block: eliminate byte concat copies in peek() with a memoryview

[kroketio]

When writing a 19mb file to disk (Volume(...); volume.inode_at("/test_file"); f.write(inode.open().read()) I noticed the file writing was slow, so I profiled it, and most time was spent in block.py:peek() (21 seconds)

python3 -m pstats profile.out
Welcome to the profile statistics browser.
profile.out% sort cumulative
profile.out% stats 20
Tue Mar 10 10:21:09 2026    profile.out

         195853 function calls (194557 primitive calls) in 21.915 seconds

   Ordered by: cumulative time
   List reduced from 1057 to 20 due to restriction <20>

   ncalls  tottime  percall  cumtime  percall filename:lineno(function)
     88/1    0.002    0.000   21.915   21.915 {built-in method builtins.exec}
        1    0.000    0.000   21.915   21.915 /home/dsc/foo/bar/test.py:1(<module>)
       42    0.000    0.000   21.848    0.520 /home/dsc/foo/bar/ext4/block.py:93(read)
       42   21.759    0.518   21.848    0.520 /home/dsc/foo/bar/ext4/block.py:105(peek)
     4689    0.008    0.000    0.088    0.000 /home/dsc/foo/bar/ext4/block.py:38(__getitem__)
     4689    0.012    0.000    0.061    0.000 /home/dsc/foo/bar/ext4/extent.py:42(__getitem__)
     98/2    0.000    0.000    0.040    0.020 <frozen importlib._bootstrap>:1165(_find_and_load)
     98/2    0.000    0.000    0.040    0.020 <frozen importlib._bootstrap>:1120(_find_and_load_unlocked)
     89/3    0.000    0.000    0.040    0.013 <frozen importlib._bootstrap>:666(_load_unlocked)
     71/3    0.000    0.000    0.040    0.013 <frozen importlib._bootstrap_external>:934(exec_module)
    207/5    0.000    0.000    0.040    0.008 <frozen importlib._bootstrap>:233(_call_with_frames_removed)
     4731    0.005    0.000    0.027    0.000 /home/dsc/foo/bar/ext4/volume.py:167(read)
        1    0.000    0.000    0.020    0.020 /home/dsc/foo/bar/admp_src/magic.py:1(<module>)
        1    0.000    0.000    0.019    0.019 /home/dsc/foo/bar/ext4/__init__.py:1(<module>)
     9930    0.008    0.000    0.018    0.000 /home/dsc/foo/bar/ext4/extent.py:39(__contains__)
     4815    0.015    0.000    0.015    0.000 {method 'read' of '_io.BufferedReader' objects}
       77    0.013    0.000    0.013    0.000 {method '__exit__' of '_io._IOBase' objects}
        1    0.011    0.011    0.011    0.011 {method 'write' of '_io.BufferedWriter' objects}
        1    0.000    0.000    0.011    0.011 /home/dsc/foo/bar/ext4/volume.py:1(<module>)
  370/364    0.003    0.000    0.009    0.000 {built-in method builtins.__build_class__}

With this commit file writing is a lot faster, 100ms for 50mb on my machine.

Summary by CodeRabbit

• Refactor

  • More efficient block-data assembly and safer handling of missing/partial blocks, improving memory use and read/peek reliability for large or sparse files.

• Tests

  • Expanded test coverage and stress scenarios: increased iteration counts, more generated test files and attribute operations, and additional runtime validations for block I/O peek/seek behavior.

[coderabbitai]

Warning

Rate limit exceeded

@Eeems has exceeded the limit for the number of commits that can be reviewed per hour. Please wait 21 minutes and 36 seconds before requesting another review.

⌛ How to resolve this issue?

After the wait time has elapsed, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

We recommend that you space out your commits to avoid hitting the rate limit.

🚦 How do rate limits work?

CodeRabbit enforces hourly rate limits for each developer per organization.

Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout.

Please see our FAQ for further information.

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 24519102-9ab0-4137-8d26-487d2b43fd8a

📥 Commits

Reviewing files that changed from the base of the PR and between 7d8f8b7 and 14b0d0a.

📒 Files selected for processing (2)

• _test_image.sh
• test.py

📝 Walkthrough

Walkthrough

Adds a typed constructor and internal _null_block to BlockIOBlocks, changes block lookup to fall back to _null_block, refactors BlockIO.peek to build results from memoryview slices, and extends tests and the test image script to exercise larger ranges and more attribute operations.

Changes

Cohort / File(s)	Summary
BlockIO memoryview refactor ext4/block.py	Adds def __init__(self, blockio: "BlockIO") and self.blockio: BlockIO, introduces internal _null_block, updates __getitem__ to return _null_block when missing, and refactors BlockIO.peek/read logic to accumulate memoryview slices and concatenate them into final bytes. Review attention: correctness of _null_block semantics, memoryview slicing/concatenation, offset handling, and potential edge-case performance.
Expanded test script _test_image.sh	Increases initial loop range from 1..100 to 1..1000 and adds a new loop creating 100 test$i.txt files with repeated setfattr calls, increasing IO and attr operations to exercise filesystem behaviors. Review attention: test runtime, disk usage, and idempotency.
Test assertions & seeks test.py	Adds runtime checks for BlockIO behavior: asserts b.peek(0) falsy, computes size = volume.block_size + 1, asserts len(b.peek(size)) == size, and adds a seek to volume.block_size - 5. Review attention: correctness of new assertions, seek semantics relative to block boundaries.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• Fix #14 #16 — also modifies ext4/block.py's BlockIO.peek; related work on block accumulation and EOF/clamping behavior.

Poem

🐰 I hopped on blocks and nudged a null,
Collected views where bytes were full,
Sliced the start and joined the rest,
Overflow tested, attributes pressed—
A tiny rabbit's file-system lullaby. 🥕

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately and specifically describes the main change: replacing byte concatenation copies in peek() with a memoryview-based approach for performance improvement.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[Eeems]

When writing a 19mb file to disk (Volume(...); volume.inode_at("/test_file"); f.write(inode.open().read()) I noticed the file writing was slow, so I profiled it, and most time was spent in block.py:peek() (21 seconds

I assume you mean file reading was slow, unless you are working on something rather interesting?

[kroketio]

When writing a 19mb file to disk (Volume(...); volume.inode_at("/test_file"); f.write(inode.open().read()) I noticed the file writing was slow, so I profiled it, and most time was spent in block.py:peek() (21 seconds

I assume you mean file reading was slow, unless you are working on something rather interesting?

sorry yes, reading, working on something boring; https://git.maemo.org/sanderfoobar/admp

cheers!

[Eeems]

... working on something boring; https://git.maemo.org/sanderfoobar/admp

cheers!

@kroketio You should include the license files in the vendored dependency folders to be compliant with the MIT license terms for each of them. Interesting project though!

[kroketio]

@Eeems done, thanks.