Anyway, the scroll icons come later, as I want them to appear only if scrolling is possibe, and with only the "Create New Slot" bar, not much need for scrolling

Fortunately I don't have to get out of the house today anymore

It's raining here....

Running

Up-to-date

For now it shows what I want to see

Sidebar added

That is news worthy of hitting all the front papers, you know!

I saw what I wanted to see

So I fixed that!

ScrollY did indeed have to be static... Silly me

Yeah, what, did you listen to Eric Idle's advice? Always look on the bright side of life?

I did get an error, but it's no longer a .NET error, so progress was still booked after all

No more crashes.... so maaaaaybe.....

(I mean, praying to Microsoft DOES feel pretty awkward, you know).

(and hopefully I don't have to pray).

Let's test this....

Well, with these kinds of errors ANYTHING is possible so then the sherlock holmes method, to elimitate all the impossible to come to the truth, won't work, you know....

????

Oh wait, I think I found out what's wrong.... (I hope)

Guess.....

Let's see if stuff works properly now .... :-/

I guess Microsoft still operates deep back down, so I guess it's always a possibility anyway

I think I fixed the issue, but this should NEVER have resulted into a .NET user error.... :-/

When creating a box for the side bar it already goes wrong.... Oh odd, since that routine doesn't call .NET

Let's test once more!

It's sad this is what it takes

All steps broken down in CSAY statements

Apparently that was not it....

And let's test, to see what happens!

Running my compiling script... again

From what it looks like, this is because the 'exists' API had to look through a directory that doesn't exist, so it couldn't look up if a certain file was there either. Since this is the most likely error that can pop up in this process, I'll just throw a warning in the console, but let NALA move on....\nThe API will always return 'false' then, since if the directory doesn't exist, the file cannot exist either, eh?

that means "something is wrong and I am not in the mood to tell you what, asshole!"

".NET error in user code"

And antoher test coming up!

NIL callbacks only work if they are declared global... dummy!

I've at least linked the save game management flow to the debug console, so I can at least be sure that I can easily get out of the game!

And what I see is.... NOTHING AT ALL

Yeah, NIL can have a few cons, but they don't outweight the pros, or at least, not this one!

Forgotten declaration

Testing once more!

And another one

Code Typo

And now with a few quick "$" I could fix another error... This time maybe more work than intended, but in the future.....

Rebuilt NALA so the game should now support this too, and hopefully this'll save me some work

jpbubble/NIL-isn-t-Lua#43

Well that that does it for the with stuff.... yay!

For classes this now works for non-static methods.... Static methods (which are for me the most important, actually) are an issue, but I think I got a solution.... But that will take some effort

Right, with is now recursive and in groups this works too... Only classes are a bit of an eyesore, as "$" doesn't want to work in classes

Nope, groups don't work either...

The "with" keyword in NIL does at least work, however classes and groups should also take care of this, but that doesn't appear to happen. At least not in classes, I need to check groups though...

I've set the basic scope starter in NIL to see _G as the 'width' variable if no proper width has been set

I'm first gonna set something up as this has been annoying me for far too long, so I need to set it right!

Forgot to note... I've added a few arrow arrow aliases for quicker access to a few features later in the game

Let's see now

File not found!!!! (404) :P

All direct parser issues accounted for, as far as I can see so far

(And I hope it'll work then, as I'm getting hungry)

Of course, another test iis in order....

Having to modify the engine itself a lot is normal for such a brand new engine seeing its first real action...
As the game progresses, these kinds of issues will get less and less so I can focus on the content more and more.... In other words, nothing has happened that is out of the ordinairy, but annoying it surely is!

It is one that obligates me to recompile NALA.... AGAIN!

Code typo

And Time for another test

Recompiling NALA

I think I found why a certain line which was merely a method call was seen as a delegate definition and even an invalid one too

Take... Wait... I forgot to count... Oh, crap!

So let's do that, and try again!

A few issues with the garbage collector pop up, but likely this is due to not yet re-compiling NALA

Another test

Group mixup

A basic setup for the interface has been done.... Point is due to lack of savegames, this is now hard to test, but at least I can do something... well I suppose...

