Why is noone on 1.12?

[Gwledig]

Just wondering but why are all Multiplayer games on 1.10, noone seems to be using 1.12 even yet?

[Heindal]

Imho i believe that many players stay with the older more stable version. I think especially new players will go with new version and old players will stay with their version. The only thing that could change that would be an automated update.

I'm developing campaigns, thats why I have several versions. As a developer I've just got to say, that it takes work to shift complicated projects into a new version. I can't really say how much time I put in testing in 1.10, but it was a lot. I can't really say if 100% of all features of my campaigns work in 1.12 and I think I'm not the only developer with that problem.

[tekelili]

Heindal, I can not be more agree with you. You have just resumed my main complaints about developing process of new BfW version:
- Lack of documentation for UMC authors about changes that can break code. Current developers expectations about how we should deal with this issue looks to be "test your add on heavily, if you spot different behavior then try figure out what changed and fix it"
... so no surprise 1.12 add on server miss many players favorite add ons

- Very tough filters for bug report. I am not going to list all inconveniences I had to report bugs, but just to name the most discouraging: Be even listened. An usual answer to bug report was "no, I dont belive that happens"
... so no surprise 1.12 is full of bugs and players prefer stand with a more stable version.

[Anonymissimus]

1.12 is just out. Way too early for switching multiplayer. Let others do the bug report work. For instance, 1.12.1 will apparently feature the "classic" UI theme, which means the battle screen it will almost look the same as in 1.10 (positioning of unit info and such) - if activated - it won't be the default it seems.

Lack of documentation for UMC authors about changes that can break code. Current developers expectations about how we should deal with this issue looks to be "test your add on heavily, if you spot different behavior then try figure out what changed and fix it"

As soon as the 1.11.x branch, leading to 1.12.x, had been created, I created this thread: http://forums.wesnoth.org/viewtopic.php?f=21&t=36057
Whenever I intentionally changed the behavior of some wml, I documented it there, to have a list that covers the changes as effectively as possible. Not only did nobody yet create such thread for 1.13.x yet, although there probably are a a few changes, there are probably also things missing in the list for 1.12 since I didn't closely follow the commit log and document it. *sigh* (That would have been the task of the developers making the changes anyway.)

[tekelili]

As soon as the 1.11.x branch, leading to 1.12.x, had been created, I created this thread: http://forums.wesnoth.org/viewtopic.php?f=21&t=36057

That is very usefull, and when not all issues I had, are listed there, certainly would have helped me a lot... if I knew it exists. Thanks a lot for create such a thread, I would suggest it should be sticked on top of list of WML workshop forum, in order to be fully usefull.

Edit: Ups, I have just noticided it is placed in the very top "pre-list", my apologizes
(may be add a reference to 1.12 in thread title?, I think I passed it for that reason)

[iceiceice]

Let others do the bug report work.

Please report bugs, don't assume that they have been reported or that others will do it...

Not only did nobody yet create such thread for 1.13.x yet, although there probably are a a few changes, there are probably also things missing in the list for 1.12 since I didn't closely follow the commit log and document it. *sigh* (That would have been the task of the developers making the changes anyway.)

AFAIK this practice is not documented anywhere in the DeveloperResources or anywhere similar... IMO there should not be "secret changelogs". If you want to have an official "UMC changelog" then maybe there should be an official umc changelog file, or we could perhaps repurpose the players changelog for this... From what I understood, all compatibility-breaking changes should go into the release notes anyways. Maybe this list could be updated with each release based on the release notes, as part of the release process.

Also FWIW, I did not intentionally make any compatibility breaking changes to WML api in 1.13 (other than what was announced for 1.12), nor did I notice that anyone else did this, and I have been following quite closely.

Lack of documentation for UMC authors about changes that can break code. Current developers expectations about how we should deal with this issue looks to be "test your add on heavily, if you spot different behavior then try figure out what changed and fix it"

Actually my suggestion was here: http://forums.wesnoth.org/viewtopic.php?f=8&t=40449

The simple fact is, there are a ton of features that can be combined in many different ways... the overall code-style of the WML API somewhat eschews encapsulation in favor of ease of implementation... it's just very difficult to be absolutely sure whether or not you might have broken something, or to test all of the possibilities, as for the most part testing must be manual. There are only so many developers, and they are only capable of so much testing... we need users to report bugs to us, preferably during beta.

The simplest way to reduce the burden on everyone is to make automatic tests. If everyone chipped in a few tests it would go a long way.

[Gwledig]

My stuff (Conq+/Space, Chaoz, Wescraft) was dev. in 1.11 or whatever that was called and I found I just needed to change a few image paths for 1.12. Given 1.11/1.12 was in development for at least a year, I think there was pretty good time for testing MP mods, however, my mods are not campaigns they are MP scenarios which can perhaps be easier to quality control, you are dealign with a bunch of core rules and cfgs/modules instead of reams of scenario stuff. I agree it's a pain when Wesnoth gets updated, but there seem to be less problems than a few years ago, I also think it's worse when you are working with someone else's WML like when I took on a version of the old Conquest and made Conquest +, this is one of the reasons I re-wrote just about all of the MWL for "my" version, I imagine Slowthinker did this for the same reason in Conquest Minus, although of course he improved the efficency of the mod a lot and added more features.

