[FeatureRequest] :LAST[mod] pass that runs right before :FINAL

[Sigma88]

I was going to ask for a new pass specifier that ran just before :FINAL

but then I thought, would it be possible to allow both a normal pass specifier and :FINAL to be used?

with the effect that:

• firstly, all pass specifiers without :FINAL are run following standard rules
• then, all the pass specifiers with :FINAL are run alphabetically
• finally, all patches that have only :FINAL are run

this will not change how current patches are applied, but open up a second pass of patches.

this would remove the need of using :FOR[zzzzzzMyMod] and should be pretty easy to implement into the current code

[Sigma88]

I don't mind taking a shot at it myself if you like the idea but don't want to write it yourself

[blowfishpro]

What about :FINAL[myMod]

[Sigma88]

Yeah that works fine for me if it's easier to implement :+1:

[sarbian]

As long as we don't start a race to the FINALer modifier I am fine with it.

[Sigma88]

I only proposed it because people are already running at the most Zful mod :D

if you feel like this would not add anything to the mod I'm fine with not getting this feature

[blowfishpro]

Now that I think about this a bit it would probably make sense to name it something else to avoid confusion. Maybe :CLEANUP[xxx] or :LAST[xxx]

[Sigma88]

my initial idea was :LAST so that works well for me 👍

[sarbian]

Yes, LAST may lower the risk of confusion

[Sigma88]

in your idea, :LAST[XXX] would not be integrated into the :FOR system?

it would just be a simple alphabetical ordering right?

[blowfishpro]

All patches marked :LAST[xxx] would run right before the :FINAL pass, in alphabetical order.

[blowfishpro]

Oh, one more thing, should :LAST[xxx] act like :BEFORE[xxx] and :AFTER[xxx] in that it will only run if mod xxx is present? Leaning toward yes, but wanted to get some input.

[Sigma88]

no I think it should be analogue to :FOR so that people can use it instead of :FOR[zzzzzzwhatever]

but it shouldn't add a mod definition. so if I do :LAST[SigmaBinary] the patch will apply regardless of SigmaBinary being installed, but if there is no :FOR[SigmaBinary] or no "SigmaBinary.dll" other patches using ":NEEDS[SigmaBinary]" will still fail

[Sigma88]

Hi guys, I've been out of the loop for a while, is there any news on this?

if you need help testing/coding feel free to let me know

[blowfishpro]

I did some work on it, then backed up to do some refactoring that would make it simpler. I can probably go back and re-work that now that the refactoring is done.

[blowfishpro]

This is done

[Sigma88]

Thanks!

[zekew11]

I was looking for documentation on this feature and this exchange is sufficient, but took some digging to find; mostly because "last" is such a poor search term. The patch ordering wiki page acknowledges the LAST pass but doesn't describe that it is analogous to the FOR pass. I'd be happy to write in an update for the page, but don't want to speak imprecisely about code I didn't write if the convention is that contributors write the doc.