we know about ASAP. It's just that ASAP isn't always right now, (we too have other bits of life to care for), and of course there are important and unimportant bugs. Your two reports that we commented on practically immediately were important. The others aren't quite as important, and to be honest, I don't even know whether they're bugs. We can argue about this all day, but fact is: bugs that don't get reported will not get fixed.

[Monkey]

I didn't say that Arma sucks and I usually report every bug that I find ASAP. Look at Launchpad and see how many of the more recently reported bugs were reported by me. I also never criticised any developers at all, it was you that criticised me.

Your two reports that we commented on practically immediately were important.

"Bug #245925 reported by Monkey on 2008-07-06"
"Manuel Moos wrote on 2008-08-01:"

Z-man: It took you a month to comment on that "immediately" and only because I told you about it elsewhere, you didn't check Launchpad. No other devs commented on it either. That is a bug that was previously marked as both "critical" and "fixed" so I take it you are counting that as one of my "important" reports. If it takes you one month to reply to an "important report" then you can't criticise me for not reporting a bug, on one occasion, within the first few days of me finding it. I also have to get more screenshots and try to do some testing of my own first to see exactly when it happens. If I put some lame, half-ass report on Launchpad, you would complain about that instead.

As far as the "two other bugs" that you aren't sure whether or not they are bugs at all: They are either bugs (accidental) or, even worse, design flaws (deliberate) and either way could be fixed quickly by a dev.

Bug1: https://bugs.launchpad.net/armagetronad/+bug/254525
Bug2: https://bugs.launchpad.net/armagetronad/+bug/253978

I have one of the strictest attitudes towards bugs that you will ever see. I believe in fixing all bugs, not just security/stability fixes, ASAP (not lazy ASAP but proper ASAP) and providing updates ASAP. Even Linux distro's annoy me in the fact that they only provide security/stability bug fixes for stable versions.

I know the score with bugs both in Arma and elsewhere. There is no way that I have a slacker attitude, than any devs, towards bugs.

I won't even begin to express my views on documentation...

[Lucifer]

I know the score with bugs both in Arma and elsewhere. There is no way that I have a slacker attitude, than any devs, towards bugs.

We do fix bugs in 0.2.8 we know about ASAP. It's just that ASAP isn't always right now, (we too have other bits of life to care for),

As an example, I don't know what Z-man's got going right now, but I just got divorced, and that has occupied all of *my* time. Right now I'm stuck at work about 14 hours a day, even though I'm not working that many hours, and don't have a development environment up there anyway. That's in case you were wondering what "other bits of life" z-man was talking about. The rest of the folks around here are mostly students, including z-man , and I'm sure you're aware of how much time that can suck up.

Be careful what you toss around, and mind you that you have been extremely guilty in the past of giving really bad bug reports, and refusing entirely to report bugs (you and I have argued about this in IRC, remember?). And, of course, comparing your work ethic to that of the developers here with a holier-than-thou attitude won't win any points with anybody.

And, having reviewed the start of this side chat, it looks like you were suggesting delaying a beta release so we can fix bugs, but the *purpose* of a beta release is to *find more bugs*, so we can fix them before we make a final release. Beta releases are like MS Windows, they're supposed to have bugs.

[Z-Man]

Monkey: You still don't get it. What drives us mad (well, me anyway) is not that you decided to report a bug later. That's fine. What is not fine is that you announce that, while making assertions of the bug's importance. Imagine going to a doctor, and the doctor says "Oh, I've found something, you're ill. But sorry, I can't tell you what exactly it is right now. Too busy. Come back next week." That's what you're doing to us.

[Monkey]

Lucifer: I know that people don't always have time.

you have been extremely guilty in the past of giving really bad bug reports, and refusing entirely to report bugs

No...and don't criticise me when your wiki documentation is as poor as poor can be.

You still don't get it

Yes I do. You lot are the ones criticising me, not the other way round. It's ok for you guys to not respond to a "critical" bug for one month but not ok for me to not report a bug that I have not had time to get all the necessary information for? Pft. I simply corrected someone's statement that there are more than "just 2 bugs to fix".

Mokey: Your feature is not important for 0.2.8. Plain simple.

The only dev that actually plays the game is epsy. How can those of you, that don't even play, know what features are "important" or not? I might as well mention it then:
The 0.3 glances are a complete change from 0.2.8. We want to keep additions to 0.2.8 to a minimum so, rather than putting the 0.3 glance code into 0.2.8, I would like to add one option to the 0.2.8 glances. At the moment, when you glance and then turn, the camera gets reset to the forward position. I would like a simple setting that allows the camera to not be reset to the forward position and, instead, continues to face in the same direction, RELATIVE TO THE CYCLE (not the same direction full stop). I'm not sure if there is this option in the 0.3 glances (if not then it needs to be added) but IMO it is *essential* to basic gameplay. I currently suffer greatly without it.

[Z-Man]

It's ok for you guys to not respond to a "critical" bug for one month

Yes, because
1. The affected versions are alphas.
2. There is a workaround, don't use display lists.
3. The bug is of a kind explicitly mentioned when the alphas were released, with hints to 2.
4. As far as I can tell, it's not even our bug.

but not ok for me to not report a bug that I have not had time to get all the necessary information for?

You haven't said so far that you need to gather information. That changes things somewhat. Incomplete bug reports, where there could be steps to reproduce, but aren't, are indeed a pain.

I simply corrected someone's statement that there are more than "just 2 bugs to fix".

