Lzma fails on large files

compressing and uncompressing using Lzma fails unexpectedly with the following snippet:
```kotlin
myBigStringFileContents.encodeToByteArray().compress(Lzma).uncompress(Lzma)
```

With the following error:

```java
java.lang.Exception: Error in data stream
at korlibs.io.compression.lzma.Lzma.uncompress(Lzma.kt:29)
at korlibs.io.compression.CompressionMethodKt.uncompress(CompressionMethod.kt:43)
at korlibs.io.compression.CompressionMethodKt$uncompress$2.invokeSuspend(CompressionMethod.kt:56)
at korlibs.io.compression.CompressionMethodKt$uncompress$2.invoke(CompressionMethod.kt)
at korlibs.io.compression.CompressionMethodKt$uncompress$2.invoke(CompressionMethod.kt)
at korlibs.io.async.RunBlockingNoSuspensionsKt.startCoroutineUndispatched(runBlockingNoSuspensions.kt:73)
at korlibs.io.async.RunBlockingNoSuspensionsKt.runBlockingNoSuspensions(runBlockingNoSuspensions.kt:40)
at korlibs.io.compression.CompressionMethodKt.uncompress(CompressionMethod.kt:55)
at korlibs.io.compression.CompressionMethodKt.uncompress(CompressionMethod.kt:63)
at korlibs.io.compression.CompressionMethodKt.uncompress$default(CompressionMethod.kt:63)
```

Start by reproducing the Kotlin snippet and inspect korlibs.io.compression.lzma.Lzma.kt at line 29, then follow the call through CompressionMethod.kt lines 43 and 56. Done means large-file content can be compressed and uncompressed successfully without the data stream error.

Written by the indexing model from the issue text.