2026-08-26 - fpPS4: theirs, and one line
Sleep(INFINITE) in print_stub, the handler
bound to every unresolved function import. Not a crash and not a bug - a
debugging aid, with DebugBreak and readln
commented out beside it. Unattended it costs the whole run silently,
because the process stays alive and simply stops speaking. Every "fpPS4
seems stuck" observation was this.
Nothing on our side could have avoided it: a call that never returns
cannot be detected by its caller, and the dangling try
naming strspn is the mitigation working, not failing.
Patched to return zero - a function returning
QWORD, not a procedure, because the trampoline hands the
guest whatever is left in rax and garbage there crashes a
pointer- returning function far from the cause. Same mistake that
crashed Kyty at entry once.
36 records and a permanent hang → 36,631 records, 27/27 sections, 515/515 checks, 2 seconds, surviving 181 unimplemented-import calls. Stock would have needed ~362 runs to converge at two per blocker; fpPS4 is now the fastest loader in the sweep. (D196)