Children whose parent dies are reparented to the guest launchd, and its job
manager only reaps PIDs it has a job record for, so every leftover mldr/CEF
process from a killed client tree stayed a zombie until shutdown. Reap zombies
that launchd owns but does not track on a 5 second timer; tracked jobs keep
their exit status for job_reap(). proc_listpids() is not implemented in Darling
yet, so the scan enumerates the guest PID namespace through /proc.
recover-stale-client now signals the stale tree deepest-first so a live parent
can still reap its own children, and prepare.py repairs interrupted checkouts
and missing Git LFS objects instead of publishing pointer text as a library.
After a pull that changed a patch, prepare.py stopped: the patch neither
applied nor reverse-applied to a tree holding its earlier version. Now
it looks for the earlier version applied in that component -- the copy
it records in build/applied-patches, or a version from this
repository's history -- reverts exactly it and applies the new one.
It never resets a repository; when nothing matches it stops and names
the repository with other edits, and it restores the earlier version
if the new one does not apply.
scripts/update.sh runs build.sh, the base install and
install-runtime.sh, and refuses to start while a Riot Client or League
process is open, since the install steps stop Darling.
Checked on a sandbox Indium checkout holding the previous patch: the
result matches a clean apply, a second run is a no-op, an unrelated
local edit survives and a conflicting one stops with the error.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The pinned src/external/swift/.lfsconfig points lfs.url at the retired
git-lfs.darlinghq.org, so a fresh `git submodule update --init --recursive`
blocks with no output right after src/external/sqlite is checked out: swift is
the next submodule in sorted order and its checkout waits on that host.
Override lfs.url to GitHub's endpoint, which still serves the objects (verified
44 objects / 16 MB, real Mach-O dylibs), and print explicit submodule progress so
the last completed submodule is not mistaken for the stalled one.
lucannez64