lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steven Rowe (JIRA)" <>
Subject [jira] [Commented] (LUCENE-4584) Compare the LZ4 implementation in Lucene against the original impl
Date Sun, 02 Dec 2012 00:33:58 GMT


Steven Rowe commented on LUCENE-4584:

bq. Steven: We are a Java project and we have no platform independent way to run C code. I
think Adrien's idea is enough, maybe use 2 or 3 different length examples. Just like Apache
TIKA parading MS Office docs, they also don't start MS Office while testing ♥

I agree Adrien's idea is enough, and in any case is better than what we have now (pnp), but
I still think cross-impl testing would be nice.

cpp-tasks is used to compile NativeUnixDirectory.cpp, so there is precedent for this in our

bq. I'm a little worried about adding the sources of LZ4 and a task to compile C code to the
lucene-core build (or maybe I could do these LZ4 tests in a dedicated module so that it doesn't
make lucene-core depend on native code?).

I think it would be fine to include test-only native code in lucene-core tests, as long as
compilation and testing were both fast.

bq. I think we should also compare the output bytes to make sure the compressor is efficient.
[...] Given that the LZ4 output depends on the endianess of the machine, this can unfortunately
only be done with static files.

You wouldn't need static files if you compared output lengths, though correctness would of
course no longer be implied.

> Compare the LZ4 implementation in Lucene against the original impl
> ------------------------------------------------------------------
>                 Key: LUCENE-4584
>                 URL:
>             Project: Lucene - Core
>          Issue Type: Task
>            Reporter: Adrien Grand
>            Assignee: Adrien Grand
>            Priority: Blocker
>             Fix For: 4.1
> We should add tests to make sure that the LZ4 impl in Lucene compresses data the exact
same way as the original impl.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message