Class DocxCompare

Namespace

DocToolkit

Assembly

DocToolkit.Docx.dll

Compares two versions of a document and returns the later one with the differences marked as tracked changes (A118).

public static class DocxCompare

Inheritance

object

DocxCompare

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

The result is an ordinary .docx carrying revisions, not a report — so Word shows it the way it shows any tracked-changes document, and DocxReview reads and applies it without knowing it came from here: Inspect(byte[]) returns a DocxReviewReport whose Revisions lists them, and AcceptRevisions(byte[]) / RejectRevisions(byte[]) resolve them.

Every member named above is a cref, not a code span, on purpose. This paragraph shipped naming DocxEditor.Revisions — a type that has no such member — and nothing caught it: a wrong name inside <c> is prose, while a wrong cref is CS1574 and this repository builds with TreatWarningsAsErrors. The package README had it right the whole time, which is the usual shape of the drift.

Paragraph text is what gets compared, and everything else is REPORTED rather than silently skipped. Tables, lists and formatting changes are named in the warnings of CompareWithReport(byte[], byte[], string). A comparison that quietly mis-marked a table would be worse than one that says what it did not look at — the same contract DocToDocxConverter offers through LegacyDocOptions.AllowContentLoss.

A formatting-only change is not detected at all, and is not reported as a text change either. DocxRevisionKind has no formatting member, and inventing one to describe something this does not measure would be worse than the gap.

Methods

Compare(byte[], byte[], string)

Returns revised with its differences from original marked as tracked insertions and deletions.

public static byte[] Compare(byte[] original, byte[] revised, string author)

Parameters

original byte[]

The earlier version.

revised byte[]

The later version, which the result is built from.

author string

The name recorded against each revision.

Returns

byte[]

Remarks

Only paragraph text is compared. Use CompareWithReport(byte[], byte[], string) to find out what was not — this overload discards that, and a caller who does not know whether the documents contain tables should not be using it.

Comparing a document with itself produces no revisions, rather than a document marked entirely rewritten.

The cost is dominated by the differing text rather than by document size: the shared prefix and suffix are removed first, so two revisions of one document are cheap and two unrelated documents are the bounded worst case.

Exceptions

ArgumentNullException

Any argument is null.

ArgumentException

Either document is empty, or author is blank.

DocumentConversionException

Either package could not be opened or edited.

CompareWithReport(byte[], byte[], string)

Compares two documents and reports what the comparison did not look at.

public static ConversionResult<byte[]> CompareWithReport(byte[] original, byte[] revised, string author)

Parameters

original byte[]

revised byte[]

author string

Returns

ConversionResult<byte[]>

The marked-up document, with a warning for every construct present but not compared. HasLoss is true whenever anything was skipped, which is the signal that the verdict covers less than the document.

Remarks

Only paragraph text is compared. Use CompareWithReport(byte[], byte[], string) to find out what was not — this overload discards that, and a caller who does not know whether the documents contain tables should not be using it.

Comparing a document with itself produces no revisions, rather than a document marked entirely rewritten.

The cost is dominated by the differing text rather than by document size: the shared prefix and suffix are removed first, so two revisions of one document are cheap and two unrelated documents are the bounded worst case.