aliens

Author Topic: OpenXcom 1.1  (Read 26189 times)

Offline Meridian

  • Global Moderator
  • Commander
  • *****
  • Posts: 6996
  • OXCE developer
    • View Profile
    • My Wiki
Re: OpenXcom 1.1
« Reply #15 on: February 09, 2015, 09:17:15 pm »
At least we know about it. Better than not knowing ;-)

Offline SupSuper

  • Lazy Developer
  • Administrator
  • Commander
  • *****
  • Posts: 2163
    • View Profile
    • Email
Re: OpenXcom 1.1
« Reply #16 on: February 09, 2015, 09:40:41 pm »
Foolish decision.
SupSuper, your problem is care not for the game, but for the community.
Ok you pick a suitable milestone now that we're half-way into TFTD and that won't just get us tons of comments asking where's TFTD.

Offline volutar

  • Colonel
  • ****
  • Posts: 351
  • Vanilla digger & Quality assistant
    • ICQ Messenger - 695047
    • View Profile
Re: OpenXcom 1.1
« Reply #17 on: February 09, 2015, 09:56:23 pm »
and that won't just get us tons of comments asking where's TFTD.
Why should that bother me?
When there are two ways:
1. Ironing out code, fixing bugs, vanillifying, and THEN adding features.
2. Adding features, multiplying bugs geometrically, and THEN trying to fix them (in some unseen future).
I would pick first. And noone's comments and requests for features gonna change that.
There is a programmer's rule. And it shouldn't be broken. Not even if the Pope is asking.
« Last Edit: February 09, 2015, 10:01:46 pm by volutar »

Offline Yankes

  • Commander
  • *****
  • Posts: 2526
    • View Profile
Re: OpenXcom 1.1
« Reply #18 on: February 09, 2015, 10:25:50 pm »
Foolish decision.
SupSuper, your problem is care not for the game, but for the community.
Did you read different post that I? SupSuper did exactly what you ask him to do, don't stop and go speed ahead to TFTD support.
I would preferred different decision, but I will support it.

And second point, without community would be no OXC.


Offline Tarvis

  • Colonel
  • ****
  • Posts: 111
    • View Profile
Re: OpenXcom 1.1
« Reply #19 on: February 09, 2015, 11:52:30 pm »
The way I see it, adding TFTD support changes the internal structure enough that exclusively fixing bugs might end up being a waste of time since many areas end up getting rewritten anyways.

Besides, it's not like SupSuper is the only coder only working on TFTD. There is also bugfixing at the same time; they just merely don't warrant an official release.

The same justification applies to mod compatibility - the quest for TFTD also will alter the mod code, so going from a theoretical stable release between now and the TFTD release would just offset the compatibility complaints further down the line as more TFTD features get added to the nightlies. But once TFTD support is more or less finished, there isn't much more in the way of significant new features to add, so mod support will be less prone to breaking changes in future versions.
« Last Edit: February 09, 2015, 11:54:21 pm by Tarvis »

Offline Shoes

  • Commander
  • *****
  • Posts: 503
    • View Profile
Re: OpenXcom 1.1
« Reply #20 on: February 10, 2015, 01:52:48 am »
Did nobody notice this thread? https://openxcom.org/forum/index.php/topic,3287.0.html

Pick a commit that is stable, fork it, have the community crown it as 1.1. Cherry-pick some commits that fix relevant bugs if need be.

Offline volutar

  • Colonel
  • ****
  • Posts: 351
  • Vanilla digger & Quality assistant
    • ICQ Messenger - 695047
    • View Profile
Re: OpenXcom 1.1
« Reply #21 on: February 10, 2015, 04:47:51 am »
SupSuper did exactly what you ask him to do, don't stop and go speed ahead to TFTD support.
Why do you think I asked that? I wasn't ASKing. I was accounting on making wise decisions. And ramming forward towards features by the cost of accumulating bugs is not wise. Or you can end up with the version more buggy than vanilla TFTD.

