Spring RTS Engine

Replay does not save state. Replay saves messages.

Can such replay-loads be possibly handled from within mod lua?

KingRaptor wrote:I did write a ZK gadget for the new save/load system a month or so back, but it's still very primitive (only handles units and their commands, no features, RulesParams or info specific to other gadgets; doesn't even remember which way units were facing).
Never really had time to work on it further, though I do now for a couple of months at least.

Are you interested in the result (having save/load in ZK), or the process (having fun coding in Lua)?

I have a vague notion Tobi said something along the lines of save being there after an overhaul, but nobody implemented the callbacks in the lua gadgets

lolwut

Save the heightmap, not the units.
As I said previous page.

zwzsg wrote:

KingRaptor wrote:I did write a ZK gadget for the new save/load system a month or so back, but it's still very primitive (only handles units and their commands, no features, RulesParams or info specific to other gadgets; doesn't even remember which way units were facing).
Never really had time to work on it further, though I do now for a couple of months at least.

Are you interested in the result (having save/load in ZK), or the process (having fun coding in Lua)?

Both, really. Save/load is kinda expected of any game that claims to have SP support, which in turn is on the list of things that ZK needs/wants to gain a larger playerbase.
That doesn't mean I can't have fun implementing it, though.
Why do you ask, though?

I think for missions it would be important to save everything, variables in gadgets and all that so that the mission functions properly.
Also won't native AIs lose their "memory" on save/load unless they save/load it themself?

there always was a save and a load call-in for AIs, plus they should be able to initialize mid-game anyway, which i consider the better way to do it, as it should be both: easier to code, and more versatile/useful, and most likely also more robust.

KingRaptor wrote:Save/load is kinda expected of any game that claims to have SP support, which in turn is on the list of things that ZK needs/wants to gain a larger playerbase.

Yes, but there are many ways of handling it. Many games only have functionality to save between the missions, or use a code that you have to enter instead of saving to a file.

Saving as a code would be kinda cool anyway. You rember those:

JHFJDS-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-XXU

Jools wrote:

KingRaptor wrote:Save/load is kinda expected of any game that claims to have SP support, which in turn is on the list of things that ZK needs/wants to gain a larger playerbase.

Yes, but there are many ways of handling it. Many games only have functionality to save between the missions, or use a code that you have to enter instead of saving to a file.

Saving as a code would be kinda cool anyway. You rember those:

JHFJDS-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-URRTR-JHFJDS-XXU

How about a quasi-save system where the game analyses your total metal invested and resources and so on, and you get a boost corresponding to that (obv scaled down a lot) if you have to quit mid-mission and restart it later. It would be easy to do and give people a bit of continuity.

KingRaptor wrote:Why do you ask, though?

To know if I should bother proposing my code, or if you prefer recoding all by yourself.

BaNa wrote:How about a quasi-save system where the game analyses your total metal invested and resources and so on, and you get a boost corresponding to that (obv scaled down a lot) if you have to quit mid-mission and restart it later. It would be easy to do and give people a bit of continuity.

BaNa, you're into BA right? How about you use my existing system instead of asking a poorer new one?

Anarchid wrote:Can such replay-loads be possibly handled from within mod lua?

If you mean the replay-load as descripted by Google_Frog on previous page, I guess that could be automated.
Maybe would need to save replay length in an extra file so the Lua knows how much to /skip ahead.

zwzsg wrote:How about you use my existing system

People do not want save&load, they just want to talk about it and dream about playing really epic missions and campaigns that never get made. As can be seen by threads like this going over multiple pages every time, while nobody comments if something actually working is posted.

Well i am glad people are talking about this vital subject.
This game is the most engaging strategy game ever made. There are so many layers to the units and way that they effect the battlefield. It makes for a realistic feel to all Big Strategic plans and Tactical fights. This is such a perfect game and if saving was possible would make it playable for single players.

Everything is so perfect and is just a shame it is to hard to make it work. Would play 10000 hours on this if there was good and different AIs and saving. Cause i like the huge battles that take a long time and that just cannot be done at one sitting.

Thanks for a perfect concept though.

We have what amounts to a save/load system that works for every game, it just doesn't have a nice UI.

To Save:
So is this a actual way to save and load a big epic campaign?? I have to try this game on a better computer. So would wish there was some way to go to the middle of a epic battle. Is this one by GOOGLE a real working option???




