fix(stdlib): string writers terminate their own output - #1817
Open
ghaith wants to merge 2 commits into
Open
Conversation
Build Artifacts🪟 Windows
From workflow run 🐧 Linux
From workflow run |
The *_TO_STRING writers relied on the caller zero-filling the whole result buffer: they wrote content without a trailing terminator, and TIME_TO_STRING_EXT even located its write position by scanning for the first zero byte. STRING_TO_WSTRING_EXT, CHAR_TO_STRING and WCHAR_TO_WSTRING had the same gap; the string_functions writers already terminate through the encoders. Every writer now terminates what it writes (via a cursor-tracking write_terminated helper for the formatted conversions), so results stay well-formed regardless of the destination buffer's prior contents. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
ghaith
force-pushed
the
fix/stdlib-string-writers-terminate
branch
from
July 21, 2026 11:39
6dab235 to
6d4efe1
Compare
Port the terminated-write fix onto the TIME/LTIME split introduced by #1801: the write_time/dt/date/tod_to_string helpers now use write_terminated, and the regression test drives the L-variant entry points. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem
The stdlib string writers assume the caller hands them a zero-filled result buffer: the formatted
*_TO_STRING_EXTconversions write their content without a trailing terminator,TIME_TO_STRING_EXTeven locates its write position by scanning for the first zero byte, andSTRING_TO_WSTRING_EXT,CHAR_TO_STRINGandWCHAR_TO_WSTRINGleave the terminator to the pre-zeroed buffer. Any caller that hands over a dirty buffer gets unterminated or corrupted results.Fix
Every writer now terminates its own output. The formatted conversions go through a cursor-tracking
write_terminatedhelper that appends the terminator after the written content;TIME_TO_STRING_EXTtracks its write offset instead of scanning for zero bytes and terminates up front so a zero duration yields an empty string.Testing
New unit tests call every affected writer on a
0xAA-filled buffer and assert the result reads back correctly up to its terminator.🤖 Generated with Claude Code