fix: make uproot.recreate physically truncate an existing file

[ariostas]

🤖 AI text below 🤖

Addresses finding 3 of "PR 1" in #1688. This is a regression, with a clear origin.

Before the fsspec migration, uproot.recreate truncated the path itself:

if isinstance(file_path, str):
    with open(file_path, "w"):
        pass
    sink = uproot.sink.file.FileSink(file_path)

#1016 / #1034 moved that into FileSink.__init__, which only truncates when the file does not already exist:

https://github.com/scikit-hep/uproot5/blob/main/src/uproot/sink/file.py#L52-L58

so recreating over an existing file opens it "r+b" and leaves every byte past the new fEND in place. Recreating a 100 kB path as an empty ROOT file left the physical size at 100 kB with fEND at 1658.

Restores the truncation in recreate, which is also the path create delegates to.

• uproot.update is unaffected and keeps its "r+b" behavior.
• uproot.create still raises FileExistsError before reaching this point.
• Recreating from a file-like object is unchanged — it was not truncated before the migration either. Happy to extend it there too if you'd prefer.

Tests

tests/test_1688_recreate_truncates.py: physical size after recreating over a large non-ROOT file and over a larger ROOT file, plus guards that recreate still creates missing files and parent directories, that create still refuses to overwrite without touching the file, and that update does not truncate. 2 of the 5 fail on main.

Full suite passes locally (1028 passed, 90 skipped).

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 74.88%. Comparing base (e598b61) to head (d60e1b1).

Additional details and impacted files

Files with missing lines	Coverage Δ
src/uproot/writing/writable.py	79.89% <100.00%> (+0.05%)	⬆️