[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: PATCH fix 510970, 529551, 530541



On Mon, 2009-11-23 at 10:41 -0500, Chris Lumens wrote:
> > >Cool, then the question is at what amount of ram should this kick in at,
> > >I went with MIN_GUI_RAM.
> > >
> > 
> > That sounds reasonable. Maybe want to do use something like MIN_GUI_RAM + 100MB,
> > so as to have atleast MIN_GUI_RAM free when the install.img (which is approx
> > 100Mb) lives in RAM.
> 
> I don't like this one bit.  Sure, today it's ~ 100 MB, but what about in
> two years when it's more like ~ 150 MB and no one remembers this needs
> updating?  Because I guarantee, no one will remember.

How about something like MIN_GUI_RAM > memAvailable? Nothing is really
using that function now, it returns memory less what is used by /tmp.
That would do away with the hard-coding, and scale into the future. The
only variable going forward would be the size of install.img
 
> 
> Also, you're introducing another memory limit besides MIN_GUI_RAM that
> will mean a different set of behavior happens at some random memory
> threshold.
> 

You have that now with a difference between http/ftp/hd and media
installs when you test iutil.memInstalled() < isys.MIN_GUI_RAM with out
taking into account what might be in /tmp, maybe that should be using
memAvailable? 

Jerry








[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]