The way I see it, adding TFTD support changes the internal structure enough that exclusively fixing bugs might end up being a waste of time since many areas end up getting rewritten anyways.
TFTD support doesn't touch battlescape code yet. And yet, battlescape has lots of bugs to fix and things to polish, and will affect xcom1 and any other played mod. But instead - new bug-o-features gonna be added to geoscape, which will postpone battlescape fixing even more.
« Last Edit: February 10, 2015, 05:09:05 am by volutar »

Offline robin

  • Commander
  • *****
  • Posts: 1142
  • ULTIMATE ROOKIE
    • View Profile
    • Email
Re: OpenXcom 1.1
« Reply #22 on: February 10, 2015, 12:30:37 pm »
We've given up on 1.1 because:

- Everyone is waiting on TFTD. Everyone. Nobody cares about little fixes and improvements, we've gotten much more requests about TFTD than issues in the bug tracker in the last year. Everyone wants the shinies.

- Modders keep attaching themselves to the latest nightly, which triggers more fixes and requests, which makes the "stable" constantly shift forward. Everyone wants the shinies.

So for us, trying to reach stability is pointless. You've forced our hand, it's full speed ahead until TFTD (2.0). There will probably be a pre-2.0 to iron everything out but that's it. If you wanna pick a "nightly" to stick to, feel free.
I'm a hungry dog circling the table, ready to chew any piece of food that somehow drops from it.
1. I think things.
2. I implement what I can.
3. I check the Nightlies section and see that the latest one adds "cool feature x" and "amazing stuff y".
4. I use it right away so I can implement more/ better the things I thought initially.
5. Rinse and repeat.

Also asking is a collateral activity of what I'm doing; doesn't mean though that you have to do what I ask or rush to it.

I feel I'm in no position to declare a potential "stable", I  don't even know well the development roadmap to tell when is a good time to stop the train, I don't even know the TFTD ETA to tell if it makes sense to have "pit-stop" or rush.

I think the key point if you decide to have 1.1, is not to release "shinies" for a bit after it, so modders won't immediately have new shiny things to flock to (they will; I will).

(Modders at this point are sticking to the latest nightly also because they are ephemeral, the nightly you choose to stick to could disappear from download page relatively soon).

Offline hellrazor

  • Commander
  • *****
  • Posts: 1966
  • Deep Ruleset Digger & Bughunter
    • View Profile
    • Github Account
Re: OpenXcom 1.1
« Reply #23 on: February 10, 2015, 02:54:27 pm »
When there are two ways:
1. Ironing out code, fixing bugs, vanillifying, and THEN adding features.
2. Adding features, multiplying bugs geometrically, and THEN trying to fix them (in some unseen future).
I would pick first. And noone's comments and requests for features gonna change that.
There is a programmer's rule. And it shouldn't be broken. Not even if the Pope is asking.

I can only agree here with volutar. First the iron, then the bugs -> stable well running code.
Then add new features on a solid an properly working codebase, iron them, fix their bugs, then next feature.

Build code in steps testing each segment to the heart, and when you know it is working perfectly then the next.
In the end your overall codebase, will be more stable.
I rather tend to fix bugs right away then let them queue up in the long run, because even if you have written the code yourself, not looking at specific sections (because they are considered working) will only bring you more pain, more time consumption on fixing bugs etc..

anyway my 2 cents.

Offline Arthanor

  • Commander
  • *****
  • Posts: 2490
  • XCom Armoury Quartermaster
    • View Profile
Re: OpenXcom 1.1
« Reply #24 on: February 10, 2015, 07:10:46 pm »
I actually quit modding when the map rework was introduced. I was busy, and it caused problems of many kinds, which I did not know how to address or even identify.

When I work on project, I tend to do it in an incremental way: Introduce Feature, Test, Debug, Fix -> Release new version -> Restart.

There have been a few features that would have warranted new releases since I have been interested in OpenXCom, the main ones being: Two weapon tanks and the new map stuff. Tanks were really wanted by users and Maps made big changes. Shortly after their introduction, when sure that things work, releasing v1.1 (Tanks) and v1.2 (Maps) would have been great.

Mods could then say: "Works with v1.*" and everybody would know what to use them for. My mod collection works with "v1.1" or rather what I defined as "v1.1" myself: A certain build that had 2 weapons HWPs but not the crazy map stuff.

For a certain time, no mod worked with "v1.2" if it had anything to do remotely with maps (including crafts). Mod authors would need to do nothing, since their mod would states "Works with v1.1", it is implied that it does not with v1.2. Much better than a slew of bug reports from people trying to use the mod with a nightly build and being entirely unaware that something changed overnight. Having the nightly gives some people the false impression that they have something stable, it's the nightly! After a while people would have gotten around to updating their mod and could apply the "Works with v1.2" seal or, if they are really daring, "Works in nightly" which I personally would never use.

