account for completed utf-8 sequence in parse_string size limit

[Ramya-9353]

Repro: feed a multi-byte UTF-8 string to stream_parser one byte at a time, so each character completes across a write_some boundary; max_string_size/max_key_size is not applied to it and on_string_part/on_key_part report a total that omits those bytes.
Cause: the str8 resume path in parse_string (completing a UTF-8 sequence split across a chunk) emits seq_.length() bytes to the handler without bounds-checking them or adding them to total. do_str1 and parse_escaped both account before their handler calls; str8 was the gap.
Fix: on the str8 path, check seq_.length() against the remaining budget and add it to total before emitting, matching do_str1.

Regression test in test/limits.cpp drives split multi-byte strings and keys through stream_parser a byte at a time; the size limits are skipped before the change and enforced after.

[cppalliance-bot]

An automated preview of the documentation is available at https://1185.json.prtest2.cppalliance.org/libs/json/doc/html/index.html

If more commits are pushed to the pull request, the docs will rebuild at the same URL.

2026-08-13 06:54:42 UTC

[cppalliance-bot]

GCOVR code coverage report https://1185.json.prtest2.cppalliance.org/gcovr/index.html
LCOV code coverage report https://1185.json.prtest2.cppalliance.org/genhtml/index.html
Coverage Diff Report https://1185.json.prtest2.cppalliance.org/diff-report/index.html

Build time: 2026-08-13 06:57:59 UTC

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.92%. Comparing base (3dd600f) to head (5a3cb6c).
⚠️ Report is 4 commits behind head on develop.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #1185      +/-   ##
===========================================
- Coverage    93.92%   93.92%   -0.01%     
===========================================
  Files           91       91              
  Lines         9290     9297       +7     
===========================================
+ Hits          8726     8732       +6     
- Misses         564      565       +1     

Files with missing lines	Coverage Δ
include/boost/json/basic_parser_impl.hpp	98.67% <100.00%> (+<0.01%)	⬆️

... and 1 file with indirect coverage changes

---

Continue to review full report in Codecov by Harness.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 3dd600f...5a3cb6c. Read the comment docs.

🚀 New features to boost your workflow:

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

[cppalliance-bot]

Benchmark test results. More info at https://benchmark.cppalliance.org/jsonbenchmarks-pullrequests/1185/pullrequest.html

[cppalliance-bot]

Benchmark test results. More info at https://benchmark.cppalliance.org/jsonbenchmarks-pullrequests/1185/pullrequest.html

[grisumbras]

So, coverage checks fail on this PR here: https://1185.json.prtest2.cppalliance.org/diff-report/include/boost/json/basic_parser_impl.hpp.html#NL1125

This shows that we do not have a test for when a key has a UTF-8 multibyte character, the parser suspends within it and then resumes. Can you please add a test for that as a separate commit, so that the PR's CI is green?

[Ramya-9353]

Added in a separate commit: same feed shape as the key overflow test, but with string::max_size() - 2 filler bytes so the split character completes on the str8 resume path exactly at the limit and the key parses cleanly. That runs the on_key_part call the diff report flagged as uncovered. Passes locally under ASan/UBSan.

[cppalliance-bot]

Benchmark test results. More info at https://benchmark.cppalliance.org/jsonbenchmarks-pullrequests/1185/pullrequest.html

[grisumbras]

Merged as 5fad60c and c527374.