[cmaloney]

Truncate beyond current size of I/O objects is documented to change the amount of allocated space. BytesIO supported truncate to shrink but not truncate to extend. Update BytesIO to support extension and fully implement the IOBase.truncate() API.

• Issue: io.BytesIO.truncate does not work as advertised #71448

[read-the-docs-community]

Documentation build overview

📚 cpython-previews | 🛠️ Build #35004900 | 📁 Comparing 7c36430 against main (285d96d)

  🔍 Preview build  

425 files changed · + 12 added · ± 412 modified · - 1 deleted

+ Added

• builtins/constants.html
• builtins/exceptions.html
• builtins/functions.html
• builtins/index.html
• builtins/stdtypes.html
• builtins/threadsafety.html
• builtins/time-complexity.html
• library/asyncio-tools.html
• library/time-complexity.html
• library/tkinter.dialogs.html
• and 2 more...

± Modified

• glossary.html
• index.html
• license.html
• py-modindex.html
• c-api/arg.html
• c-api/bool.html
• c-api/buffer.html
• c-api/bytearray.html
• c-api/bytes.html
• c-api/capsule.html
• and 402 more...

- Deleted

• library/asyncio-policy.html

[github-actions]

This PR is stale because it has been open for 90 days with no activity.

[cmaloney]

@eendebakpt could you take a look at this one?

[eendebakpt]

Could you check the following cases?

data = b"abcdef"
b = io.BytesIO(data)   # buf is `data` itself (exact bytes, no copy)
b.truncate(3)          # 3 < alloc 6: quick exit, still shared
b.truncate(5)          # 5 < alloc 6: quick exit, memset
data                   # b'abc\x00\x00f'  -- an immutable constant mutated

and

b = io.BytesIO(b"abc")
b.truncate(10**9)      # Segmentation fault?

[cmaloney]

Created gh-160781 for some underlying pieces which should be backported that found while updating this

[cmaloney]

Closing until gh-160781 is done. Making this work well on large sizes is a lot easier with that. Will make a new PR after that which covers the extended cases here.

Ended up refactoring resize_buffer_lock_held into two helpers which de-duplicates code and makes the extend easier, but that needs the large size edge cases first (and pulling into this PR is a lot more work then opening a new one)

[cmaloney]

Replacement PR is up: #160933. Properly unshares branches, handles large truncates gracefully.