Hello Jordi,

Jordi Pujol:
> uloop shows various advantages over other forms of copying the filesystem to 
> disk or to RAM.
> With that there is no delay at starting the system, since the cache is 
> created 
> when the Operating System needs new parts from the original media and the 
> space used by the cache grows according to the reading from the original 
> device.
> 
> Thank you for that tool !
> and all the rest of the assemble !
> because my Debian Live uses squasfs+lzma and aufs also.

Glad to hear that.
Uloop used be one of my favorite, but I don't use it recently.


> Observing that tool, The original filesystem, by now, only could be a block 
> device and "http" URLs.  
> 
> A suggestion, I propose the creation of the tool "ulofile":

As I describe in the readme file, ulobdev.c and ulohttp.c are just
samples. And you can easily apply it to various type of backends.

Ah, you already did it...
I read your patch in another mail...
One comment. It is better to issue fstat(2) after open(2), instead of
lstat(2) + open(2). Probably, to close(2) or not to close(2) in the
error case doesn't matter.

I'll describe about your patch and trial somewhere under sample/uloop.
Thank you.


Junjiro Okajima

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone

Reply via email to