> This is kind of annoying because implementing io_map on the server side in
> not
> necessarily trivial I think. How can this be implemented in a server that
> doesn't use libpager (I don't think there are such examples in the source
> tree)?
It is easy to add. I one wrote a small test program t
On Tue, Nov 12, 2002 at 06:21:04PM +0100, Ludovic Courtès wrote:
> On Tue, Nov 12, 2002 at 01:41:47PM +0100, Marcus Brinkmann wrote:
> > They should fall back to normal read/write if mmap fails.
>
> Gcc does not actually fall back to normal read/write. Should we ask them to
> fix that?
I think t
On Tue, Nov 12, 2002 at 01:41:47PM +0100, Marcus Brinkmann wrote:
> They should fall back to normal read/write if mmap fails.
Gcc does not actually fall back to normal read/write. Should we ask them to
fix that?
> Why can't you use libpager?
This is a netfs-based filesystem (tarfs). So far, I'
"Ludovic Courtès" <[EMAIL PROTECTED]> writes:
> Although the GNU Coding Standards specify that user applications shouldn't
> rely on mmap() to read files, it turns out that gcc sometimes relies on it
> (depending on the source file size I think). Should we ask the gcc team to
> change that? ;)
g
On Tue, Nov 12, 2002 at 10:59:16AM +0100, Ludovic Courtès wrote:
> Hi,
>
> Although the GNU Coding Standards specify that user applications shouldn't
> rely on mmap() to read files, it turns out that gcc sometimes relies on it
> (depending on the source file size I think). Should we ask the gcc t
Hi,
Although the GNU Coding Standards specify that user applications shouldn't
rely on mmap() to read files, it turns out that gcc sometimes relies on it
(depending on the source file size I think). Should we ask the gcc team to
change that? ;)
This is kind of annoying because implementing io_ma
