file on the wrong spot

Corona

Linked to the spell Corona

Corona

NO! This has nothing to do with a certain virus!

Corona Art

Take II

Maybe this works better?

The code I used was very extremely DIRTY you know...

My next fight against Sheck-Lock will have to show if this all works the way it was intended

Linked to the action module

Characters replaced after being targeted will be purged out of the target chain when the action begin

Oh no, you don't!

Oh no, you don't!

#155

#155 - Take III

Referrence error

#155 - Take II

Forgotten End

#155

#155

After I had some rest I'm gonna look at #155

Well, let's see...

More chance for poison on Onyx members (except Jeracko who'll remain 100% immune).

Order of Onyx

Missing CANCELCOUNTER scenario in Dark Storage, which can cause Rondomo to crash the game

Let's see how things go... For now!

Merya skill crash on item conservation

At least I can continue to work things out outside of VisualStudio now!

I REFUSE to call the crash on savegame issue fixed.... At least I prevented it from happening again... I hope for a more elegant solution in the future...

#214

I gotta see if stuff works properly

Important to note is that this approach does NOT allow me to center or align to the right... Hopefully that's only a small price to pay

I'm doing a little experiment on real-time text... The way I had was very demanding on .NET's memory management, and may even be the reason behind the slowdowns I was experiencing, particularly in combat...

I think this will have to do, and I guess I'm reliant on Igor Pavlov to get this issue fixed... But since Igor Pavlov has like better things to do, I'll have to take it the way it is.... :-/

Still OUT OF MEMORY!

Kirana nerfed.... She killed my entire party in one blow, which I can never allow!

Take IV

Fixed some occurances

The many peaks in VisualStudio's RAM usage graphic made me fear that lzma was STILL being used somewhere, down there.... And using a JCR6 file viewer actually confirmed that

And Take III can start... at last

Data in sync once more

No it still isn't done, yet!

The Backup is simply handling tons of data now, so that's the reason behind the lag

I need to re-run all builders, synchronizers and stuff to make sure everything is in order now!

Everything went lagging, so I cannot fully trust the game Data I Have

I will finish the test of Kirana's section and Rondomo's section, hopefully in one go, through the VisualStudio debug mode, and I'll save as much as possible to make sure the "Out of Memory" error won't occur now...

Data Sync

Updated

Updated

I made a rehaul of the complete IAA folder.... A lot of junk couldbe living there over time, and deleting and regenerating was the only thing I could do to make sure all shit was gone there....

As a "reward" though the points you need to gain a level will also increase... haha!

More skill exp for using abilities

Hyper weakened you in stead of strengthening you

Recompiling NALA

It pains me, but it's clear I had no other choice. For the time being I've set the savegame feature not to pack at all.

Savegames that are still packed will still be readable, since they are simply JCR files, JCR6 will auto-detect any packing. and unpack accordingl. The lzma unpacker is not the problem... the packer is.... Like I said before lzma is notorious for being very RAM unfriendly when packing, but not when unpacking....

Well, so it seems it *WAS* the lzma compression routine after all that causes me to suffer.... The debug report VisualStudio gave out couldn't be more clear on that matter.... Now meddling with the existing lzma module seems like a bad idea to me, so I'm afraid the best I can do is... accept that I'll have to use non-compression for savegames...

This should take of those "Auto" crashes

A crash totally unrelated to the memory issues ruins everything.... AGAIN!

Take II

This fix did require me to restart the game, also destroying what I set up so far on things... Oh well....

Blockmap issue

It would leave me with two options... either I force the garbage collector to activate after every damn entry saved (I don't like that, but hey, if that is what it takes).... or I strip the lzma compression from the savegames (that would by the way also fasten the savegames)....

I'm wondering if the fact that the save game routine uses the lzma compression method to pack all data that lives in the savegame file is what causes it to collapse.... lzma is notorious for it's huge use of memory usage during packing (and needing almost none during unpacking)...

Let's sync this in visual studio and see what will happen if I play on there, as this is really getting on my nerves, and if this doesn't get fixed, the project will be a failure and after all this work... I don't wanna risk that....

It's still a complete mystery why "Out of Memory" exceptions are thrown after making multiple savegames, though....

No audio

Take IX

recompile to make sure

? Nala version conflict ?

Take VIII

At least all parse errors have been accounted for.... Which counts for something!

Inferno had only 1 flame.... Now that's 20 till 50

Take VII

And this is one of the reasons why I began the Neil project... In order to get rid of issues like THOSE!

Case error!

Let's go... again.... Take VI

Declaration error

Take V

.....

Take V

Return error

Take IV

Code Typo

Take III

Forgotten "end"

Take II

Syntax error

Take I

(whish me luck)

Well, during the test of the Dark Storage I will see if the Inferno spells gets animated the way it should....

Linked to the actions causing an inferno

Inferno

Take V

Not sure, but here goes...

Take IV

Not sure about the projectile stuff though

No money no message system works

Take III

I hope this debug line can shine a light on matters

Fix ignored!

Take II

If nothing is dropped, even no money, don't chaching and don't prompt it either!

I hope I fixed that!

Well first time this stuff was used, so let's be kind here.... But the calculation of the way the projectile should go is..... ES ACHE I TEA!

So here goes.... Take I, and let's see where this takes me :)

Good!

Okay, a quick snack first....