Epsy mentions five problems, which, according to the bugs that actually were reported, was quite correct.

Mokey: Your feature is not important for 0.2.8. Plain simple.

The only dev that actually plays the game is epsy. How can those of you, that don't even play, know what features are "important" or not?

Even though we're not too active right now, the rest of us do have a lot of play time on 0.2.8.

About your glancing option, the problem there is that the camera code is a complete mess, especially the glancing. Adding that feature on 0.2.8 would be easy, but it won't merge to the trunk at all, so it would need to be redone there and probably behave slightly differently. People are really sensitive to that kind of stuff. There definitely should not be a new menu item for it. A regular config item may be possible, but don't hold your breath for it. As said, the camera code is ugly and suffering from featureitis and otpionitis already.

[Monkey]

Yes, because 1. The affected versions are alphas.

The bug I have found affects only alphas too. If you assumed that it affects all versions then you shouldn't have.

You haven't said so far that you need to gather information.

I also have to get more screenshots and try to do some testing of my own first to see exactly when it happens. If I put some lame, half-ass report on Launchpad, you would complain about that instead.

Yes I did. Even so, it should really be obvious. Many bugs require an effort to gather information and reproduce.

I simply corrected someone's statement that there are more than "just 2 bugs to fix".

Epsy mentions five problems, which, according to the bugs that actually were reported, was quite correct.

I didn't mean that epsy was being incorrect, which he was not, to his knowledge. I meant that the facts were incorrect and I was simply informing people.

Even though we're not too active right now, the rest of us do have a lot of play time on 0.2.8.

You certainly don't spend much time on the network servers on the master list. Unless you are having LAN parties that I am not aware of, I can't see how you can fully experience what the game needs, to play competitive, network play against other people.

Adding that feature on 0.2.8 would be easy, but it won't merge to the trunk at all, so it would need to be redone there and probably behave slightly differently.

I am aware of that. That is why I suggested it only for 0.2.8. I am separating 0.2.8 and 0.3 glances entirely.

A regular config item may be possible, but don't hold your breath for it.

I won't. When I am more ready to program, in the near future I hope, I may just write a patch for it myself. We can go from there and decide whether or not to include it in the official code.

[Tank Program]

This is spiraling out of control. This is not a Monkey vs developers battle. This is supposed to be us announcing our intention to push 0.2.8.3 forward. Hurrah! Good news! Everyone celebrate! We're doing a release for the first time in two years.

We of course will try and fix bugs, that's what the release process is about. But it's not just about your bugs. We appreciate your feedback but we can't do something about everything. We also of course take the community under advisement and do things for them, but we also try and push forward our own innovations which we think will improve the experience for everyone.

No one has the right to decide what's best for someone else, saying this or that is more important or shouldn't be here. However, us official developers are in a position where we can do this because of the time and effort we've invested in this project. It needs to stay fun to us as well as to everyone else, that's why we're not so serious. The community gets its kicks playing- we get it developing. (Or really I do it webmastering- same difference.)

Long story short, I'd really like to see this argument stop and work proceed, because all I think this is doing is making it discouraging to start working. Why bother if people complain? I'm not trying to shut you up here, but give it some time. Let's see what other feedback there is during the release process and we'll go from there.

[Monkey]

us official developers are in a position where we can do this because of the time and effort we've invested in this project.

Don't forget the time and effort that some people have put into the community that does not involve coding.

Why bother if people complain?

If you look at my first post Tank, you will see that I was just posting some information which the devs then decided to pounce upon. I just defended myself as any self-respecting person would. I think it has been mostly resolved and I don't want to carry it on either.

On with the release cycle...

[Z-Man]

New build, 0.2.8_alpha20090113, up on sourceforge. It should also be up on aabeta, but svn update notifications seem broken, so the site hasn't been updated yet. I think all known non-obscure bugs have been fixed in this build. Obscure bugs left:

Lucifer and dlh have seen odd cycle wall remnants that stick around and don't vanish like they should. Recordings very much required for a fix.

While I've made double and triple sure everything is allright with our OpenGL calls and display lists, enabling display lists still may cause crashes. If so, please report your exact hardware and post your user.cfg so we can blacklist their usage.

My Mac client build has very sporadic crashes They have consistent callstacks (cycle update->eGameObject::Move()), but they're not exactly reproducible, and dlh's 0.3.1 build doesn't have them. I suspect it's the compiler screwing up somewhere (wouldn't be the first time). I'm using Xcode 2.5. I'll be playing around with the compiler settings and do longterm tests.

Bug reports welcome.

[DDMJ]

My Mac client build has very sporadic crashes They have consistent callstacks (cycle update->eGameObject::Move()), but they're not exactly reproducible, and dlh's 0.3.1 build doesn't have them. I suspect it's the compiler screwing up somewhere (wouldn't be the first time). I'm using Xcode 2.5. I'll be playing around with the compiler settings and do longterm tests.

Z-Man, try my build here:

http://durkas.info/mac/

I built it 2 days ago, but I'm curious to know if the crashes continue (since I haven't noticed them). Then we'll know that it's probably your compiler.

[Z-Man]

Thanks. I'll run my tests on it (it's basically a local game with some AIs and very high speed at disabled vsync). In the meantime, I compiled with -O1, that seemed to have helped, too. I uploaded them instead of the old builds now.

[newbie]

thanks for the good work, i'm not sure if anyone has mentoined it yet but the difference between 2821 and the latest alpha is about 200fps in ct team sumo server with 12 players