ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

install: six round-five review fixes across provisioner, setup, uninstall

Browse source 
Fifth review round; each item traced through the live scripts before fixing.

A provisioner fresh clone that failed to produce a server was left behind as
a markerless git tree; under a custom UNSLOTH_STUDIO_HOME the next run's
ownership assert refuses the unmarked dir and aborts the whole install until
the user deletes it by hand. _restore_prev now removes a clone this script
created when no server came out of it (backed-up dirs restore as before).

The CUDA provision gate ignored --with-llama-cpp-dir linked mode, so a
linked user tree with a CPU-only server could be checked out to a pinned
ref, rebuilt in place, or moved aside entirely and replaced by a fresh
clone. The gate now skips linked local dirs.

uninstall.sh removed the CUDA build artifacts without stopping a running
detached build; _pkill_studio only matches Studio roots, so live cmake/nvcc
kept burning thermals, recreated build files, and defeated the trailing
rmdir. The runner, provisioner, and llama.cpp-path processes now get
TERM-then-KILL with the same escape helper and grace the Studio kill uses.

The worker's memory-fraction guard classified Spark purely from device
props, so UNSLOTH_FORCE_DGX_SPARK=1 on an unlisted name got no fraction
guard (and the fraction env was dead), while FORCE=0 could not disable it;
the guard now honors the same force semantics as the detectors.

UNSLOTH_LLAMA_TAG / UNSLOTH_LLAMA_PR were interpolated into the runner
script's single-quoted exports unvalidated while every sibling forward has
an allow-list; they now get the INSTALL_REF ref allow-list and a digits-only
check respectively (own-machine robustness, not a trust boundary).

On WSL-fallback success with a custom UNSLOTH_STUDIO_HOME, the installer
deleted the rolled-aside custom-root venv right after telling the user that
root is not used by the WSL install; a custom root now restores the previous
venv instead (the WSL shim does not depend on the Windows venv), while the
default root keeps dropping the vestigial backup.

Verified: bash -n on all three shell scripts, AST parse on worker.py,
PowerShell AST parse on install.ps1, icon suites pass, sh battery matches
the branch baseline. Two resurfaced anchors (build/bin backup, --package
forwarding) confirmed already fixed at head.
This commit is contained in:
Daniel Han 2026-07-18 12:08:39 +00:00
commit af459f4673
5 changed files with 42 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

7
studio/scripts/provision_llama_cuda.sh
View file

@ -212,10 +212,16 @@ fi
# Back up any existing (e.g. CPU-only) llama.cpp: restored on any failure exit,
# dropped only once the fresh build yields a server -- never leave NO server.
_LLAMA_BAK=""
_FRESH_CLONE=0
_restore_prev() {
if [ -n "$_LLAMA_BAK" ] && [ -e "$_LLAMA_BAK" ]; then
rm -rf "$LLAMA_DIR" 2>/dev/null
mv "$_LLAMA_BAK" "$LLAMA_DIR" 2>/dev/null && log "restored previous llama.cpp install"
elif [ "$_FRESH_CLONE" = "1" ] && [ ! -x "$SERVER" ]; then
# We created this clone and produced no server. Leaving a markerless git
# tree under a custom STUDIO_HOME bricks reruns: setup.sh's ownership
# assert refuses the unmarked dir and aborts the whole install.
rm -rf "$LLAMA_DIR" 2>/dev/null
fi
}
if [ ! -d "$LLAMA_DIR/.git" ]; then
@ -231,6 +237,7 @@ if [ ! -d "$LLAMA_DIR/.git" ]; then
if [ "$_clone_ok" -ne 1 ]; then
git clone --depth 1 https://github.com/ggml-org/llama.cpp "$LLAMA_DIR" >/dev/null 2>&1 && _clone_ok=1
fi
[ "$_clone_ok" -eq 1 ] && _FRESH_CLONE=1
if [ "$_clone_ok" -ne 1 ]; then
log "git clone failed"
_restore_prev