Re: Debbugs: The Next Generation
On Tue, Aug 07, 2001 at 05:26:48PM -0700, Sean 'Shaleh' Perry wrote:
> > Feedback?
>
> questions really.
>
> Why a language interface? Most of us access the bugs via bugs.d.o.
> How much overhead does the addition of the sql daemon add to the server?
To maximize code sharing between the various pieces of the bug reporting
system, including:
- bugs.debian.org CGIs
- mail-handling programs
- command-line tools
- various special-purpose scripts, such as:
- bugscan
- the "testing" scripts, whatever they're called
Some of these will probably still have to access the database, because an API
that allowed enough flexibility for their reporting needs would be nearly as
complex as SQL itself. However, they can pull bug IDs out of the database
themselves, then use the API to collect everything into meaningful bug objects
instead of doing that on their own.
As for postgresql, master is already running it, though it doesn't look very
busy:
postgres 6481 0.0 0.1 8084 1660 ? S Aug06 0:00 /usr/lib/postgresql/bin/postmaster -D /var/lib/postgres/data
I think it would be safe to say that CPU and I/O requirements would, overall,
be considerably less than the current debbugs for an equal set of features.
The prototype runs great on my workstation, which makes master look like a
beast. Enhancements such as full text searching would have to be tested for
resource requirements before unleashing them, of course.
--
- mdz
Reply to: