Darren J Moffat <[EMAIL PROTECTED]> wrote:

> > I poinTed him the the most recent code - please get yourself informed!
>
> That link is just a tar file.  There is no indication at all that is an 
> updated version of the code already in ONNV.
>
> So I had to make an assumption, and given it was you the assumption is 
> that it is a schily version since you have produced alternate and 
> portable versions of many things and often point people to your 
> alternate versions.

I did also mention that this is a more recent version. I would not do that for 
inofficial code.

> I'd happily be informed but a pointer to a tar file alone with no 
> additional info on what to expect isn't useful.  Maybe a simple readme 
> on that page would help rather than just a tar file.

I did put it out for a differente reason (to give Sun employees access to the 
current code) and just send a pointer to the location I created 3 weeks ago.

> Glad to see you are working on improving the hsfs in OpenSolaris, thanks.

Adding code to ON is a hard job.... it takes a lot more time than it takes
to write the code.

FYI: I did write the Joliet & ISO-9660:1999 extensions last year and the
"correct hardlink" / Rock Ridge V-1.12 extension that now are waiting for 
integration since approx. 9 months. The next code will add support for ISO-9660
files > 4 GB (this feature is called "multi-extent" files). I did just publish 
the related code to enhance mkisofs in cdrtools:

ftp://ftp.berlios.de/pub/cdrecord/alpha/cdrtools-2.01.01a32.tar.bz2

which is needed to be able to have "multi-extent" hot filesystem images for 
testing.

The code to support Joliet and ISO-9660:1999 did already eliminate a few costly
string compares and the code to support multi-extent files (> 4 GB) will also
change the way directories are read. There is a hack from Moinak Ghosh that
adds read-adead for hsfs _file content_ but with this extension, hsfs still 
causes a lot of seeks on the slow CD/DVD media. If someone has a good idea on
how to find where excatly the most important bottle-neck in hsfs is located, I 
would be happy.

I was aking because I did suspect that working on the way directory reads and 
file meta data caching is done could help more to speed up hsfs.

Jörg

-- 
 EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin
       [EMAIL PROTECTED]                (uni)  
       [EMAIL PROTECTED]     (work) Blog: http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code

Reply via email to