fix(install,studio): address Codex round-4 review (6 of 7 comments real)

- install.sh: gate the new aarch64 bitsandbytes block on SKIP_TORCH=false --
  with --no-torch/UNSLOTH_NO_TORCH (GGUF-only install) it would have pulled
  torch back into the venv through bitsandbytes' dependencies.

- studio worker: in the new Spark OOM-guard section, decide
  PYTORCH_CUDA_ALLOC_CONF (expandable_segments) BEFORE the guard's first CUDA
  touch -- get_device_properties initializes the CUDA allocator, after which
  the env var is ignored, and the later `import unsloth`
  (patch_dgx_spark_memory_config) is too late for the worker process. Uses
  the same CUDA-free nvidia-smi name sniff, append-don't-override, and
  UNSLOTH_NO_EXPANDABLE_SEGMENTS opt-out as the library patch. Live-verified
  on the N1X: env set while torch.cuda.is_initialized() is still False.

- uninstall.ps1: only run `fuser -k 8888/tcp` in a probed WSL distro when an
  Unsloth install actually exists there (checked BEFORE the rm deletes the
  marker) -- an unrelated listener on 8888 (e.g. Jupyter) in a clean distro
  must survive a Windows-side uninstall. The Unsloth-specific pkills stay
  unconditional.

- install.ps1 + uninstall.ps1: persist the chosen WSL distro to
  %LOCALAPPDATA%\Unsloth\wsl-distro.txt at install; uninstall reads it
  (before removing the directory) and prepends it to the cleanup candidates,
  so a custom UNSLOTH_WSL_DISTRO install is cleaned without the env var
  being set again at uninstall time.

- provision_llama_cuda.sh: honor UNSLOTH_LLAMA_PR (numeric-validated,
  best-effort fetch of pull/N/head after clone) so a provisioned tree
  matches a PR pin the way setup.sh does; and require only llama-server in
  the main cmake build (mirroring setup.sh), building the helper targets
  (llama-cli/quantize/mtmd-cli/gguf-split) best-effort afterwards -- an
  older UNSLOTH_LLAMA_TAG pin lacking a newer helper target no longer fails
  the whole provision.

Not changed: the "--tauri rejection doesn't restore the venv rollback"
comment is incorrect -- the rejection returns through Exit-InstallFailure,
which itself calls Restore-StudioVenvRollback.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

studio/scripts/provision_llama_cuda.sh

@@ -143,6 +143,20 @@ if [ ! -d "$LLAMA_DIR/.git" ]; then
         _restore_prev
         exit 0
     fi
+    # Honor a llama.cpp PR pin (UNSLOTH_LLAMA_PR, the same var setup.sh supports)
+    # so a provisioned tree matches the user's request instead of silently building
+    # the default branch. Best-effort: a failed fetch keeps the default branch.
+    case "${UNSLOTH_LLAMA_PR:-}" in
+        ''|*[!0-9]*) ;;
+        *)
+            if git -C "$LLAMA_DIR" fetch --depth 1 origin "pull/${UNSLOTH_LLAMA_PR}/head:_unsloth_pr_${UNSLOTH_LLAMA_PR}" >/dev/null 2>&1 \
+                    && git -C "$LLAMA_DIR" checkout "_unsloth_pr_${UNSLOTH_LLAMA_PR}" >/dev/null 2>&1; then
+                log "checked out llama.cpp PR #${UNSLOTH_LLAMA_PR} (UNSLOTH_LLAMA_PR)"
+            else
+                log "could not fetch llama.cpp PR #${UNSLOTH_LLAMA_PR}; building the default branch"
+            fi
+            ;;
+    esac
 fi
 cd "$LLAMA_DIR" || { _restore_prev; exit 0; }
 
@@ -190,8 +204,16 @@ _NICE=""
 command -v nice   >/dev/null 2>&1 && _NICE="nice -n 19"
 command -v ionice >/dev/null 2>&1 && _NICE="$_NICE ionice -c 3"
 _cmake_build() {
-    $_NICE cmake --build build -j"$JOBS" --target \
-        llama-server llama-cli llama-quantize llama-mtmd-cli llama-gguf-split >/dev/null 2>&1
+    # Only llama-server is REQUIRED (mirrors setup.sh's source path): an older
+    # UNSLOTH_LLAMA_TAG pin may predate newer helper targets (llama-mtmd-cli,
+    # llama-gguf-split), and those missing must not fail the whole provision.
+    $_NICE cmake --build build -j"$JOBS" --target llama-server >/dev/null 2>&1
+}
+_cmake_build_extras() {
+    # Helper targets unsloth-zoo's GGUF exporter also uses -- best-effort each.
+    for _t in llama-cli llama-quantize llama-mtmd-cli llama-gguf-split; do
+        $_NICE cmake --build build -j"$JOBS" --target "$_t" >/dev/null 2>&1 || true
+    done
 }
 if ! _cmake_build; then
     # An interrupted build (e.g. a thermal/power shutdown mid-compile, which this
@@ -203,6 +225,7 @@ if ! _cmake_build; then
     _cmake_configure || { log "cmake configure failed"; cd /; _restore_prev; exit 0; }
     _cmake_build || { log "cmake build failed"; cd /; _restore_prev; exit 0; }
 fi
+_cmake_build_extras
 
 if is_cuda_server "$SERVER"; then
     log "CUDA llama-server ready: $SERVER"