Close Spring
Go to spring\demos and rename the most recent replay


To Load:

Drag your replay file onto Spring.exe
Once ingame type '/cheat'
Type '/team 0', in some cases you may have to replace 0 with a different number
Type '/skip 10000' until you are at the point you left off
Type '/cheat'


For hour long games this may take a while, it is not a quicksave.

lol that isn't a save.

Hello,
I’m new to this forum, just registered to add my mustard on this topic (german idiom, to add one’s two cents ).

I have played NOTA a few moths before and came back to Spring a few days ago. The reason I quit the first time, is that the save system was not working.

For me, it’s a major annoyance, ruining singleplayer because I like lengthy battles over several hours (I think I have made some witch took over a dozend hours in Warzone2100, would have to check it to be sure)

I’m not into game development, but I claim to have at least a bit experience about programming collected in – eh … four years (?) – of development in C++.
Thus, the most important questions in this regard are for me:
Why does nobody bother to fix it – is development time really that scarce? (Slow releases suggest this)
And why is it that hard to fix at all?
Why does the engine not handle saving and loading?

I haven’t looked into the code by now, but I guess that there are several data structures where every important piece of information is saved.
I further guess that every entity knows about its information and what would have to be saved.
Also, I assume that the game is tick based (what else?) and has some means of telling its parts what to do (callbacks, messages, etc.)
So, why don’t you just implement saving so that the engine notifies every part of it, which in turn propagates this message down? Wouldn’t be that hard, I guess.
As far as I know, the relevant information has just to be serialized (there are librarys for it, boost::serialize should be fine from what I’ve overheard) and passed to the engine, which writes it to disk.
It would be perfectly fine, as far as I can tell, so I really don’t know what the problem is. (Information on that topic would also be appreciated)

Oh, and please, save projectiles, too. I guess it would suck if you launched 20 nuclear missiles and they were all gone at once.


In the meantime, or as an alternative, I think I will get a copy of the source-code and disable the watchdog, as it seems to prevent me from just suspending the game’s process in ProcessExplorer (better TaskManager for Windows) and resume it later on. (Derived from the messages in the chatlog).


With best regards (hope that doesn’t sound wrong),
Fast2

soo...
you basically guess it is all done already, but it is just not done yet, even though very easy and little work...
without having had a single look at the source.

guess what image you leave behind this way.

Fast2 wrote:I further guess that every entity knows about its information and what would have to be saved.

.. and how did you guess so? cause C++ has built in reflection i guess?

Of cause.

Maybe its that this comunity is angry. Angry because everybody has poured his whole work in, created awesome stuff, and still people keep showing up, who are more interested in a game of *historic* value, whos best parts (GUI, physix) have become part of every game thats available for spring.

So, instead of going retro, why not try something that is less out of TAouch?

Like Evo.
http://www.youtube.com/watch?v=k6CJDMm8l4w

Or gundam.
http://www.youtube.com/watch?v=3XhgJvFnrUQ

Or ZeroK.
http://www.youtube.com/watch?v=vkZaLLyhEgI

Or the Cursed. Or KernelPanic. Or or or or or.. ad infinitum.

Its not that we hate TA. Its just, that the best pieces are absorbed. Eaten,and now integral part of every Game that you can get here. So besides nostalgia, there is really no reason to go for the allready known. If it has to be TA-ish at any price, check for ZeroK.

Ah, that was faster than I thought it would be :)
Probably, I just chose the wrong words.

I tried to express that every – yeah … thing (class, for example) knows what must be saved by itself, while it can also propagate the order further down the line. Think of it as the mechanism of how an object is destroyed in C++, maybe this makes my intention clearer.

By the way, what can be done with RTTI could be called reflection, I guess, but I don’t suggest using it

hoijui wrote:you basically guess it is all done already, but it is just not done yet, even though very easy and little work...

Aren’t you contradicting yourself?
It’s not done, otherwise it would be working and I was playing some game instead of writing this post ( )
Additionally, I didn’t say it was little work. For sure, coding a save-function for every class* will take some time.
The single part of what you claim I said which is actually at least partly true, is that it should be easy work. I think once done for one part, it should be boilerplate for the others.

*(I’m not sure whether everything in the game is a class, that’s why I chose entity above, which doesn’t seem to be the right word for it)

With kind regards,
Fast2

Edit:
I didn’t look into the source code yet, because I don’t think I will have the time to contribute in a useful way at all in the next few weeks anyway.
I just wanted to suggest an approach to fix saving (and to show that there is at least one person more who would like to save his progress) and hear what you think about this proposal.