So this makes the temporary link at least available, and functional

Command set up to activate the save game screen

Link up code to actually enter the SaveGame menu.... (Since no save spots exists now, this code is not yet accesible, duh!)

A link up code for loading the save game manager from the field... This will of course be needed when accessing save spots (although, mostly this load process already took place due to loading a savegame, but since you can also start a New Game, I may never assume this!)

A basic setup has been done to generate a slot to create a new file in.... This is all in preparational form, though so no big results can be expected.... YET!

Set up a checklist in #33, so I won't forget what data a save game file needs.

I've made it possible for the BUBBLE API to load text files in "lines" (Please note, it only takes Unix based text files with only LF based line ends, and files must be in full ASCII).

And so far for now

Both converted to PNG

They are both made in DPaint, so I got them in PCX format for now... This is however easily replaced

Arrow Down

Arrow Up

Well, a neutral tag this time.... An error did pop up, BUT, it could only do so, because no longer in use Kthura maps are INDEED disposed of! COOL!

MOAR Kthura case error fixings

The folder that I intend to use for "dynamic data" (save games, swap files and main config files) is correctly set

Kthura case error

I swear... I should make NIL classes case INsensitive :P

workdir issue

Although on second thought.... i do think it can wait... I've a better idea

I just realized that before I can get properly on the savegames I need to set the global config right. The global config will contain data that will always be in order no matter which savegame has been loaded... This file will allow the NG+ to start, or the True Abyss and what to put in the JukeBox and not to mention some vital data about the savegames themselves as well (not the content of the savegames, but some general data)

Skeleton for the save game interface

SaveSpot Red

SaveSpot Green

SaveSpot Blue

Well, later!

#29

THAT WORKS!!!!

I've set the texture of the actor to update automatically... This should allow Irravonia to turn around in accordance of the direction she faces

I've worked out stuff a bit in the NIL glue code on actors... Hopefully this can prevent memory leaks in the future, but I cannot yet be sure

Added tag OPTIMIZATION

Right now Irravonia won't turn her face depending on the direction she goes... This is not a bug, this is simply because I did not yet put in the code required to make tha happen.

YAY! All issues fixed

I think I found the "evil".... A correction on the x-coord where it had to be the y-coord.... Could make perfect sense in the first go, since Kthura exit spots are (especially in my forest levels) rarely bound to the grid (something I often do for fun) :P And thus correction is needed, and the y coordinate is very low in the start position.....

I need to try this.....

I did remove some warnings from C#, but that will not do much

Added tag CLEANUP

No visible effect

Let's see if this clears things up a little.... :-/

The first walk instructruction will however still cause the sprite to go haywire, however from that moment on things appear to work fine... I wonder why

I worked out the length property and yes, now it works as it should

Let's see what happens now....

98% of these warnings are mostly crap, so you can't blame me for ignoring them most of the time, but on rare occasions, I guess they do serve a purpose, eh?

Hmmm... One warning I see in my compiler report might solve it all.... :-/

More crap added to narrow things down

The debug appears to confirm what I thought.... Immediately after the first node has been handled (which merely will confirm the point where the hero was already standing), the chain appears to be terminated... I need to sort out why this is

Well the broken message has been found, and like i thought... it had a non-ascii character, but I fixed that...

Added tag RECOVERED

Indeed, now it compiles.... Some warnings are thrown, but they don't matter in this case, and I should take care of them over time.... :-/

This devlog system does not allow non-ASCII character, so perhaps one snuck in somehow.... Likely the broken note was not important, but it should still be in my history

Little site issue, it seems?

Well, fixed... is a big word, but I took care of at least one warning

I did see a little parse error, and due to this the debugging code was not executed in my test before... The new version of NALA was never compiled so the old version did the work... Not very good, eh?

Right, back in the office

I'm afraid this is gonna take some VERY serious debugging sessions, but they will require a lot of time and focus.... Focus I have... time I have not, as I'm expecting someone to arrive any minute now.... But I could already do more than I expected, so, let's not complain.... too much....

