Bump version to 3.14.2 and write release notes

README.md

@@ -1,6 +1,6 @@
 [![CI](https://github.com/amrali-eg/EncodingChecker/actions/workflows/ci.yml/badge.svg)](https://github.com/amrali-eg/EncodingChecker/actions/workflows/ci.yml)
 
-# EncodingChecker v3.14.1
+# EncodingChecker v3.14.2
 
 EncodingChecker is a Windows tool for finding, checking, and safely converting text-file encodings. Use the GUI for everyday work or the command line for repeatable jobs.
 

docs/RELEASE-NOTES-v3.14.2.md

@@ -0,0 +1,105 @@
+# EncodingChecker v3.14.2
+
+Five fixes to what EC reports when a run is cancelled or interrupted. None change what
+EC detects, converts, or refuses — every fix here is about the record left behind
+afterward being accurate.
+
+## An interrupted run no longer claims to have done work it didn't
+
+The GUI's confirmation dialog runs a "decide" pass before you ever click Convert, and
+that pass marks eligible files `Converted` — the label it uses for "would convert," so a
+preview can show you the plan. If you then cancelled the review, hit a stale plan (a file
+changed underneath it), or the plan couldn't be built at all, that label was never
+corrected. Exporting the CSV afterward could report those same untouched files as
+`Converted`.
+
+The gap was wider than the confirmation dialog: cancelling while the decide pass itself
+was still running — a plain Ctrl+C on a large folder, before you're even asked anything —
+escaped uncaught and left the same stale label behind.
+
+Every one of these paths now marks the affected rows `NotAttempted` before returning,
+reusing the marker already used for a write pass stopped mid-batch. `Run()` wraps the
+whole decide/confirm/write sequence in one cancellation handler for exactly this reason.
+
+The same class of bug existed on the command line. An interrupted `-Apply` or direct
+scan now marks every row the write pass never reached as `NotAttempted` in both the
+console summary and the exported CSV, instead of `-Apply` calling them "unchanged" and a
+direct scan's CSV/journal having no way to tell an interrupted run from a complete one.
+Journal schema **6** adds an explicit `Interrupted` flag for this. A direct scan's journal
+still only covers the files it reached — a direct scan was never a promise to cover the
+whole folder, so there's no "remainder" to mark.
+
+A cancelled scan also used to skip writing its CSV report entirely, discarding evidence
+for files it had already scanned; it now saves the partial report before returning. A
+cancelled preflight (the pass that decides what a `-Plan` would contain) leaves any
+existing plan file untouched and says so on stderr, rather than silently producing
+nothing.
+
+**BL-31, BL-32, BL-33** in the defect ledger record these.
+
+## The CSV export now names the source that was actually used
+
+You can override automatic detection for a file EC refused. Conversion and the journal
+already used your chosen encoding; the CSV kept printing the encoding EC originally
+detected. The file converted correctly — the exported record of what it was converted
+*from* was wrong. The CSV writer now uses the source label captured when the file was
+actually read, for both the GUI export and the CLI `-Report`.
+
+A refused row is the deliberate exception: it still reports the originally detected
+source, because nothing was converted and the diagnostic on that row refers to detection,
+not to a chosen-but-rejected encoding.
+
+## Cancelling a direct scan and `-Apply` now agree on exit codes
+
+An interrupted direct scan returned exit code 4 unconditionally, even when a file it had
+already reached failed to convert before Ctrl+C landed — while an interrupted `-Apply` in
+the same situation returned 3. Processing failure now wins over cancellation on both
+cancellable paths, matching the documented precedence in `docs/CLI.md`, which was itself
+missing the sentence saying so until this release.
+
+## What else is in this release, and why you will not notice it
+
+- **A cancellation timeout no longer drops the last refusal.** If a "would you like to
+  cancel?" prompt timed out right after a button press, the real refusal was replaced
+  with a literal empty value, misreporting a refused press as no button having been
+  found at all. This only affects the GUI smoke suite's own diagnosis of a stuck run;
+  `EncodingChecker.exe` is unaffected.
+- **The smoke suite's own instrumentation got three narrow fixes**: a transient
+  automation failure now leaves a status flag unknown instead of forcing it false,
+  matching how its siblings behave; scoring an empty phase list is now `Inconclusive`
+  instead of vacuously `Passed`; and the suite's own CI-gate self-test now launches
+  `pwsh`, matching what CI actually runs. **EC-32** records the timeout fix.
+- **A regression-test hook was added to the converter** (`BeforeVerifyTemporaryOutput`)
+  so a test can simulate storage corruption at the one point that matters — after writing
+  finishes, before verification and installation — without needing a real disk fault.
+
+## Compatibility
+
+**The journal schema moves from 5 to 6.** The only change is the additive `Interrupted`
+flag; EC never re-reads its own journals, so nothing rejects an older one — this is
+informational, for anyone else's tooling that parses them. The plan schema stays at
+**6**, conversion semantics at **7**. A saved plan from any 3.14.x release still applies
+unchanged.
+
+The one behavior change with a visible before/after:
+
+| Situation | Before | After |
+|---|---|---|
+| Direct scan cancelled, with a reached file failure | exit 4 | exit 3 |
+| GUI cancel/stale-plan/mid-decide cancellation, then CSV export | row: `Converted` | row: `NotAttempted` |
+| CLI `-Apply`/direct scan interrupted, then report/journal | unreached rows: `Unchanged` or absent | unreached rows: `NotAttempted` |
+| CSV source column for an explicit-source conversion | detected encoding | encoding actually used |
+
+Every reason code, report column, and exit code keeps its existing meaning; detection and
+conversion behave exactly as they did in v3.14.1.
+
+## Verification
+
+- 826 tests pass, none skipped (was 756); release build with no warnings
+- The GUI smoke suite passes all ten phases against an ordinary Release build
+- Every regression test added for the cancellation-reporting fixes was mutation-checked:
+  the fix was reverted, the new assertion confirmed to fail with the exact defect
+  described, then the fix restored and the test confirmed passing again
+- `docs/Test-DefectBacklog.ps1` passes at **77 findings — 66 fixed, 7 open**
+- **No four-corpus audit was run.** This release changes no detection or conversion
+  policy, so the checklist does not require one

sources/EncodingChecker/Properties/AssemblyInfo.cs

@@ -41,5 +41,5 @@
 // You can specify all the values, or you can default the Build and Revision Numbers
 // by using the '*' as shown below:
 // [assembly: AssemblyVersion("1.0.*")]
-[assembly: AssemblyVersion("3.14.1.0")]
-[assembly: AssemblyFileVersion("3.14.1.0")]
+[assembly: AssemblyVersion("3.14.2.0")]
+[assembly: AssemblyFileVersion("3.14.2.0")]