portsearch: approach to finding ports that provide a given file

[Vasil Dimov]

On Tue, Jan 17, 2006 at 03:44:34PM +0100, Jean-Yves Lefort wrote:
> On Tue, 17 Jan 2006 11:48:05 +0200
> Vasil Dimov <vd at datamax.bg> wrote:
> 
> > Some ports generate their plists after being installed, making it
> > impossible to record the plist without fetching the source tarballs,
> > compiling and installing the software. Currently for these ports empty
> > or flawed plists are recorded. One way to fix this problem is to
> > retrieve the plists for these ports from the precompiled packages
> > available at ftp.freebsd.org but the whole packages need to be fetched,
> > causing (large) amount of unefficient network traffic.
> > Any ideas for resolving the problem are welcome.
> 
> The package building scripts (Tools/portbuild) could be modified to
> create pkgname-metadata.tbz files and make them available on the
> FreeBSD ftp site. An utility which would retrieve these files and
> maintain a database could then be created. Said utility could be run
> by end users on their home machines, as well as by some machine on the
> FreeBSD cluster which would maintain a database queryable through
> www.FreeBSD.org. This approach requires coordination with the package
> building and www people.

Here are some numbers:

Currently the list of all the plist files is 64M, without any trace
which port each file belongs to. When each plist file is prepended with
port number (1-14004) the list becomes 72M, when bzip2ed it becomes 6M,
but then split into 14000 separate files and each one bzip2ed separately
then all 14000 archives are 28M.

-- 
Vasil Dimov
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 155 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-ports/attachments/20060117/9a1f5d6e/attachment.bin