[taylor]

A cross-platform ini file is going to replace the registry, a new in-game configuration screen is going in to replace most general config options, and a new cross-platform launcher, which is only needed if you want advanced options, is being made.  Things like resolution & color depth, D3D/OGL modes, and the most used visual/game options (like glow/spec/env mapping, ambient factor, etc.) will be settable in-game and not require the launcher to set.  

The launcher will not be required, or even useful, unless you want a GUI method of setting the more advanced options.  There will also be mod specific ini files which can set options specific to them automatically (such as -tbpwarpeffects for TBP) and those ini files can be distributed in the mods VPs.

[karajorma]

Originally posted by taylor
A cross-platform ini file is going to replace the registry.

The launcher will not be required, or even useful, unless you want a GUI method of setting the more advanced options.  There will also be mod specific ini files which can set options specific to them automatically (such as -tbpwarpeffects for TBP) and those ini files can be distributed in the mods VPs.

I believe RTs plan was to allow the mod.ini file to carry each mods preferences for this. So that for instance if you run Wing Commander and FS2 from the same folder you can turn on the WCP or TBP options without having to do it from the launcher every time you swap between them.

[Tolwyn]

Originally posted by Galemp
Isn't that what RT's checkboxes are for in the Launcher? I wish that was editable, so we can add new tickboxes for command line params.

I wish that we could remove a few boxes as well, as several tabs change the gameplay of total conversions in an undesired way

[taylor]

Originally posted by karajorma
I believe RTs plan was to allow the mod.ini file to carry each mods preferences for this. So that for instance if you run Wing Commander and FS2 from the same folder you can turn on the WCP or TBP options without having to do it from the launcher every time you swap between them.

The current notion of mod.ini files won't work very well cross-platform (not without wierd hacking, which isn't going to happen) and will NOT be supported in the new code.  Something similiar, yes, but I don't give a crap about mod.ini.  And before anyone says anything, this plan of moving to cross-platform ini files and a new launcher has been in the works for over a year.  It's nothing new and I even made a note of it when RT first brought up the idea of mod.ini in the forums.

Originally posted by Tolwyn
I wish that we could remove a few boxes as well, as several tabs change the gameplay of total conversions in an undesired way

About 95% of the cmdline options will be removed in the new code and replaced by ini settings.  Mods will have to ability to NOFLAG particular options so that they can either be turned off or the user setting overwritten by the mods preference to avoid problems like this.  Not every option will be available to do this (such as -ambient_factor which can be a system dependant preference) but most will.  It will also open the door to per mission option settings with it resetting to the startup config after the mission has ended.

[karajorma]

Originally posted by taylor
The current notion of mod.ini files won't work very well cross-platform (not without wierd hacking, which isn't going to happen) and will NOT be supported in the new code.  Something similiar, yes, but I don't give a crap about mod.ini.  And before anyone says anything, this plan of moving to cross-platform ini files and a new launcher has been in the works for over a year.  It's nothing new and I even made a note of it when RT first brought up the idea of mod.ini in the forums.

Got a link? I'm getting more and more interested in the idea of using the launcher to download and install mods. But I don't want to start programming and have all my code become obsolete halfway through.

[taylor]

Originally posted by karajorma
Got a link? I'm getting more and more interested in the idea of using the launcher to download and install mods. But I don't want to start programming and have all my code become obsolete halfway through.

No link to anything working.  The code is in a billion pieces at the moment and only a few parts are acutally working.  That's why it's taking so long to get this going, it depends on a ton of new code.  Everything from CFILE updates, complete ini/cmdline/option handling rewrites, completely new pilot file code, new configuration screens and a bunch or new artwork, UI code updates to handle the new and modified controls, graphics code updates and new io handling to revamped audio code.  Nothing small and nothing simple.  Every part depends at least some on another which is why all of it has to be developed at once.  I've been too busy with other projects and bug fixing for 3.6.7 to have worked on this stuff for nearly 5 months now.  There is no ETA for when anything gets into public hands since some of the code pieces haven't even gotten past the mental stage yet.  The delay in getting 3.6.7 out the door has pushed it back too far to guess at this point.

This is also why I hated seeing the mod.ini thing starting up.  Given how little RT seemed to be working on anything, and the fact that the Launcher CVS had been broken for months, I didn't think anyone would ever get to really use it.  For that reason, and quite a few others I might add, I was an extremely unhappy person when 5.3 got released.  None of this is to say that RT can't modify the current Launcher when the new code hits CVS but it does mean there will be two Launchers and until RT updates his Launcher it can't be used with the new code.  The mod.ini stuff will not be supported on anything but Windows and the new Launcher (from me) will not handle downloading and installing new mods, only setting advanced game options.  A seperate app would be needed for that which can properly handle it on all platforms from a central repository or something.

