Skip to main content

GZipDecompressFile

Function GZipDecompressFile( const aFileName : string) : TBytes

Example​

procedure ScriptEvent(var Value: variant);
begin
  { Current behaviour: this encodes the original .gz file bytes, not its payload. }
  Value := Base64EncodeBytes(GZipDecompressFile('C:\Example\input.gz'));
end;

Usage​

Do not use GZipDecompressFile when expanded content is required. Despite its name, current Velox returns the file's original compressed bytes unchanged.

Use GZipDecompress(LoadFileToBytes(Path)) instead, after applying appropriate file-size, expansion-size and trust limits. The replacement reads and expands the complete content in memory.

Parameters

NameTypeDescription
aFileNameString, constPath of the file to read under the executing Velox process identity. The extension is not inspected.

Returns

The original file bytes in a new TBytes value. This is current defective behaviour, not the intended decompressed result.

Behaviour

For a stable, readable, non-empty file, the result is a byte-for-byte copy of the file. A .gz extension has no special effect. The source is not modified.

Errors

File-open, access, read, allocation, stream-construction and range errors propagate. A well-formedness error in the gzip data normally does not occur because the gzip bytes are never read through the decompressor.

Usage notes

Treat scripts using this identifier as candidates for remediation. A safe replacement pattern is GZipDecompress(LoadFileToBytes(Path)), but it still inherits the custom decompressor's unbounded-expansion and truncated-input loop, so it requires explicit source trust and expected-size controls.

Additional Technical Info​

Warning - current implementation defect: GZipDecompressFile does not decompress the file in the current product source. It returns the bytes stored in the nominated file unchanged. Do not use this function where expanded content is required. Load the file bytes and call GZipDecompress only after applying the size and trust controls described on that page.

The fictional example demonstrates the actual result so an existing script can be recognised and corrected. It is source-reviewed and is not executed by the documentation workflow.

Implementation

The direct registration calls vxCompression.GZipDecompressFile. It opens the file for shared reading and creates an empty TBytesStream. It then constructs a TZDecompressionStream over that empty byte stream, but never reads from the decompression stream. Instead, it copies lFile.Size bytes directly from the file into the underlying byte stream and copies those same bytes to the result. Freeing the unused decompression stream cannot retroactively transform the bytes.

Edge cases and quirks

  • The implementation defect is unconditional in the traced source: no decompression read occurs.
  • The file is opened with fmShareDenyNone, so another process can write it while the copy is running. The requested copy count is the size observed before copying.
  • A full copy is held in the byte stream and then copied again to the result.
  • The file, byte stream and decompression stream are constructed before entering the protected block. Constructor failures can leak earlier resources.
  • For a zero-byte file, the exact-size Move still indexes element zero of both empty arrays. Current Velox project configurations enable range checking, so an empty file raises a range-check exception even though the move count is zero.

Side effects

Opens and reads the nominated file. It performs no destination write.

Performance and concurrency

The operation is a synchronous whole-file memory copy with an unnecessary decompression-stream allocation. Concurrent writers can make the read inconsistent. No shared product state is changed.

Related entries

External references

Created 2026-07-15