Let's test it out!

I've added more debug lines which could shine a light (hopefully)

I did clean Kitty's litterbox, and I feel like I poisoned myself in the process.... Damn that thing reeks!

Now Irravonia doesn't walk at all.... I think a proper debug breakdown is now REALLY in order)

This should fix PART of the issue, but we can never be sure prior to a test, can we?

I've altered the Move values at the "WALKTO" routine, making sure they will be the same as the player position when the walking starts

At least I fixed the crash caused when a route failed

The values that my MYPOS commands gives, do seem to be in line with what Dijkstra throws me, so everything looks in order, but that does not come out in the way Irravonia walks... So something must be wrong with the way Kthura calculates back the values Dijkstra gives it into the actor... Which is odd at best....

One "null" reference in case of impossible walks

I've added a debug line which should give more confirmation about matters

I've made a quick calculation of the ending value, and at least that value appears to be in order.... The other walk values do also appear to make sense to me

Let's find out more about this matter

yeah, that did fix stuff, but not really in the way expected, so I still got a lot of work to do

Oh yeah, forgot to tell you, Kitty enjoyed her meal by now :P

Of course, i gotta test this.... :-/

I think I fixed this issue.... Now all updates will happen AFTER the actors have moved, and my means of optimization ONLY if they have moved.

I think I solved the mystery... I think the automapper is to blame for this, which immediately activates on ANY sort of modification to any Kthura object during drawing, which is PRECISELY what the moving engine does.... Now having these remappings would only lead to slowdowns anyway, so I can best make sure they are turned off during the drawing iteration

The path was correctly generated, so the error does not lie there.... However, now I get an error of the most dubious kind, clearly thrown by .NET itself, and which should not be possible no matter how you put it.... I wonder how or where this error was caused....

Oh, and I missed a C# parse error too... that also explains....

Let's do that again, then

Forgot to save my source code prior to calling the compiler

Neither success nor failure are given... This would denote that the Dijkstra linkup code was not called.... That would be the only possibility left... however the earlier out-of-bounds errors given earlier denote that the request was done, and those errors are impossible without that call... I'm at a loss now....

Let's test

Let's narrow down the possibilities, after all when you eliminate all the impossible, whatever remains, however improbable, must be the truth, after all... This debug line can rule out 50% of the now open possible scenarios to the impossible, so only 50% that can be true will then remain.... That is a huge progression, doncha think?

Apparently, that was not the only thing that wasn't right

Of course, a little test will have to confirm I'm right or not....

I think I solved the issue why little Irravonia didn't walk.... Not only the "to" coords had to be adepted to the blockmap grid... The "from" coordinates had to be done so, as well

At least due to the earlier "out-of-bounds" errors thrown by Dijkstra, I can at least be rule out the possibility that Dijkstra wasn't called at all, and you know the saying: "When you eliminate all the impossible, whatever remains, however improbable, must be the truth!"

At least the "out of bounds" protection is operational... Now I gotta find out why little Irravonia doesn't walk, when instructed to....

In the meantime, I'll let my builder scripts do their work, so I can immediately start testing while Kitty is eating... ;)

Kitty (my cat) is begging me to fill her food bowl, so I guess I should do that.... It's about time too, so she has the right to do so....

The Dijkstra out-of-bounds protector needed a "return" command, since it was missing the code causing the error was still executed...

All fixes ignored???

A little NIL translation issue

Indeed, Dijkstra and Kthura were not in one mind when it came to how to use booleans for passible or impassible areas.... I made an extra routine in Kthura to make the reverse, but I also kept the original.... This to prevent compatibility issues with other stuff....

In the meantime my backup has finished so I can get back to "business as usual" mode

What I thought is true, Dijkstra did have an out-of-bounds protection during the route calculation, but not during the check of the starting and end coordinates.

