Mailing-List: contact issues-help@commons.apache.org; run by ezmlm
Precedence: bulk
Reply-To: issues@commons.apache.org
Date: Mon, 11 Aug 2014 05:48:12 +0000 (UTC)
From: "Stefan Bodewig (JIRA)" <jira@apache.org>
To: issues@commons.apache.org
Message-ID: <JIRA.12725219.1404380956707.53379.1407736092474@arcas>
In-Reply-To: <JIRA.12725219.1404380956707@arcas>
References: <JIRA.12725219.1404380956707@arcas>
Subject: [jira] [Commented] (COMPRESS-285) checking of availability of XZ
 compression is expensive - result should be reused
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable


    [ https://issues.apache.org/jira/browse/COMPRESS-285?page=3Dcom.atlassi=
an.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D14=
092467#comment-14092467 ]=20

Stefan Bodewig commented on COMPRESS-285:
-----------------------------------------

Thanks Sebb, I think your two suggestions are good ideas and will see to im=
plementing them the coming week, in particular you will only pay for the fa=
iled XZ check if you are really trying to uncompress XZ streams.

The additional constructor won't help Wojciech since he's using Compress be=
hind Tika, Tika would need to get adapted to the new constuctor and in the =
end implement its own logic which would also need to take OSGi contexts int=
o account.  I think it might be a good idea to add an explicit flag whether=
 the result is cacheable and make that flag default to true unless BundleEv=
ent can be loaded - Wojciech would then need to set the flag explicitly.

> checking of availability of XZ compression is expensive - result should b=
e reused
> -------------------------------------------------------------------------=
--------
>
>                 Key: COMPRESS-285
>                 URL: https://issues.apache.org/jira/browse/COMPRESS-285
>             Project: Commons Compress
>          Issue Type: Improvement
>          Components: Compressors
>    Affects Versions: 1.5, 1.6, 1.7, 1.8
>         Environment: linux 64-bit, java 7, glassfish, solr, tika
>            Reporter: Wojciech =C5=81ozowicki
>            Priority: Minor
>              Labels: performance
>
> I use solr with apache tika for indexing documents. Tika uses commons-com=
press to handle compressed files. Using sampler (jvisualvm) I have seen tha=
t quite a lot of time (5-7%) during my tests is spent in XZUtils.isXZCompre=
ssionAvailable because of unavailable XZ compression (I guess for each time=
 classloaders spend some time looking for unavailable classes, then NoClass=
DefFoundError).
> I think the result of the first check should be stored and reused.
> Here is the stacktrace (just to show the way tika is using commons-compre=
ss):
> org.apache.commons.compress.compressors.xz.XZUtils.isXZCompressionAvailab=
le(XZUtils.java:52)
> =09at org.apache.commons.compress.compressors.CompressorStreamFactory.cre=
ateCompressorInputStream(CompressorStreamFactory.java:140)
> =09at org.apache.tika.parser.pkg.ZipContainerDetector.detectCompressorFor=
mat(ZipContainerDetector.java:95)
> =09at org.apache.tika.parser.pkg.ZipContainerDetector.detect(ZipContainer=
Detector.java:81)
> =09at org.apache.tika.detect.CompositeDetector.detect(CompositeDetector.j=
ava:61)


--
This message was sent by Atlassian JIRA
(v6.2#6252)