apache-solr: proposal for an enhanced portfile with a japanese variant

[humem (humem)]

I present an enhanced version of apache-solr Portfile including ja variant for Japanese texts.

Major changes are:

• install not only 'example' directory but also 'dist', 'contrib', 'docs', 'client' and info files
• add ja variant for Japanese texts: install lucene-gosen in contrib; create example/sor-ja; modify and add config files
• enhance bin/solr to enable clustering and to include the ja option

Minor changes are:

• modify the categories to include textproc instead of search
• add me to the maintainers
• remove the check for JavaVM 1.6 because solr seems to run on 1.5
• add notes of instructions
• modify livecheck.url

[hayato_1980@…]

I have some questions and sugesstion.

Major changes are:

• install not only 'example' directory but also 'dist', 'contrib', 'docs', 'client' and info files

I agree.

• add ja variant for Japanese texts: install lucene-gosen in contrib; create example/sor-ja; modify and add config files
• enhance bin/solr to enable clustering and to include the ja option

I suggest separating 3 ports.

• apache-solr pure apache-solr

• lucene-gosen lucene-gosen jar file

• solr-ja config files for japanese and shell script. depend on apache-solr and lucene-gosen

Because, lucene-gosen and configfiles for japanese are not used in general.

Minor changes are:

• modify the categories to include textproc instead of search

please tell me background.

• add me to the maintainers
• remove the check for JavaVM 1.6 because solr seems to run on 1.5
• add notes of instructions

ok.

• modify livecheck.url

please tell me background.

[humem (humem)]

Thank you for your replying! I attach modified portfiles for apache-solr and lucene-gosen.

I suggest separating 3 ports.

• apache-solr pure apache-solr

• lucene-gosen lucene-gosen jar file

• solr-ja config files for japanese and shell script. depend on apache-solr and lucene-gosen

Because, lucene-gosen and configfiles for japanese are not used in general.

It seems to be reasonable to create lucene-gosen port, because it can be used not only solr but also lucene and other java projects. In addition, the license of lucene-gosen is different from that of solr: lucene-gosen is licensed under LGPL although solr is under Apache.

However, it is difficult for me to separate a Japanese config port from the solr one. The Japanese config should be created based on a copy of the original solr config, and I prefer to implement solr-ja as a variant in the solr port.

• modify the categories to include textproc instead of search

please tell me background.

Search category contains nothing but your apache-solr port. Some search applications are categorized in textproc. For example, please see hyperestraier.

• modify livecheck.url

please tell me background.

I think livechek.url should refer to the main host server rather than a mirror.

[humem (humem)]

portfile for lucene-gosen

[humem (humem)]

apache-solr contains Portfile (Porfile2.diff), and patch-solr-ja.diff (patch-solr-ja.2.diff), solr-ja.xml, solr.in (solr.in.2.diff) and velocity.properties in files.

lucene-gosen has Portfile, and mapping-japanese.txt, stoptags_ja.txt and stopwords_ja.txt in files.

Please take a look.

[hayato_1980@…]

Thanks for your reply and answer.

I like separating ports and define variant approach. I feel good your patches.

[humem (humem)]

Committed in r83879 for lucene-gosen, and r83880 for apache-solr.