You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Oct 23, 2026. It is now read-only.
Repository navigation
This repository was archived by the owner on Oct 23, 2026. It is now read-only.
Blocker: ccleste and RommPS stall on main in the klog stderr loop (fix PR boykopovar/AnyPS5#4536); ccleste Presents is from a local branch #44
On main 0ef9dc92 (Windows 11, RTX 4080), ccleste and RommPS run 90 s without a frame. The line above is their last, no error is printed, and no swapchain is created. In gdb the title's klog reader thread is stuck in sceKernelDebugOutText, writing into the pipe that the title's dup2(pipe, 2) made descriptor 2, and the thread creating the Vulkan device waits for the stderr lock. On Linux (container, lavapipe) ccleste stops at the same line. Details are in #4536.
With #4536, both create their swapchain and present at about 60 frames a second: ccleste shows its title screen, RommPS its UI.
fix(libkernel): keep sceKernelDebugOutText off a redirected guest stderr boykopovar/AnyPS5#4536 is open and now a blocker (pr-4536, fix PR #4536) at the end of the chains of ccleste (PPSA59322) and RommPS (PPSA76677), as the stall after Vulkan device ready on main 0ef9dc92. It carries bf510143, the second commit of the closed #1870. That commit is not on main (HostStandardError is absent from File.hpp and Stdio.cpp), and #2478 (merged) added dup2 without it.
ccleste's Presents came from TheAggie's comment on #1732 (2026-10-10, local build with 0 registered shaders), before that PR merged. A stage review now downgrades it to Starts on main (Windows, NVIDIA), quoting #4536: "On main, neither saves a frame in 90 s."
RommPS: ctime_nid_postfix is on main (libc/src/Time.cpp), though #1871 itself was closed unmerged, so no ctime step was added. I could not check the gdb stacks or the 60 frames a second myself; those are your measurements in #4536.
roadmap.py --validate passes and 75 tests pass. An offline build shows ccleste at Starts with #4536 in its chain. I committed only data/blockers.json, as in f9883e7.
Activity
Add #4536 as blocker of ccleste and RommPS, correct ccleste stage (#44)
AlxcBTW commentedon Oct 12, 2026
Thanks @orishamir. Fixed in 7f5c5b2.
pr-4536, fix PR #4536) at the end of the chains of ccleste (PPSA59322) and RommPS (PPSA76677), as the stall afterVulkan device readyon main 0ef9dc92. It carries bf510143, the second commit of the closed #1870. That commit is not on main (HostStandardErroris absent fromFile.hppandStdio.cpp), and #2478 (merged) addeddup2without it.main, neither saves a frame in 90 s."ctime_nid_postfixis on main (libc/src/Time.cpp), though #1871 itself was closed unmerged, so noctimestep was added. I could not check the gdb stacks or the 60 frames a second myself; those are your measurements in #4536.roadmap.py --validatepasses and 75 tests pass. An offline build shows ccleste at Starts with #4536 in its chain. I committed onlydata/blockers.json, as in f9883e7.AI-assisted: yes (Claude Code)
AlxcBTW commentedon Oct 12, 2026
Fixed in 7f5c5b2.