Well, let's try this in stead then.... :-/

WTF is happening...

Something makes NALA take too much memory, but whatever it is that causes it.... Who knows....

Well let's test this again, to see if stuff works now....

Now I did get a crash message from GIMP, but I wonder if that was the source of all "evil"

Something's wrong here... Everything froze

let's gather some info here... :-/

The counter value the debugger returns confirms this!

My hunch appears confirmed... Two sessions get into the memory in stead of one!

Let's do this to find out or rule out

Ah yeah, the start at position 2 is explained due to the OnLoad chunk containing one "end" instruction so that was at least NOT a bug!

Indeed, I had to form the call in the boxtext routine differently, so that issue has been resolved....
Doesn't answer why it takes two instructions at once though.....

Or maybe.....

The clue is, everything works the way it should except for the fact that no results are produced... Can you still follow?

More debugging lines as things do get stranger and stranger....

A little break, and boy I needed that!

Clearly... NOT!

I've added a debug line to the boxtext linker... This line will indicateif the boxtext initiator was called at all!

MOAR debug stuff to get to the bottom of this.... Are all commands skipped, or simply never executed?

Arrays in Lua are based on 1 ... not 0

Fixed the SASKIA script... I think....

Let's test!

I think I fixed this now....

STRONT!

Some more lines added to get some answers....

the data generate rule out the possibility of this error... Further investigation is required, I guess...

And I removed an older no longer needed debug line.

Let's see what this extra debug line will do

And now it seems its sees the parameter, but deems it.... empty????

Asserted added due to another crash

Session parameter missing

Whitelines will now be ignored... completely!

and once again WHITELINES are at evil

An extra debug line's been added in order to investigate that one!

Instruction is ""?

But in my defense I must note that I couldn't test SASKIA until the moment it would see action

BOZO! 'ins' is a class, not a string variable... Add the field .command to it...

Getting somewhere, but it's not yet enough

is "true" as bugged as "nil"?

I think I found the reason why this is happening!

Yes, the cycle does NOT allow the script to be executed, as I suspected.... The question now is.... WHY!

#14

A temp debug line to investigate #12

Grim Reaper

issued as #12

I cannot yet be sure, but it seems like the SASKIA scripts are being ignored... Also given by the fact the script wanted to unlink to soon.... I need to fully investigate this!

Function so ()

Added FLOW command, as I don't trust things at the moment....

Case Error?

Fixed the label issue

Ah... stupid... I didn't include the start label yet, as I was supposed to add it later, and I forgot its earlier omission

MOAR debugging crap!

I've added an extra line that should answer a few questions

I've set another thing for a new stack, and I hope that fixes something :-/

I think I fixed it

But the label parser is completely bogus...

On-Load label not found

"Annoy" can NEVER be 'nil' without NIL not crashing on it in translation, however the fact that the eror was thrown means something right, as the infinitely loop is now broken...

I added one, and that should fix things!

No incrementor?

For some reason the traceback doesn't show!

Wrong class call

I know BAD PUN!

More missing "self awareness" issues

A missing "self" issue in the datetime library

Now that SASKIA requires the datetime library and the game itself does not, I could check if libraries requiring other libraries worked properly, and that it does... so at least this cockroach was good for something!

I've put on a few time-out error enforcers to find out where the infinite loop could be

I've set up a quick datetime routine and a timeout checker... Hopefully this will shine a light on things!

Another infinite loop pops up... wtf is this?

I think I fixed a nil concatenation

Issue fixed!

Indeed that's it!

I think I found the issue

I've set a Time-Out feature, which may hopefully provide me some answers..... :-/

NOPE! Nothing happened!

And now to test this "fix"

Some last minute shopping

My investigation does show the parser is responsible for the infinite loops, though.... And I even think I fixed this....

I'd only wish the wasn't an auto-report system to Microsoft.... MicroSoft has nothing to do with this matter, so they'd better butt out!

I'm investigating the cause of the loop

Infinite loops are always a pretty nasty thing to deal with, as there is no telling what the responsible factor is... So further investigation will need to be done.

That issue's been fixed however somehow an infinite loop now appears....

So I added a temporary line to SASKIA to sort this out, as the claim given by the error is NOT possible!

I really need to get to the bottom of this

Code typo

has been expanded with 'full assert' to force traceback, as this is getting on my nerves!

Nil for Trim parameters

#10 should be fixed, but more trouble is on the way now!

And that should fix #10 (I hope).

This link has now been established

Fortunately I already designed SASKIA to be able to read from both JCR and real file system or whatever file system you can think of as long as you write a proper kind of driver for that.

That solves the mystery

I realize now that I did no tyet properly attach SASKIA to JCR6, so it's now trying to find SASKIA scripts in the real file system....

Time to investigate #10

See ya later!

running

No time for that now

#10

FOUND IT (I think!)

I read the error message wrong... Still this is not how it should be!

