https://bugs.winehq.org/show_bug.cgi?id=50252
--- Comment #2 from Rafał Mużyło galtgendo@o2.pl --- (In reply to Alexandre Julliard from comment #1)
I don't see any difference in behavior. Please explain how to reproduce the regression.
Well, the additional element might be bug 42606, but I've yet to hear about a better solution there, as that theme setting can't be volatile, like some of the other locale dependent settings are, yet the current way caches charset across locales, which is - as the bug explains - broken.
Other than that, it's all standard - just font replacement keys set up between respective IPA Mona and MS Mincho/Gothic families.
(if the log doesn't make it clear, besides bug 42606 adjustment, MenuFont is still standard Tahoma)
(PS: to make it clear: my primary locale is pl_PL.utf8, bug 42606 happens once the secondary one (ja_JP.utf8) is used; this bug also happens in the secondary locale, as in primary, wine's Tahoma has the sufficient coverage - so if you were trying to reproduce in a prefix that was set up for ja_JP.utf8 as as *primary* locale, if bug 42606 actually *does* play a role, chances are things appeared to work fine)
(the other, non-menu font used in regedit looks fine; as it seems to be a linked font too ("MS Shell Dlg", going by regedit.rc), this might suggest that the issue here might be not the changes in handling linked fonts, but those in font caching...perhaps DEFAULT_CHARSET confuses font comparison helpers and instead of DEFAULT_CHARSET, it should always explicitly translated to the charset bound to ACP ?)