[tekelili]

Given 1.11/1.12 was in development for at least a year, I think there was pretty good time for testing MP mods

That is true... in theory. In practice I started developing/maintaining World Conquest a year ago, and had to do all stuff for 1.10 if I wanted to do any test (not my bad start so "late", if I didnt do it, campaign was probably "abandoned"). Once I had some reasonable Beta for 1.10 I run to develope in 1.12 to not being caught by surprise with conversion problem... and what I found were some severe bugs that broken scenario (like a total crash just for recruit a unit). I had to do intensive "cut code in pieces and test" in order to detect bug and report it in a way it was admited, and then wait for the fix. During that time, that basically lasted until 1.12 realese, I could not even dream with get testers for the add on (who the hell wants to play something tha becomes broken at first recruit?). Without testers, testing multiplayer safety was too big task. With no debug commands, I was basically forced to play campaign alone controlling several sides from diferent clients, spending hours in games I was really no interested to play (I better spent my time enjoying testing in 1.10 with other players help).

As I often complaint, I want point I do valorate what developers do for BfW, but I think is important comunicate these problems from someone that also wants contribute to improve it

[Anonymissimus]

Please report bugs, don't assume that they have been reported or that others will do it...

That's the developer speaking, not the player. The sad truth is that vast majority of "testers" will not even bother with a new version until the new stable series is out. Especially multiplayer people. "Preferably during beta phase" is wishful thinking.

From what I understood, all compatibility-breaking changes should go into the release notes anyways.

Changes in the wml or lua API ? I never added them there nor did I hear they should be there.

The various existing changelogs are all useless for a UMC author for various reasons. There had been such threads in the past as well, and I found them very useful for updating addons over several major versions so I kept the practice. The point is to have a compact list showing exactly only everything what you need to search for in your addon when updating, without testing everything. I'm kinda sure that there's more use for it than for the player's changelog. And even the full changelog is often forgotten by devs.

[iceiceice]

That's the developer speaking, not the player. The sad truth is that vast majority of "testers" will not even bother with a new version until the new stable series is out. Especially multiplayer people. "Preferably during beta phase" is wishful thinking.

Okay, but it's also the only game plan that we have... if no users reported bugs we wouldn't know about or be able to fix most of them. This is not different from many other open source projects, ultimately to have a quality project we need to actually have users (check) and they need to report bugs (check?). Fortunately some (many) of them actually do, although certainly not most but that's fine.

Changes in the wml or lua API ? I never added them there nor did I hear they should be there.

It was our practice in 1.11.13, and 1.11.15 release notes to announce all compatibility breaking changes that are new in the release, in a colored section. It seemed to be consensus on irc that a compatibility-breaking change to WML api is always RELEASE_NOTES-worthy.

The various existing changelogs are all useless for a UMC author for various reasons. There had been such threads in the past as well, and I found them very useful for updating addons over several major versions so I kept the practice. The point is to have a compact list showing exactly only everything what you need to search for in your addon when updating, without testing everything. I'm kinda sure that there's more use for it than for the player's changelog. And even the full changelog is often forgotten by devs.

Yeah but we should coordinate stuff like this, since it's important... I think zookeeper said he also had made such a list for 1.12 also. There's no point in each dev making a separate list, there should just be one and it should be described on the wiki when it should be updated.

[Anonymissimus]

It was our practice in 1.11.13, and 1.11.15 release notes to announce all compatibility breaking changes that are new in the release, in a colored section. It seemed to be consensus on irc that a compatibility-breaking change to WML api is always RELEASE_NOTES-worthy.

So all wml/lua API compatibility-breaking changes ever done should have gone into the respective release notes, then ? Or was this just during beta or release candicate phase ?
Anyway, going through all of the release notes of the previous development cycle while updating some addon (I did it somewhere during beta phase) is impractical.

Yeah but we should coordinate stuff like this, since it's important... I think zookeeper said he also had made such a list for 1.12 also. There's no point in each dev making a separate list, there should just be one and it should be described on the wiki when it should be updated.

Zookeeper edited the list I linked above, so I hope he has merged his list into it and hope it's complete, thanks to his usual quality-work.

[Gwledig]

I just thought I'd see if there was still people on 1.10 and I'm looking right now at boat loads of people playing on that server/client. This really is not good, I don't know if there are other old versions of the servers running but this is surely fragmenting the player base, for one thing its harder to get a MP game up. Surely the old server needs turning off at some point? The 1.10 client isn't even easily available from the downloads page any more, what is going on? Are us mod developers really expected to develop stuff for multiple instances of the game?

[Ravana]

Surely the old server needs turning off at some point?

Based on what happened with 1.8 server, I believe this wont happen anytime soon.

Are us mod developers really expected to develop stuff for multiple instances of the game?

That is your personal preference. I stopped dealing with 1.10 immediately after 1.12 was announced. While 1.12 introduces new bugs in user interface, [modification] support is vital for me. And every single addon I maintain, I do so for myself just as well.

[Andrettin]

Imho i believe that many players stay with the older more stable version. I think especially new players will go with new version and old players will stay with their version. The only thing that could change that would be an automated update.

This, pretty much. It is a bit cumbersome to have to redownload the entire game to update to the latest version, hence why players will tend to put it off until later. It would be great seeing Wesnoth on Steam and making use of automatic updates, exactly for that reason.

[Gwledig]

Imho i believe that many players stay with the older more stable version. I think especially new players will go with new version and old players will stay with their version. The only thing that could change that would be an automated update.

Not really, the admins can put up an announcement in the MP lobby like they did the other day on 1.12.1 telling people they need to upgrade (there was a security vulnerability so 1.12.2 was released. They could also disallow MP games or just turn off the server. I think that older servers should go offline after a year tops.

Some points

* Developers don't have the time to update their WML for multiple versions, a lot of time can be needed to iron out problems caused by new Wesnoth versions and they can't really be expected to update multiple versions to handle updates, fixes and new content

* Fragmentation of the MP base, if loads of people are sticking with old server/client then this will damage the viability of the game overall, because no one can get a MP game up when half the players are on another server

* There's less incentive for developers to update to the new server/client

* Ultimately compatibility problems with devices and OSs where the new Wesnoth server/client is compatible

and that's just a few