Class Kthura still 'nil'... for some reason...

I think I solved it... That boolean expression did look faulty...

This is the second time a class is declared "nil" without any form of reason...

Forgotten )

Fired her up, and I expect TONS of bugs!

Great! I've killed what was beautiful code, and dirty code straight from hell (ruining all rules of programming even my own) seems to work... TERRIBLE!

It's getting stranger....

Let's try a few things.... Whatever is happening here... It should not be possible....

Now let's see if I get this straight... Class CL_SASKIA is nil, while doing a request from CL_SASKIA is actually running.... WTF is going on here?

I've linked the SASKIA cycler to the Update callback in the field... Now things can get to work... (I hope).

Coolvetica 20 has been linked to the BoxText routine as standard font

A little reorderning of code... Would be very important to do before the BoxText routine kicks in!

All crashes accounted for!

Flow error

I think I fixed the damn scope error

#8

#8

jpbubble/NIL-isn-t-Lua#42

local local?

And that should take care of all parse error crap

Parse error :(

I think I fixed this.... (I hope)....

Aha... hahaha... I think I got it :P

A strange error, of which I cannot fully tell if this is another bug or still the same one refusing to be fixed.

And another parse error killed

This SHOULD fix it

Yup, and NOW I found it... Now I've added an extra parameter in the classes tracker, so it can see that it's a group.... This does not yet fix the issue, but I wanna see if this is being registered!

And hopefully I can study this now at last!

I fixed that now

Ah... dupe vars are the reason behind the nonsensical data

How can I fix this, if Lua refuses me to create data that makes sense?

Nil?

let's set this out more properly as this is getting more and more vague

So let's try it again!

A few issues in setting this study material up

Some more study

That did not shine a light

Let's see what I can do to make sure I can rule groups out here... :-/

Aha... Now I see... The parser didn't understand the difference between a group and a class, and as the "." wasn't used, it threw an error... That should be fixable, though...

And now it's just playing dumb on me

Forgotten commas

Some function misdefinitions

Well THAT is fixed now

That syntax would NEVER work, Jeroen! Neither in NIL nor in Lua

I refuse to call this fixed, but at least this should prevent trouble (for now).

I see private classes are still bugged (was to be expected, but still)

VAGUE error

A test

But for now I'm too tired... I cannot think clearly anymore, so it's better if I call it!

I've set a link to the BoxText library and the game. This has not yet been tested, and it must be very fragile as a result. I will have to set this out properly later

New Anyway.nil file generated

And a link established to the 'scenario' library, which will only be used in the BoxText routine (at least for now, as a few sub-events may also use this)

Skeleton for BoxText

I've converted the scenario library into NIL format and I've made it extensible for future engines

Spicing my chicken for my dinner tonight

I'm taking a break now

Running

(Of course since I don't have to set up any scenario any more, except for testing for typos, things are a bit easier on me).

Of course I wasted an entire day on such an imbecile thing, but at least it works now... Next step will be the BoxText script, which will also be one hell of a job....

GOTCHA!!!! Idiots

So here's that test!

This should now be fixed, but I need to make another test to be 100% sure

Leave it the guys at MicroSCHOFT (not a typo, just look up the Dutch word "schoft") to make simple actions into a mess!

I did deliberately place Brendor out of the prison walls to see if he'd appear, and he does.... So what is clear is that the objects are (for some reason) drawn in the wrong order... This would imply dominance maps being ignored, somehow

HARSH TEST!

The textures ARE being loaded... The debug round did confirm that one!

Hopefully this debug line can provide me some answers, as I'm really getting clueless now...

The crasher is NOT activated, meaning the textures in question ARE LOADED. But what goes wrong then?

In order to make sure I've made sure the game will crash out. I've done this by using delegates to make sure the EDITOR will not crash on things like these.

My guess is something goes wrong with loading picture bundles

Indeed, the objects are loaded... Question is why they do not appear... :-/

I guess I've to eat more fish... If it wasn't so revolting and expensive!

I KNEW that whatever it was, it had to be stupid!

NIL VALUE!!! Not possible unless faulty script has been produced by the underlying API (which should in turn also not be possible, but let's check)

If this game ever gets finished it'd be a disaster, eh?

Code Typo

again....

F.U.

Why does Lua have loadstring and load... Makes NIL communication harder, ya know

And a command for the console to see the results... Now I wanna know what I get to see... :-/

Linked to the script VMs

Easy to write anyway :P

I've added a Kthura Object counter

Parameters wrongly received

almost there...

AGAIN!

ANOTHER Code typo

Anyway, it's been fixed, so let's try that again, shall we?

Typos in code are lethal, you know!

It seems fixed, but now I gotta deal with a TYPO!

This should take care of it!

And this one is for ME!

Let's try { ... } catch this one, shall we?

Now that CLEARLY didn't work out the way it should