[Dovecot] doveadm fts optimize undocumented
Subject says all. Timo, would you write something up where the world will find it? Also, it would be great if doveadm fts was covered in the doveadm manpage.
Thanks,
-- Dave Abrahams BoostPro Computing Software Development Training http://www.boostpro.com Clang/LLVM/EDG Compilers C++ Boost
On 4.12.2012, at 18.46, Dave Abrahams wrote:
Subject says all. Timo, would you write something up where the world will find it? Also, it would be great if doveadm fts was covered in the doveadm manpage.
It runs the Lucene's optimize function with fts-lucene backend. Does nothing with others. Basically I think it just gathers multiple Lucene index files together and packs them into a single file, dropping all deleted data (Lucene documentation has more details I think). Similar to what "doveadm purge" does to mdbox files. fts-lucene runs the optimize also internally automatically once in a while, so I'm not sure if that doveadm command should necessarily even exist.
on Tue Dec 04 2012, Timo Sirainen <tss-AT-iki.fi> wrote:
On 4.12.2012, at 18.46, Dave Abrahams wrote:
Subject says all. Timo, would you write something up where the world will find it? Also, it would be great if doveadm fts was covered in the doveadm manpage.
It runs the Lucene's optimize function with fts-lucene backend. Does nothing with others. Basically I think it just gathers multiple Lucene index files together and packs them into a single file, dropping all deleted data (Lucene documentation has more details I think). Similar to what "doveadm purge" does to mdbox files. fts-lucene runs the optimize also internally automatically once in a while, so I'm not sure if that doveadm command should necessarily even exist.
I think the answer lies in your perception of benefits vs maintainance (including documentation) costs... ;-)
-- Dave Abrahams BoostPro Computing Software Development Training http://www.boostpro.com Clang/LLVM/EDG Compilers C++ Boost
On 4.12.2012, at 18.59, Dave Abrahams wrote:
on Tue Dec 04 2012, Timo Sirainen <tss-AT-iki.fi> wrote:
On 4.12.2012, at 18.46, Dave Abrahams wrote:
Subject says all. Timo, would you write something up where the world will find it? Also, it would be great if doveadm fts was covered in the doveadm manpage.
It runs the Lucene's optimize function with fts-lucene backend. Does nothing with others. Basically I think it just gathers multiple Lucene index files together and packs them into a single file, dropping all deleted data (Lucene documentation has more details I think). Similar to what "doveadm purge" does to mdbox files. fts-lucene runs the optimize also internally automatically once in a while, so I'm not sure if that doveadm command should necessarily even exist.
I think the answer lies in your perception of benefits vs maintainance (including documentation) costs... ;-)
Based on the awesome explanation I just gave, I wouldn't mind if someone converts something based on it to a man page patch!
on Tue Dec 04 2012, Timo Sirainen <tss-AT-iki.fi> wrote:
On 4.12.2012, at 18.59, Dave Abrahams wrote:
on Tue Dec 04 2012, Timo Sirainen <tss-AT-iki.fi> wrote:
On 4.12.2012, at 18.46, Dave Abrahams wrote:
Subject says all. Timo, would you write something up where the world will find it? Also, it would be great if doveadm fts was covered in the doveadm manpage.
It runs the Lucene's optimize function with fts-lucene backend. Does nothing with others. Basically I think it just gathers multiple Lucene index files together and packs them into a single file, dropping all deleted data (Lucene documentation has more details I think). Similar to what "doveadm purge" does to mdbox files. fts-lucene runs the optimize also internally automatically once in a while, so I'm not sure if that doveadm command should necessarily even exist.
I think the answer lies in your perception of benefits vs maintainance (including documentation) costs... ;-)
Based on the awesome explanation I just gave, I wouldn't mind if someone converts something based on it to a man page patch!
The lack of a working Git mirror really raises the barrier for me to contribute. I was just wondering if you had tried to reinstate it with a recent version of the hggit plugin? It appears to be officially under development at https://bitbucket.org/durin42/hg-git...
-- Dave Abrahams BoostPro Computing Software Development Training http://www.boostpro.com Clang/LLVM/EDG Compilers C++ Boost
On 4.12.2012, at 20.24, Dave Abrahams wrote:
Based on the awesome explanation I just gave, I wouldn't mind if someone converts something based on it to a man page patch!
The lack of a working Git mirror really raises the barrier for me to contribute. I was just wondering if you had tried to reinstate it with a recent version of the hggit plugin? It appears to be officially under development at https://bitbucket.org/durin42/hg-git...
Based on a quick read of the web page, that looks to me more like a reverse of what you want. Or alternatively it looks like something you would install yourself. In any case I'm not really planning on installing an official git mirror.
on Tue Dec 04 2012, Timo Sirainen <tss-AT-iki.fi> wrote:
On 4.12.2012, at 20.24, Dave Abrahams wrote:
Based on the awesome explanation I just gave, I wouldn't mind if someone converts something based on it to a man page patch!
The lack of a working Git mirror really raises the barrier for me to contribute. I was just wondering if you had tried to reinstate it with a recent version of the hggit plugin? It appears to be officially under development at https://bitbucket.org/durin42/hg-git...
Based on a quick read of the web page, that looks to me more like a reverse of what you want. Or alternatively it looks like something you would install yourself.
I did, actually...
In any case I'm not really planning on installing an official git mirror.
Oh, I thought you had one and it had to be taken offline because it stopped working... maybe I'm thinking of a different project.
I'll see what I can do with that plugin.
-- Dave Abrahams BoostPro Computing Software Development Training http://www.boostpro.com Clang/LLVM/EDG Compilers C++ Boost
participants (2)
-
Dave Abrahams
-
Timo Sirainen