And that should link it all together #111

Adepted ability #111

I'M NOT THERE YET!!!

Done

Of course, now to script it all together.... Which should be easy, but let's see how stuff pops out, eh?

"Pew" sound for #111

Since "Out of Memory" errors still seem to occur, I've set NALA to force the Garbage Collector to react immediately when Audio objects are released. Most smaller audio objects are pre-loaded and never unloaded anyway, but the music is the different story, and the music appears the big evil here....

Don't come whining to me talking about bad code (amateur), as I am more concerned about Out of Memory errors than a bit of a lag due to pre-mature garbage collection.

Apparently I overesitmated the C# garbage collection system, as it only appears to take the number of allocations and releases in mind to activate the garbage collector, but not the ammount of data... So that's why I took this measure

Laser Projectile Art - for #111

#205

Tested floor 1-6 science facility

The full backup is still running, which was to be expected....

Quick Backup completed

All kids disappear accordingly

I'll have to test this, but my system lags, so I hope I can get it done!

The boys were more complicated, but I think I got them!

The girls should be taken care off

Working on #153

At this moment the system lags like crazy, but that's only natural when running 2 backups in once

It appears some harsher measures are required to deal with the non-disappearing kids

Well, let's test shall we?

Achievement "C'est la vie" linked to the proper event

But I've also done a "Total Redo" backup... The latter is bound to take more than one day, but will at least be fully up-to-date!

First in quick backup setting

And with the notice that the backup was outdated, let's immediately make a new one!

The Identify text was lost... So I had to recover that... that is, rewrite it from scratch

And I also see the scenario database inmy backup was outdated.... URGH!

A bug damaged the scenario database

Linked the two together

Trophy

C'est la vie!

IT WORKS!!!!

So let's do it!

I've tried to cover #88 , but I really cannot tell if it works, so I gotta test that

BlockMap issue in the south section of Rayal Palace, and Scharum's chamber

I've been bothered by #88 for a LOOOOONG time now, so I wanna take care of that today

Issued #213

I cannot recollect what happened in anyway... Of course, this can also be a general hardware communication failure in the underlying OS... I hope it was that, as then it's not an error on my side, but we'll have to see....

.NET error during saving.... Not a clue what the fuck happened there.... But I guess I gotta find out....

I think I found the "evil"....

NOW this is getting mysterious.....

Some hardcore debug as things are only getting stranger here!

.NET error?

It's not yet "elementairy", but I think I'm closing in, my dear Watson!

I've added some more debug crap to make sure about that....

This is gonna be a rough one.... I do see instances I don't want, and I'm sure the C# garbage collector refuses to release the music because of that....

I need to test this all out!

More debug....

I can indeed confirm that instances are not being released.... I do not know why

I've tried to set a few things up here.... This is pretty complicated stuff

For some reason the released music data lingers

Moar testing

BlockMap issue

I don't know why moves on 1 foe target the caster when learning... This should not happen, but for some odd reason it does... The source code does not deem this possible, though....

Fixed it!

Why this is shown... Not a clue, but let's find out!

Let's verify that!

THIS should fix the music stack error

A few extra debug lines might enlighten me on that course

Pushing and popping music doesn't work

Aziëlla is a bit bugged now, and the time has come to find out why....

Let's test Frundarmon's mansion....

#158 - Although I cannot yet test this!

What the logs show me now looks hopeful

Take V

All cool?

Take IV

Nil indexed error

Take III

And also something as trivial as a code typo

Take II

Of course I had to encounter something as trivial as a case error along the way

I need to pull all this to the test of course

Made a quick new music handler from scratch (well I only used the old code to make sure all functions are covered or the game could crash needlessly)... This one is by far less advanced, but should at least make sure that audio buffers are released (if not then they are hard-cover overwritten by C# itself, and at the mercy of the C# garbage collector, which is a lot safer than the one in Lua, I tell ya).

It appears I was correct... Not a single release call is being made, and that is.... BOTHERSOME at best...

And let's see what happens now....

Recompiling

AUDIOMEM debug command added

I've added an extra command which should help me on debugging

I've set up more text, as this is getting suspicious..... Hopefully this will enlighten me more....

The logs do make painfully clear that not a single people of audio is released when it should be done....

Debug bug (yeah, debugging *IS* harder than coding!)

Take III

And once again I must recompile NALA (sigh)

lack of self

And onto take II

Recompiling

NALA link up script error

I'll have to run a quick test to see what happens under the hood (I love my logs)

But first, I need to eat something... Due to my illness I could not eat so well, so I'm starving now!

If worst comes to worst I can write a simpler approach for the music routine... I do not prefer this, but hey, it's always better than tons of "Out of Memory" errors....

It should be noted that NIL relies on Lua when it comes to garbage collecting, and Lua's garbage collector is bugged on a few points, requiring a Lua coder to be extra alert on a few things, and when using a kind of shell language like NIL this can be more of an issue... Especially when the instances that have to be removed have to trigger the garbage collector in C#... Now the C# garbage collector does appear to work perfectly (as far as I can tell so far), but since Lua doesn't send the unlink command (which it should) because the records in question are not being unlinked, we got a problem... Of course, what I deem possible is because the data handled by Lua itself is pretty little and Lua may think the time is not right to release, as Lua cannot know how much data it has resposibility for.... I guess that's the flaw of using scripting languages with underlying APIs.....

I need to recompile NALA now