I hate the Launcher, it's tech from the early to mid '90s and we shouldn't be continuing to use a crutch introduced by FS1.  We should fix the source to handle things right and have the Launcher as this strange little accessory which most people will never need.

[karajorma]

I meant a link to the discussion with RT.

As for the launcher. How do you plan to handle mods without it? You'd need a start up screen in FS2_open before anything else like even the splash screen is used to make it work as far as I can see.

[taylor]

Originally posted by karajorma
I meant a link to the discussion with RT.

Ooops.  Here is one (in which I went pretty much ignored) but I think there was another thread someone too.  Plus this was discussed in some developer stuff last year a bit.

http://www.hard-light.net/forums/index.php/topic,30993.0.html

As for the launcher. How do you plan to handle mods without it? You'd need a start up screen in FS2_open before anything else like even the splash screen is used to make it work as far as I can see.

I'm not 100% sure, ideas are welcome.  I want it easy and simple for people to handle.  Having to use the launcher to change between mods seems kinda stupid to me.  It should be as simple just having an icon per mod for switching between them, maybe going as far as having a mod switcher which could hande changing/managing mods.  Probably roll that into the mod installer and have it mimic as much as the mod.ini stuff as possible since I do like the idea of it, just not how it was done.  Either that or having the pilot code handle some internal switching using the upcoming CFILE changes.

I don't know, I'm months away from having anything useful so there is time to figure out what method is best, or most wanted/needed.  With the amount of code changes I'm making though there will be plenty of possibilities.

[karajorma]

Originally posted by taylor
Ooops.  Here is one (in which I went pretty much ignored) but I think there was another thread someone too.  Plus this was discussed in some developer stuff last year a bit.

I'll have a poke about in the internal later.

Originally posted by taylor
I'm not 100% sure, ideas are welcome.  I want it easy and simple for people to handle.  Having to use the launcher to change between mods seems kinda stupid to me.  It should be as simple just having an icon per mod for switching between them, maybe going as far as having a mod switcher which could hande changing/managing mods.  Probably roll that into the mod installer and have it mimic as much as the mod.ini stuff as possible since I do like the idea of it, just not how it was done.

You object to mod.ini simply because it's hard to make multiplatform rather than on any fundemental basis then? I see.

What I don't get is the difference between the mod switcher/downloader you're talking about and a crossplatform version of the launcher. I can sort of see why you'd want to strip the configuration stuff out of the launcher as much as possible but I don't see a huge difference between the way the mod choosing functions work at the moment and the way they'd work in your app. Are we simply talking about changes that are needed for cross-platform compatability or something else?

Originally posted by taylor
Either that or having the pilot code handle some internal switching using the upcoming CFILE changes.

I'd love to see that but I didn't suggest it cause I didn't fancy getting my head bitten off

[taylor]

Originally posted by karajorma
You object to mod.ini simply because it's hard to make multiplatform rather than on any fundemental basis then? I see.

For a little more than just the multiplatform thing, but on it's basic level, yes.  The basic idea of what mod.ini is supposed to do is great and a big help to everyone.  The real problem for me is that it's abstracted away from the game itself and handled by a separate program, in this case the launcher.  This is something that should be handled by the game itself in order for us to simplify and standaradize configuration and get away from this notion of a launcher.  The launcher boils down to nothing more than a bad hack that's been continually reinforced by us.  It's time that was fixed.

What I don't get is the difference between the mod switcher/downloader you're talking about and a crossplatform version of the launcher. I can sort of see why you'd want to strip the configuration stuff out of the launcher as much as possible but I don't see a huge difference between the way the mod choosing functions work at the moment and the way they'd work in your app. Are we simply talking about changes that are needed for cross-platform compatability or something else?

Your still looking at it from the launcher point of view.  Think of it with the notion that nearly all configuration can be done in-game on a system/pilot/campaign basis and more directly tied into the current detail settings that you make in the game.  At that point the launcher has been seriously stripped of most of it's use and many people will no longer have any use for it.  Mod support is still relegated to the launcher which is an empty shell of what it was but still a hacked up substitute for what should have been.

