System.IO: Fix StringStream reads with empty encoder fallback

[jozkee]

Summary

Fix StringStream.Read so it does not return 0 while encoded input remains after an encoder consumes characters without producing bytes.

With UTF-8 and an empty encoder replacement fallback, invalid UTF-16 characters can be consumed while producing no output. The small-buffer streaming path previously interpreted that zero-byte conversion as EOF even when later characters could still produce bytes.

The streaming path now continues conversion while:

• no bytes have been produced,
• source characters remain, and
• the encoder consumed characters.

It stops after producing output or if the encoder makes no progress. The existing single-shot fast path and async delegation remain unchanged. This follows the retry-until-output-or-EOF precedent used by TranscodingStream.

Tests

Added coverage for:

• synchronous and asynchronous reads,
• one-byte streaming and large-buffer fast paths,
• two and four invalid surrogate characters before valid output,
• ReadByte, and
• input whose characters all encode to zero bytes in both paths.

Validation

• .\build.cmd clr+libs -rc checked — succeeded, 0 warnings, 0 errors
• .\build.cmd clr.corelib+clr.nativecorelib+libs.pretest -rc checked — succeeded, 0 warnings, 0 errors
• dotnet build src\libraries\System.Runtime\tests\System.IO.Tests\System.IO.Tests.csproj /t:Test '/p:XUnitOptions=-class System.IO.Tests.StringStreamTests_String_Misc' — 16 passed, 0 failed
• dotnet build src\libraries\System.Runtime\tests\System.IO.Tests\System.IO.Tests.csproj /t:Test /p:testnobuild=true — 2,424 passed, 0 failed

Note

This pull request description was generated with GitHub Copilot.

[azure-pipelines]

Azure Pipelines:
Successfully started running 3 pipeline(s).
13 pipeline(s) were filtered out due to trigger conditions.
There may be pipelines that require an authorized user to comment /azp run to run.

[dotnet-policy-service]

Tagging subscribers to this area: @dotnet/area-system-io
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

This PR fixes System.IO.StringStream.Read to avoid returning 0 (EOF) when the encoder consumed input characters but produced zero bytes (e.g., UTF-8 with an empty EncoderReplacementFallback), ensuring reads continue until bytes are produced or the encoder makes no progress. This aligns Read / ReadAsync / ReadByte behavior with expected stream semantics where 0 indicates end-of-stream.

Changes:

• Update StringStream.Read(Span<byte>) to loop when no bytes have been produced yet but the encoder consumed characters, preventing premature EOF.
• Add targeted tests covering sync/async reads, ReadByte, fast-path vs 1-byte streaming path, and “all input skipped” cases under empty encoder replacement fallback.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
src/libraries/System.Private.CoreLib/src/System/IO/StringStream.cs	Adjusts the streaming conversion loop to retry when characters were consumed but no bytes were produced, avoiding false EOF.
src/libraries/System.Runtime/tests/System.IO.Tests/StringStream/StringStreamTests_String.cs	Adds regression tests for empty encoder replacement fallback scenarios across sync/async paths and ReadByte.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated no new comments.

[ViveliDuCh]

LGTM, thanks for catching this!