sup talk - Oct 2008 - bug: reproducable Sup crash, ferret-less rewrite?

I can reproduce now the following:

"Sending...
  /var/lib/gems/1.8/gems/sup-0.6/lib/sup/index.rb:367: [BUG] Segmentation
fault
  ruby 1.8.6 (2007-09-24) [x86_64-linux]
  Aborted"

Googling around showed me a similar problem [1] and particulary an
enigmatic answer [2] by William Morgan:

"These are both Ferret index corruption errors, which means I can either
 debug Ferret, or rewrite Sup to not use it. (I''ve chosen the
latter.)"

I can not find any trace of the latter, i.e. Sup''s source code
repository at Gitorious [3] doesn''t show any ferret amputation. Are
you, William Morgan working on a Sup version that doesn''t show that
supposedly ferret induced crashiness? Possibly your "private" branch
is not ready yet and not accessible?
*t

[1] http://www.mail-archive.com/sup-talk at rubyforge.org/msg01217.html
[2] http://www.mail-archive.com/sup-talk at rubyforge.org/msg01218.html
[3] http://gitorious.org/projects/sup

"Tomas Pospisek ML" <tpo2 at sourcepole.ch> @ 2008-10-31 6:52
PM:
> I can not find any trace of the latter, i.e. Sup''s source code
> repository at Gitorious [3] doesn''t show any ferret amputation.
Are
> you, William Morgan working on a Sup version that doesn''t show
that
> supposedly ferret induced crashiness? Possibly your "private"
branch
> is not ready yet and not accessible?
See these blog posts about "Sup: The Server!"
http://all-thing.net/search/label/sup

-- 
Christopher Warrington <chrisw at rice.edu>
Jones College

On 11/1/2008, "Christopher Warrington" <chrisw at rice.edu>
wrote:
>"Tomas Pospisek ML" <tpo2 at sourcepole.ch> @ 2008-10-31
6:52 PM:
>
>> I can not find any trace of the latter, i.e. Sup''s source code
>> repository at Gitorious [3] doesn''t show any ferret
amputation. Are
>> you, William Morgan working on a Sup version that doesn''t show
that
>> supposedly ferret induced crashiness? Possibly your "private"
branch
>> is not ready yet and not accessible?
>
>See these blog posts about "Sup: The Server!"
>http://all-thing.net/search/label/sup
Very interesting, thanks. So, while William Morgan is pursuing his
ambitous project [1]:
- has anybody tried to swap Ferret with any other indexer [also in [1]]?
- has anybody tried to fix the index corruption / ferret crash bug [2]?
*t

[1] http://rubyforge.org/pipermail/sup-talk/2008-July/001593.html
[2] http://ferret.davebalmain.com/trac/ticket/279

Reformatted excerpts from Tomas Pospisek ML''s message of
2008-11-01:
> - has anybody tried to swap Ferret with any other indexer [also in
> [1]]?
Not AFAIK. There are a couple indexing options with Ruby bindings:
Sphinx, Xapian, Solr, and HyperEstraier, off the top of my head. When I
last looked, they all had their problems: Sphinx requires a daemon, is
overly tied to databases, and has no incremental indexing, Xapian is
confusing and doesn''t seem to support document fields, Solr is Java,
HyperEstraier doesn''t seem to have anyone using it, etc.
> - has anybody tried to fix the index corruption / ferret crash bug [2]?
I''d love for someone to do this! Ferret has the best API and document
model, by far, of any of the alternatives.
-- 
William <wmorgan-sup at masanjin.net>