Trunk Code revision 8190: No rotating radar dishes

[Vermithrax]

In trunk code versions 8178 and 8190, radar dishes no longer exist. Thus they no longer rotate.

I'm sure someone will claim the buildings in question weren't complete, but they are.

[Zarel]

:[

What is with people committing these kinds of changes to trunk?

[Vermithrax]

Oh, you want real life examples? Like right before software release, a code freeze has been instituted, and some twit checks in something that crashes on startup, and if they'd bothered to do a get from the frozen codebase they'd have known about it? Or worse, they did a get, made the change and checked it in anyway?

How about a whole team of software engineers that refused to test their code on an actual system for eight f'ing years because they'd have to go into a cleanroom? That kept me in a state of constant irritation.

[cybersphinx]

What is with people committing these kinds of changes to trunk?

So those changes should have been put into a ticket for a week, be ignored by everyone, and then committed to break things? This way it's faster, at least.

[Per]

What is with people committing these kinds of changes to trunk?

Not everyone can be omniscient!

[Zarel]

So those changes should have been put into a ticket for a week, be ignored by everyone, and then committed to break things? This way it's faster, at least.

I dunno, but whatever change that can break this kind of thing sounds like it would be pretty major, so I expected at least some ML discussion.

And I actually discuss patch tickets if I have something to say about them. Especially now that they get e-mailed to me.

[cybersphinx]

I dunno, but whatever change that can break this kind of thing sounds like it would be pretty major, so I expected at least some ML discussion.

Turns out it was just a little "else" too much - fixed in r8195.

[i-NoD]

So those changes should have been put into a ticket for a week, be ignored by everyone, and then committed to break things? This way it's faster, at least.

A "faster" way to break things I guess? And somehow I assumed that nobody have checked consistency of texAnim.playbackRate and texAnim.nFrames in all .pie files before relying on shape's "weird" global versions. This most likely will result in less noticeable bugs.

[cybersphinx]

A "faster" way to break things I guess?

I meant a faster way to find out that it breaks things, and fix the problems.

And somehow I assumed that nobody have checked consistency of texAnim.playbackRate and texAnim.nFrames in all .pie files before relying on shape's "weird" global versions. This most likely will result in less noticeable bugs.

Per said no PIEs actually use that.

[-Kosh-]

What is with people committing these kinds of changes to trunk?

So those changes should have been put into a ticket for a week, be ignored by everyone, and then committed to break things? This way it's faster, at least.

What is the policy for patches? I noticed some of the devs post patches in trac and others do not. Does this deal with what the patch does?

I also agree with cybersphinx, most of the tickets I have read that are patches are ignored as he said and in the end it serves no real purpose.

I would also like to point out that having a test suite in this example would not have picked up this hiccup. It requires someone to actually sit down and notice the missing dish.

[i-NoD]

I also agree with cybersphinx, most of the tickets I have read that are patches are ignored as he said and in the end it serves no real purpose.

I understand that this whole conversation is pointless because the rules are often broken by their authors, still are there many tickets patches that were simply ignored? Or they were buggy or their author was not not 'Bruce The Almighty' and someone hinted him a better way of doing things?

[Zarel]

We're talking about patches by devs.

De facto policy is commit small patches directly. Larger patches go on the issue tracker. Sometimes, they get commented on, sometimes, they don't. Either way, with no major objections, they eventually get committed. I often add a "if no one comments, I'm committing in four days" or something. If you really want comments, you can go on IRC and ask for them. Otherwise, I've commented all of the patches I have something to say on, so as far as I know, the policy works as it should.

Also, i-NoD, be on IRC more! D:

[Per]

There is still the http://developer.wz2100.net/wiki/CommitGuidelines