3.3.0 is imminent

[Magnate]

As jens noted in the "3.3 Release" thread, the original scheduled date for 3.3.0 was this coming weekend.

This weekend just gone, I pulled down a ton of outstanding pull requests from various people. It occurred to me that although most of them were minor tweaks and adjustments, the cumulative effect of merging them all into 3.3 would be to trigger a requirement for another significant period of testing (and quite possibly further bugfixes or tweaks).

So in order to stay on schedule, I didn't release a new RC this weekend. I am thinking that the existing RC has had only minor bugs reported, and that we should release it as 3.3 so that all the wonderful contributors can get on with making 3.4.

So what do people think? Of course the last RC isn't perfect - no release ever will be. RL has made all three currently active committers (me, d_m and myshkin) busy over the past couple of weeks, so time for hunting down and fixing minor or hard-to-reproduce bugs is limited. My instinct is to release on schedule, even though I know it's not perfect. The alternative I fear is a long delay and risk of stagnation. As changes are made for 3.4, I will backport any fixes for 3.3 bugs.

So, this week, please post *in this thread* if you think there are any bugs in the latest RC (03 Jul 2011 at 20:14 UTC, revision fd51c57) which ought to prevent its release. I am happy to postpone release by a couple of weeks if it will make a serious difference.

Thanks.

[Nomad]

Not a bug, but I just posted an updated version of my tile set that I don't think has been included yet.

(I also just designed a few new pit types, which wouldn't require any code changes to slot in, but I'm not sure they should go into a proper release without time for playtesting as there could be balance issues.)

[jens]

I can only think of one issue that comes close: delayed level feelings. Maybe not showstopper quality, but also very easy to test. And possibly squelching bad as default... Also easy to change without rigorous testing...

You'll still get your head spinning though, because I'd say that most of the tweaks in the pipe line are 3.3 tweaks that should go into 3.3.1.

[Magnate]

I can only think of one issue that comes close: delayed level feelings. Maybe not showstopper quality, but also very easy to test. And possibly squelching bad as default... Also easy to change without rigorous testing...

You'll still get your head spinning though, because I'd say that most of the tweaks in the pipe line are 3.3 tweaks that should go into 3.3.1.

No, they're not. 3.3.x is for bugfixes only, no tweaks.

I will revert the squelch bad default before release, that's on my list. Level feelings I'm not bothered about. If people don't like them there is an option not to have them. If you want them, you have them.

[jens]

No, they're not. 3.3.x is for bugfixes only, no tweaks.

If that is the case, then I'd prefer to have the whole lot in the pipe line added (except stealth). All of those are tweaks to the current game, not changes planned for 3.4.

[dos350]

plz no unfinished release, it can wait~ plz no raging

[fizzix]

plz no unfinished release, it can wait~ plz no raging

Well sure, right after you define what is a "finished release."

My definition is a game that plays well and has no significant bugs. I think the current 3.3 RC fits those conditions and should be released.

[CunningGabe]

Well sure, right after you define what is a "finished release."

My definition is a game that plays well and has no significant bugs. I think the current 3.3 RC fits those conditions and should be released.

I agree. Any tweaks that don't make it into 3.3 will soon be in the nightlies anyway -- it's not like people are going to have to wait until the release of 3.4 if they want to play with them.

[jens]

But the rest of the world, i.e. those not hanging around here at the forum will forever have a slightly inferior version of 3.3. For that matter we will also forever have a slightly inferior version of 3.3...

[CJNyfalt]

But the rest of the world, i.e. those not hanging around here at the forum will forever have a slightly inferior version of 3.3. For that matter we will also forever have a slightly inferior version of 3.3...

With that argument we never get 3.3 out, for there will always a some slight improvements that could be done forever.

[jens]

Well I did not say we should stay at it forever, just to use the stuff that has already been commited. Just so you know, I don't feel strongly either way, just felt that the issue had to be heard :-)

[PowerWyrm]

Just a little glitch discovered while implementing stuff for my variant. In monster2.c, line 3159:

my_strcpy(race_name, r_ptr->name, sizeof(buf));

should be:

my_strcpy(race_name, r_ptr->name, sizeof(race_name));

Not a major stopper, but would take a second to fix...

[Magnate]

Just a little glitch discovered while implementing stuff for my variant. In monster2.c, line 3159:

my_strcpy(race_name, r_ptr->name, sizeof(buf));

should be:

my_strcpy(race_name, r_ptr->name, sizeof(race_name));

Not a major stopper, but would take a second to fix...

Excellent catch, thank you - unfortunately CunningGabe has just nuked monster2.c with a big refactor, so I have to find out where this line is now! What function was it in?

[PowerWyrm]

Probably flush_monster_messages().

Oh and another one... assertion failed when using stairs while hallucinating (cave.c, line 837):

assert((int)g->m_idx < cave->mon_max);

This is due to g->m_idx set to 1 while hallucinating... and cave->mon_max equals 1. Fix is easy... don't assert if hallucinating, since the value of g->m_idx is irrelevant in that case.