fix: rollback pending state on tempfile failure in prepare_add

[srpatcha]

Summary

\prepare_add\ updated pending size before \ empfile_in(), which never rolled back on failure. This caused \make_space()\ to panic later when encountering negative available space.

Moved state update after successful tempfile creation and replaced .expect()\ with ?\ operator.

This is split from #2679 per reviewer feedback to keep changes focused.

Changes

• \src/lru_disk_cache/mod.rs: Move pending size update after tempfile success
• \src/server.rs: Replace .expect()\ with ?\ operator

Signed-off-by: Srikanth Patchava spatchava@meta.com

[codecov-commenter]

Codecov Report

❌ Patch coverage is 93.33333% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 74.34%. Comparing base (d11e2e0) to head (0a5cfd8).
⚠️ Report is 11 commits behind head on main.

Files with missing lines	Patch %	Lines
src/lru_disk_cache/mod.rs	96.55%	1 Missing ⚠️
src/server.rs	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2689      +/-   ##
==========================================
+ Coverage   74.17%   74.34%   +0.16%     
==========================================
  Files          70       70              
  Lines       39207    39396     +189     
==========================================
+ Hits        29083    29288     +205     
+ Misses      10124    10108      -16     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[sylvestre]

would it be possible to have a unit test to cover this change? thanks

[srpatcha]

Hi @sylvestre, thank you for the suggestion! Added a regression test in commit 0a5cfd8:

test_prepare_add_rollback_on_tempfile_failure exercises the failure path by:

1. Creating an LruDiskCache normally
2. Removing the cache root directory underneath it
3. Calling prepare_add(), which now fails because tempfile_in() can't create a file in a missing dir
4. Asserting that c.size() == baseline_size (no pending state leaked)

Before the fix, the test would fail with size() == 10 because the old code incremented pending_size before the tempfile creation. The test serves as a regression guard for #2689.

[sylvestre]

please run rustfmt on the change