If you request starting and ending coordinates outside the bouds of the blockmap, the path calculation will now always result into a failed empty route

A few things are clear, though

1. The coordinates are correctly adapted to the blockmap grid, so at least that is not the issue
2. It is also clear that the protection for walking out of bounds is not really operative... The test program did make the route touch the edges (on purpose) to see if that would properly block things.... So I must fix Dijkstra itself to always reject out-of-bounds walking requests
3. I do not know if Irravonia not walking is an actual failure in the route calculation, or the blockmap, but I think I got an idea on this though, as I think Dijkstra and blockmap have reversed meanings on True and False, but I need to verify that one

An extra debug line will have to provide me some information on the reason WHY!

This is a very common bug I had with the pathfinder in the original Dyrt as well, also when I was on Star Story, and my 63 fires of lung.... Only TFT REVAMPED didn't suffer here, but that was only because TFT REVAMPED and Star Story use the same engine, so the issue was fixed before that project began XD

Wherever I click NOTHING happens at all...

Take II

Illegal group call

Let's see what happens.. Take I

And now the command to make Irravonia actually walk has been given, meaning she should walk now to the location the player directs her to, *if* in fact she can reach it... Please note, due to the way the blockmap has been set up, Little Irravonia will keep herself to a 32x32 grid

So I can start my test anyway, as I don't wanna waste time waiting for my backupper to finish its job (and I really need to fix some issues with that tool asap).

Well luckily Powershell grants me a few quick commands to "void" that....

Unfortunately my backup tool got a bit stuck in the same window from where I run my building scripts... I hate it when that happens.... :-/

And let's see what that'll do!

For now I've limited myself to making sure the system understand where you want the hero to walk to

I've Made a "dummy" field... It does nothing yet, but at least this makes it easier on me to make 100% sure I can easily do alternate movements, like clickable objects, and that is very important for NPCs, savespots, signs and other things you can find in your everyday RPG dungeon or city.

I'll make a few worksouts in preparation to use Dijkstra though.... I got some time before I really must quit my session today

I needed a little break! Since the walking around routine is also more complex than just setting up a Dijkstra calculation whenever the player clicks something (as there are also clickable objects in the field), this may be something that won't be done until after the weekend, but when I say such things, you never know...

You should all know by know how quickly I can change my mind when it comes to these things....

I did a test and no parse errors came up, which is so far a good thing

All i can do now is to make sure all parse errors are taken care of... As the game doesn't yet have any controles allowing you to move, the Dijkstra routines are not called or anything yet.

#27 (And it doesn't look important, as this causes under-the-hood-issues, the player may never see, but memory leaks (even when they are resolved as soon as the game ends) are always a bad thing)

And the NIL link code has been created

Or rather a 3rd as a WalkToSpot was also there

The C# WalkToCoords has been written.... That is half of the contract

Of course, now I need to write the proper link code for Bubble, which is in this case for NALA

Moving in general and also in accordance with Dijkstra

Kthura is now able to calculate the route to walk over the Kthura Map, thanks to the underlying Dijkstra unit....
I must note though, that the route will be calculated, but the result will be ignored as I do not have th proper walking update stuff available yet.... (can't do all in once you know).

(This will require me to include Dijkstra also in the Kthura Map Editor source code, but that is a later concern).

I've now set the path finding variables to work with my own Dijkstra module

Dijkstra Has now been linked to NALA, so I can use it in Kthura

I'v deliberately created a track that was impossible to get from start from finish to see if this would indeed get a 0 path and a "false" for success rate... This happens to be the case

Time is however not on my side to actually attach this code to Kthura and to Bubble, but that doesn't matter right now... It works, and that's important ;)

And how come I never realized how simple this algorithm was..... :-/

I've been working things out on the Nodes parents, and NOW it works.... And both the demo map and the breakdown confirm that.... WOOOHOOOO!

.git folder form JCR6 in the test project

But seeing this mapped out, with a complete breakdown list, this no longer makes sense.... Some nodes have been correctly generated, yes, but been linked to the wrong parent....

I could not yet see the actual route, but.... What I did see was that the route was calculated to be 24 units long, and guess what, when I walk the route manually as a human, I come to the same calculation.... This could already be a good indication, the Dijkstra routine works....

Testing stuff