> It's been rumoured that Jonathan Corbet said:
> >
> > I've been messing with 1.3.5, trying to decide if I can really move my
> > family finances over and also pondering a new LWN article. Things are
>
> I still haven't seen the article, but I gather it was compilmentary.
> Thanks.
>
> > But...I've noticed that my X server seems to grow without bound while I'm
>
> If its the X server, then its an X server bug. X clients have no
> imediate or direct effect on memory management inside the x server.
> The idea is security: no single x client, no matter how malicious,
> is supposed to be able to crash or otherwise mess up the x server.
>
> Normally, the x server malloc's memory to store pixmaps that the client
> asked to have displayed. It allocs memory for fonts; lots of other
> things. When a client goes away, that memory is 'reclaimed'. That
> does not (necessarily) mean that X11 calls the free() subroutine;
> instead, it keeps that memory and puts it in an internal pool of
> available memory for reuse at a later time. In part, it does this to
> avoid 'memory fragmentation' which is the memory analogue of
> what happens to MSDOS file systems. So in fact, the x server may have a
> pile of free memory inside of it; its just that the linux kernel & the
> other memory reporting tools don't know/haven't been told this.
>
> (BTW this is also why X11 is client-server: if an X11 client crashes,
> the server is still there, running, and able to clean up after the dead
> client. The down side is, of course, that the x server (like the
> kernel) has to be 'perfect' and 'bug free'.)
Is there a tool to find out how the X server is using its memory?
Maybe we could at least see where it's going wrong.
dave
--
Gnucash Developer's List
To unsubscribe send empty email to: [EMAIL PROTECTED]
