Remove unnecessary InputStream.close call

[eric-maynard]

#1942 changed the way that the bootstrap init script is handled, but it added an extra InputStream.close call that shouldn't be needed after the BufferedReader here is closed. This PR removes that extra call.

[snazy]

The current code is correct. This change leaves the risk of not closing the input stream.

[eric-maynard]

What makes you say that, @snazy? Have you found a JDK that behaves that way?

Oracle documents BufferedReader#close as [Closing] the stream and releases any system resources associated with it. here and InputStreamReader#close the same way here. I believe this comes from the javadoc of Reader itself.

Looking at OpenJDK, the BufferedReader will close the InputStreamReader, and then InputStreamReader will close the stream.

[snazy]

@eric-maynard you see that a) the BufferedReader is created somewhere inside some function calls and b) that the given stream should be closed nontheless, right?

[eric-maynard]

a) the BufferedReader is created somewhere inside some function calls

What function calls? It's created inside a try-with-resources which is visible in the diff here.

that the given stream should be closed nontheless, right?

That is what the try-with-resources is for

[dimas-b]

Regardless of the .close() call, I just realised that runWithinTransaction() has re-tries... So if this code is called again, the reading of the scriptInputStream will lead to invalid data. WDYT?

[eric-maynard]

Yeah, to fix that I think we need to go back to having two try-with-resources. Still, the extra close call can be removed. Just pushed a fix.

[dimas-b]

Thx for the fix @eric-maynard ! It LGTM 👍