Faction defs used as Slayer on weapon | Other defs section not working in IF statements

[AtillaG1]

Faction defs used as Slayer on weapon | Other defs section not working in IF statements

[xwerswoodx]

Okay I found out the issue, it's same as I wrote before, because _Evaluate_Conditional_EvalSingle returns as fVal = bool(Exp_GetVal(ptcSubexpr)); while Exp_GetVal is 32bit integer it causes this issue. Changing it Exp_GetLLVal solve the problem.

[AtillaG1]

[cbnolok]

Possible solution, from a discord discussion:
nolok
—
15/07/2023 20:08
split the faction prop in two?
Gladie
—
15/07/2023 20:09
what do you mean?
nolok
—
15/07/2023 20:10
a better idea even
we have npcgroups and factions
some factions are exclusive to an npcgroup
still, every faction is in a big bitmask, thus we have enormous numbers
eg:

[DEFNAME npc_factions]
// Fey group (opposed to Abyss Group)
faction_fey 01 // Super Slayer

// Elemental Group (opposed to Abyss Group)
faction_elemental 02 // Super Slayer
faction_air_elemental 04
faction_blood_elemental 08
faction_earth_elemental 010
faction_fire_elemental 020
faction_poison_elemental 040
faction_snow_elemental 080
faction_water_elemental 0100

// Abyss group (opposed to Elemental and Fey Groups)
faction_demon 0200 // Super Slayer
faction_gargoyle 0400 // Super Slayer

we could instead start from 1 when beginning a new group
eg:

[DEFNAME npc_factions]
// Fey group (opposed to Abyss Group)
faction_fey 01 // Super Slayer

// Elemental Group (opposed to Abyss Group)
faction_elemental 01 // Super Slayer
faction_air_elemental 02
faction_blood_elemental 04
faction_earth_elemental 08

// Abyss group (opposed to Elemental and Fey Groups)
faction_demon 01 // Super Slayer
faction_gargoyle 02 // Super Slayer

that means that we have to change the code to check for both group and faction, but that will solve our problem with no additional memory footprint
or core complexities
Gladie
—
15/07/2023 20:16
but then would be 01 usefull twice?
because the faction system in core is done that if you have only the faction_elemental, you deal double damage to all the elementals of this group, but if you have only faction_air_elemental, you only deal tripple dmg to air elemental
maybe i dont get fully what you mean
nolok
—
15/07/2023 21:02
Yes
More than twice
Then we would need to split that into faction_group and faction_species/subgroup/something in the source
I have no better ideas honestly
And if the faction species is 0 but not the group, you will deal more damage to every monster in the group
The other thing that might be worth adding in the case of a rework is to enable to slay a specific defname
Maybe with another property like faction_extended, if 1 you check for additional tags containing defnames
That's an idea, maybe not the best one

[xwerswoodx]

The problem is not a range of values if I understand the issue correctly. Because 2^62 == 4,611,686,018,427,387,904 so that mean you can set values until 0x4000000000000000

The problem is evaluate not works perfectly for llong values, even faction_wolf not working correctly

FACTION faction_wolf
SERV.LOG <FACTION>
IF (<FACTION>&<def.faction_wolf>)
  SERV.LOG X
ELSE
  SERV.LOG NO-X
ENDIF
IF (<FACTION>&faction_wolf)
  SERV.LOG Y
ELSE
  SERV.LOG NO-Y
ENDIF
IF <EVAL <FACTION>&0800000000>
  SERV.LOG Z
ELSE
  SERV.LOG NO-Z
ENDIF```

Returns;

```Resync complete!
19:09:(defs_faction.scp,78)0800000000
19:09:(defs_faction.scp,82)NO-X
19:09:(defs_faction.scp,87)NO-Y
19:09:(defs_faction.scp,92)NO-Z
19:09:1:'xwerswoodx' commands 't'=1```

But faction_bird works well;

```[FUNCTION t]
FACTION faction_bird
SERV.LOG <FACTION>
IF (<FACTION>&<def.faction_bird>)
  SERV.LOG X
ELSE
  SERV.LOG NO-X
ENDIF
IF (<FACTION>&faction_bird)
  SERV.LOG Y
ELSE
  SERV.LOG NO-Y
ENDIF
IF <EVAL <FACTION>&080000000>
  SERV.LOG Z
ELSE
  SERV.LOG NO-Z
ENDIF```

Returns

```Resync complete!
19:12:(defs_faction.scp,78)080000000
19:12:(defs_faction.scp,80)X
19:12:(defs_faction.scp,85)Y
19:12:(defs_faction.scp,90)Z
19:12:1:'xwerswoodx' commands 't'=1```

080000000 equals to 2,147,483,648 which is 32bit integer max value not llong.

We probably need to change the evaluation system from int to llong to update that for 64-bit integers.

[Jhobean]

But there no eval on the problematic if. You mean server us eval on internal?

[xwerswoodx]

Yes, I meant the IF statements evaluate function like _Evaluate_Conditional_EvalSingle and QvalConditional for QVAL function not the EVAL function itself. EVAL, FVAL, HVAL etc already set 64-bit in CScriptObj.cpp but IF and QVAL have forgotten in 32-bit int. I changed both of them to 64-bit, I am testing right now but honestly it should work fine.

[Jhobean]

Now with the new action on github, .exe will be create when you create a PR. Everyone can test it easy.

[AtillaG1]

works, thanks.