dotnet/runtime · error · RuntimeError

Missing JIT

Error message

Missing JIT

What it means

Raised by upload_command in jitrollingbuild.py during the 'upload' subcommand. After computing the JIT filename via determine_jit_name(host_os) (clrjit.dll on Windows, libclrjit.so on Linux, libclrjit.dylib on macOS), the script checks os.path.isfile(jit_path) under coreclr_args.product_location and aborts if the primary JIT binary is absent. The upload cannot proceed because the essential artifact is missing.

Solutions

  1. Build the JIT for the selected arch/build_type/host_os first (e.g. `build.cmd clr+libs -arch x64 -c Checked`) so clrjit exists in the product directory.
  2. List product_location and confirm the file reported in the 'Couldn't find JIT at <path>' log line actually exists.
  3. Re-run upload with -arch / -build_type / -host_os matching the tree you actually built.

Example fix

// before
python jitrollingbuild.py upload -git_hash abc123   # raises 'Missing JIT'
// after (build first, then upload with matching arch/build_type)
build.cmd clr+libs -arch x64 -c Checked
python jitrollingbuild.py upload -git_hash abc123 -arch x64 -build_type Checked
Defensive patterns

Strategy: validation

Validate before calling

import os
from jitutil import determine_jit_name

jit_name = determine_jit_name(host_os)
jit_path = os.path.join(product_location, jit_name)
if not os.path.isfile(jit_path):
    raise SystemExit(f"No JIT at {jit_path}; build clr+libs for {host_os} first")
# safe to proceed to upload

Prevention

When it happens

Trigger: Running `jitrollingbuild.py upload -git_hash <hash>` before the JIT has been built, or with product_location / -arch / -build_type / -host_os that do not match the built tree, so the expected clrjit.* file is not on disk at the computed path.

Common situations: Forgetting to run the CLR build (e.g. build.cmd clr+libs -arch x64 -c Checked) before uploading; mis-set CORE_ROOT/product_location; selecting Release when only Checked was built; cross-arch uploads without the matching arch JIT.

Related errors


AI-assisted analysis of dotnet/runtime@60108ba66e (2026-08-10). Data as JSON: /api/errors/b75ee692c19a1dba. Report an issue: GitHub.

Appendix: source

Thrown at src/coreclr/scripts/jitrollingbuild.py:388

    #
    # We could also upload debug info, but it's not clear it's needed for most purposes, and it is very big:
    # it increases the upload size from about 190MB to over 900MB for each roll.
    #
    # For reference, the JIT debug info is found:
    #    a. For Windows, in the PDB subdirectory, e.g. PDB\clrjit.pdb
    #    b. For Linux .dbg files, and Mac .dwarf files, in the same directory as the jit, e.g., libcoreclr.so.dbg

    # Target directory: <root>/git-hash/OS/architecture/build-flavor/
    # Note that build-flavor will probably always be Checked.

    files = []

    # First, find the primary JIT that we expect to find.
    jit_name = determine_jit_name(coreclr_args.host_os)
    jit_path = os.path.join(coreclr_args.product_location, jit_name)
    if not os.path.isfile(jit_path):
        logging.error("Error: Couldn't find JIT at {}".format(jit_path))
        raise RuntimeError("Missing JIT")

    files.append(jit_path)

    # Next, look for any and all cross-compilation JITs. These are named, e.g.:
    #   clrjit_unix_x64_x64.dll
    #   clrjit_universal_arm_x64.dll
    #   clrjit_universal_arm64_x64.dll
    # and so on, and live in the same product directory as the primary JIT.
    #
    # Note that the expression below explicitly filters out the primary JIT since we added that above.
    # We handle the primary JIT specially so we can error if it is missing. For the cross-compilation
    # JITs, we don't bother trying to ensure that all the ones we might expect are actually there.
    #
    # We don't do a recursive walk because the JIT is also copied to the "sharedFramework" subdirectory,
    # so we don't want to pick that up.

    if coreclr_args.host_os == "osx":
        allowed_extensions = [ ".dylib" ]

View on GitHub (pinned to 60108ba66e)