nnsql.el ?

nnsql.el ?

[Julian Assange]

Has anyone considered writing a mail-indexing sql backend to replace
nnml? I'm a *heavy* nnml user (8k groups, 15k messages in my personal
mail folder etc) and unfortunately dealing with this morass is starting
to show nnmls limitations.

Cheers,
Julian.

Re: nnsql.el ?

[Simon Josefsson]

Julian Assange <proff@q.org> writes:

> Has anyone considered writing a mail-indexing sql backend to replace
> nnml? I'm a *heavy* nnml user (8k groups, 15k messages in my personal
> mail folder etc) and unfortunately dealing with this morass is starting
> to show nnmls limitations.

Does profiling show that nnml is the bottleneck?

When I profiled it, nnml wasn't the biggest bottleneck so other things
should be optimized first. Like gnus-range-*.

Re: nnsql.el ?

[Julian Assange]

Simon Josefsson <jas@pdc.kth.se> writes:

> Julian Assange <proff@q.org> writes:
> 
> > Has anyone considered writing a mail-indexing sql backend to replace
> > nnml? I'm a *heavy* nnml user (8k groups, 15k messages in my personal
> > mail folder etc) and unfortunately dealing with this morass is starting
> > to show nnmls limitations.
> 
> Does profiling show that nnml is the bottleneck?
> 
> When I profiled it, nnml wasn't the biggest bottleneck so other things
> should be optimized first. Like gnus-range-*.

The bottle neck(s) are derived from loading in, then parsing (in one
way or another) huge .overview files. With an sql backend the amount
of data passed up to gnus can be minimised. Expirations and many other
functions can be simplified.

There's a looking elisp postgresql client that was recently posted to
gnu.emacs.sources that looks ideal.

Cheers,
Julian.

Re: nnsql.el ?

[Lars Magne Ingebrigtsen]

Julian Assange <proff@q.org> writes:

> Has anyone considered writing a mail-indexing sql backend to replace
> nnml? I'm a *heavy* nnml user (8k groups, 15k messages in my personal
> mail folder etc) and unfortunately dealing with this morass is starting
> to show nnmls limitations.

I think an sql-based backend would be nice.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: nnsql.el ?

[Greg Stark]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Julian Assange <proff@q.org> writes:
> 
> > Has anyone considered writing a mail-indexing sql backend to replace
> > nnml? I'm a *heavy* nnml user (8k groups, 15k messages in my personal
> > mail folder etc) and unfortunately dealing with this morass is starting
> > to show nnmls limitations.
> 
> I think an sql-based backend would be nice.

Hm, perhaps this could be written in perlmacs and use the DBI database driver
switch so it could use any database. And this wouldn't require writing an SQL
driver for elisp.

-- 
greg

Re: nnsql.el ?

[Lars Magne Ingebrigtsen]

Greg Stark <gsstark@mit.edu> writes:

> Hm, perhaps this could be written in perlmacs and use the DBI database driver
> switch so it could use any database. And this wouldn't require writing an SQL
> driver for elisp.

Uhm -- perlmacs?  Is that what I think it is?  Please say it ain't so!
:-)

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: nnsql.el ?

[Randal L. Schwartz]

>>>>> "Lars" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

Lars> Uhm -- perlmacs?  Is that what I think it is?  Please say it ain't so!
Lars> :-)

... http://search.cpan.org/search?dist=Emacs-Lisp

yes, if you're afraid of that, be very afraid. :)

-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn@stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Perl/Unix/security consulting, Technical writing, Comedy, etc. etc.
See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training!

Re: nnsql.el ?

[Bruce Stephens]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Greg Stark <gsstark@mit.edu> writes:
> 
> > Hm, perhaps this could be written in perlmacs and use the DBI
> > database driver switch so it could use any database. And this
> > wouldn't require writing an SQL driver for elisp.
> 
> Uhm -- perlmacs?  Is that what I think it is?  Please say it ain't so!
> :-)

OK, it isn't so.  

And you don't want to read <URL:http://john-edwin-tobey.org/perlmacs/>

Re: nnsql.el ?

[Lars Magne Ingebrigtsen]

Bruce Stephens <bruce@cenderis.demon.co.uk> writes:

> And you don't want to read <URL:http://john-edwin-tobey.org/perlmacs/>

Oh.  My.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: nnsql.el ?

[William M. Perry]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Greg Stark <gsstark@mit.edu> writes:
> 
> > Hm, perhaps this could be written in perlmacs and use the DBI database driver
> > switch so it could use any database. And this wouldn't require writing an SQL
> > driver for elisp.
> 
> Uhm -- perlmacs?  Is that what I think it is?  Please say it ain't so!
> :-)

You haven't run across that yet?  It's freaky!

-Bill P.

Re: nnsql.el ?

[Randal L. Schwartz]

>>>>> "William" == William M Perry <wmperry@aventail.com> writes:

>> Uhm -- perlmacs?  Is that what I think it is?  Please say it ain't so!
>> :-)

William> You haven't run across that yet?  It's freaky!

Oddly enough, as a heavy Perl user, and a heavy Emacs user, I'm
still not using that yet. :)

-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn@stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Perl/Unix/security consulting, Technical writing, Comedy, etc. etc.
See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training!