Oddity in Solr artifact size differences now that docs are no longer included

[Shawn Heisey <ap...@elyograg.org>]

One of the changes mentioned for Solr 6.5.0 is the removal of the
javadocs from the download.  Due to what I observed about the size of
the docs and how slow they were to extract from the .zip format, I've
been looking forward to this.  Unfortunately, it doesn't seem to have
helped the .tgz download size.

Here are the sizes shown in the download archive for 6.4.2:

solr-6.4.2.tgz: 140M
solr-6.4.2.zip: 150M

These are the download sizes for 6.5.0:

solr-6.5.0.tgz: 142M
solr-6.5.0.zip: 143M

The change in the zip size (minus 7 megabytes) is about what I expected
... but the tgz archive has gotten BIGGER.  After a ton of fiddling
where I uncompressed and recompressed a whole bunch of different things
from both versions, I FINALLY figured out *why* there is such a difference.

There's little difference in the compression ratio between the two
formats for everything that the 6.5.0 archive contains, but when I
compressed *just* the docs directory from 6.4.2 to both formats, this is
what I ended up with when I ran "du" on the resulting files (numbers in KB):

5896    d1.tgz
13852   d1.zip

The zipped version of the docs is nearly 8 MB larger than the compressed
tarball, so the zip archive was affected more by the removal than tgz
was.  The tgz archive lost about 6 MB from docs, but that loss was more
than made up by additional dependency jars and increases in the sizes of
the jars that we build.

There's nothing we can do about this, but I thought others might find
the information to be interesting and possibly helpful if a user asks.

Thanks,
Shawn


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org

Re: Oddity in Solr artifact size differences now that docs are no longer included

[Alexandre Rafalovitch <ar...@gmail.com>]

The conclusion I came to when I saw that was that we now ship several
more 3rd party libraries that are quite big. I think most of them are
due to the JDBC implementation changes.

Regards,
   Alex.
----
http://www.solr-start.com/ - Resources for Solr users, new and experienced


On 7 April 2017 at 17:33, Shawn Heisey <ap...@elyograg.org> wrote:
> One of the changes mentioned for Solr 6.5.0 is the removal of the
> javadocs from the download.  Due to what I observed about the size of
> the docs and how slow they were to extract from the .zip format, I've
> been looking forward to this.  Unfortunately, it doesn't seem to have
> helped the .tgz download size.
>
> Here are the sizes shown in the download archive for 6.4.2:
>
> solr-6.4.2.tgz: 140M
> solr-6.4.2.zip: 150M
>
> These are the download sizes for 6.5.0:
>
> solr-6.5.0.tgz: 142M
> solr-6.5.0.zip: 143M
>
> The change in the zip size (minus 7 megabytes) is about what I expected
> ... but the tgz archive has gotten BIGGER.  After a ton of fiddling
> where I uncompressed and recompressed a whole bunch of different things
> from both versions, I FINALLY figured out *why* there is such a difference.
>
> There's little difference in the compression ratio between the two
> formats for everything that the 6.5.0 archive contains, but when I
> compressed *just* the docs directory from 6.4.2 to both formats, this is
> what I ended up with when I ran "du" on the resulting files (numbers in KB):
>
> 5896    d1.tgz
> 13852   d1.zip
>
> The zipped version of the docs is nearly 8 MB larger than the compressed
> tarball, so the zip archive was affected more by the removal than tgz
> was.  The tgz archive lost about 6 MB from docs, but that loss was more
> than made up by additional dependency jars and increases in the sizes of
> the jars that we build.
>
> There's nothing we can do about this, but I thought others might find
> the information to be interesting and possibly helpful if a user asks.
>
> Thanks,
> Shawn
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: dev-help@lucene.apache.org
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org