The other thing is: I have no idea where to get the info about what's new in a nightly and how to cope with it. Sometimes the XCom1Ruleset.rul changes and you can guess at what's needed, other times you can't. Is there a resource somewhere with informative content as to what was committed for those who can't read the OpenXCom code?

Offline Tarvis

  • Colonel
  • ****
  • Posts: 111
    • View Profile
Re: OpenXcom 1.1
« Reply #25 on: February 12, 2015, 03:37:18 am »
Having the nightly gives some people the false impression that they have something stable
I don't really understand this. It says on the download page for the nightlies that they are prone to having the latest bugs. Plus this is true of any other programming project out there.

If anything, I think it's mostly the fault of modders. I understand the need to use the biggest and best features, but it's the modders and not the devs in this case that push people to keep up with the nightlies rather than for testing purposes. In most other modding communities, people generally are reluctant to officially release mods using features unavailable in latest stable versions. If they do, they also make clear what build the mod is intended to work with. This prevents all the confusion you mention.

The other thing is: I have no idea where to get the info about what's new in a nightly and how to cope with it. Sometimes the XCom1Ruleset.rul changes and you can guess at what's needed, other times you can't. Is there a resource somewhere with informative content as to what was committed for those who can't read the OpenXCom code?
Yeah, the changelog on the nightly download page describes changes, and so does the commit tracker on Github

I'm still not seeing the "adding features before fixing bugs" tendency that some of you mention the devs have. Look at the commits yourself. Almost every other commit is a bugfix. Much of the time the reason the bugs don't get fixed until later is that it is a while before they are even found, and usually are in fact uncovered by the addition of new features. Looking over code relentlessly does not yield anywhere near the amount of discovered bugs that users and testers are able to find.
« Last Edit: February 12, 2015, 03:46:51 am by Tarvis »

Offline volutar

  • Colonel
  • ****
  • Posts: 351
  • Vanilla digger & Quality assistant
    • ICQ Messenger - 695047
    • View Profile
Re: OpenXcom 1.1
« Reply #26 on: February 12, 2015, 05:08:19 am »
I'm still not seeing the "adding features before fixing bugs" tendency that some of you mention the devs have. Look at the commits yourself. Almost every other commit is a bugfix.
And it should be ratio of like dozen(fixes) to one(feature).
Most of recent gameplay fixes was mine and dozen should be made. But surprise - the bugfixing has been moved away from the path towards buggy tftd. I can't understand this. And can't figure fair excuse.
« Last Edit: February 12, 2015, 05:09:55 am by volutar »

Offline Arthanor

  • Commander
  • *****
  • Posts: 2490
  • XCom Armoury Quartermaster
    • View Profile
Re: OpenXcom 1.1
« Reply #27 on: February 12, 2015, 04:52:25 pm »
I don't really understand this. It says on the download page for the nightlies that they are prone to having the latest bugs. Plus this is true of any other programming project out there.
Well, this comes from reading posts in mods (or troubleshooting) where users say they encountered an issue and when asked what they use just reply "the nightly" without any associated date or other way of finding out what they are actually using. IT then often turns out  that they are using either an old nightly that doesn't work for the mod, or one that is too recent for which the mod is not ready..

Quote
If anything, I think it's mostly the fault of modders. I understand the need to use the biggest and best features, but it's the modders and not the devs in this case that push people to keep up with the nightlies rather than for testing purposes. In most other modding communities, people generally are reluctant to officially release mods using features unavailable in latest stable versions. If they do, they also make clear what build the mod is intended to work with. This prevents all the confusion you mention.
The problem is that in most other games, you don't go straight from 1.0 to 2.0, you get 1.1, 1.2, etc. as well. In OpenXCom, that did not happen. It very much looks like we will go from EU to TftD with nothing in between even though the game itself has changed a lot from 1.0.

If you want a new feature, you have to use the nightly. And obviously, people want the new features and modders want people to want to use their mods. We could wait for the next 1.* to release a new version of a mod, but not for 2.0. So far, it seems rather difficult to "pick" a nightly because, since there is no shared decision on which to pick, you would get a series of mods that all use different versions of the game. That's totally useless, so modders either stick with 1.0 (ie the latest stable version) or chase the nightly but inevitably fall behind when life and a big update happen simultaneously. And there's some people who think we should all use the nightly to help find bugs and support the project on top of that...

