Use of database for SCM

Michael Ellerman michael at ellerman.id.au
Fri Jun 24 15:23:23 BST 2005


On Fri, 24 Jun 2005 23:14, tnorth at bluewin.ch wrote:
> Here is a question that I really have to ask :) It may be stupid, because
> of my lack of knowledge in SCM, but let`s go anyway...

No it's a reasonable question.

> I was just wondering why nobody (as far as I know) uses a database like
> SQLite to stock patches and/or patches informations.

Subversion uses an in-process database as its backend. Although recently a 
file-based backend has been developed and seems to be quite popular.

The DB backend is known to cause some problems, with databases sometimes 
getting corrupted and otherwise in a funny state.

Good discussion here: 
http://svnbook.red-bean.com/en/1.1/ch05.html#svn-ch-5-sect-1.3.1

> Is it because it is easier to work with a file tree ?
> or because Unix` philosophy is "file-centric" ?

There's a bunch of reasons. Certainly the UNIX philosophy is one reason. 
Although it might seem silly, I think a lot of users like to be able to look 
inside their SCM and understand how it works a little, that's a lot harder 
with a DB.

> I know the use of a DB would certainly bring more complications, but what
> about speed ? (is maybe the only profit we could have)
> SQLite seems to be very very quick...

Where are your benchmarks? :D The only example I'm aware of is Subversion with 
bdb and that is *not* very fast.

cheers

-- 
Michael Ellerman
IBM OzLabs

email: michael:ellerman.id.au
inmsg: mpe:jabber.org
wwweb: http://michael.ellerman.id.au
phone: +61 2 6212 1183 (tie line 70 21183)

We do not inherit the earth from our ancestors,
we borrow it from our children. - S.M.A.R.T Person
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20050625/3154c6ae/attachment.pgp 


More information about the bazaar mailing list