[PC-BSD Testing] Long standing redundancy in ports

Kris Moore kris at pcbsd.com
Thu Sep 4 07:52:53 PDT 2008

DragonFlier wrote:
> Network connection doesn't make any difference...right now the image of
> already extracted ports is written to disk on an install but not any INDEXes
> (those found in /usr/ports/)
> -rw-r--r--     1 root  wheel  10441465 Sep  2 21:02 INDEX
> -rw-r--r--     1 root  wheel  17370624 Sep  2 21:02 INDEX-5
> -rw-r--r--     1 root  wheel  18467854 Sep  2 21:02 INDEX-6
> -rw-r--r--     1 root  wheel  18468202 Sep  2 21:02 INDEX-7
> Even though the ports tree exists after an install if you use the PCBSD
> ports manager or go from the command line you still have to do a 'portsnap
> fetch extract' (pfe) before you can do any further 'portsnap fetch updates'
> (pfu). The pfe downloads an entire snapshot of the ports (currently 52+ MBs)
> and extracts this snapshot over the already existing ports tree. The pfu
> merely gets what's not current and updates the port tree. If it were
> possible to include the snapshot during the install (on CD) and then extract
> it creating your INDEXes as part of the install, any further updates to the
> port tree would only include those items that were out of date and not the
> entire 52+ MBs again.

I've been watching this thread, and I'm happy to say that I've fixed 
this on the build server. It now uses portsnap instead of csup, and will 
include all the INDEX's that are needed, so you should only have to grab 
the latest updates now :)


Kris Moore
PC-BSD Software

More information about the Testing mailing list