proposal: drop Mac support due to difficulties

[o11c]

Mac OS X currently has the oldest prebuilt binaries of any currently supported OS - older even than the BSD source ports, and nobody cares about BSD.

The reasons for this are varied, but it comes down to two core problems:

• Lack of TMW and Mana developers who actually use Mac OS X
• Intrinsic difficulties in packaging for that platform, that do not occur on any other platform.

Therefore, I propose that the TMW project official drop support (i.e. the prebuilt binary) for Mac OS X.

The Mana project will likely not drop support, so users *can* still build the client from source. We would not close the Mac OS X support forum.

Here's a more verbose reason for not supporting Mac OS X.

• It is not possible to build or test Mac OS X packages from any other platform. It is trivial to build and test Windows packages on Linux, or Linux packages on Windows.
• There is no distribution support. Every Linux distribution has a "games team" that really helps with the uptake. In addition, distributions often supply patches to make Mana a better-behaving client.
• Only 3 people seem to have any interest at all in working on a Mac package of the client:
  • trapdoor has produces packages in the past, but is notoriously inactive.
  • Frost only follows instructions provided by other people and reports that they don't work.
  • Socapex is the best chance for continued Mac support, but he still hasn't got a package built. Plus, he's new and we don't know how long he'll last - even if he gets a 0.6.1 build up, will he be around for 0.7 ?
• Apple makes it difficult to build packages that work across OS X versions. (Admittedly, I lack information about for what "minor versions" of Mac OS X 10.x people actually use, but this issue *has* been raised)
• undoubtedly many other reasons that escape my mind at the moment

[Big Crunch]

i would say that we continue supporting it, on the hopes that socapex pans out. If he doesnt end up sticking around, then nothing lost, but what happens if he does stick around and does make a binary or whatever is required? This game cant afford to refuse a growing population of computer users the opportunity to play.

BC

[Frost]

• Intrinsic difficulties in packaging for that platform, that do not occur on any other platform.

Packaging for the Mac requires knowledge of how to do so, and a Mac. Are those the "intrinsic difficulties" you mean, or are there others?

Therefore, I propose that the TMW project official drop support (i.e. the prebuilt binary) for Mac OS X.

The TMW project doesn't produce any clients, so what support do you mean? Stop hosting the package on the download page? Stop including Mac clients when testing changes on the server?
Would this basically put the Mac client on par with manaplus?

It is not possible to build or test Mac OS X packages from any other platform. It is trivial to build and test Windows packages on Linux, or Linux packages on Windows.

I'm curious. How would I build and test a Linux package on my Windows computer? I guess one would use WINE to build and test a Windows package on Linux?

There is no distribution support. Every Linux distribution has a "games team" that really helps with the uptake. In addition, distributions often supply patches to make Mana a better-behaving client.

This is a big deal. If Linux distributions provide help with porting packages, that reduces the workload on Mana developers. (Does it also reduce workload on TMW developers?)

Only 3 people seem to have any interest at all in working on a Mac package of the client

Ignoring the jabs, I agree that we don't have a lot of expertise in developing on the Mac.

Apple makes it difficult to build packages that work across OS X versions.

Apple does a better job of this than any consumer OS I've used except Windows. I think your perception comes from the previously noted thin expertise.

• The XCode IDE includes a drop-down box to select the target version and architecture (PPC, x86, x64). The same tool will build for IOS.
• dynamic libraries are contained in the app, and referenced by relative paths. Because of this, apps have minimal dependencies.
• MacOS supports a "universal binary" format where executables for both PPC and intel architecture are contained in the same file. PPC is so old that this might be of little concern now.
• It's quite easy to distribute multiple binaries, and easy for the user to identify which they should use. Mac software is typically distributed in the "DMG" virtual disk format, which can contain many files. If a particular version of MacOS cannot use an app, it displays the universal "No" symbol of a diagonal line across the application icon. Throw all the apps you wish on the same DMG and let the user choose which to run.

[Crush]

I'm curious. How would I build and test a Linux package on my Windows computer?

In a virtual machine, for example.

[Frost]

I'm curious. How would I build and test a Linux package on my Windows computer?

In a virtual machine, for example.

We could debate whether this qualifies as "trivial", but you're right.

However, to run Windows in a VM, you must purchase Windows.
To run MacOS in a VM, you must purchase OSX Server.

