https://bugs.winehq.org/show_bug.cgi?id=42741
--- Comment #73 from Claudio sick_soul@yahoo.it --- (In reply to Maciej Stanczew from comment #72)
But… it's not actually a Wine regression. This behavior was introduced in 64-bit versions of Blizzard games, not in Wine itself. Original Warcraft III still works without issues, it's the Reforged client that has this problem.
This is not how I see the events unfolding at all.
Yes, the applications changed. But this is _NOT_ the problem I am talking about here. Read all the (duplicate?) bugs opened on this, they are eye opening.
See https://bugs.winehq.org/show_bug.cgi?id=49525 .
" Since wine-staging 5.12, the Battle.net launcher is failing to launch some games and is getting error "04c4:err:module:LdrInitializeThunk "ClientSdk.dll" failed to initialize, aborting". With wine-staging 5.3, 5.9 - 5.11, both Warcraft III and Starcraft I are able to launch. Vanilla wine 5.1, 5.9 - 5.12 all appear to fail with the same ClientSdk.dll error.
[...]
Comment 2Olivier F. R. Dierick 2020-07-06 05:28:52 CDT
Hello, This is a resurgence of STAGED bug 42741. It was covered by wine-staging patchset 'winebuild-Fake_Dlls', but that patchset is disabled in 5.12 due to changes in ntdll.
https://github.com/wine-staging/wine-staging "
So what happened here is:
1) overarching changes to ntdll broke / caused the patchset in 5.12 staging to be disabled.
2) regression bug reports _ignored_, ntdll changes not reverted. Screw them apps, development must go on!
3) some developers trying to hunt enablement again.
This is not how to make application enablement sustainable. This is how to have a constant state of brokeness.
As soon as a regression is detected the general changes should have been stopped/reverted until the application had been _unbroken_.
And yes, the few frustrated people are "wine developers", and?
The process and priorities are backwards, this is my point here. Does not matter.
And yes, there are attempts to "fix this downstream", but I am not sure it can really work out. A real solution for which users can reasonably expect applications to work "one day in the future" is if upstream takes an application focus, and regressions on enablement seriously.
Maybe Valve will end up fixing this, maybe not, we'll see.