Class LegacyDocOptions

Namespace

DocToolkit

Assembly

DocToolkit.Primitives.dll

Controls how DocToDocxConverter treats content a Word 97-2003 binary document holds but a .docx package cannot be given.

public sealed class LegacyDocOptions

Inheritance

object

LegacyDocOptions

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

There is one setting, and it defaults to the safe answer. A legacy .doc stores pictures, drawings and form fields in a binary Data stream that the import can see but cannot carry across. By default the conversion refuses rather than silently producing a .docx missing those payloads — set AllowContentLoss to accept the loss deliberately.

This is the common case, and the measurement is now at scale. Across 111 real .doc files from a public .gov crawl, measured 2026-08-17, the default succeeded on 12 of them

• 11%, while the opt-in succeeded on 99 (89%). A .doc holding a table carries the stream; plain text, bold runs and headings do not. Tables are ordinary, so expect the opt-in to be needed rather than to be an exception - and prefer DocToDocxConverter.ConvertWithReport, which returns the same bytes and names what was dropped.

Reading is never affected. DocToDocxConverter.ExtractText takes no options and never refuses: text is not what the binary stream holds, so there is nothing for a policy to decide.

Properties

AllowContentLoss

Convert even when the source holds content the .docx cannot carry, instead of throwing. Defaults to false.

public bool AllowContentLoss { get; init; }

Property Value

bool

Remarks

What is lost is the unprojected binary payload — pictures, drawings and form fields. Measured 2026-08-16: text, tables (every cell), and character formatting such as bold all survive the conversion intact, so this accepts a specific, bounded loss rather than a general "best effort".

Prefer DocToDocxConverter.ConvertWithReport when setting this: it returns the same bytes and tells you exactly what was dropped, so the loss is recorded rather than merely permitted.