[o11c]

However, to run Windows in a VM, you must purchase Windows.
To run MacOS in a VM, you must purchase OSX Server.

WINE works for Windows emulation; Linux is open-source and thus freely-downloadable. While it's allegedly possible to run Mac OS binaries on Linux, this only works if you have all the system libraries (i.e. you've purchased the OS), and I don't know if graphical programs would work anyway.

[4144]

wine also can work on intel cpu based Macs and this allow run windows client version.

[Nard]

WINE works for Windows emulation; Linux is open-source and thus freely-downloadable. While it's allegedly possible to run Mac OS binaries on Linux, this only works if you have all the system libraries (i.e. you've purchased the OS), and I don't know if graphical programs would work anyway.

It is possible to make any X application, freebsd based, work on mac. And it is how the client works at the moment through macports or fink. Some users use manaplus with Wine, mac version.

Nard

[Frost]

I'd still like to hear what support would be "dropped" in this proposal.
I'm not aware that TMW provides direct support for any clients.

[o11c]

I'd still like to hear what support would be "dropped" in this proposal.
I'm not aware that TMW provides direct support for any clients.

We

1. offer the client package on our download page, and
2. make sure content is compatible with that version

[Socapex]

Hey guys been a while.

I WILL stay around, but school has started again and its getting pretty rough. I find it sad that you would drop support for an OS that is based on BSD just because you don't own it. Making a hackintosh is pretty easy. But from what I gather the project doesn't actually want OS X support... So why bother? I'll pop up once in a while and hopefully we can get it to build, but a package is out of the question right now. Especially when the project manager hates that platform so much.

o11c: Why don't you just remove everything about mac on this site. You DO NOT need ppls agreement...

edit: And BTW I'm waiting for 0.7 before putting my heart into it. I want to support both TMWathena and manasource.

[Jenalya]

But from what I gather the project doesn't actually want OS X support... So why bother? I'll pop up once in a while and hopefully we can get it to build, but a package is out of the question right now. Especially when the project manager hates that platform so much.

o11c: Why don't you just remove everything about mac on this site. You DO NOT need ppls agreement...

Hi Socapex, it's nice to see you still around.

There seems to be a missunderstanding here. o11c is neither the project manager of tmw nor does tmw have a single person taking important decisions. We have people being in charge of certain aspects of tmw, as e.g. o11c being Source Lead and me being Content Lead. That means we're the persons to talk to about bugs, feature requests or ideas regarding those aspects, and we decide things related to those aspects more or less independently.
But for bigger decisions we usually try to reach consent within the admins, GM's and senior developers (or at least nobody wholeheartedly disagreeing). We also try to consider player opinions as much as possible, as well as trying to be open to new contributors.

Regarding the question of supporting macs, I think supporting more platforms widens our possible playerbase, so dropping support should only be done if absolutely necessary.

[Crush]

Making a hackintosh is pretty easy.

It's also illegal in some jurisdictions. And it can get pretty expensive when the hardware you have laying around isn't MacOS-compatible.

Besides, it would be unnecessary when the Mac player community would stand on its own and provide patches to ensure compatibility.

[Frost]

But for bigger decisions we usually try to reach consent within the admins, GM's and senior developers (or at least nobody wholeheartedly disagreeing). We also try to consider player opinions as much as possible, as well as trying to be open to new contributors.

Regarding the question of supporting macs, I think supporting more platforms widens our possible playerbase, so dropping support should only be done if absolutely necessary.

Jenalya's point is essential. This project is focused on serving players.

Within the project, people are focused on their particular areas. For example, I like to joke that as programmer, o11c's job is to change as many things as possible; as admin, my job is to avoid all change.

When a decision has to be made at the project level, then the broad picture is considered. We want to improve the experience for players. Sometimes, improvements to the game require changes in the client. If it appears that a certain game client is no longer supported by its developers, then TMW must balance the benefits of certain incompatible changes against the cost of losing valued players. Such a decision is not taken lightly, and it's not taken in the dark of night.

As I understand it, this discussion about how to support Mac players is intended to expose concerns and (hopefully) find solutions.

[Ablu]

edit: And BTW I'm waiting for 0.7 before putting my heart into it. I want to support both TMWathena and manasource.

Hi.

0.7 is still a bit away i think. Though it should not be hard to adapt the stuff to 0.7.
In emergency you can always allow manaserv support via cmake option for 0.6.1 release as well.

Regards
Ablu