nothing manacles and ursa furiosa plugsets differ between implementation and manifest

[deangaudet]

for random rolls, two exotics seem to use a different internal plugsets compared to the manifest. in the case of ursa furiosa it's a minor issue, but in the case of nothing manacles it's fairly severe and noticeable. i've been tracking exotic rolls since the exotic roll changes in season of the deep, and the issues have been on-going since then, but before reporting this i wanted to be sure that the issues persisted today in echoes act 1.

images below are from the community tool d2armorpicker investigation tab

nothing manacles:
nothing manacles is the more severe issue: it appears the internal implementation was not updated when the rest of the exotics had their rolls improved back in season of the deep. in the manifest, nothing manacles appears to use two copies of plugset 1094127323, which is a discipline focused plugset. 1094127323 contains plugs with at least 14 total stats, yet a random drop of nothing manacles from a nightfall this week has two plugs with 13 total stats and which do not appear in 1094127323.

ursa furiosa:
the issue here is more benign, and it is likely only an error in the API manifest. ursa stands out as the only exotic in which sockets 8 and 9 have different plugsets. the manifest lists 3111243035 and 3147432585 respectively. 3111243035 is what i would call the "unfocused dis/int/str plugset", whereas 3147432585 i would call the "int focused plugset". in all other stat-focused exotics there are two copies of the stat-focused plugset. in practice, ursa rolls plugs which suggest it is internally using the 3147432585 int-focused plugset in socket 8 as well as 9; so it's only the manifest which is in error.

thanks
-dean

p.s. while i have your attention, there is a third issue which i'm not sure is a bug or if it's intended. the strength focused plugset 2360091748 appears to more heavily weight 14-stat plugs than the other five focused stats. it contains 43 plugs in total, whereas the other five focused plugsets contain only 34 plugs (compare 3147432585 int-focused for example). random str-focused exotics appear to implement this difference, so the manifest and implementation seem to agree here... which makes it hard to say if it's a bug or intended.