Quote
Yeah, the changelog on the nightly download page describes changes, and so does the commit tracker on Github

I'm still not seeing the "adding features before fixing bugs" tendency that some of you mention the devs have. Look at the commits yourself. Almost every other commit is a bugfix. Much of the time the reason the bugs don't get fixed until later is that it is a while before they are even found, and usually are in fact uncovered by the addition of new features. Looking over code relentlessly does not yield anywhere near the amount of discovered bugs that users and testers are able to find.

I am well aware that looking at code doesn't do you nearly as much good as trying to run it in terms of finding bugs, but it is not something that we can all do. I can code, I spend most of my work doing that, but I can't code in C++, neither can I really make sense of much of what changed in the commits.

What I was asking is whether there was a source of information on how to cope with the changes/what they did/what their impact is for modders who can read yaml and are familiar with OpenXCom but not its code. This recent commit is great: it tells modders how to cope with it. Now take look at most other commits that may impact on gameplay/mods and it says rather little. The big map script commit being an obvious offender (which has thankfully been taken care of since, I think).

Offline volutar

  • Colonel
  • ****
  • Posts: 351
  • Vanilla digger & Quality assistant
    • ICQ Messenger - 695047
    • View Profile
Re: OpenXcom 1.1
« Reply #28 on: February 12, 2015, 06:52:56 pm »
If you want a new feature, you have to use the nightly. And obviously, people want the new features and modders want people to want to use their mods. We could wait for the next 1.* to release a new version of a mod, but not for 2.0. So far, it seems rather difficult to "pick" a nightly because, since there is no shared decision on which to pick, you would get a series of mods that all use different versions of the game.
That is why I got really upset about. There was a definite way to 1.* version (1.1 or 1.2 or 1.5), and suddenly that changed. "WTF!" was the first thought seeing that.

There was a plan of getting that intermediate versions, at least one, with dozens of gameplay fixes and changes being made (not only crashes), and know what? - they gonna never be released until TFTD! "F that", I would say! It's not they way how things should be done! That really pissed me off, and STILL doesn't let go.

There should be really "stable" release with all noticed issues being fixed. You know, there was no stable releases. At all. 1.0 was full of hilarious bugs.
For the God sakes, let LPers and modders rely on some milestone stable release for the next 1/2 years until 2.0 come...

Sorry.
« Last Edit: February 12, 2015, 07:38:14 pm by volutar »

Offline Arthanor

  • Commander
  • *****
  • Posts: 2490
  • XCom Armoury Quartermaster
    • View Profile
Re: OpenXcom 1.1
« Reply #29 on: February 12, 2015, 07:38:01 pm »
Don't apologize volutar, you're standing up for the community ;) I, for one, fully agree!

I currently use a nightly from the fall (after 2 weapon tanks, before map scripts) that seems to do things pretty well for what I want. So should I label my mod as "Works in nightly 2014-10-22"?

You can't download anything older than the 31st of January (but for good ol' 1.0 of course) so there's absolutely no point on doing that. I might as well be the sole user of my mod and not publish it instead of requiring something 99% of users won't be able to find.

And would users go through the trouble of getting an old nightly instead of grabbing the most recent one and benefiting from its most recent developments (along with the mods that can manage to follow)? It is harder to get an old nightly than the most recent one, after all...

If there were a 1.*, it would be easy to grab for users (the milestone is much easier to install than any nightly) and it would change much less frequently (so it would be easier to keep on top of things for modders).

Plus, encouraging the users to run the latests milestone would mean that developers get relatively reports of recent bug (the old ones having been purged from the milestone), instead of things that happen in 1.0. So it helps with the development as well, whereas 1.0 users don't really.

Seems like a Win/Win/Win scenario to me...

Did nobody notice this thread? https://openxcom.org/forum/index.php/topic,3287.0.html

Pick a commit that is stable, fork it, have the community crown it as 1.1. Cherry-pick some commits that fix relevant bugs if need be.

I just replied there (just got back to OpenXCom from a few months break). It indeed doesn't really have to be picked by the devs, but it should be released on the main website as 1.1 as well. There's no point in modders picking a nightly to use as 1.1 if regular users coming in still install 1.0 and the only nightly available for Windows only go back 2 weeks (and I have no idea how to pick for Mac or Linux..).

The next milestone, regardless of who picks it (devs or modders) has to be widely distributed so that users will use it instead of other version. The whole point of publishing mods is for people to use them. If only modders use the modders' milestone, then it's pointless.