Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 30 Nov 2010 11:15:09 -0800
From:      Chuck Swiger <cswiger@mac.com>
To:        David Southwell <david@vizion2000.net>
Cc:        freebsd-ports@freebsd.org
Subject:   Re: Ports - installation & upgrade history
Message-ID:  <500E66A0-FC50-4867-80E4-D188E45B3B83@mac.com>
In-Reply-To: <201011301858.11175.david@vizion2000.net>
References:  <201011301824.15550.david@vizion2000.net> <201011301841.35462.david@vizion2000.net> <3B286D1E-E7F8-4FF5-85E9-29309A8693C6@mac.com> <201011301858.11175.david@vizion2000.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On Nov 30, 2010, at 10:58 AM, David Southwell wrote:
> Seems to me that a comprehensive record would be extremely useful on a local system. I am wondering how difficult it might be to collect data from applications such as: 
>  
> cvsup of ports tree
> portupgrade/portmaster 
> changes to /var/db/ports
> changes to /usr/ports/distfiles

It's not difficult.  The normal way people track changes to filesystems over time is by making backups (or snapshots, or other equivalents).  The normal way people track process execution is accton / sa.

> The results could be held in a mysql database.

I suppose...MySQL isn't particularly efficient at dealing with large numbers of BLOBs, which is what importing filesystem-tree changes would probably become.  You'd likely end up with a MySQL database which grows to be many orders of magnitude bigger than the size of /var/db/ports + whatever under /usr/ports.

Things like CVS or Subversion better understand how to represent the list of deltas representing the changes than MySQL does.  Good backup software which understands dedup'ing, things like mbox format, etc can also track changes more efficiently than a naive method of keeping around a copy of every file every time it changes.

-- 
-Chuck




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?500E66A0-FC50-4867-80E4-D188E45B3B83>