atweaks & scs in EET installs.

[southfla79]

From your thread on the beamdog forums....SCS should go after EET_End now. Does this mean aTweaks should as well? I usually use both together but I thought the general rule of thumb is that majority of mods should go before EET_End. I guess that leads to the next question of is v32 still compatible with atweaks?

[DavidW]

Most mods should go before EET_end, though for BG2 mods that don't mess with NPCs, it doesn't much matter. aTweaks should be fine after EET_end, unless I'm missing something. As far as I know aTweaks is compatible with SCS v32 though I haven't run tests on it.

[southfla79]

thanks for the quick reply!

[K4thos]

Quote

From your thread on the beamdog forums....SCS should go after EET_End now. Does this mean aTweaks should as well? I usually use both together but I thought the general rule of thumb is that majority of mods should go before EET_End. I guess that leads to the next question of is v32 still compatible with atweaks?

regardless of what other readme files say installing EET_end as the last mod will never break anything in your game (or at least it won't have negative effects compared to installing it before other mods, I can't promise that some regression won't slip up to the code at some point). Instead of thinking which mod may be installed after and risk problems, you can install other mod earlier and don't worry about it.

btw. unreleased IWD-in-EET mod optionally appends chunk of code to EET_end that I think would conflict with SCS and other tweak mods, if it's installed before it (but won't if installed after) - it's a code that converts all class related stuff in all resources to use bitwise stat checks instead of current hardcoded class implementation (why it's needed has been disused here). There will likely be more code related to IWD-in-EET to ensure that all mod added content works with this mod changes to the game.

Edited June 14, 2019 by K4thos

[DavidW]

7 hours ago, K4thos said:

regardless of what other readme files say installing EET_end as the last mod will never break anything in your game (or at least it won't have negative effects compared to installing it before other mods, I can't promise that some regression won't slip up to the code at some point). Instead of thinking which mod may be installed after and risk problems, you can install other mod earlier and don't worry about it.

This is probably correct for mods that are designed to edit specific resources in BG2EE or BGEE (i.e., most mods). It is not reliably correct for mods that read data from the game and edit on the basis of that, because the data may not be correct for EET before EET_end is installed, and specifically, it is definitely incorrect for SCS. As a concrete example (the main example that makes me enforce SCS after EET_end), SCS detects NPCs by reading PDIALOG.2DA. That table is not populated fully until EET_end is installed, so not all NPCs are detected if SCS is run before EET_end. That in turn breaks the 'level and customize NPCs' component. I think there are probably other such issues: certainly I had a range of odd bugs reported by players testing SCS release candidates on EET that I couldn't reproduce, at least some of which I suspect are about these sorts of issues. (SCS reads the game files very extensively, and gets upset if the game isn't in a playable state with all those files correct.) I wouldn't be surprised if similar things happen with some components of Tweaks Anthology, which operates somewhat similarly.

Of course, there's a price I pay for installing after EET_end: I have to allow for EET's changes rather than just letting EET_end do it procedurally. For instance, one SCS component has Minsc interject into a ToB conversation, and I need to make sure I patch MINSC25J on non-EET installs and MINSCJ on EET installs. But that's not especially problematic. (I already have to write code that copes with being installed on BGTutu, BGT, BG2 vanilla, BGEE, BGEE-SoD, and BG2EE - what's one more game system? ) 

[K4thos]

Quote

PDIALOG.2DA. That table is not populated fully until EET_end is installed, so not all NPCs are detected if SCS is run before EET_end.

same is true for vanilla BG:EE+SoD installation (BDDIALOG.2DA), so in this instance I think it would be more reliable to just read campaign variants PDIALOG file names from CAMPAIGN.2DA. Same is true for other 2da files modified by EET_end (all have vanilla SoD variants, so a code to read them is needed anyway, if the mod supports BG:EE).

If it helps with compatibility I can add code that appends all SoD and BG1 PDIALOG.2DA entries during main component installation instead of waiting with it until EET_end (will be done in the next version).

Quote

 I think there are probably other such issues: certainly I had a range of odd bugs reported by players testing SCS release candidates on EET that I couldn't reproduce, at least some of which I suspect are about these sorts of issues.

I really doubt it, considering the nature of EET_end changes (described here). But if anyone is aware about such conflicts please let me know.

Edited June 14, 2019 by K4thos

[DavidW]

I take that point (I treat SoD as a black box because I haven't played it myself, and don't make any pledge that SCS affects anything in the SoD-specific bit of the game. But at some point I'll get around to playing it, and then I'll need to allow for this.) I don't necessarily object to taking that route around the issue here too. (Am I right in thinking that prior to EET_end you've just moved BGEE into a separate campaign slot, and EET_end does the work required to integrate them?)

BUT on a more general point I'm just not sure what the case is for saying 'EET_end must be after everything'. Doing complicated procedural things on the understanding that they have to give the right answer not immediately but after another complicated procedural thing is done is a nightmare both for code development and for code testing.