Re: [Monotone-devel] mtn: error: sqlite error: database is locked

[Juan Jose Comellas]

Having a single header certainly helps to make the distribution easier, but it makes modifications much harder, especially if you're not Sergei. :)

Anyway, don't get me wrong, I really liked OTL. It's incredible what the library can do with a single header and how efficient it is. It worked very well for me.

On 1/3/08, Craig L. Ching <address@hidden> wrote:

Hi Jose,

I think OTL is a great idea, in fact, I'm sure that Kelly Hickel here
has mentioned it before.  We employ the developer of OTL and use it in
our products, it is definitely a good tool.

As for the distribution being in one big header file, that actually
shouldn't put anyone off.  It uses templates pretty heavily and he does
this so that it's easy to build on multiple platforms and it's easier
for him to redistribute.  Honestly, I'd think it would be easier to
include it in monotone this way.

Anyway, if anyone has any questions, feel free to ask.  Sergei is very
helpful through his normal channels, but we'd be happy to make sure that
if you did decide to look at it, that you got special attention ;-)

Cheers,
Craig

> -----Original Message-----
> From: monotone-devel-bounces+cching=address@hidden
> [mailto: address@hidden] On
Behalf
> Of Juan Jose Comellas
> Sent: Thursday, January 03, 2008 6:09 AM
> To: address@hidden
> Subject: Re: [Monotone-devel] mtn: error: sqlite error: database is
locked
>
> If you're going to work on improving Monotone's database layer it
might be
> worth it to use a library that can provide a generic interface to
multiple
> databases. There is one I've used successfully with Oracle (via OCI),
> PostgreSQL (via ODBC) and Informix (also via ODBC), that also supports
> SQLite. It's called OTL and you can download it from
> http://otl.sourceforge.net.
>
> The abstraction it presents is very comfortable to use (something like
> iostreams for the database world) but the way the library is
distributed
> (a huge header file with lots of templates in it) can be a put down.
>
>
>
> On 12/27/07, Alvaro Herrera <address@hidden > wrote:
>
>       Zack Weinberg wrote:
>       > At this point I'm going to have to ask you to back up your
claims
> with
>       > actual design, if not actual code.  Please look at the
existing
>       > database layer in Monotone (database.cc , schema.sql) and the
> higher
>       > layers where most of the application-specific consistency
> constraints
>       > are currently enforced ( revision.cc, cset.cc, roster.cc,
>       > roster_merge.cc).  Please tell us in detail how you would take
>       > advantage of SQL features that we are currently not using, to
> enable
>       > additional concurrency.  You can totally redesign the database
> tables
>       > if you want.  If at all possible, please distinguish between
> things
>       > that could be done with SQLite and things that would require
> either
>       > extending SQLite or switching to a new database
implementation.
>
>       I'll have a look.
>
>       While fetching the source I noticed that
>       http://www.venge.net/mtn-wiki/SelfHostingInfo still mentions
> venge.net
>       as the Monotone server, but that host is not listening on the
> netsync port.
>       Changing it to monotone.ca works great.
>
>       --
>       Alvaro Herrera
> http://www.PlanetPostgreSQL.org/
>       One man's impedance mismatch is another man's layer of
abstraction.
>       (Lincoln Yeoh)
>
>
>       _______________________________________________
>       Monotone-devel mailing list
>       address@hidden
>       http://lists.nongnu.org/mailman/listinfo/monotone-devel
> < http://lists.nongnu.org/mailman/listinfo/monotone-devel>
>
>