Now look at an application whose sole purpose is to manage mods for the game from the users perspective.  You can download and install new mods, remove them, upgrade them, get basic mod information like website and forum links, readmes, etc.  The mod would come with all of the in-game configuration stuff using the new mod based configuration ini files which can set what options are on or off so this "mod app" wouldn't have to have any configuration stuff in it.  When you install a mod through this app it makes a shortcut/link to jump right to that mod when you want to play.  Dependancies can be handled by the mod app itself and if a mod requires another mod or data then it will say so and the mod app should take care of installing that at the same time.

The idea here is simple, make it easy for the user to take advantage of.  Mods are still just something that got glued on at the moment.  Some of that is getting fixed with the new pilot code which is being redone from scratch to be used for mods rather than having that support basically hacked in.  What's left is to make the mods easier to use so that we can get people away from just unziping into data/ and being done with it.  That only causes problems for the user down the road and makes SCP's job harder.

I'd love to see that but I didn't suggest it cause I didn't fancy getting my head bitten off

 I don't know how feasible it is from a "good thing to do" perspective but I think it's possible.  I'm so anal with the new pilot code that it's almost scary.  It's what needed to be done though and what I should have done in the first place.  The only question of whether this will be a good thing or not is how far I go with the CFILE changes.  As I get things done and start plugging it all together I'll be able to figure out what works best and possibly rewrite some things to go even further.  We'll see.

[karajorma]

Originally posted by taylor
Your still looking at it from the launcher point of view.  Think of it with the notion that nearly all configuration can be done in-game on a system/pilot/campaign basis and more directly tied into the current detail settings that you make in the game.  At that point the launcher has been seriously stripped of most of it's use and many people will no longer have any use for it.  Mod support is still relegated to the launcher which is an empty shell of what it was but still a hacked up substitute for what should have been.

How feasible is that though? I still don't know enough C++ to tell for myself but everything I've heard from the guys who can is that swapping CFILE stuff after the game has initiated them is a nightmare as it's used all over the place.

Originally posted by taylor
The idea here is simple, make it easy for the user to take advantage of.

Hey. You're preaching to the choir here. I spend an enormous amount of time troubleshooting problems users have installing mods. That's why I was fully behind RT's launcher plans. This does sound more interesting though. Even if it does deprive me of the launcher as a good place to start practicing C++

[taylor]

Originally posted by karajorma
How feasible is that though? I still don't know enough C++ to tell for myself but everything I've heard from the guys who can is that swapping CFILE stuff after the game has initiated them is a nightmare as it's used all over the place.

I assume that you quoted the wrong part of my message since it had nothing to do with your comment.

I commented in another thread about the CFILE changes actually.  I didn't think it was feasible at first either but that was before I remembered a horrible CFILE hack I did to get Linux ${HOME} working in the icculus.org version.  The hack was totally useless for what I was trying to do and didn't work right but it did allow some play with search orders, since that was it's basic design.  That is CFILE search order changing though, but the interesting thing is that we can add VPs after CFILE has been initted.  Simply adding them doesn't help at all but with the ability to change the search order as well, without a reinit, I think we can pull it off.

The key is that ultimately CFILE will always fall back on one function for finding a file.  This gives us a central location to intercede in it's operation and with some extra code we can change/dictate the search order of all cached file roots and do it without actually changing their init order (which was always the problem).

Hey. You're preaching to the choir here. I spend an enormous amount of time troubleshooting problems users have installing mods. That's why I was fully behind RT's launcher plans. This does sound more interesting though. Even if it does deprive me of the launcher as a good place to start practicing C++

You deffinitely do a ton of work in the support department for everyone and I came up with this basic idea after reading so many user complaints/issues.  I've still got a lot of work to do though so as things start coming together I'll be back to ask questions of everyone and try and hammer out the final functionality based on what can be agreed upon.

But with my time so limited, and with the **** load code needed, it could easily be 5 months before I have anything usable by the public.  I don't want you, or anyone else, wasting their time on something that will get replaced, but RT's launcher is here now.  You can still work on it in the meantime and at least use it to test what works and what doesn't with the mod.ini system.  If there is more info on what works well and what doesn't (or what would work better) then that just makes the new code come out more useful in the end.  If it helps people in the short-term, gives you more coding experience and provides better data on how the end user and modders will use such a thing then I don't think that's a waste.

[karajorma]

Odd. I did quote the wrong thing. I don't think my brain has been working too well this last week.

I may dabble with the code from RT's launcher but since I'm still playing about learning C++ I doubt I'll have anything much until you've finished the changes anyway