fix: correct newline handling for p and x commands

[pocopepe]

Fixes #306 and #143

• Hold space was initialized without has_newline: true, causing 1x to output nothing instead of an empty line when input has no trailing newline
• p was unconditionally appending \n, breaking -n p on no-newline input

Fixed by initializing hold space as newline terminated and adding a pending_newline mechanism in OutputBuffer that defers newline injection until the next write, so -n mode never triggers it
Also incidentally fixes the g/G empty hold space behavior from #266 (overlaps with open PR #263 same bug, different approach)

[codecov]

Codecov Report

❌ Patch coverage is 78.57143% with 21 lines in your changes missing coverage. Please review.
✅ Project coverage is 82.25%. Comparing base (7b204c0) to head (22b5356).

Files with missing lines	Patch %	Lines
src/sed/fast_io.rs	79.34%	18 Missing and 1 partial ⚠️
src/sed/processor.rs	0.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #378      +/-   ##
==========================================
- Coverage   82.25%   82.25%   -0.01%     
==========================================
  Files          13       13              
  Lines        5438     5522      +84     
  Branches      304      310       +6     
==========================================
+ Hits         4473     4542      +69     
- Misses        963      977      +14     
- Partials        2        3       +1     

Flag	Coverage Δ
macos_latest	82.94% <93.90%> (+0.23%)	⬆️
ubuntu_latest	83.04% <93.90%> (+0.23%)	⬆️
windows_latest	0.00% <0.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.

[codspeed-hq]

Merging this PR will degrade performance by 6.47%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 1 improved benchmark
❌ 3 regressed benchmarks
✅ 7 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	access_log_no_op	747.7 ms	781.1 ms	-4.27%
❌	access_log_append	634.7 ms	669.7 ms	-5.23%
⚡	access_log_all_del	541.5 ms	523.7 ms	+3.39%
❌	no_op_short	964 ms	1,030.6 ms	-6.47%

---

_{Comparing pocopepe:issue-306-143 (22b5356) with main (7b204c0)}

[dspinellis]

LGTM. Please squash the commits into a single one.